LSTM END-TO-END SINGLE-LEAD ELECTROCARDIOGRAM CLASSIFICATION METHOD

摘要

Provided is an LSTM end-to-end single-lead electrocardiogram (ECG) classification method, comprising: S1: preprocessing as experimental data an MIT normal sinus rhythm database and an MIT arrhythmia database; S2: using a deep LSTM and a fully connected layer neural network model to process and classify the preprocessed experimental data. In implementing the ECG classification method, it is unnecessary to excessively process the ECG to ensure the accuracy and completeness of the data, and the ECG data can be accurately classified.