Detection of epileptic seizures on EEG signals using Decision Tree, KNN, SVM and ensemble classifiers

Abstract

Epilepsy is a neurological condition resulting to brain cell stimulation. According to the findings of the research, an electroencephalogram (EEG) can identify epileptic episodes in patients who have epilepsy. Performance evaluations based on EEG detection of epilepsy require feature extraction methods. As a result of our research, we were able to identify a number of different feature extraction approaches. These methodologies were dependent upon nonlinear, wavelet-based entropy characteristics, time - frequency domain characteristics, and a few statistical traits. An additional in-depth examination that made use of cutting-edge machine learning classifiers as well as numerous factors was carried out. When evaluating kernels for support vector machines, the multiclass kernel as well as the box constraint levels are both useful tools. In a similar manner, we computed the various distance measures, neighbour weights, and neighbour relationships for the K-nearest neighbours (KNN). In a similar manner, we altered the decision trees of the paramours depending on maximal splits as well as split criteria, and we evaluated the ensemble classifiers based on a variety of ensemble approaches and learning rates.