Combined neural network electrocardiogram classification method and system

The invention relates to a combined neural network electrocardiogram classification method. The method comprises the steps that original electrocardiogram data are acquired and filtered; intercepting the data into sample data according to the timestamps and recording corresponding labels; sorting the sample data into a training set and a verification set; sequentially sending the training set into an Inception module and a one-dimensional convolutional neural network module for learning, wherein the Inception module comprises a plurality of one-dimensional convolutional neural networks connected in parallel; sequentially sending the data output by learning into an LSTM module and a Softmax function for learning and outputting categories; and verifying the trained and learned model by using the verification set to obtain a label prediction value, and obtaining the accuracy of the classification model. According to the invention, an Inception module is applied in front of the one-dimensional convolutional neura