Deep learning approach for active classification of electrocardiogram signals

详细信息    查看全文

• 作者：M.M. Al Rahhal^a ; Yakoub Bazi^a ; ^{yakoub.bazi@gmail.com" class="auth_mail" title="E-mail the corresponding author} ; ^{ybazi@ksu.edu.sa" class="auth_mail" title="E-mail the corresponding author} ; Haikel AlHichri^a ; ^{hhichri@ksu.edu.sa" class="auth_mail" title="E-mail the corresponding author} ; Naif Alajlan^a ; ^{najlan@ksu.edu.sa" class="auth_mail" title="E-mail the corresponding author} ; Farid Melgani^b ; ^{melgani@disi.unitn.it" class="auth_mail" title="E-mail the corresponding author} ; R.R. Yager^c ; ¹ ; ^{yager@panix.com" class="auth_mail" title="E-mail the corresponding author}
• 关键词：ECG signal classification ; Feature learning ; Denoising autoencoder (DAE) ; Deep neural network (DNN) ; Active learning (AL)
• 刊名：Information Sciences
• 出版年：2016
• 出版时间：1 June 2016
• 年：2016
• 卷：345
• 期：Complete
• 页码：340-354
• 全文大小：1451 K

文摘

In this paper, we propose a novel approach based on deep learning for active classification of electrocardiogram (ECG) signals. To this end, we learn a suitable feature representation from the raw ECG data in an unsupervised way using stacked denoising autoencoders (SDAEs) with sparsity constraint. After this feature learning phase, we add a softmax regression layer on the top of the resulting hidden representation layer yielding the so-called deep neural network (DNN). During the interaction phase, we allow the expert at each iteration to label the most relevant and uncertain ECG beats in the test record, which are then used for updating the DNN weights. As ranking criteria, the method relies on the DNN posterior probabilities to associate confidence measures such as entropy and Breaking-Ties (BT) to each test beat in the ECG record under analysis. In the experiments, we validate the method on the well-known MIT-BIH arrhythmia database as well as two other databases called INCART, and SVDB, respectively. Furthermore, we follow the recommendations of the Association for the Advancement of Medical Instrumentation (AAMI) for class labeling and results presentation. The results obtained show that the newly proposed approach provides significant accuracy improvements with less expert interaction and faster online retraining compared to state-of-the-art methods.