Explainable ECG Classification Using Deep Learning and Large Language Models

Short Description

This paper presents a system for multi-label arrhythmia classification that predicts five diagnostic superclasses: Normal, Myocardial Infarction, ST/T-Wave Changes, Conduction Disturbance, and Hypertrophy. The proposed ap-proach uses a hybrid deep learning model combining a CNN-Transformer-MLP architecture with a retrieval augmented generation (RAG) module, which pro-duces clinically grounded explanations for the predicted labels.

Keywords

Countries

Romania