An Improved Particle Swarm Optimization Algorithm Based K-means Clustering Analysis

摘要

The improved Particle Swarm Optimization (PSO) algorithm is proposed to optimize the K-means clustering analysis problem in this study. The improved PSO algorithm aimed to prevent the convergence premature and to get more quickly convergence velocity in the optimization clustering process. Due to the clustering characteristics of the K-means based PSO, a new adjustable searching strategy of PSO is presented firstly. Subsequently, a fitness function is denned to guide the searching procedure to enhance the accuhave been undertaken with a test bench of synthetic and real life datasets. The results reflect the superiority of the proposed algorithm in terms of accuracy, convergence speed and robustness.