Koch_Case_Study.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "mount_file_id": "1gRN82nlbadQj4VCF4Ib5c0RAUKhBpbWS",
      "authorship_tag": "ABX9TyO1VDO0kh0fdnbWAnAkVliE",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "cd4da867b21649eab062750137ca3696": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_8718c3563e0041709bf7b4c33d844883",
              "IPY_MODEL_3047972f4b774bb6bf65f74963f90816",
              "IPY_MODEL_f3cc47e629cc4b30891c096e9f70c643"
            ],
            "layout": "IPY_MODEL_707679c695124c2397293e1c59991fc8"
          }
        },
        "8718c3563e0041709bf7b4c33d844883": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c6209706c3fb45b1b1553070bea026dc",
            "placeholder": "​",
            "style": "IPY_MODEL_9661b02fc1af4cf1a4e0e978d6f4de74",
            "value": "100%"
          }
        },
        "3047972f4b774bb6bf65f74963f90816": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_9eef03528a9242cda497487293ef9b9d",
            "max": 3,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_24d411e7b13c41c3b8dd80b6b0b7a1c0",
            "value": 3
          }
        },
        "f3cc47e629cc4b30891c096e9f70c643": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_8e099eccd74a423596b899490c5c47ef",
            "placeholder": "​",
            "style": "IPY_MODEL_88b9370604ed4d088a2079f9ff3a3113",
            "value": " 3/3 [00:00&lt;00:00, 39.82it/s]"
          }
        },
        "707679c695124c2397293e1c59991fc8": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c6209706c3fb45b1b1553070bea026dc": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "9661b02fc1af4cf1a4e0e978d6f4de74": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "9eef03528a9242cda497487293ef9b9d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "24d411e7b13c41c3b8dd80b6b0b7a1c0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "8e099eccd74a423596b899490c5c47ef": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "88b9370604ed4d088a2079f9ff3a3113": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/yifanwu0909/Yifan-s-Docs/blob/master/Koch_Case_Study.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Author\n",
        "\n",
        "Yifan Wu  \n",
        "yw515@georgetown.edu  \n",
        "2064340079  \n",
        "linkeIn: https://www.linkedin.com/in/yifan-wu-193050a9/\n",
        "\n",
        "# Introduction\n",
        "\n",
        "Predicting stock markets has been an endeavor a lot of people have chased.  Using Machine Learning (ML) and past price data to predict the next periods price or direction in the stock market is not new, neither does it produce any meaningful predictions. Since stock price is a stochastis(non-stationary) process, the characteristics of price movement don’t stay the same with time. Hence, predicting future prices based on historical patterns isn't theoretically possible.\n",
        "\n",
        "\n",
        "I propose a simple yet effective classification problem where only the direction of the stock is predicted for the next day. That seems much more doable than predicting price. \n",
        "\n",
        "\n",
        "# Methodology\n",
        "\n",
        "The company I chose is Exxon.\n",
        "XGboost binary classification model is trained with historic Exxon stock index data.\n",
        "\n",
        "# Result\n",
        "\n",
        "The model prediction reached 98% accuracy on test data and outperformed S&P500 stock index(67% accuracy).\n"
      ],
      "metadata": {
        "id": "C7yYD5pm8NnD"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Python Implementation\n",
        "\n",
        "## Import necessary packages\n"
      ],
      "metadata": {
        "id": "s92b2IC6JECf"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bVtUV-1R7W_d",
        "outputId": "0b875b5b-f504-4164-f5be-5405dc2c2038"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "ERROR:numba.cuda.cudadrv.driver:Call to cuInit results in CUDA_ERROR_NO_DEVICE\n"
          ]
        }
      ],
      "source": [
        "import os\n",
        "file_dir = \"/content/drive/MyDrive/Colab Notebooks/\"\n",
        "os.chdir(file_dir)\n",
        "import math\n",
        "import matplotlib\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import skopt\n",
        "from datetime import date\n",
        "from matplotlib import pyplot as plt\n",
        "from pylab import rcParams\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from tqdm import tqdm_notebook\n",
        "from xgboost import XGBClassifier\n",
        "import pandas_datareader as pdr\n",
        "from tsfresh.utilities.dataframe_functions import roll_time_series\n",
        "from tsfresh import extract_features\n",
        "import pandas as pd\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "from copy import deepcopy\n",
        "import matplotlib as mpl\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "from statsmodels.graphics.tsaplots import plot_acf,plot_pacf\n",
        "from sklearn.model_selection import GridSearchCV,RandomizedSearchCV,train_test_split\n",
        "from skopt import BayesSearchCV\n",
        "from sklearn.metrics import accuracy_score,f1_score,roc_auc_score,confusion_matrix,roc_curve\n",
        "import xgboost as xgb\n",
        "import time\n",
        "import warnings\n",
        "%matplotlib inline"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pip install scikit-optimize"
      ],
      "metadata": {
        "id": "ajKdHtg1Fgf4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "pip install -U tsfresh"
      ],
      "metadata": {
        "id": "SY5Cd-1Wyyl2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "pip install scikit-plot"
      ],
      "metadata": {
        "id": "YsbPeYUI52gl"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#pip install fix-yahoo-finance==0.1.30\n",
        "!pip install --upgrade pandas-datareader"
      ],
      "metadata": {
        "id": "VA6suC6iZTyC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Import data\n",
        "\n",
        "Data source: [Yahoo finance](https://finance.yahoo.com/quote/Xom?ltr=1)"
      ],
      "metadata": {
        "id": "MwgJ3hpyTxik"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Convert Date column to datetime\n",
        "df=pdr.get_data_yahoo('XOM',start = '2015-02-28', end = '2022-02-28')\n",
        "df=df.reset_index()\n",
        "df.loc[:, 'Date'] = pd.to_datetime(df['Date'],format='%Y-%m-%d')\n",
        "# Change all column headings to be lower case, and remove spacing\n",
        "df.columns = [str(x).lower().replace(' ', '_') for x in df.columns]\n",
        "# Sort by datetime\n",
        "df.sort_values(by='date', inplace=True, ascending=True)\n",
        "df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "gtiNytw1LdRS",
        "outputId": "b45a5557-8f67-4e76-f7cb-529877efa821"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        date       high        low       open      close      volume  \\\n",
              "0 2015-03-02  88.519997  87.349998  88.339996  88.040001  13793700.0   \n",
              "1 2015-03-03  88.019997  87.129997  87.930000  87.620003  13493800.0   \n",
              "2 2015-03-04  87.959999  86.930000  87.959999  87.180000  13467600.0   \n",
              "3 2015-03-05  87.190002  86.519997  87.070000  86.739998  12527600.0   \n",
              "4 2015-03-06  86.599998  85.320000  86.410004  85.629997  16070500.0   \n",
              "\n",
              "   adj_close  \n",
              "0  61.355690  \n",
              "1  61.063007  \n",
              "2  60.756363  \n",
              "3  60.449726  \n",
              "4  59.676163  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-3d3faa42-bcc7-467d-ae8f-2f1957f28b7a\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>date</th>\n",
              "      <th>high</th>\n",
              "      <th>low</th>\n",
              "      <th>open</th>\n",
              "      <th>close</th>\n",
              "      <th>volume</th>\n",
              "      <th>adj_close</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2015-03-02</td>\n",
              "      <td>88.519997</td>\n",
              "      <td>87.349998</td>\n",
              "      <td>88.339996</td>\n",
              "      <td>88.040001</td>\n",
              "      <td>13793700.0</td>\n",
              "      <td>61.355690</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2015-03-03</td>\n",
              "      <td>88.019997</td>\n",
              "      <td>87.129997</td>\n",
              "      <td>87.930000</td>\n",
              "      <td>87.620003</td>\n",
              "      <td>13493800.0</td>\n",
              "      <td>61.063007</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2015-03-04</td>\n",
              "      <td>87.959999</td>\n",
              "      <td>86.930000</td>\n",
              "      <td>87.959999</td>\n",
              "      <td>87.180000</td>\n",
              "      <td>13467600.0</td>\n",
              "      <td>60.756363</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2015-03-05</td>\n",
              "      <td>87.190002</td>\n",
              "      <td>86.519997</td>\n",
              "      <td>87.070000</td>\n",
              "      <td>86.739998</td>\n",
              "      <td>12527600.0</td>\n",
              "      <td>60.449726</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2015-03-06</td>\n",
              "      <td>86.599998</td>\n",
              "      <td>85.320000</td>\n",
              "      <td>86.410004</td>\n",
              "      <td>85.629997</td>\n",
              "      <td>16070500.0</td>\n",
              "      <td>59.676163</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-3d3faa42-bcc7-467d-ae8f-2f1957f28b7a')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-3d3faa42-bcc7-467d-ae8f-2f1957f28b7a button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-3d3faa42-bcc7-467d-ae8f-2f1957f28b7a');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot adjusted close over time\n",
        "rcParams['figure.figsize'] = 10, 8 # width 10, height 8\n",
        "\n",
        "ax = df.plot(x='date', y='adj_close', style='b-', grid=True)\n",
        "ax.set_xlabel(\"date\")\n",
        "ax.set_ylabel(\"USD\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 482
        },
        "id": "CTGgFMvIlf9A",
        "outputId": "f39f4c96-15ec-4b9e-cc4f-fba0d8475ebf"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'USD')"
            ]
          },
          "metadata": {},
          "execution_count": 3
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Once the data has been downloaded and stored into a new environment I clean the data up a little, put all the lists into a single data frame, compute the daily returns for each asset and create the up or down direction which will be what the classification model will try to predict. We want to use the adjusted closing prices column to generate direction(adjusted for stock splits, which can make it appear that a stock is performing poorly if a split is included)."
