EEG signals classification based on wavelet packet and ensemble Extreme Learning Machine

To solve the problem of unstable predicted results and poor generalization ability when a single extreme learning machine is treated as a classifier, this paper puts forward a classification algorithm using ensemble Extreme Learning Machine based on linear discriminant analysis. The main idea is applying linear discriminant analysis on each subset of the training samples generated by bootstrapping. By this way, a subset of the larger diversities can be got, which increases the diversity between each machine and reduces the ensemble generalization error and redundant data. Wavelet packet is used to extract features, and the proposed algorithm is used for EEG signal classification. The experiments results with the UCI datasets and another publicly available datasets show that compared with traditional methods and others, the proposed method can significantly improve the classification accuracy and stability, and produce better generalization performance.