Comparative and Analysis Study of normal and epileptic seizure EEC signals by using various classification Algorithms

摘要

Epilepsy is defined as a brain activity disorder which is characterized by epileptic seizures. Electroencephalogram (EEG) signals are one of the most pre ferable and basic ways o f diagnosing epileptic seizures due to their practicality and simplicity. On the other hand, interpreting those signals is not an easy task, because of the non-linear and variable signal properties. In this paper -we present a data mining classification approach by applying machine learning algorithms to detect normal and epileptic seizure from EEG brain signals, we are using t-SNE t-Distributed Stochastic Neighbor Embedding Algorithm for preprocessing as adimensionality reduction algorithm on the dataset then we applied three algorithms on the original dataset and on the preprocessed dataset to classify normal and epileptic seizure, and evaluate the performance of these three different classifiers (Support Vector Machine, K nearest Neighbor and Random Forest), so that, the classifier with the best performance is selected to implement such system. Thus, classifiers with high accuracy than other classifiers, proposed in previous studies, are evaluated in this study, so that, better predictions accuracies are provided by these classifiers. As a result, the best classification precision was obtained at 98.52%. when the Random Forest was used.