Analysing Epileptic EEG Signals Based on Improved Transition Network

Abstract

The epileptic automatic detection was very significance in clinical. The nonlinear time series analysis method based on complex network theory provided a new perspective understand the dynamics of nonlinear time series. In this paper, we proposed a new epileptic seizure detection method based on statistical properties of improved transition network. First, we improved the transition network and electroencephalogram (EEG) signal was constructed into the improved transition network. Then, based on the statistical characteristics of improved transition network, the mathematical expectation of node distribution in a network was extracted as the classification feature. Finally, the performance of the algorithm was evaluated by classifying the epileptic EEG dataset. Experimental results showed that the classification accuracy of proposed algorithm is 97%.