文献类型：期刊文献

英文题名：An unsupervised transfer learning model based on joint information for sleep stage classification

作者：Li, Ruichen[1]; Wei, Huan[1]; Wang, Bei[1]

机构：[1] School of Information Science and Technology, East China University of Science and Technology, Department of Automation, Shanghai, 200237, China

年份：2022

外文期刊名：Proceedings - 2022 15th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics, CISP-BMEI 2022

收录：EI(收录号：20230313390351)

语种：英文

外文关键词：Bioelectric phenomena - Biomedical signal processing - Classification (of information) - Electroencephalography - Learning systems

摘要：Automatic sleep staging methods took an important role to improve the efficiency for sleep stage scoring. The performance of most methods were based on the assumption of consistent data distribution. However, The actual recorded neuro-physiological signals have individual differences and the recording conditions are also different. In this study, an unsupervised transfer learning model based on joint information is proposed for automatic sleep stage classification. The aim is to avoid the affect of data distribution difference and improve the classification performance. Several time and frequency features are extracted from electroencephalograms and electrooculogram. Features with contextual time information on the source and target domains are obtained through bidirectional long short-term memory network. An unsupervised transfer learning model was developed and implemented by a cascaded sleep stage classification scheme. The sleep recordings from the Sleep-EDFx database, including both of normal subjects and patients with sleep disorders, were adopted for testing. Comparing with the baseline models, the classification accuracy achieved 83.7% improved by 5% which demonstrate the effectiveness of the developed unsupervised transfer learning model with joint information for sleep staging. ? 2022 IEEE.