Param302 / Heart-Disease-predictor

Heart Disease Predictor

A Heart Disease predictor GUI made in Python using Neural Network ML model to predict the probability of Heart Disease.

Project details

Working

It takes $8$ parameters and do prediction based on them

• Age
• Sex (M / F)
• Chest Pain (4 types)
• Chest Pain during exercise (yes / no)
• trestbps - Resting blood pressure
• Maximum Heart Rate
• Blood sugar above 120mg/dl (yes / no)
• Cholesterol

Model details

Model made using Neural Networks with accuracy of $98.26%$ consists of total 3 layers

• $1^{st}$ hidden layer $18$ neurons (relu)
• $2^{nd}$ hidden layer $9$ neurons (relu)
• $3^{rd}$ output layer $1$ neuron (linear)

Used BinaryCrossentropy() loss function with from_logits=True, and convert predictions into sigmoid using tf.nn.sigmoid(preds) and Adam optimizer

Clone

git clone https://github.com/Param302/Heart-Disease-predictor.git

Software used

Python 3.10.2

Setup before run

• Make sure to install all required libraries and module before using it:

pip install -r requirements.txt

• Run main.py file:

python main.py