      ],
      "metadata": {
        "id": "2v7-EoPbQxig"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df['ret'] = df['adj_close'].pct_change(periods = 1)\n",
        "df.dropna(inplace=True)\n",
        "df.reset_index(drop=True, inplace=True)\n",
        "df['chng'] = df['ret'].map(lambda v: 1 if v > 0 else 0)\n",
        "df.head(10)\n",
        "#a random sample of observations of the data looks like:"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 363
        },
        "id": "rgPhZOATLqDz",
        "outputId": "2ea5edfc-3a91-45fb-f183-e1d209d31021"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "           date       high        low       open      close      volume  \\\n",
              "1224 2020-01-10  69.959999  69.120003  69.930000  69.139999  11995600.0   \n",
              "430  2016-11-11  86.730003  84.889999  86.529999  85.669998  13812200.0   \n",
              "1438 2020-11-13  36.250000  35.380001  35.400002  36.080002  29241500.0   \n",
              "632  2017-09-01  76.849998  76.320000  76.370003  76.570000   7361100.0   \n",
              "497  2017-02-21  82.150002  81.300003  82.120003  81.889999  16419600.0   \n",
              "504  2017-03-02  84.250000  82.790001  82.800003  83.300003  14170300.0   \n",
              "627  2017-08-25  77.129997  76.430000  76.559998  76.720001   6844900.0   \n",
              "1210 2019-12-19  70.050003  69.239998  69.989998  69.389999  13273200.0   \n",
              "1674 2021-10-22  63.220001  62.349998  62.770000  63.119999  14266800.0   \n",
              "555  2017-05-15  83.180000  82.650002  83.000000  82.800003  10881500.0   \n",
              "\n",
              "      adj_close       ret  chng  \n",
              "1224  58.016151 -0.008888     0  \n",
              "430   63.455635 -0.015853     0  \n",
              "1438  32.692989  0.024127     1  \n",
              "632   58.334358  0.003144     1  \n",
              "497   61.210453  0.001590     1  \n",
              "504   62.264370  0.003373     1  \n",
              "627   58.448647  0.005109     1  \n",
              "1210  58.225933 -0.006870     0  \n",
              "1674  59.906467  0.006859     1  \n",
              "555   62.475090  0.003028     1  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-e08237b4-7152-4159-ae6c-21d9edc32744\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>date</th>\n",
              "      <th>high</th>\n",
              "      <th>low</th>\n",
              "      <th>open</th>\n",
              "      <th>close</th>\n",
              "      <th>volume</th>\n",
              "      <th>adj_close</th>\n",
              "      <th>ret</th>\n",
              "      <th>chng</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1224</th>\n",
              "      <td>2020-01-10</td>\n",
              "      <td>69.959999</td>\n",
              "      <td>69.120003</td>\n",
              "      <td>69.930000</td>\n",
              "      <td>69.139999</td>\n",
              "      <td>11995600.0</td>\n",
              "      <td>58.016151</td>\n",
              "      <td>-0.008888</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>430</th>\n",
              "      <td>2016-11-11</td>\n",
              "      <td>86.730003</td>\n",
              "      <td>84.889999</td>\n",
              "      <td>86.529999</td>\n",
              "      <td>85.669998</td>\n",
              "      <td>13812200.0</td>\n",
              "      <td>63.455635</td>\n",
              "      <td>-0.015853</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1438</th>\n",
              "      <td>2020-11-13</td>\n",
              "      <td>36.250000</td>\n",
              "      <td>35.380001</td>\n",
              "      <td>35.400002</td>\n",
              "      <td>36.080002</td>\n",
              "      <td>29241500.0</td>\n",
              "      <td>32.692989</td>\n",
              "      <td>0.024127</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>632</th>\n",
              "      <td>2017-09-01</td>\n",
              "      <td>76.849998</td>\n",
              "      <td>76.320000</td>\n",
              "      <td>76.370003</td>\n",
              "      <td>76.570000</td>\n",
              "      <td>7361100.0</td>\n",
              "      <td>58.334358</td>\n",
              "      <td>0.003144</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>497</th>\n",
              "      <td>2017-02-21</td>\n",
              "      <td>82.150002</td>\n",
              "      <td>81.300003</td>\n",
              "      <td>82.120003</td>\n",
              "      <td>81.889999</td>\n",
              "      <td>16419600.0</td>\n",
              "      <td>61.210453</td>\n",
              "      <td>0.001590</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>504</th>\n",
              "      <td>2017-03-02</td>\n",
              "      <td>84.250000</td>\n",
              "      <td>82.790001</td>\n",
              "      <td>82.800003</td>\n",
              "      <td>83.300003</td>\n",
              "      <td>14170300.0</td>\n",
              "      <td>62.264370</td>\n",
              "      <td>0.003373</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>627</th>\n",
              "      <td>2017-08-25</td>\n",
              "      <td>77.129997</td>\n",
              "      <td>76.430000</td>\n",
              "      <td>76.559998</td>\n",
              "      <td>76.720001</td>\n",
              "      <td>6844900.0</td>\n",
              "      <td>58.448647</td>\n",
              "      <td>0.005109</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1210</th>\n",
              "      <td>2019-12-19</td>\n",
              "      <td>70.050003</td>\n",
              "      <td>69.239998</td>\n",
              "      <td>69.989998</td>\n",
              "      <td>69.389999</td>\n",
              "      <td>13273200.0</td>\n",
              "      <td>58.225933</td>\n",
              "      <td>-0.006870</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1674</th>\n",
              "      <td>2021-10-22</td>\n",
              "      <td>63.220001</td>\n",
              "      <td>62.349998</td>\n",
              "      <td>62.770000</td>\n",
              "      <td>63.119999</td>\n",
              "      <td>14266800.0</td>\n",
              "      <td>59.906467</td>\n",
              "      <td>0.006859</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>555</th>\n",
              "      <td>2017-05-15</td>\n",
              "      <td>83.180000</td>\n",
              "      <td>82.650002</td>\n",
              "      <td>83.000000</td>\n",
              "      <td>82.800003</td>\n",
              "      <td>10881500.0</td>\n",
              "      <td>62.475090</td>\n",
              "      <td>0.003028</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-e08237b4-7152-4159-ae6c-21d9edc32744')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-e08237b4-7152-4159-ae6c-21d9edc32744 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-e08237b4-7152-4159-ae6c-21d9edc32744');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Feature Engineering\n",
        "\n",
        "We will generate the following features on a rolling bases: collapse the time series data down by taking the previous N days of data and calculating the feature functions on it. Here I chose N = 7. \n",
        "\n",
        "In two of my mail reference project for this project(see foot note), each of them chose a specific set of time series features. Immediately the question popped into my head was: why these features? And is there more features that we can unpack? Through my research I found 10+ features that I can implement but soon I was getting overwhelmed. Lukily I discovered a package call tsfresh, which generate thousands of time series features. It also provide feature selection so we only use the influential ones for training. "
      ],
      "metadata": {
        "id": "wAWuFuNUvbVf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# feature importance https://stackoverflow.com/a/50848452 \n",
        "from tsfresh import extract_features\n",
        "# tsfresh:https://towardsdatascience.com/automate-time-series-feature-engineering-in-a-few-lines-of-python-code-f28fe52e4704\n",
        "df['id'] = 'XOM'\n",
        "X = df.drop([\"chng\", \"ret\", \"adj_close\"], axis=1)\n",
        "\n",
        "# Roll time series data with lag number N = 7, we are aggregating information over the pass 7 days, for every day, we calcullate some features using data from day N, N - 1, ... , N - 7\n",
        "df_rolled = roll_time_series(X, column_id=\"id\", column_sort=\"date\", max_timeshift = 7)\n",
        "df_features = extract_features(df_rolled, column_id=\"id\", column_sort=\"date\")\n",
        "skopt.dump(df_features, \"df_features.pkl\")\n",
        "df_features.head(3)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 324
        },
        "id": "5eJeuE4y1fvg",
        "outputId": "d1e04159-3a8a-4523-d75b-2752e7f32d87"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Rolling: 100%|██████████| 1762/1762 [00:03<00:00, 458.70it/s]\n",
            "Feature Extraction: 100%|██████████| 8810/8810 [08:27<00:00, 17.36it/s]\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                high__variance_larger_than_standard_deviation  \\\n",
              "XOM 2015-03-03                                            0.0   \n",
              "    2015-03-04                                            0.0   \n",
              "    2015-03-05                                            0.0   \n",
              "\n",
              "                high__has_duplicate_max  high__has_duplicate_min  \\\n",
              "XOM 2015-03-03                      0.0                      0.0   \n",
              "    2015-03-04                      0.0                      0.0   \n",
