A Hybrid Approach to Continuous Valued Datasets Classifying based on Particle Swarm Optimization, Variable Precision Rough Set Theory and Modified Huang-index Function

This paper proposed a new hybrid method, designated as PSOVPRS-index method, for partitioning and classifying continuous valued datasets based on particle swarm optimization (PSO) algorithm, Variable Precision Rough Set (VPRS) theory and a modified form of the Huang-index function. In contrast to the Huang-based index method which simply assigns a constant number of clusters to each attribute and in which the Rough Set (RS) theory is applied, this method could not only cluster the values of the individual attributes within the dataset and achieves both the optimal number of clusters and the optimal classification accuracy, but also extends the applicability of classification using VPRS theory. The validity of the proposed approach is investigated by comparing the classification results obtained for a real-world dataset containing stock market information with those obtained by PSORS-index method and pseudo-supervised decision-tree classification method. There is good evidence to show that the proposed PSOVPRS-index method not only has a better classification performance than the considered methods, but also achieves a more reliable basis for the extraction of decision-making rules. Key-Words: Particle Swarm Optimization, Variable Precision Rough Set theory, PSOVPRS-index method, classification, discretization, pseudo-supervised classification method