An AI model using various machine learning algorithms like KNN and SVM for detecting Left Bundle Branch Block (LBBB) in ECG signals. The project includes data preprocessing, feature extraction, model training, evaluation, and deployment through a GUI application.
The data preprocessing step involves:
- Removing noise from the ECG signals using a Butterworth bandpass filter.
- Normalizing the signals to a standard range to ensure consistent feature extraction.
Features are extracted from the preprocessed signals using wavelet transforms.
Statistical features such as:
- Mean
- Standard deviation
- Skewness
- Kurtosis
are calculated from the wavelet coefficients.
Various machine learning models are trained and evaluated on the extracted features, including:
- K-Nearest Neighbors (KNN)
- Support Vector Machine (SVM)
- Random Forest
- Decision Tree
- Naive Bayes
The best model is selected based on accuracy and other evaluation metrics.
The best-performing model (KNN) is deployed using a GUI application.
This application allows users to:
- Input ECG signals
- Receive a classification result: Normal or LBBB
| Model Name | Train Accuracy | Test Accuracy |
|---|---|---|
| K-Nearest Neighbors | 99% | 100% |
| Naive Bayes | 100% | 53% |
| Support Vector Machine (SVM) | 100% | 47% |
| Decision Tree | 100% | 35% |
| Random Forest | 100% | 35% |
