AUTHOR=Prabhakar Sunil Kumar , Won Dong-Ok 

TITLE=Performance comparison of bio-inspired and learning-based clustering analysis with machine learning techniques for classification of EEG signals

JOURNAL=Frontiers in Artificial Intelligence

VOLUME=Volume 6 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2023.1156269

DOI=10.3389/frai.2023.1156269

ISSN=2624-8212

ABSTRACT=A comprehensive analysis of an automated system for epileptic seizure detection is explained in this work. When the seizure occurs, it is quite hectic to differentiate the non-stationary patterns from the discharges occurring in a rhythmic manner. The proposed approach deals with it efficiently by clustering it initially for the sake of feature extraction by using six different techniques categorized under two different methods such as Bio-inspired clustering and Learning Based clustering. Learning based clustering includes K-Means clusters and Fuzzy C-Means (FCM) clusters while bio-inspired clusters include Cuckoo search clusters, Dragonfly clusters, Firefly clusters and Modified Firefly clusters. The clustered values are then classified with ten suitable classifiers and after the performance comparison analysis of the EEG time series, the results prove that this flow of methodology has achieved a good performance index and a high classification accuracy. A comparatively higher classification accuracy of 99.48% and a Performance Index (PI) of 98.93% is achieved if Cuckoo search clusters are utilized with Linear Support Vector Machine (SVM) for epilepsy detection. A high classification accuracy of 98.96% and a performance index of 97.87% is obtained when K-Means Clusters are classified with Naïve Bayesian Classifier (NBC) and Linear SVM, and similar results are also obtained when FCM clusters are classified with Decision Trees yielding the same values. The comparatively least classification accuracy of 75.5% is obtained when Dragonfly clusters are classified with K-Nearest Neighbour (KNN) classifier and a second least classification accuracy of 75.75% is obtained when Firefly clusters are classified with NBC.