Extract Features Using Gaussian Filter and Its Application for Improvement of an Accuracy of the Epileptic Detection Using SVM  
Author Ha Manh Hoang

 

Co-Author(s) Quang Tung Ha Hoang; Ngoc Giau Ho

 

Abstract In diagnostic of Epilepsy, an important issues is to infer epileptic seizure presence. Until recently, the SVM is the main tool for the epileptic seizure recognition purpose. The accuracy of this approach base on features extraction ability of the pre-processing. An important issue is to extract a phenomena that correspond to the thorn of epileptic. This paper introduces method for improving the robustness of classification problems using one dimensional Gaussian filter. The experimental results show that our method has advantage in case of complicated epileptic seizure.

 

Keywords EEG, WT, ANN, SVM
   
    Article #:  DSBFI19-79
 
Proceedings of ISSAT International Conference on Data Science in Business, Finance and Industry
July 3-5, 2019 - Da Nang, Vietnam