The aim of this project is to utilize the Cleveland heart disease dataset to understand how medical attributes of a person contribute to heart-related diseases. We perform a detailed data analysis and create a machine learning model capable of predicting whether a person has heart disease or not, based on their medical attributes.
This is a binary classification task. Detailed explanations of the medical attributes are provided in the Jupyter Notebook.
This project is a milestone in the Data Science/ML course by ZTM.
The dataset contains medical records of 303 patients collected during their follow-up period, with each patient profile having 13 clinical characteristics and a target variable indicating the presence of heart disease.
- Officially available at UCI ML Repository.
- Also available in this repository here.
| Attribute | Description |
|---|---|
| Age | Age of a patient [years] |
| Sex | Gender of the patient [1: Male, 0: Female] |
| ChestPain | Chest pain type [0: Typical Angina, 1: Atypical Angina, 2: Non-Anginal Pain, 3: Asymptomatic] |
| RestingBP | Resting blood pressure in mm Hg (Normal blood pressure - 120/80 Hg) |
| Cholesterol | Serum cholesterol level in blood in mg/dL (Normal cholesterol level below for adults 200mg/dL) |
| FastingBS | Fasting Blood Sugar (1 if > 120 mg/dL, 0 otherwise) |
| RestingECG | Resting electrocardiogram results [0: Normal, 1: ST-T wave abnormality, 2: Probable or definite left ventricular hypertrophy] |
| MaxHR | Maximum heart rate achieved [Numeric value between 60 and 202] |
| ExerciseAngina | Exercise-induced angina [1: Yes, 0: No] |
| Oldpeak | ST depression induced by exercise relative to rest |
| ST_Slope | The slope of the peak exercise ST segment [0: Upsloping, 1: Flat, 2: Downsloping] |
| CA | Number of major vessels (0-3) colored by fluoroscopy |
| Thal | Thallium stress test result [1, 3: Normal, 6: Fixed defect, 7: Reversible defect] |
| HeartDisease | Output class [1: heart disease, 0: Normal] |
Read more in the data dictionary in the Jupyter Notebook.
Here is an overview of the procedural steps performed throughout this project:
- Problem Definition: Defining the objective of predicting heart disease.
- Data Collection & Exploratory Data Analysis (EDA): Gathering data and performing initial analysis.
- Data Preparation & Preprocessing: Cleaning and preparing data for modeling. The data was already preprocessed.
- Model Selection, Training, and Validation: Choosing appropriate models and validating their performance.
- Hyperparameter Tuning & Model Filtering: Fine-tuning model parameters for optimal performance.
- Final Model Evaluation: Evaluating the final model on the test dataset.
- Prediction and Interpretation: Making predictions and interpreting the results.
Our final model, evaluated on an unseen test dataset, yields considerable performance metrics:
