Submit Manuscript  

Article Details

Implementation of Bagged SVM Ensemble Model for Classification of Epileptic States Using EEG

[ Vol. 20 , Issue. 9 ]


Arshpreet Kaur*, Karan Verma, Amol P. Bhondekar and Kumar Shashvat   Pages 755 - 765 ( 11 )


Background and Objective: To decipher EEG (Electroencephalography) intending to locate inter-ictal and ictal discharges for supporting the diagnoses of epilepsy and locating the seizure focus is a critical task. The aim of this work is to find how the ensemble model distinguishes between two different set of problem which are group 1: inter-ictal and ictal, group 2: controlled and inter-ictal using approximate entropy as a parameter.

Method: The work is divided to conquer two problem groups the first being classifying inter ictal and ictal for which case 1(C-E), case 2(D-E) are considered which is Group 1. The other group (Group 2) is classifying between inter-ictal and controlled for which case 3 (A-C), case 4(B-C), case 5 (A-D) and case 6(B-D) are considered. To divide the EEG into sub-bands DWT (Discrete Wavelet Transform) is used and approximate Entropy is extracted out of all the five sub-bands of EEG for each case. Bagged SVM is used to classify between different groups considered.

Results: Highest accuracy for Group 1 using Bagged SVM Ensemble model for case 1 is 96.83% with testing data; which is similar to 97% achieved by using training data. For case 2 (D-E) 93.92% with training and 84.83% with testing data has been obtained. For Group 2 there is a large disparity between SVM and Bagged Ensemble model where with the latter 76%,81.66%,72.835% and 71.16% for case 3 , case 4 , case 5 and case 6 has been obtained. While for training data set 92.87%, 91.74 %, 92% and 92.64% are attained respectively. The results obtained by SVM for Group 2 were at huge difference from the highest accuracy achieved by bagged SVM for both training and the testing data.

Conclusion: Bagged Ensemble model has outperformed SVM model for every case with huge difference with both training as well as testing dataset for Group 2 and marginally better for Group 1.


EEG classification, approximate entropy, discrete wavelet transform, bagged SVM, ensemble model, epileptic states.


National Institute of Technology, New Delhi, Delhi, National Institute of Technology, New Delhi, Delhi, Central Scientific Instruments Organization, Chandigarh, National Institute of Technology, New Delhi, Delhi

Read Full-Text article