Document Type : Research Paper
Authors
1
Department of Electronics and Communication Engineering, Vaageswari College of Engineering, Karimnagar, Telangana, India
2
Department of Electronics and Communication Engineering, Jyothishmathi Institute of Technology and Science, Karimnagar, Telangana, India.
3
Department of Electrical and Electronics Engineering, Kaveri University, Gouraram, Siddipet, Telangana, India.
4
Department of Computer Science and Engineering, New Horizon College of Engineering, Bengaluru, Karnataka, India.
5
Sr Cybersecurity Engineer at VISA, Inc, Foster, San Mateo County 94404, CA, USA.
6
Department of Computer Science and Engineering, Koneru Lakshmaiah Education and Foundation, Vaddeswaram, Guntur (Dt), AP, India.
7
Faculty of Business, Middlesex University Dubai, Dubai, United Arab Emirates.
10.30772/qjes.2026.166722.1797
Abstract
For Brain Computer Interface (BCI) applications, this research suggests an ideal feature selection framework for Electroencephalography (EEG)-based emotion recognition. In order to improve signal quality, EEG signals are first preprocessed using an adaptive notch filter to reduce power-line interference and artifacts. Then, in order to capture discriminative emotional patterns, time–frequency features, frequency domain and time-domain, are extracted. A Modified Artificial Bee Colony algorithm with inertia weight (MABC-IW) is used to choose the best subset of features that minimizes dimensionality and maximizes classification performance in order to handle feature redundancy and overfitting. A Support Vector Machine (SVM) is employed to categorize the chosen features. The suggested MABC-IW–SVM framework works better than traditional feature selection techniques, according to experimental assessment carried out in MATLAB with classification accuracy of 94%, precision of 91%, recall of 100 % and F1-score of 95%. The suggested method reduces the feature set while increasing accuracy when compared to baseline SVM without optimization. These findings demonstrate that the suggested optimum feature selection approach greatly improves the ability to recognize emotions from EEG signals.
Keywords
Subjects
