AIML -Volume 9

This paper presents a hybrid data mining approach for knowledge extraction and classification in databases. The aim of this work is two fold. First, we study the performance of Support Vector Machine (SVM) approach in the classification of well known datasets. Second, we present a hybrid classification system based on Particle Swarm Optimization (PSO) to improve the performance of the SVM classifier. The idea is to classify the datasets using SVM with dot and Radial Basic Function (RBF) kernels and optimize it using PSO.          10-fold cross validation is applied in order to validate and evaluate the provided solutions. The performance of SVM classifier with that of PSO-SVM is studied on benchmark datasets of various sizes. The obtained results show that the hybrid PSO-SVM approach finds interesting patterns and provides improved classification performance in terms of accuracy even though the execution time is increased.

 Keywords: Data mining, Support Vector Machines, Particle Swarm Optimization,