Heart Disease Prediction Project

Heart Disease Prediction Project

Project Overview

This project aims to predict heart disease stages using a machine learning model deployed through a Flask web application. It involves data analysis, model development, and a user-friendly interface for inputting patient data and viewing predictions.

Project Steps

1. Data Analysis

Objective: Analyze and clean the UCI Heart Disease dataset, handling missing values efficiently.

Missing Values: Instead of deleting rows (which would reduce the dataset from 920 to 299 rows), imputation with median and mode was used.

EDA (Exploratory Data Analysis):

Used functions like describe(), info(), and value_counts() to understand data distribution.

Checked unique values in each column and explored feature relationships using plots.

2. Model Development

Model Creation: Developed a predictive model with steps including encoding, scaling, and data splitting.

Performance and Saving: The model's performance was evaluated, and the following were saved:

Prediction model: heart_disease_model.keras

Encoders: heart_disease_encoder.pkl and heart_disease_OneHotEncoder.pkl

Standard Scaler: heart_disease_standard_scaler.pkl

Integration: Provided examples to ensure input data compatibility with the Flask app.

3. Flask Application

Purpose: Collect inputs from users, process them, and use the model for predictions.

Input Consistency: The app processes inputs using the same steps as the model development phase.

4. Templates

index.html: A basic HTML template with input fields matching model features, a prediction button, and a section to display results.

Saved Files

heart_disease_encoder.pkl - Saved Label Encoder.

heart_disease_model.keras - Saved Prediction Model.

heart_disease_OneHotEncoder.pkl - Saved OneHot Encoder.

heart_disease_standard_scaler.pkl - Saved Standard Scaler.

heart_disease_uci.csv - Dataset in CSV format.

Screenshots

Input Page (0.png): Displays how users enter feature values.

Prediction Result (1(1).png): Shows JSON output containing:

Predicted Category: The category number with the highest probability.

Predicted Label: The meaning of the predicted category.

Prediction Probabilities: Probabilities for each class in the prediction.

How to Run the Project

1. Set Up the Conda Environment

Create and activate a conda environment with the required Python version:

conda create --name heart_disease_env python=3.8

conda activate heart_disease_env

2. Install Required Libraries

Create a requirements.txt file with the following content:

TensorFlow==2.13.0

Keras==2.13.1

Scikit-learn==1.3.2

Pandas==2.0.3

Matplotlib==3.7.5

Seaborn==0.13.2

Imbalanced-learn==0.12.3

Joblib==1.4.2

NumPy==1.24.3

Flask==3.0.3

Install the libraries using:

pip install -r requirements.txt

3. Run the Application

Run the Flask app using:

python app.py

Ensure that the directory contains:

Model: heart_disease_model.keras

Label Encoder: heart_disease_encoder.pkl

OneHot Encoder: heart_disease_OneHotEncoder.pkl

Standard Scaler: heart_disease_standard_scaler.pkl

Include a templates folder with the index.html file.