Particle Swarm Optimization for Data Classification

A novel data classification method, Particle Swarm Optimization for classification (PSOC), was put forward. After pretreating the data, the classification rules were discovered by particle swarm optimization algorithm based on the training samples, and then the data was classified by the discovered rules. The convergence of the new algorithm was analyzed based on Bayes’s theorem and stochastic transform process. Experimental study on UCI machine learning repository and the remote sensing image data shows the proposed algorithm obtains good performances.