Research on Intelligent Detection Technology for Bundle Branch Conduction Block

[摘要] To help doctors diagnose heat disease effectively, an Intelligent detection method of bundle branch conduction block(BBB) heart rhythm is proposed in this paper. Firstly, continuous wavelet transform is used to preprocess experimental data obtained from MIT/BIH arrhythmia database, and then, some features of ECG signals are extracted from the time domain, frequency domain, and wavelet domain. After normalization, these features are fed into the BP neural network optimization algorithm based on genetic algorithm(GABPNN) to train the network. The training error, testing error, and F-1 score are evaluation criterions for classification in this paper. Compared with BP neural network algorithm(BPNN), GABPNN algorithm behaves better in accuracy. The accuracy of 99.502% was achieved, which is better than the results of some literature.

[发布机构] College of Electrical Engineering and Information Technology, Sichuan University, Chengdu, China^1

[效力级别] 计算机科学

[关键词] BP neural networks;Conduction block;Continuous Wavelet Transform;Evaluation criteria;Frequency domains;Intelligent detection;Intelligent detection methods;Training errors