Electroencephalogram signal analysis based on the improved k-nearest neighbor network

In this paper, we propose a method to analyze epileptic electroencephalogram based on time series that is transformed from improved k-nearest neighbor network. The study of complex networks has become a hot research of electroencephalogram signal. Electroencephalogram time series generated by the network keeps node information of network, so researching the time series from the network can also achieve the purpose of studying epileptic electroencephalogram. The results of this experiment show that studying power spectrum of time series from the network is more easily than the power spectrum of time series directly generated from brain data to distinguish between normal and epileptic patients. In addition, studying the clustering coefficient of improved k-nearest neighbor network is also able to distinguish between normal and patients with epilepsy. This study can provide an important reference for the study of epilepsy and clinical diagnosis.