Abstract
Epilepsy is a neurological condition that affects the central nervous system. While its effects are different for each person, they mostly include abnormal behaviour, periods of loss of awareness and seizures. There are various traditional methods used to analyse EEG signals for epilepsy detection, which concludes to be time-consuming. Recently, several automated seizure detection frameworks using machine learning algorithms have been proposed to replace conventional methods. In this paper, more emphasis has been given to develop SPPCA and SUBXPCA dimensionality reduction algorithms to increase the classification accuracy of various machine learning models. Firstly, Discrete Wavelet Transform (DWT) is applied to EEG signals for extracting the time-frequency domain features of epileptic seizures such as the energy of each sub-pattern, spike rhythmicity, Relative Spike Amplitude (RSA), Dominant Frequency (DF) and Spectral Entropy (SE). The features obtained after performing DWT on an EEG signal are extensive in number, to select the prominent features and to retain their properties, correlation feature sub-pattern-based PCA (SPPCA), and cross sub-pattern correlation-based PCA (SUBXPCA) are used as a dimensionality reduction techniques. To validate the proposed work, performance evaluation parameter such as the accuracy of the time-frequency domain features from different combinations of the dataset has been compared with the latest state-of-the-art works. Simulation results show that the best accuracy of 97% is achieved for SPPCA algorithm by CatBoost classifier. And the best accuracy of 98% for SUBXPCA is achieved by random forest classifier, which clearly outperformed the other related works both in terms of accuracy and computational complexity.