A Machine Learning-Based Big EEG Data Artifact Detection and Wavelet-Based Removal: An Empirical Approach. (7th October 2021)

• Record Type: Journal Article
• Authors: Stalin, Shalini; Roy, Vandana; Shukla, Prashant Kumar; Zaguia, Atef; Khan, Mohammad Monirujjaman; Shukla, Piyush Kumar; Jain, Anurag
• Other Names: Pereira A. M. Bastos Academic Editor.
• Is Part Of: Mathematical problems in engineering. Volume 2021(2021)
• Publication Date: 2021-10-07
• Subjects: Engineering mathematics -- Periodicals; 510.2462
• DOI: 10.1155/2021/2942808
• Languages: English
• ISSNs: 1024-123X

Abstract

Abstract : The electroencephalogram (EEG) signals are a big data which are frequently corrupted by motion artifacts. As human neural diseases, diagnosis and analysis need a robust neurological signal. Consequently, the EEG artifacts' eradication is a vital step. In this research paper, the primary motion artifact is detected from a single-channel EEG signal using support vector machine (SVM) and preceded with further artifacts' suppression. The signal features' abstraction and further detection are done through ensemble empirical mode decomposition (EEMD) algorithm. Moreover, canonical correlation analysis (CCA) filtering approach is applied for motion artifact removal. Finally, leftover motion artifacts' unpredictability is removed by applying wavelet transform (WT) algorithm. Finally, results are optimized by using Harris hawks optimization (HHO) algorithm. The results of the assessment confirm that the algorithm recommended is superior to the algorithms currently in use.