Title: Implemented OBL-DE assisted Tasmanian devil optimisation for selecting the optimal features using EEG signal for stress detection

Authors: Dipali Nilesh Dhake; Yogesh Suresh Angal

Addresses: JSPM's Rajarshi Shahu College of Engineering, Tathawade, Pune, Maharashtra 411033, India; Pimpri Chinchwad College of Engineering and Research, Ravet, Pune, Maharashtra, India ' Department of Electronics and Telecommunication Engineering, JSPM's Bhivarabai Sawant Institute of Technology and Research, Wagholi, Pune – 412207, India

Abstract: This article introduces a stress detection framework called opposition learning differential evaluation assisted Tasmanian devil optimisation (OBLDETO)-based hybrid classifier for stress detection (OBLDETO-HC) using EEG signal. The model includes four stages like: 1) pre-processing; 2) feature extraction; 3) feature selection; 4) classification. The source EEG signal is subjected to pre-process. Further, from the pre-processed signal. Features are extracted in terms of Stockwell transform, proposed common spatial pattern and DWT features. From the extracted features, optimal features will be chosen. For optimal feature selection, this paper introduces an OBL-DE-assisted Tasmanian devil optimisation (OBLDETO) model. These chosen features will be provided as the source of the detection phase. The classification process takes place via hybrid classification that combines a gated recurrent unit (GRU) and an improved deep belief network (IDBN). Here, the final classification result will be determined by the improved score level fusion.

Keywords: deep belief network; gated recurrent unit; GRU; OBLDETO.

DOI: 10.1504/IJAHUC.2024.142712

International Journal of Ad Hoc and Ubiquitous Computing, 2024 Vol.47 No.4, pp.240 - 257

Received: 03 Apr 2024
Accepted: 30 May 2024
Published online: 18 Nov 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article