Open AccessSeizure Detection using Deep Multiset Canonical Correlation Analysis and Bayesian Optimization
Author(s) -
Xuefeng Bai,
Lijun Yan
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1748/2/022034
Subject(s) - multiset , computer science , artificial intelligence , bayesian optimization , bayesian probability , pattern recognition (psychology) , canonical correlation , feature (linguistics) , deep learning , epileptic seizure , electroencephalography , feature extraction , set (abstract data type) , focus (optics) , machine learning , bayesian network , mathematics , psychology , neuroscience , physics , combinatorics , optics , programming language , linguistics , philosophy
As one of the most common neurological diseases in the world, epilepsy seizure is difficult to ignore. Seizure detection is receiving more and more attention from researchers. Feature extraction is one of the key steps in automatic seizure detection. Lots of features have been proposed to detect seizure using EEG signal. However, few works focus on feature fusion. In this paper, deep multi set CCA is explored for seizure detection. Since deep neural network architecture has a great impact on performance of deep multiset CCA, bayesian optimization is employed to search architecture parameters automatically. Preliminary experiments show it is effective for seizure detection using deep multiset CCA and bayesian optimization. Satisfactory seizure classification results are achieved with little manual intervention.