Effective and Extensible Feature Extraction Method Using Genetic Algorithm-Based Frequency-Domain Feature Search for Epileptic EEG Multi-classification

摘要

In this paper, a genetic algorithm-based frequency-domain feature search(GAFDS) method is proposed for the electroencephalogram (EEG) analysis ofepilepsy. In this method, frequency-domain features are first searched and thencombined with nonlinear features. Subsequently, these features are selected andoptimized to classify EEG signals. The extracted features are analyzedexperimentally. The features extracted by GAFDS show remarkable independence,and they are superior to the nonlinear features in terms of the ratio ofinter-class distance and intra-class distance. Moreover, the proposed featuresearch method can additionally search for features of instantaneous frequencyin a signal after Hilbert transformation. The classification results achievedusing these features are reasonable, thus, GAFDS exhibits good extensibility.Multiple classic classifiers (i.e., $k$-nearest neighbor, linear discriminantanalysis, decision tree, AdaBoost, multilayer perceptron, and Na\"ive Bayes)achieve good results by using the features generated by GAFDS method and theoptimized selection. Specifically, the accuracies for the two-classificationand three-classification problems may reach up to 99% and 97%, respectively.Results of several cross-validation experiments illustrate that GAFDS iseffective in feature extraction for EEG classification. Therefore, the proposedfeature selection and optimization model can improve classification accuracy.