Feature Extraction and Classification for Graphical Representations of Data

摘要

The barycentre graphical feature extraction method of the star plot is proposed, based on the graphical representation of multi-dimensional data. Because for the different feature order the same multi-dimensional data lead to the different star plots, and extract the different barycentre graphical features, which affect the classification error of the classifiers. The novel feature order method based on the improved genetic algorithm (GA) is proposed. Meanwhile the traditional feature order method based on the feature selection is researched and the traditional vector feature extraction methods are researched. The experiments results of the 4 real data sets show the classification effectiveness of the new graphical representation and graphical features.