              "    2015-03-05                      0.0                      0.0   \n",
              "\n",
              "                high__has_duplicate  high__sum_values  high__abs_energy  \\\n",
              "XOM 2015-03-03                  0.0         88.019997       7747.519809   \n",
              "    2015-03-04                  0.0        175.979996      15484.481248   \n",
              "    2015-03-05                  0.0        263.169998      23086.577774   \n",
              "\n",
              "                high__mean_abs_change  high__mean_change  \\\n",
              "XOM 2015-03-03                    NaN                NaN   \n",
              "    2015-03-04               0.059998          -0.059998   \n",
              "    2015-03-05               0.414997          -0.414997   \n",
              "\n",
              "                high__mean_second_derivative_central  high__median  ...  \\\n",
              "XOM 2015-03-03                                   NaN     88.019997  ...   \n",
              "    2015-03-04                                   NaN     87.989998  ...   \n",
              "    2015-03-05                                -0.355     87.959999  ...   \n",
              "\n",
              "                volume__permutation_entropy__dimension_6__tau_1  \\\n",
              "XOM 2015-03-03                                              NaN   \n",
              "    2015-03-04                                              NaN   \n",
              "    2015-03-05                                              NaN   \n",
              "\n",
              "                volume__permutation_entropy__dimension_7__tau_1  \\\n",
              "XOM 2015-03-03                                              NaN   \n",
              "    2015-03-04                                              NaN   \n",
              "    2015-03-05                                              NaN   \n",
              "\n",
              "                volume__query_similarity_count__query_None__threshold_0.0  \\\n",
              "XOM 2015-03-03                                                NaN           \n",
              "    2015-03-04                                                NaN           \n",
              "    2015-03-05                                                NaN           \n",
              "\n",
              "                volume__matrix_profile__feature_\"min\"__threshold_0.98  \\\n",
              "XOM 2015-03-03                                                NaN       \n",
              "    2015-03-04                                                NaN       \n",
              "    2015-03-05                                                NaN       \n",
              "\n",
              "                volume__matrix_profile__feature_\"max\"__threshold_0.98  \\\n",
              "XOM 2015-03-03                                                NaN       \n",
              "    2015-03-04                                                NaN       \n",
              "    2015-03-05                                                NaN       \n",
              "\n",
              "                volume__matrix_profile__feature_\"mean\"__threshold_0.98  \\\n",
              "XOM 2015-03-03                                                NaN        \n",
              "    2015-03-04                                                NaN        \n",
              "    2015-03-05                                                NaN        \n",
              "\n",
              "                volume__matrix_profile__feature_\"median\"__threshold_0.98  \\\n",
              "XOM 2015-03-03                                                NaN          \n",
              "    2015-03-04                                                NaN          \n",
              "    2015-03-05                                                NaN          \n",
              "\n",
              "                volume__matrix_profile__feature_\"25\"__threshold_0.98  \\\n",
              "XOM 2015-03-03                                                NaN      \n",
              "    2015-03-04                                                NaN      \n",
              "    2015-03-05                                                NaN      \n",
              "\n",
              "                volume__matrix_profile__feature_\"75\"__threshold_0.98  \\\n",
              "XOM 2015-03-03                                                NaN      \n",
              "    2015-03-04                                                NaN      \n",
              "    2015-03-05                                                NaN      \n",
              "\n",
              "                volume__mean_n_absolute_max__number_of_maxima_7  \n",
              "XOM 2015-03-03                                              NaN  \n",
              "    2015-03-04                                              NaN  \n",
              "    2015-03-05                                              NaN  \n",
              "\n",
              "[3 rows x 3945 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-406bb3d2-a6a4-4bc9-a0dc-cca5534a90ff\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th>high__variance_larger_than_standard_deviation</th>\n",
              "      <th>high__has_duplicate_max</th>\n",
              "      <th>high__has_duplicate_min</th>\n",
              "      <th>high__has_duplicate</th>\n",
              "      <th>high__sum_values</th>\n",
              "      <th>high__abs_energy</th>\n",
              "      <th>high__mean_abs_change</th>\n",
              "      <th>high__mean_change</th>\n",
              "      <th>high__mean_second_derivative_central</th>\n",
              "      <th>high__median</th>\n",
              "      <th>...</th>\n",
              "      <th>volume__permutation_entropy__dimension_6__tau_1</th>\n",
              "      <th>volume__permutation_entropy__dimension_7__tau_1</th>\n",
              "      <th>volume__query_similarity_count__query_None__threshold_0.0</th>\n",
              "      <th>volume__matrix_profile__feature_\"min\"__threshold_0.98</th>\n",
              "      <th>volume__matrix_profile__feature_\"max\"__threshold_0.98</th>\n",
              "      <th>volume__matrix_profile__feature_\"mean\"__threshold_0.98</th>\n",
              "      <th>volume__matrix_profile__feature_\"median\"__threshold_0.98</th>\n",
              "      <th>volume__matrix_profile__feature_\"25\"__threshold_0.98</th>\n",
              "      <th>volume__matrix_profile__feature_\"75\"__threshold_0.98</th>\n",
              "      <th>volume__mean_n_absolute_max__number_of_maxima_7</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th rowspan=\"3\" valign=\"top\">XOM</th>\n",
              "      <th>2015-03-03</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>88.019997</td>\n",
              "      <td>7747.519809</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>88.019997</td>\n",
              "      <td>...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2015-03-04</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>175.979996</td>\n",
              "      <td>15484.481248</td>\n",
              "      <td>0.059998</td>\n",
              "      <td>-0.059998</td>\n",
              "      <td>NaN</td>\n",
              "      <td>87.989998</td>\n",
              "      <td>...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2015-03-05</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>263.169998</td>\n",
              "      <td>23086.577774</td>\n",
              "      <td>0.414997</td>\n",
              "      <td>-0.414997</td>\n",
              "      <td>-0.355</td>\n",
              "      <td>87.959999</td>\n",
              "      <td>...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>3 rows × 3945 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-406bb3d2-a6a4-4bc9-a0dc-cca5534a90ff')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-406bb3d2-a6a4-4bc9-a0dc-cca5534a90ff button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-406bb3d2-a6a4-4bc9-a0dc-cca5534a90ff');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Almost 4000 features are created from my dataset! Here is the full list: https://tsfresh.readthedocs.io/en/latest/text/list_of_features.html"
      ],
      "metadata": {
        "id": "p8NNPaXDb0bK"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_features = skopt.load(\"df_features.pkl\")\n",
        "df_features = df_features.reset_index(level=[0,1])\n",
        "df_features = df_features.dropna(axis=1, how='any')\n",
        "df_features = df_features.rename({'level_0': 'id', 'level_1': 'date'}, axis = 1)"
      ],
      "metadata": {
        "id": "G0sN9cYo5Usk"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## More features \n",
        "\n",
        "Create more time series rolling features manually with a new N. This time I chose N = 3.\n",
        "\n",
        "* Mean 'adj_close' of each month\n",
        "* Difference between high and low of each day\n",
        "* Difference between open and close of each day\n",
        "* Mean volume of each month"
      ],
      "metadata": {
        "id": "JfnfroTRZT_U"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Get difference between high and low of each day\n",
        "df['range_hl'] = df['high'] - df['low']\n",
        "\n",
        "# Get difference between open and close of each day\n",
        "df['range_oc'] = df['open'] - df['close']\n",
        "\n",
        "# Add a column 'order_day' to indicate the order of the rows by date\n",
        "df['order_day'] = [x for x in list(range(len(df)))]\n",
        "\n",
        "# merging_keys\n",
        "merging_keys = ['order_day']\n",
        "\n",
        "# List of columns that we will use to create lags\n",
