Medical Insurance Premium Prediction Project

Project Overview

The Medical Insurance Premium Prediction project aims to build a machine learning model that predicts the annual premium cost for customers based on their health-related parameters. By leveraging this model, we can help individuals and insurance companies make informed decisions about medical coverage and its pricing.

Problem Context

The rising cost of healthcare services has made it essential for individuals to plan their medical expenses carefully. Understanding the factors influencing insurance premiums can:

• Enable customers to budget for their healthcare.
• Help insurance companies design personalized insurance plans.
• Provide insights into health-related risks and their impact on premium costs.

This project uses a dataset of health parameters provided voluntarily by nearly 1,000 individuals to create a predictive model for insurance premiums. The premium prices are represented in INR (₹) for yearly coverage.

Dataset Description

The dataset includes the following attributes:

• Age: The age of the individual (in years).
• Diabetes: Whether the individual has diabetes (Yes/No).
• BloodPressureProblems: Whether the individual has a history of blood pressure issues (Yes/No).
• AnyTransplants: Whether the individual has undergone any organ transplants (Yes/No).
• AnyChronicDiseases: Whether the individual has chronic diseases (Yes/No).
• Height: The height of the individual (in cm).
• Weight: The weight of the individual (in kg).
• KnownAllergies: Whether the individual has known allergies (Yes/No).
• HistoryOfCancerInFamily: Whether there is a family history of cancer (Yes/No).
• NumberOfMajorSurgeries: The number of major surgeries the individual has undergone.
• PremiumPrice (Target Variable): The annual medical insurance premium (in INR ₹).

Goals

1. Exploratory Data Analysis (EDA):

   • Analyze and visualize data to uncover patterns and relationships.
   • Identify key features that influence premium costs.
   • Handle missing values and detect outliers.

2. Model Development:

   • Build and evaluate regression models to predict the premium cost.
   • Compare different algorithms to determine the best-performing model.
   • Optimize the model to improve accuracy and generalization.

3. Insights and Impact:

   • Highlight the most significant factors influencing premiums.
   • Provide actionable insights for both customers and insurance providers.

Run predict.py

To run this code from the console, follow these steps:

1. Ensure you have Python installed on your system. This code requires Python 3.x.
2. Ensure you have the ridge_model.bin file in the same directory as predict.py. This file should contain the serialized model and test data.
3. Open a terminal or command prompt and navigate to the directory where app.py is located.
4. Run the application by executing the following command:

   python predict.py
   

5. The Flask application will start, and you should see output indicating that the server is running on http://0.0.0.0:9696.
6. You can now send POST requests to http://0.0.0.0:9696/predict with JSON data containing customer information to get price predictions.

Notebook: predict_test.ipynb

The predict_test.ipynb notebook is designed to demonstrate the prediction capabilities of the trained machine learning model on new, unseen data. This notebook includes the following sections:

1. Loading the Model:

   • Load the pre-trained model from the saved file.
   • Ensure all necessary dependencies and libraries are imported.

2. Preparing Test Data:

   • Load the test dataset containing new individuals' health parameters.
   • Perform any required preprocessing steps to match the training data format.

3. Making Predictions:

   • Use the loaded model to predict the insurance premiums for the test dataset.
   • Store and display the predicted premium values alongside the test data.

This notebook serves as a practical guide for applying the developed model to real-world data, ensuring that the predictions are accurate and reliable.

To build and run the Docker image for the Medical Insurance Prediction project, follow these steps:

1. Build the Docker Image:

   • Navigate to the project directory where the Dockerfile is located.
   • Run the following command to build the Docker image and tag it as insurance:

     docker build -t insurance .

2. Run the Docker Container:

   • After building the image, run the container interactively and remove it after it exits:

     docker run -it --rm insurance

This will start the server and then you can make post request to it.

3. Execute the Jupyter Notebook:
   • Navigate to the notebook named predict_test.ipynb.
   • Open the notebook and run all cells to perform the predictions.