Electroencephalogram Signals Classification Based on Feature Normalization

Abstract

Data standardization is a fundamental process in which binary or multi- classification systems incorporate it as a sub-system in the classification- based question. Standardization can be called a mapping function moving from one space to another. Depending on the quality of the data, various types of normalization methods have been suggested. Research is underway recently on whether this approach is actually necessary. In this article, the various standardization methods efficiency is measured for the purpose of categorizing signal-based emotion with EEG. Binary classifier based on Naïve Bayes Classifier to classify the emotions. Only various kernel functions are considered for Naïve Bayes Classifier. While the experimental results may not show a substantial difference in performance between varrious types of normalization, the process of normalization generally improves emotion recognition classification efficiency.