|Autors: Neshov, N. N., Tonchev, K., Velchev, Y. S., Manolova, A. H., Poulkov, V. K.|
Title: SoftVotingSleepNet: Majority Vote of Deep Learning Models for Sleep Stage Classification from Raw Single EEG Channel
Keywords: Deep Learning; Electroencephalography; Sleep Stage Classification; Temporal Data Analysis
Abstract: This work proposes a new neural network architecture named SoftVotingSleepNet for classification of sleep stages based on single EEG channel. This network consists of two individual classifiers, using deep learning algorithms that are trained individually on the same dataset. Their individual decision is combined into single one using Soft voting. Based on the experimental results of comparison to the state of the art, can be concluded that the combination using majority voting mechanism, of the class probabilities predicted by the two relatively simple architectures, achieves similar or better results. Furthermore, it performs better, compared to more complex architectures containing similar components.
Вид: публикация в международен форум, публикация в реферирано издание, индексирана в Scopus и Web of Science