Research on Fast Clustering Algorithm Based on Improved Particle Swarm Optimization

Traditional clustering algorithm is sensitive to the initial center, and the convergence result is easy to fall into local optimum, and the execution efficiency is low. In order to solve the problems, a fast clustering algorithm based on improved particle swarm optimization is proposed. In this algorithm, the sample data set is implemented with the clustering division based on high density and threshold value analysis. The particle swarm initial particle position is generated. The weight value mapping coefficient and information entropy of the particle is calculated. The particle weight value is adjusted, and the self adaptive degree value of each particle is calculated. The local extreme and global extreme of particle are updated. Finally, the iterative clustering is taken based on the particle position and velocity update mechanism. The clustering center is optimized. Simulation result shows that this algorithm has good operation efficiency, the convergence speed is remarkable, and the cluster precision is improved greatly.