Electrocerebral non-linear dual-measure feature extraction and fusion processing method

The invention discloses an electrocerebral non-linear dual-measure feature extraction and fusion processing method, and relates to the field of electrocerebral signals. The method comprises the following steps: acquiring an electrocerebral signal; performing data preprocessing on the electrocerebral signal; extracting an LZC (Leucocyte Zinc Content) complexity feature from preprocessed data; extracting a sample entropy feature, and performing feature analysis on the LZC complexity feature and the sample entropy feature; establishing a classification model with an SVM (Support Vector Machine) classifier by jointly taking the LZC value of the electrocerebral signal and a sample entropy value which are remarkably different as a feature parameter in order to classify and identify people in different psychological states. By adopting the method for classifying and identifying the electrocerebral signals of people in different psychological states, higher classifying accuracy can be achieved, and an objective evaluation index can be provided for the partition of people in different psychological states.