Analysis of EEG Based Emotion Detection of DEAP and SEED-IV Databases using SVM
Thejaswini S1, K M Ravikumar2, Jhenkar L3, Aditya Natraj4, Abhay K K5
1Mrs. Thejaswini S, Assistant Professor, Department of Telecommunication Engineering, B M S Institute of Technology and Management, Bengaluru (Karnataka), India.
2Dr. K M Ravikumar, Principal and Professor, Department of Electronics & Communication Engineering, S J C Institute of Technology, Kothanoor (Karnataka), India.
3Jhenkar L, Student, Department of Telecommunication Engineering, B M S Institute of Technology and Management, Bengaluru (Karnataka), India.
4Abhay K K, Student, Department of Telecommunication Engineering, B M S Institute of Technology and Management, Bengaluru (Karnataka), India.
5Aditya Nataraj, Student, Department of Telecommunication Engineering, B M S Institute of Technology and Management, Bengaluru (Karnataka), India.
Manuscript received on 22 May 2019 | Revised Manuscript received on 12 June 2019 | Manuscript Published on 27 June 2019 | PP: 207-211 | Volume-8 Issue-1C May 2019 | Retrieval Number: A10360581C19/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The Affective computing is one of the fast-growing areas which has inspired research in the field of emotion detection for many applications. This paper briefs out the related work on EEG based emotion detection using publicly available data and a proposed method to detect inner emotion-states. A supervised machine learning algorithm is developed to recognize human inner emotion states in two-dimensional model. The electroencephalography signals from DEAP and SEED-IV database are considered for emotion detection. Discrete Wavelet Transforms are applied on preprocessed signals to extract the desired 5 frequency bands. Some features like Power, energy, differential entropy and time domain are extracted. Channel wise SVM classifier is developed and channel combiner is done to detect the appropriate emotion state. The classification rate for four classes are 74%, 86% ,72% & 84% for DEAP database and 79%, 76%,77% & 74% for SEED-IV database.
Keywords: BCI, DEAP, DWT, EEG, SEED-IV, SVM.
Scope of the Article: Predictive Analysis