        "lag_cols = ['adj_close', 'range_hl', 'range_oc', 'volume']\n",
        "\n",
        "N = 3  \n",
        "shift_range = [x+1 for x in range(N)]\n",
        "\n",
        "for shift in tqdm_notebook(shift_range):\n",
        "    train_shift = df[merging_keys + lag_cols].copy()\n",
        "    \n",
        "    # E.g. order_day of 0 becomes 1, for shift = 1.\n",
        "    # So when this is merged with order_day of 1 in df, this will represent lag of 1.\n",
        "    train_shift['order_day'] = train_shift['order_day'] + shift\n",
        "    \n",
        "    foo = lambda x: '{}_lag_{}'.format(x, shift) if x in lag_cols else x\n",
        "    train_shift = train_shift.rename(columns=foo)\n",
        "\n",
        "    df = pd.merge(df, train_shift, on=merging_keys, how='left') #.fillna(0)\n",
        "    \n",
        "del train_shift"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49,
          "referenced_widgets": [
            "cd4da867b21649eab062750137ca3696",
            "8718c3563e0041709bf7b4c33d844883",
            "3047972f4b774bb6bf65f74963f90816",
            "f3cc47e629cc4b30891c096e9f70c643",
            "707679c695124c2397293e1c59991fc8",
            "c6209706c3fb45b1b1553070bea026dc",
            "9661b02fc1af4cf1a4e0e978d6f4de74",
            "9eef03528a9242cda497487293ef9b9d",
            "24d411e7b13c41c3b8dd80b6b0b7a1c0",
            "8e099eccd74a423596b899490c5c47ef",
            "88b9370604ed4d088a2079f9ff3a3113"
          ]
        },
        "id": "dPzrF8gSDAFY",
        "outputId": "85c9f644-717f-4c24-9e6b-7956168d15f2"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/3 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "cd4da867b21649eab062750137ca3696"
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "I merge the auto-generated features and manually generated dfeatures together as our feature candicate pool. This is all features that there is that COULD build a accurate model. But I do not know which ones are really informative and which ones are just weakly relavent. Yet!"
      ],
      "metadata": {
        "id": "3gY2mSXMb_41"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_all_features = df.merge(df_features, left_on = \"date\", right_on = \"date\")\n",
        "df_all_features = df_all_features.drop([\"chng\", \"ret\", \"adj_close\"], axis=1)\n",
        "df_all_features = df_all_features.dropna(axis=1, how='any')"
      ],
      "metadata": {
        "id": "L1TXBTB5guRX"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "tsfresh have built in feature filtering function. To limit the number of irrelevant features, tsfresh deploys the fresh algorithm (fresh stands for FeatuRe Extraction based on Scalable Hypothesis tests). It is an efficient, scalable feature extraction algorithm, which filters the available features in an early stage of the machine learning pipeline with respect to their significance for the classification or regression task, while controlling the expected percentage of selected but irrelevant features. [(reference)](https://tsfresh.readthedocs.io/en/latest/text/feature_filtering.html)"
      ],
      "metadata": {
        "id": "4a_JMuOvclv0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from tsfresh import extract_features\n",
        "from tsfresh.feature_selection.relevance import calculate_relevance_table\n",
        "relevance_table = calculate_relevance_table(df_all_features, df['chng'])\n",
        "relevance_table = relevance_table[relevance_table.relevant]\n",
        "relevance_table.sort_values(\"p_value\", inplace=True)\n",
        "relevance_table = relevance_table.reset_index(drop = True)\n",
        "relevance_table\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "SBmBVkrJCYE3",
        "outputId": "54da52cd-0a88-442e-d8d7-79f48a0e0454"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                              feature  type        p_value  \\\n",
              "0                                            range_oc  real  1.641771e-177   \n",
              "1   close__energy_ratio_by_chunks__num_segments_10...  real   1.599245e-82   \n",
              "2                            close__number_peaks__n_1  real   3.685120e-46   \n",
              "3   close__change_quantiles__f_agg_\"mean\"__isabs_F...  real   1.308765e-36   \n",
              "4   low__energy_ratio_by_chunks__num_segments_10__...  real   7.598428e-36   \n",
              "..                                                ...   ...            ...   \n",
              "76      low__time_reversal_asymmetry_statistic__lag_2  real   2.998441e-04   \n",
              "77  open__change_quantiles__f_agg_\"mean\"__isabs_Fa...  real   3.128749e-04   \n",
              "78  open__change_quantiles__f_agg_\"var\"__isabs_Tru...  real   3.239422e-04   \n",
              "79  close__energy_ratio_by_chunks__num_segments_10...  real   3.280027e-04   \n",
              "80  high__change_quantiles__f_agg_\"var\"__isabs_Tru...  real   5.905615e-04   \n",
              "\n",
              "    relevant  \n",
              "0       True  \n",
              "1       True  \n",
              "2       True  \n",
              "3       True  \n",
              "4       True  \n",
              "..       ...  \n",
              "76      True  \n",
              "77      True  \n",
              "78      True  \n",
              "79      True  \n",
              "80      True  \n",
              "\n",
              "[81 rows x 4 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-6d645655-ec56-4719-a0b4-eb30dcafdfd0\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>feature</th>\n",
              "      <th>type</th>\n",
              "      <th>p_value</th>\n",
              "      <th>relevant</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>range_oc</td>\n",
              "      <td>real</td>\n",
              "      <td>1.641771e-177</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>close__energy_ratio_by_chunks__num_segments_10...</td>\n",
              "      <td>real</td>\n",
              "      <td>1.599245e-82</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>close__number_peaks__n_1</td>\n",
              "      <td>real</td>\n",
              "      <td>3.685120e-46</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>close__change_quantiles__f_agg_\"mean\"__isabs_F...</td>\n",
              "      <td>real</td>\n",
              "      <td>1.308765e-36</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>low__energy_ratio_by_chunks__num_segments_10__...</td>\n",
              "      <td>real</td>\n",
              "      <td>7.598428e-36</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>76</th>\n",
              "      <td>low__time_reversal_asymmetry_statistic__lag_2</td>\n",
              "      <td>real</td>\n",
              "      <td>2.998441e-04</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>77</th>\n",
              "      <td>open__change_quantiles__f_agg_\"mean\"__isabs_Fa...</td>\n",
              "      <td>real</td>\n",
              "      <td>3.128749e-04</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>78</th>\n",
              "      <td>open__change_quantiles__f_agg_\"var\"__isabs_Tru...</td>\n",
              "      <td>real</td>\n",
              "      <td>3.239422e-04</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>79</th>\n",
              "      <td>close__energy_ratio_by_chunks__num_segments_10...</td>\n",
              "      <td>real</td>\n",
              "      <td>3.280027e-04</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>80</th>\n",
              "      <td>high__change_quantiles__f_agg_\"var\"__isabs_Tru...</td>\n",
              "      <td>real</td>\n",
              "      <td>5.905615e-04</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>81 rows × 4 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-6d645655-ec56-4719-a0b4-eb30dcafdfd0')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-6d645655-ec56-4719-a0b4-eb30dcafdfd0 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-6d645655-ec56-4719-a0b4-eb30dcafdfd0');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The above table is the relavent features selected by the filtering algorithm based on their statistical significance. This final 81 features are the ones I am going to train the model on. "
      ],
      "metadata": {
        "id": "-NAGD5lldsMY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Organize the feature data\n",
        "df_rlvt_features = df_all_features[['date'] + relevance_table['feature'].to_list()]\n",
        "df_merge = df.merge(df_rlvt_features, left_on = \"date\", right_on = \"date\")\n",
        "df_merge = df_merge.dropna(axis=1, how='any')\n",
        "df_merge['month'] = df_merge['date'].dt.month\n",
        "df_merge = df_merge.drop([\"ret\", \"adj_close\", \"id\", 'order_day'], axis=1)"
      ],
      "metadata": {
        "id": "XncRW8Ppsmie"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Create training, dev, and test dataset\n",
        "\n",
        "* Training Dataset: The sample of data used to fit the model.\n",
        "* Validation Dataset: The sample of data used to provide an unbiased evaluation of a model fit on the training dataset while tuning model hyperparameters. \n",
        "* Test Dataset: The sample of data used to provide an unbiased evaluation of a final model fit on the training dataset.\n",
        "![Screen Shot 2022-11-20 at 6.28.38 PM.png](data:image/png;base64,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)\n"
      ],
      "metadata": {
        "id": "DoNKi1xg98Rl"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from datetime import datetime\n",
        "df_merge.sort_values(by='date', inplace=True, ascending=True)\n",
        "\n",
        "num_cut1 = int(len(df_merge) * 0.8)\n",
