Chi-Square Discretization Algorithm of Continuous Attribute Based on Information Entropy
Download as PDF
Wei Yanming, Gan Xusheng, Li Huaping
Rough Set (RS) theory has been widely concerned and studied in the related fields of uncertain information processing with its unique ability of data reduction, and the discretization of continuous attributes is an important link in RS method and other induction learning system. For this reason, a Chi-square (Chi2) Discretization Algorithm Based on Information Entropy is proposed. In the algorithm, the information entropy is used to replace the inconsistency rate in the traditional Chi2 algorithm, and the combination algorithm based on RS and wavelet neural network is established. The simulation result shows that the proposed algorithm is effective and superior using UCI data set.
Wavelet Neural Network, Chi2 Algorithm, Attribute Reduction, Discretization.