Chronic Heart Failure Detection

This repository contains datasets and Python scripts using machine learning libraries to detect Chronic Heart Failure (CHF) and classify heart sounds as normal or abnormal.

---

📂 Contents

• data/ — Datasets with clinical and physiological indicators related to heart failure.
• heart_sounds/ — Audio recordings and extracted features from phonocardiograms (PCGs).
• models/ — Python scripts implementing ML and DL models using popular libraries.
• notebooks/ — Jupyter notebooks for training, evaluation, and visualization.
• README.md — Project documentation (this file).

---

🎯 Objectives

• Detect Chronic Heart Failure using clinical and demographic data.
• Classify heart sounds using signal processing and deep learning.
• Provide a pipeline for healthcare data analysis and model experimentation.

---

🧠 Models Implemented (via Python Libraries)

The models are built using code, not pre-trained files. Libraries used include:

• Scikit-learn for:

  • Logistic Regression
  • Decision Trees
  • Random Forest
  • Support Vector Machines (SVM)

• TensorFlow / Keras or PyTorch for:

  • Convolutional Neural Networks (CNNs) for heart sound classification
  • LSTM-based models for time series audio signals
  • Hybrid models using MFCC + deep learning

---

📈 Sample Results

Task	Best Algorithm	Accuracy
Heart Failure Detection	Random Forest	91%
Heart Sound Detection	CNN + MFCC	94%

(Performance may vary depending on data preprocessing and hyperparameters.)

---

🛠 Dependencies

pip install scikit-learn numpy pandas librosa matplotlib seaborn tensorflow