        "num_cut2 = int(len(df_merge) * 0.9)\n",
        "\n",
        "split_date_1 = df_merge['date'][num_cut1]\n",
        "split_date_2 = df_merge['date'][num_cut2]\n",
        "\n",
        "print(str(split_date_1) + '      ' + str(split_date_2)) \n",
        "print(str(num_cut1) + '                     ' + str(num_cut2)) "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D-gltkfCgvQ4",
        "outputId": "9040ed48-c37b-4c4e-fd47-17262e35b380"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2020-10-05 00:00:00      2021-06-17 00:00:00\n",
            "1409                     1585\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Split into train, cv, and test\n",
        "train = df_merge[:num_cut1]\n",
        "cv = df_merge[num_cut1:num_cut2]\n",
        "train_cv = df_merge[:num_cut2]\n",
        "test = df_merge[num_cut2:]\n",
        "print(\"train.shape = \" + str(train.shape))\n",
        "print(\"cv.shape = \" + str(cv.shape))\n",
        "print(\"train_cv.shape = \" + str(train_cv.shape))\n",
        "print(\"test.shape = \" + str(test.shape))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LibBIcmvhEQe",
        "outputId": "46606b17-2a5c-4579-ecd4-651f6dd8af1f"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "train.shape = (1409, 91)\n",
            "cv.shape = (176, 91)\n",
            "train_cv.shape = (1585, 91)\n",
            "test.shape = (177, 91)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Feature scaling\n",
        "\n",
        "Machine Learning models don't generalize as well on data with high scale variance, so I will adjust the features to a common scale. This process is called Standardizing and will scale data into mean value of 0 and a standard deviation of 1. It's more useful and common for classification tasks.\n",
        "\n",
        "Example:\n",
        "\n",
        "\n",
        "![Screen Shot 2022-11-20 at 6.42.26 PM.png](data:image/png;base64,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)\n",
        "\n",
        "\n",
        "[(reference)](https://www.cnblogs.com/quinn-yann/p/9808247.html)"
      ],
      "metadata": {
        "id": "qS8JKnkYlxt1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Get rid of final target and columns that does not need scaling.\n",
        "cols_to_scale = list(df_merge.columns)\n",
        "# items to be removed\n",
        "no_scale = ['date', 'chng', 'month']\n",
        "cols_to_scale = [ele for ele in cols_to_scale if ele not in no_scale]"
      ],
      "metadata": {
        "id": "QuOJCijiVWrZ"
      },
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Here we only scale the train dataset, and not the entire dataset to prevent information leak. Data leakage is when information from outside the training dataset is used to create the model. This additional information can allow the model to learn or know something that it otherwise would not know and in turn invalidate the estimated performance of the mode being constructed. Thus we should not standardize the entire dataset. \n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "a0zM8BtkuRpz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Do scaling for train set\n",
        "scaler = StandardScaler()\n",
        "train_scaled = scaler.fit_transform(train[cols_to_scale])\n",
        "# Convert the numpy array back into pandas dataframe\n",
        "train_scaled = pd.DataFrame(train_scaled, columns=cols_to_scale)\n",
        "train_scaled[no_scale] = train.reset_index()[no_scale]\n",
        "train_scaled.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 386
        },
        "id": "YEFLg2xgiV6g",
        "outputId": "b5614df1-34f0-4fe1-a77a-5260227e2068"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       high       low      open     close    volume  range_hl  range_oc_x  \\\n",
              "0  0.865029  0.878526  0.898131  0.873213 -0.128971 -0.503104    0.343832   \n",
              "1  0.860472  0.863544  0.900397  0.840105 -0.132008 -0.295412    0.919467   \n",
              "2  0.801986  0.832831  0.833144  0.806996 -0.240965 -0.829470    0.368332   \n",
              "3  0.757171  0.742938  0.783271  0.723473  0.169699  0.075470    0.919477   \n",
              "4  0.740461  0.727956  0.761357  0.688107 -0.011900  0.045793    1.139923   \n",
              "\n",
              "   range_oc_y  \\\n",
              "0    0.343832   \n",
              "1    0.919467   \n",
              "2    0.368332   \n",
              "3    0.919477   \n",
              "4    1.139923   \n",
              "\n",
              "   close__energy_ratio_by_chunks__num_segments_10__segment_focus_7  \\\n",
              "0                                         -11.719542                 \n",
              "1                                         -11.719542                 \n",
              "2                                         -11.719542                 \n",
              "3                                         -11.719542                 \n",
              "4                                         -11.719542                 \n",
              "\n",
              "   close__number_peaks__n_1  ...  \\\n",
              "0                  -2.16427  ...   \n",
              "1                  -2.16427  ...   \n",
              "2                  -2.16427  ...   \n",
              "3                  -2.16427  ...   \n",
              "4                  -2.16427  ...   \n",
              "\n",
              "   high__change_quantiles__f_agg_\"mean\"__isabs_False__qh_1.0__ql_0.6  \\\n",
              "0                                           0.141344                   \n",
              "1                                           0.141344                   \n",
              "2                                           0.141344                   \n",
              "3                                           0.038421                   \n",
              "4                                           0.038421                   \n",
              "\n",
              "   close__energy_ratio_by_chunks__num_segments_10__segment_focus_6  \\\n",
              "0                                         -13.108175                 \n",
              "1                                         -13.108175                 \n",
              "2                                         -13.108175                 \n",
              "3                                         -13.108175                 \n",
              "4                                         -13.108175                 \n",
              "\n",
              "   low__time_reversal_asymmetry_statistic__lag_2  \\\n",
              "0                                       0.063552   \n",
              "1                                       0.063552   \n",
              "2                                       0.063552   \n",
              "3                                       0.063552   \n",
              "4                                      -1.545110   \n",
              "\n",
              "   open__change_quantiles__f_agg_\"mean\"__isabs_False__qh_0.4__ql_0.0  \\\n",
              "0                                           0.078673                   \n",
              "1                                           0.078673                   \n",
              "2                                           0.078673                   \n",
              "3                                          -0.877804                   \n",
              "4                                          -0.341601                   \n",
              "\n",
              "   open__change_quantiles__f_agg_\"var\"__isabs_True__qh_0.6__ql_0.2  \\\n",
              "0                                          -0.133575                 \n",
              "1                                          -0.133575                 \n",
              "2                                          -0.133575                 \n",
              "3                                          -0.133575                 \n",
              "4                                          -0.133575                 \n",
              "\n",
              "   close__energy_ratio_by_chunks__num_segments_10__segment_focus_0  \\\n",
              "0                                          32.309569                 \n",
              "1                                          13.900584                 \n",
              "2                                           7.764672                 \n",
              "3                                           4.732933                 \n",
              "4                                           2.904148                 \n",
              "\n",
              "   high__change_quantiles__f_agg_\"var\"__isabs_True__qh_1.0__ql_0.6       date  \\\n",
              "0                                          -0.280923               2015-03-03   \n",
              "1                                          -0.280923               2015-03-04   \n",
              "2                                          -0.280923               2015-03-05   \n",
              "3                                          -0.280923               2015-03-06   \n",
              "4                                          -0.280923               2015-03-09   \n",
              "\n",
              "   chng  month  \n",
              "0     0      3  \n",
              "1     0      3  \n",
              "2     0      3  \n",
              "3     0      3  \n",
              "4     0      3  \n",
              "\n",
              "[5 rows x 91 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-4ee7322e-45d7-4639-8699-e11b8082f397\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>high</th>\n",
              "      <th>low</th>\n",
              "      <th>open</th>\n",
              "      <th>close</th>\n",
