Master's Degree Disstertation Project
The aim of this research is to develop a highly accurate deep learning model based on bagging techniques for 12-lead ECG classification, specifically targeting four types of rhythms, to improve diagnostic efficiency and accessibility in cardiac health monitoring.
The underlying hypothesis is that ECG classification can be enhanced by implementing a bagging-based ensemble approach applied to various deep learning models.
The models’ performance will be assessed for both:
- Binary Classification on the PTB dataset for two classes: Normal and Abnormal
- Multiclass Classification on the MITBIH dataset for five classes: Normal Sinus Rhythm, Supraventricular
Ectopic Beat, Ventricular Ectopic Beat, Fusion Beat, and a category for unknown beats.
#
2024.
Technologies:
- Python
- Jupyter Notebook/Google Colab
- Tensorflow/SKlearn/Keras