NovaGen Health Classification

A supervised machine learning project developed for NovaGen Research Labs to classify individuals as healthy or unhealthy based on clinical and lifestyle health indicators.

Dataset

• 9,549 patient records with 22 features
• Features include physiological measurements (BMI, Blood Pressure, Cholesterol, Glucose Level), lifestyle factors (Smoking, Alcohol, Exercise Hours, Sleep Hours), and encoded categorical variables (Diet Type, Blood Group)
• Target variable: 0 = Healthy, 1 = Unhealthy

Project Structure

NovaGen/
├── novagen_dataset.csv              # Dataset
├── novaGen.ipynb # Main ML pipeline
├── eda_overview.png                 # Exploratory data analysis charts
├── model_comparison.png             # Model performance comparison
└── best_model_analysis.png          # Feature importance and confusion matrix

Pipeline Overview

Step	Description
1	Load and inspect dataset
2	Exploratory Data Analysis (EDA)
3	Preprocessing and train/test split
4	Train 6 classification models
5	Compare models across key metrics
6	Hyperparameter tuning on best model
7	Feature importance and confusion matrix
8	Final performance summary

Models Trained

• Logistic Regression
• Decision Tree
• Random Forest
• Gradient Boosting
• K-Nearest Neighbors (KNN)
• Support Vector Machine (SVM)

Results

Model	Accuracy	F1 Score	AUC-ROC
Logistic Regression	0.8136	0.8224	0.8879
Decision Tree	0.8597	0.8665	0.9229
Random Forest	0.9366	0.9402	0.9845
Gradient Boosting	0.9199	0.9248	0.9721
KNN	0.8901	0.8947	0.9485
SVM	0.9335	0.9371	0.9776

Best Model: Random Forest (after hyperparameter tuning — max_depth=20, n_estimators=200)

• Accuracy: 0.94
• F1 Score: 0.94
• AUC-ROC: 0.9845

Requirements

numpy
pandas
matplotlib
seaborn
scikit-learn

Install with:

pip install numpy pandas matplotlib seaborn scikit-learn

Usage

python novaGen.ipynb

Ensure novagen_dataset.csv is in the same directory before running.