              "      <th>volume</th>\n",
              "      <th>range_hl</th>\n",
              "      <th>range_oc_x</th>\n",
              "      <th>range_oc_y</th>\n",
              "      <th>close__energy_ratio_by_chunks__num_segments_10__segment_focus_7</th>\n",
              "      <th>close__number_peaks__n_1</th>\n",
              "      <th>...</th>\n",
              "      <th>high__change_quantiles__f_agg_\"mean\"__isabs_False__qh_1.0__ql_0.6</th>\n",
              "      <th>close__energy_ratio_by_chunks__num_segments_10__segment_focus_6</th>\n",
              "      <th>low__time_reversal_asymmetry_statistic__lag_2</th>\n",
              "      <th>open__change_quantiles__f_agg_\"mean\"__isabs_False__qh_0.4__ql_0.0</th>\n",
              "      <th>open__change_quantiles__f_agg_\"var\"__isabs_True__qh_0.6__ql_0.2</th>\n",
              "      <th>close__energy_ratio_by_chunks__num_segments_10__segment_focus_0</th>\n",
              "      <th>high__change_quantiles__f_agg_\"var\"__isabs_True__qh_1.0__ql_0.6</th>\n",
              "      <th>date</th>\n",
              "      <th>chng</th>\n",
              "      <th>month</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.865029</td>\n",
              "      <td>0.878526</td>\n",
              "      <td>0.898131</td>\n",
              "      <td>0.873213</td>\n",
              "      <td>-0.128971</td>\n",
              "      <td>-0.503104</td>\n",
              "      <td>0.343832</td>\n",
              "      <td>0.343832</td>\n",
              "      <td>-11.719542</td>\n",
              "      <td>-2.16427</td>\n",
              "      <td>...</td>\n",
              "      <td>0.141344</td>\n",
              "      <td>-13.108175</td>\n",
              "      <td>0.063552</td>\n",
              "      <td>0.078673</td>\n",
              "      <td>-0.133575</td>\n",
              "      <td>32.309569</td>\n",
              "      <td>-0.280923</td>\n",
              "      <td>2015-03-03</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.860472</td>\n",
              "      <td>0.863544</td>\n",
              "      <td>0.900397</td>\n",
              "      <td>0.840105</td>\n",
              "      <td>-0.132008</td>\n",
              "      <td>-0.295412</td>\n",
              "      <td>0.919467</td>\n",
              "      <td>0.919467</td>\n",
              "      <td>-11.719542</td>\n",
              "      <td>-2.16427</td>\n",
              "      <td>...</td>\n",
              "      <td>0.141344</td>\n",
              "      <td>-13.108175</td>\n",
              "      <td>0.063552</td>\n",
              "      <td>0.078673</td>\n",
              "      <td>-0.133575</td>\n",
              "      <td>13.900584</td>\n",
              "      <td>-0.280923</td>\n",
              "      <td>2015-03-04</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.801986</td>\n",
              "      <td>0.832831</td>\n",
              "      <td>0.833144</td>\n",
              "      <td>0.806996</td>\n",
              "      <td>-0.240965</td>\n",
              "      <td>-0.829470</td>\n",
              "      <td>0.368332</td>\n",
              "      <td>0.368332</td>\n",
              "      <td>-11.719542</td>\n",
              "      <td>-2.16427</td>\n",
              "      <td>...</td>\n",
              "      <td>0.141344</td>\n",
              "      <td>-13.108175</td>\n",
              "      <td>0.063552</td>\n",
              "      <td>0.078673</td>\n",
              "      <td>-0.133575</td>\n",
              "      <td>7.764672</td>\n",
              "      <td>-0.280923</td>\n",
              "      <td>2015-03-05</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.757171</td>\n",
              "      <td>0.742938</td>\n",
              "      <td>0.783271</td>\n",
              "      <td>0.723473</td>\n",
              "      <td>0.169699</td>\n",
              "      <td>0.075470</td>\n",
              "      <td>0.919477</td>\n",
              "      <td>0.919477</td>\n",
              "      <td>-11.719542</td>\n",
              "      <td>-2.16427</td>\n",
              "      <td>...</td>\n",
              "      <td>0.038421</td>\n",
              "      <td>-13.108175</td>\n",
              "      <td>0.063552</td>\n",
              "      <td>-0.877804</td>\n",
              "      <td>-0.133575</td>\n",
              "      <td>4.732933</td>\n",
              "      <td>-0.280923</td>\n",
              "      <td>2015-03-06</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.740461</td>\n",
              "      <td>0.727956</td>\n",
              "      <td>0.761357</td>\n",
              "      <td>0.688107</td>\n",
              "      <td>-0.011900</td>\n",
              "      <td>0.045793</td>\n",
              "      <td>1.139923</td>\n",
              "      <td>1.139923</td>\n",
              "      <td>-11.719542</td>\n",
              "      <td>-2.16427</td>\n",
              "      <td>...</td>\n",
              "      <td>0.038421</td>\n",
              "      <td>-13.108175</td>\n",
              "      <td>-1.545110</td>\n",
              "      <td>-0.341601</td>\n",
              "      <td>-0.133575</td>\n",
              "      <td>2.904148</td>\n",
              "      <td>-0.280923</td>\n",
              "      <td>2015-03-09</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 91 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4ee7322e-45d7-4639-8699-e11b8082f397')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-4ee7322e-45d7-4639-8699-e11b8082f397 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-4ee7322e-45d7-4639-8699-e11b8082f397');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "What is being done here is calculating the parameters for standardizing data within training data itself and use those parameters to standardizing dev set and test set.[(reference)](https://machinelearningmastery.com/data-leakage-machine-learning/)\n",
        "\n",
        "Now we scale the dev set and test set onto the new distribution of training set."
      ],
      "metadata": {
        "id": "oyV6C9d3ubGc"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Do scaling for dev set\n",
        "cv_no_scale = cv[no_scale].reset_index()\n",
        "cv_scaled = scaler.transform(cv[cols_to_scale])\n",
        "cv_scaled = pd.DataFrame(cv_scaled, columns = [cols_to_scale])\n",
        "for col in no_scale:\n",
        "  cv_scaled[col] = cv_no_scale[col]"
      ],
      "metadata": {
        "id": "XTZ8dj9D7W_3"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Do scaling for test set\n",
        "test_no_scale = test[no_scale].reset_index()\n",
        "test_scaled = scaler.transform(test[cols_to_scale])\n",
        "test_scaled = pd.DataFrame(test_scaled, columns = [cols_to_scale])\n",
        "for col in no_scale:\n",
        "  test_scaled[col] = test_no_scale[col]\n"
      ],
      "metadata": {
        "id": "R-MdiX0Qpd53"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Split into X and y\n",
        "\n",
        "X: features to train on.   \n",
        "y: target {'0', '1'}, which means stock price go up(1) or down(0)."
      ],
      "metadata": {
        "id": "XOC_EuWugoJL"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "if \"month\" not in cols_to_scale:\n",
        "  cols_to_scale.append(\"month\")\n",
        "target = \"chng\"\n",
        "features = cols_to_scale.copy()\n",
        " \n",
        "# Split into X and y\n",
        "X_train_scaled = train_scaled[features]\n",
        "y_train_scaled = train_scaled[target]\n",
        "X_cv_scaled = cv_scaled[features]\n",
        "y_cv_scaled = cv_scaled[target]\n",
        "X_test_scaled = test_scaled[features]\n",
        "y_test_scaled = test_scaled[target]\n",
        "print(\"X_train_scaled.shape = \" + str(X_train_scaled.shape))\n",
        "print(\"y_train_scaled.shape = \" + str(y_train_scaled.shape))\n",
        "print(\"X_cv_scaled.shape = \" + str(X_cv_scaled.shape))\n",
        "print(\"y_cv_scaled.shape = \" + str(y_cv_scaled.shape))\n",
        "print(\"X_test_scaled.shape = \" + str(X_test_scaled.shape))\n",
        "print(\"y_test.shape = \" + str(y_test_scaled.shape))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pe6XpHe7q25m",
        "outputId": "7e6669a8-e8ac-46b3-f679-d178604cc14a"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "X_train_scaled.shape = (1409, 89)\n",
            "y_train_scaled.shape = (1409,)\n",
            "X_cv_scaled.shape = (176, 89)\n",
            "y_cv_scaled.shape = (176, 1)\n",
            "X_test_scaled.shape = (177, 89)\n",
            "y_test.shape = (177, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Train the model\n",
        "\n",
        "There are 3 state-of-the-art algorithms when it comes to time series forcasting: ARIMA model, Long Short Term Memory (LSTM) model and Extreme Gradient Boosting (XGBoost) models. \n",
        "\n",
        "In [this paper](https://escholarship.org/uc/item/0zp9s76c), all three models produced good forecast of next day’s stock price. But XGBoost can provide feature importance analysis, which aids interpretability, and it requires least time for parameter tuning and forecast calculation. Due to the time-limit and computation power limit, natually I choose to train XGBoost model for stock direction forcasting.\n",
        "\n",
        "\n",
        "Machine learning models come with default parameters: if you do not assign a specific value or string to an optional parameter, the algorithm does it automatically by a preset value or string. However, optimal parameters may vary from one data set the to other, so default parameters are usually not optimal. One needs to find the optimal parameters by grid search, where the grid represents the experimental values of each parameter (n-dimensional space)."
