Heart Attack Prediction Using Machine Learning

Project Overview

Heart Attack Prediction Using Machine Learning is an end-to-end healthcare machine learning project designed to estimate the likelihood of heart attack risk from patient medical records. The application combines data preprocessing, machine learning model training, model evaluation and an interactive Streamlit web interface.

Users can enter medical parameters such as age, diabetes status, blood pressure, cholesterol-related information and other health factors. The trained machine learning model processes the information and provides an immediate prediction result.

Available Features

• Heart attack risk prediction
• Patient medical data input
• Real-time prediction
• Streamlit web interface
• Medical dataset processing
• Data preprocessing
• Training and testing data split
• Decision Tree Classifier
• Random Forest Classifier
• Model accuracy evaluation
• Best model selection
• Pickle model serialization
• Saved trained model loading
• Automated browser launch
• Easy-to-use prediction interface

Medical Dataset

The project uses the hf.csv dataset containing medical records and a target variable called DEATH_EVENT. The dataset includes health-related attributes such as age, anaemia, creatinine phosphokinase, diabetes, ejection fraction, high blood pressure, platelet count, serum creatinine, serum sodium, sex, smoking and follow-up time.

Data Preprocessing

The heart.py training script loads the medical dataset and separates the input features from the target variable. The data is then prepared for machine learning and divided into training and testing sets using the train_test_split method.

Machine Learning Models

The project implements multiple classification algorithms for predicting outcomes from the medical dataset.

• Decision Tree Classifier: Builds a tree-based classification model from patient health records.
• Random Forest Classifier: Uses an ensemble of decision trees for classification.

The trained models are evaluated using accuracy scores on unseen testing data. The better-performing model is selected and serialized for use by the web application.

Model Serialization

The selected machine learning model is stored as a Pickle file. The Streamlit application loads the saved model instead of retraining the model whenever a prediction is requested.

Streamlit Web Application

The app.py script provides the interactive web interface using Streamlit. Patients or healthcare professionals can enter the required health parameters through the interface and receive the model's prediction in real time.

Real-Time Prediction Workflow

1. Open the Streamlit application.
2. Enter the required patient health information.
3. Submit the entered medical parameters.
4. The application processes the input values.
5. The saved machine learning model analyzes the processed data.
6. The prediction result is displayed immediately.

Model Evaluation

The project evaluates the implemented classifiers using accuracy scores calculated on the testing dataset. This allows the project to compare the trained models and select the better-performing model for deployment.

Application Components

• hf.csv: Medical dataset used for model development.
• heart.py: Data processing, model training and evaluation script.
• app.py: Streamlit web application.
• hf1.pkl: Serialized trained machine learning model.
• requirements.txt: Python dependency list.

Future Enhancements

• Integrate deep learning models.
• Expand the medical dataset.
• Improve model generalization.
• Add additional data visualization.
• Develop a mobile-friendly healthcare interface.

Technology Stack

• Python
• Streamlit
• Scikit-learn
• Pandas
• NumPy
• Pickle
• Machine Learning

Software and Tools Required

• Python
• Visual Studio Code
• Web Browser
• pip

Important Medical Disclaimer

This project is an educational machine learning prototype based on a specific medical dataset. Its output is a machine learning prediction and must not be treated as a medical diagnosis or a substitute for professional medical advice.

Watch Project Demo

Watch Heart Attack Prediction Using Machine Learning Demo

More Details and Full Documentation

View Heart Attack Prediction Using Machine Learning Details