🫀 CardioCare: Heart Failure Risk Prediction AI

📋 Abstract

Cardiovascular diseases (CVDs) are the leading cause of death globally. Early stratification of high-risk patients is critical for effective intervention.

CardioCare is a machine learning-based Clinical Decision Support System (CDSS) designed to predict mortality events in heart failure patients. This project implements a Comparative Kernel Analysis using Support Vector Machines (SVM), evaluating the efficacy of Linear, Polynomial, and Radial Basis Function (RBF) kernels on clinical data.

🔴 Launch Clinical Dashboard

🔬 Scientific Methodology

The model was developed using the UCI Heart Failure Clinical Records dataset (N=299).

1. Preprocessing & Feature Engineering

• Standardization: Applied StandardScaler to normalize features (e.g., Platelets vs. Serum Creatinine), which is mathematically critical for SVM distance hyperplanes.
• Leakage Prevention: The time variable (follow-up period) was rigorously excluded during training, as it constitutes target leakage in a predictive setting.

2. Algorithmic Approach

• Algorithm: Support Vector Classifier (SVC).
• Class Balancing: Implemented class_weight='balanced' to penalize false negatives, addressing the dataset's imbalance (Death events are the minority class).
• Hyperparameter Tuning: GridSearchCV was used to optimize $C$ (Regularization) and $\gamma$ (Kernel Coefficient).

3. Kernel Comparison Results

Kernel	Accuracy	Precision	Recall (Death)
RBF	~77%	High	Balanced
Linear	~74%	Moderate	Low
Poly	~72%	Low	Low

Conclusion: The RBF Kernel demonstrated superior capacity to model the non-linear decision boundaries of physiological data.

📊 Key Predictive Features

The model identified the following biomarkers as most significant for mortality risk:

1. Serum Creatinine: Indicator of renal function.
2. Ejection Fraction: Percentage of blood leaving the heart.
3. Age: Advanced age correlates with higher risk.

💻 Installation & Usage

Prerequisites: Python 3.8+

# 1. Clone the repository
git clone [https://github.com/Muhammad-Shahan/CardioCare-Heart-Failure-Prediction.git](https://github.com/Muhammad-Shahan/CardioCare-Heart-Failure-Prediction.git)

# 2. Install dependencies
pip install -r requirements.txt

# 3. Launch the Application
streamlit run app.py