      ],
      "metadata": {
        "id": "yy20nwMNwYuO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create the default model\n",
        "# https://xgboost.readthedocs.io/en/stable/python/python_api.html#xgboost.XGBClassifier\n",
        "model = XGBClassifier(objective='binary:logistic',\n",
        "                          booster='gbtree',\n",
        "                          eval_metric='auc',\n",
        "                          tree_method='hist',\n",
        "                          grow_policy='lossguide',\n",
        "                          use_label_encoder=False)\n",
        "\n",
        "# Train the classifier\n",
        "model.fit(X_train_scaled, y_train_scaled)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WJUPcvTSxeYu",
        "outputId": "b689f82a-b074-42b7-a6e4-426f82e7739d"
      },
      "execution_count": 144,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "XGBClassifier(eval_metric='auc', grow_policy='lossguide', tree_method='hist',\n",
              "              use_label_encoder=False)"
            ]
          },
          "metadata": {},
          "execution_count": 144
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#extracting default parameters from benchmark model\n",
        "default_params = {}\n",
        "gparams = model.get_params()\n",
        "\n",
        "#default parameters have to be wrapped in lists - even single values - so GridSearchCV can take them as inputs\n",
        "for key in gparams.keys():\n",
        "    gp = gparams[key]\n",
        "    default_params[key] = [gp]\n",
        "\n",
        "#setting grid of selected parameters for iteration\n",
        "param_grid = {'gamma': [0,0.1,0.2,0.4,0.8,1.6,3.2,6.4,12.8,25.6,51.2,102.4, 200],\n",
        "              'learning_rate': [0.01, 0.03, 0.06, 0.1, 0.15, 0.2, 0.25, 0.300000012, 0.4, 0.5, 0.6, 0.7],\n",
        "              'max_depth': [5,6,7,8,9,10,11,12,13,14],\n",
        "              'n_estimators': [50,65,80,100,115,130,150],\n",
        "              'reg_alpha': [0,0.1,0.2,0.4,0.8,1.6,3.2,6.4,12.8,25.6,51.2,102.4,200],\n",
        "              'reg_lambda': [0,0.1,0.2,0.4,0.8,1.6,3.2,6.4,12.8,25.6,51.2,102.4,200]}\n",
        "\n",
        "#No. of jobs\n",
        "bcvj = np.cumsum([len(x) for x in param_grid.values()])[-1]\n",
        "bcvj = int(bcvj)\n",
        "\n",
        "#unwrapping list values of default parameters\n",
        "default_params_xgb = {}\n",
        "\n",
        "for key in default_params.keys():\n",
        "    default_params_xgb[key] = default_params[key][0]\n",
        "\n",
        "#providing default parameters to xgbc model, before Bayesian based search cross-validation\n",
        "xgbc = xgb.XGBClassifier(**default_params_xgb)\n",
        "\n"
      ],
      "metadata": {
        "id": "jgOK-1-aFHgq"
      },
      "execution_count": 145,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Bayesian search — “an extremely powerful technique when the mathematical form of the function is unknown or expensive to compute. The main idea behind it is to compute a posterior distribution over the objective function based on the data (using the famous Bayes theorem), and then select good points to try with respect to this distribution”. [(reference)](https://cloud.google.com/blog/products/ai-machine-learning/hyperparameter-tuning-cloud-machine-learning-engine-using-bayesian-optimization)"
      ],
      "metadata": {
        "id": "6UjTo6Tc4qZg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "clf = BayesSearchCV(estimator=xgbc, search_spaces=param_grid, n_iter=bcvj, scoring='accuracy', cv=3, return_train_score=True, verbose=3)\n",
        "clf.fit(X_train_scaled, y_train_scaled.values.ravel())\n",
        "file_dir = \"/content/drive/MyDrive/Colab Notebooks/\"\n",
        "import skopt\n",
        "skopt.dump(clf, file_dir + \"xgboost_stock_price_tuned.pkl\")"
      ],
      "metadata": {
        "id": "HDcSIpg_IwUd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "clf = skopt.load(file_dir + \"xgboost_stock_price_tuned.pkl\")"
      ],
      "metadata": {
        "id": "imP7ubURTGDY"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#results dataframe\n",
        "df = pd.DataFrame(clf.cv_results_)"
      ],
      "metadata": {
        "id": "yFr4V-CceBqY"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#predictions - inputs to confusion matrix\n",
        "train_predictions = clf.predict(X_train_scaled)\n",
        "test_predictions = clf.predict(X_cv_scaled)\n",
        "unseen_predictions = clf.predict(X_test_scaled)\n",
        "    \n",
        "#confusion matrices\n",
        "cfm_train = confusion_matrix(y_train_scaled, train_predictions)\n",
        "cfm_test = confusion_matrix(y_cv_scaled, test_predictions)\n",
        "cfm_unseen = confusion_matrix(y_test_scaled, unseen_predictions)\n",
        "    \n",
        "#accuracy scores\n",
        "accs_train = accuracy_score(y_train_scaled, train_predictions)\n",
        "accs_test = accuracy_score(y_cv_scaled, test_predictions)\n",
        "accs_unseen = accuracy_score(y_test_scaled, unseen_predictions)\n",
        "    \n",
        "#F1 scores for each train/test label\n",
        "f1s_train_p1 = f1_score(y_train_scaled, train_predictions, pos_label=1)\n",
        "f1s_train_p0 = f1_score(y_train_scaled, train_predictions, pos_label=0)\n",
        "f1s_test_p1 = f1_score(y_cv_scaled, test_predictions, pos_label=1)\n",
        "f1s_test_p0 = f1_score(y_cv_scaled, test_predictions, pos_label=0)\n",
        "f1s_unseen_p1 = f1_score(y_test_scaled, unseen_predictions, pos_label=1)\n",
        "f1s_unseen_p0 = f1_score(y_test_scaled, unseen_predictions, pos_label=0)\n",
        "    \n",
        "#Area Under the Receiver Operating Characteristic Curve\n",
        "test_ras = roc_auc_score(y_cv_scaled, clf.predict_proba(X_cv_scaled)[:,1])\n",
        "unseen_ras = roc_auc_score(y_test_scaled, clf.predict_proba(X_test_scaled)[:,1])\n",
        "    \n",
        "#best parameters\n",
        "bp = clf.best_params_\n",
        "\n",
        "\n",
        "#storing computed values in results dictionary\n",
        "results_dict = {}\n",
        "results_dict['xgbc_rcv'] = {'classifier': deepcopy(clf),\n",
        "                            'cv_results': df.copy(),\n",
        "                            'cfm_train': cfm_train,\n",
        "                            'cfm_dev': cfm_test,\n",
        "                            'cfm_test': cfm_unseen,\n",
        "                            'train_accuracy': accs_train,\n",
        "                            'dev_accuracy': accs_test,\n",
        "                            'test_accuracy': accs_unseen,\n",
        "                            'train F1-score label 1': f1s_train_p1,\n",
        "                            'train F1-score label 0': f1s_train_p0,\n",
        "                            'dev F1-score label 1': f1s_test_p1,\n",
        "                            'dev F1-score label 0': f1s_test_p0,\n",
        "                            'test F1-score label 1': f1s_unseen_p1,\n",
        "                            'test F1-score label 0': f1s_unseen_p0,\n",
        "                            'dev roc auc score': test_ras,\n",
        "                            'test roc auc score': unseen_ras,\n",
        "                            'best_params': bp}\n"
      ],
      "metadata": {
        "id": "FOJzyAMwK7aF"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Predict on test set"
      ],
      "metadata": {
        "id": "fQu5qRds1NHN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Do prediction on train set\n",
        "pred_train = clf.predict(X_train_scaled)\n",
        "#Measuring accuracy on Testing Data\n",
        "from sklearn import metrics\n",
        "print(metrics.classification_report(y_train_scaled, pred_train))\n",
        "print(metrics.confusion_matrix(y_train_scaled, pred_train))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sG9LysyReefr",
        "outputId": "5e090fe2-2021-4f48-e359-193c0781562c"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      1.00      1.00       715\n",
            "           1       1.00      1.00      1.00       694\n",
            "\n",
            "    accuracy                           1.00      1409\n",
            "   macro avg       1.00      1.00      1.00      1409\n",
            "weighted avg       1.00      1.00      1.00      1409\n",
            "\n",
            "[[715   0]\n",
            " [  0 694]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Do prediction on test set\n",
        "pred_test = clf.predict(X_test_scaled)\n",
        "#Measuring accuracy on Testing Data\n",
        "from sklearn import metrics\n",
        "print(metrics.classification_report(y_test_scaled, pred_test))\n",
        "print(metrics.confusion_matrix(y_test_scaled, pred_test, labels= list(set(pred_test))))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_-xds-ZMefj8",
        "outputId": "39f97481-e778-445a-d677-8e31cab28660"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.98      0.98      0.98        84\n",
            "           1       0.98      0.98      0.98        93\n",
            "\n",
            "    accuracy                           0.98       177\n",
            "   macro avg       0.98      0.98      0.98       177\n",
            "weighted avg       0.98      0.98      0.98       177\n",
            "\n",
            "[[82  2]\n",
            " [ 2 91]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "I am very satisfied with the above result. The model is able to reach to 98% accuracy on unseen data. Let's see how is the S&P500 stock compare to this result."
