LIVE PROJECT · 2025

Heart Disease
Prediction via
Neural Networks

An ANN-based clinical decision support system that analyzes 13 patient health parameters to detect heart disease with high accuracy — enabling faster, data-driven diagnosis.

Explore Architecture → View Pipeline
13
Input Features
3
Hidden Layers
100
Max Epochs
5
Team Members
// Neural Network

Model Architecture

A deep feedforward ANN with ReLU-activated hidden layers and a sigmoid output for binary classification of heart disease presence.

INPUT
13 neurons
HIDDEN 1
32 neurons · ReLU
HIDDEN 2
16 neurons · ReLU
HIDDEN 3
8 neurons · ReLU
OUTPUT
1 neuron · Sigmoid
LOSS FUNCTION
Binary Cross-Entropy
OPTIMIZER
Adam
BATCH SIZE
32
EPOCHS
50 – 100
TRAIN / TEST SPLIT
80% / 20%
CLASSIFICATION
Binary (0 / 1)
// UCI Heart Disease Dataset

13 Input Features

Each patient record contains these clinically validated health indicators, preprocessed with normalization and encoding before feeding into the network.

01Age
02Sex
03Chest Pain Type
04Resting Blood Pressure
05Serum Cholesterol
06Fasting Blood Sugar
07Resting ECG Results
08Max Heart Rate
09Exercise-Induced Angina
10Old Peak (ST Depression)
11ST Segment Slope
12Major Vessels (Fluoroscopy)
13Thalassemia
// Evaluation Metrics
ACCURACY
🎯
PRECISION
📐
RECALL
🔍
F1-SCORE
⚖️
AUC-ROC
📈
// Project Workflow

Development Pipeline

01
Data Preprocessing
Handle missing values, one-hot encode categoricals, MinMax/Standard normalize continuous variables.
02
Model Training
Train ANN with Adam optimizer and binary cross-entropy loss over 50–100 epochs.
03
MLOps Tracking
Log experiments, hyperparameters, and runs with MLflow for reproducibility.
04
Evaluation
Analyze confusion matrix, AUC-ROC curve, and all classification metrics.
05
Deployment
Serve via Flask/Streamlit web interface with real-time prediction and monitoring.
// Expected Deliverables
🧠
Trained ANN Model
High-accuracy neural network for binary heart disease classification.
📊
Evaluation Report
Metrics, visualizations, feature analysis, and performance insights.
🌐
Web Interface
Simple frontend for clinicians to input data and receive predictions.
🎤
Final Presentation
Summary of findings, model performance, and deployment results.
// CoreMind Group

The Team

E/20/453
TEAM MEMBER
E/20/158
TEAM MEMBER
E/20/300
TEAM MEMBER
E/20/248
TEAM MEMBER
E/20/377
TEAM MEMBER