      ],
      "metadata": {
        "id": "iZvzCjX1Dvuj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Compare to S&P500\n",
        "\n",
        "Perform the same preprocessing and scaling on the sp500 stock data."
      ],
      "metadata": {
        "id": "DpvswT8ODoh1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# ^GSPC\n",
        "\n",
        "sp500=pdr.get_data_yahoo('^GSPC',start = '2015-02-28', end = '2022-02-28')\n",
        "sp500=sp500.reset_index()\n",
        "# Convert Date column to datetime\n",
        "sp500.loc[:, 'Date'] = pd.to_datetime(sp500['Date'],format='%Y-%m-%d')\n",
        "# Change all column headings to be lower case, and remove spacing\n",
        "sp500.columns = [str(x).lower().replace(' ', '_') for x in sp500.columns]\n",
        "# Sort by datetime\n",
        "sp500.sort_values(by='date', inplace=True, ascending=True)\n",
        "sp500['ret'] = sp500['adj_close'].pct_change(periods = 1)\n",
        "sp500.dropna(inplace=True)\n",
        "sp500.reset_index(drop=True, inplace=True)\n",
        "sp500['chng'] = sp500['ret'].map(lambda v: 1 if v > 0 else 0)"
      ],
      "metadata": {
        "id": "pjBp3p4I9H90"
      },
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test_date_only = test[['date']]\n",
        "sp500.loc[:, 'date'] = pd.to_datetime(sp500['date'],format='%Y-%m-%d')\n",
        "sp500 = sp500.merge(test_date_only, how='inner', on = 'date')\n",
        "len(sp500) == len(test_date_only)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UfcNrFk69JNm",
        "outputId": "a4e39dea-3d89-45a7-a6a8-649903fb0084"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(metrics.classification_report(y_test_scaled, sp500['chng']))\n",
        "print(metrics.confusion_matrix(y_test_scaled, sp500['chng']))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0Vk1oOPIK6Qw",
        "outputId": "befdcc4b-8a9a-4db7-c3a6-76e287673a11"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.65      0.62      0.63        84\n",
            "           1       0.67      0.70      0.68        93\n",
            "\n",
            "    accuracy                           0.66       177\n",
            "   macro avg       0.66      0.66      0.66       177\n",
            "weighted avg       0.66      0.66      0.66       177\n",
            "\n",
            "[[52 32]\n",
            " [28 65]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Requery Exxon stock data with adjusted closing price since previous this column is replaced by stock price direction\n",
        "df = pdr.get_data_yahoo('XOM',start = '2015-02-28', end = '2022-02-28')\n",
        "df=df.reset_index()\n",
        "df.loc[:, 'Date'] = pd.to_datetime(df['Date'],format='%Y-%m-%d')\n",
        "df.columns = [str(x).lower().replace(' ', '_') for x in df.columns]\n",
        "df.sort_values(by='date', inplace=True, ascending=True)\n",
        "test_close_adj = sp500[['date']].merge(df, how = 'inner', on = 'date')\n",
        "test_scaled_close_adj = scaler.fit_transform(test_close_adj[['adj_close']])\n",
        "test_scaled_close_adj = pd.DataFrame(test_scaled_close_adj, columns=['adj_close'])\n",
        "test_scaled_close_adj['date'] = test_close_adj['date']\n",
        "sp500_scaled = scaler.fit_transform(sp500[['adj_close']])\n",
        "sp500_scaled = pd.DataFrame(sp500_scaled, columns=['adj_close'])\n",
        "sp500_scaled['date'] = sp500['date']"
      ],
      "metadata": {
        "id": "jtMcboIDpKDC"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot adjusted close over time for SP500 and Exxon index\n",
        "\n",
        "rcParams['figure.figsize'] = 10, 8 # width 10, height 8\n",
        "\n",
        "ax = sp500_scaled.plot(x='date', y='adj_close', style='y-', grid=True)\n",
        "ax = test_scaled_close_adj.plot(x='date', y='adj_close', style='b-', grid=True, ax=ax)\n",
        "ax.legend(['sp500', 'exxon'])\n",
        "ax.set_xlabel(\"date\")\n",
        "ax.set_ylabel(\"Ajusted closing price scaled\")\n",
        "ax.set_title(\"Both scaled to reveal trend\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 506
        },
        "id": "JDLxFr6Xq0Rh",
        "outputId": "a8b9854a-cbe8-4d7d-d318-bf0c52be99e7"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'Both scaled to reveal trend')"
            ]
          },
          "metadata": {},
          "execution_count": 28
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The Xgboost yields a very good performance, out ranking SP500. "
      ],
      "metadata": {
        "id": "spNwyrp-I4Qu"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Potential Improvement\n",
        "\n",
        "\n",
        "* Choose longer time span to extend training data size.   \n",
        "* Experience with different rolling lag number N.\n",
        "* Add company fundamental data as new features.\n",
        "* Incorporate daily sentiment data. give each day a sentiment score. This can be done using tweeter text mining and sentiment analysis. \n",
        "* Experiment with different models like LSTM  \n",
        "...\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "https://iopscience.iop.org/article/10.1088/1742-6596/2161/1/012065/pdf\n",
        "sentiment analysis code: https://github.com/rania751/Stock_Market_Prediction_LSTM_Numerical_Textual_Analysis/blob/main/stock-market-prediction-lstm-numerical-textual-ana.ipynb\n"
      ],
      "metadata": {
        "id": "97Z42Fq1ex9G"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Reference\n",
        "\n",
        "two of my main inspiration projects:\n",
        "1. https://lf0.com/post/xgboost-time-series-classification-trading-strategy/xgboost-time-series-quant-trading-strategy/\n",
        "2. https://github.com/nikkizhao1202/Trading_Strategy_SP500/blob/master/XGBoost/XGBoost.ipynb\n",
        "\n",
        "Other references\n",
        "\n",
        "https://medium.com/auquan/https-medium-com-auquan-machine-learning-techniques-trading-b7120cee4f05\n",
        "\n",
        "https://lf0.com/post/xgboost-time-series-classification-trading-strategy/xgboost-time-series-quant-trading-strategy/\n",
        "https://github.com/nikkizhao1202/Trading_Strategy_SP500/blob/master/XGBoost/XGBoost.ipynb\n",
        "\n",
        "https://escholarship.org/uc/item/0zp9s76c\n"
      ],
      "metadata": {
        "id": "io17pL56L330"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "_9lOnH0dNTqx"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}