Improved Immune Programming Algorithm And Its Application In The Model Clustering

In order to overcome the traditional clustering algorithms' defects of slow convergence and easily falling into local optimal solution, an improved immune clustering algorithm was presented. In this paper, chaos optimization mechanism was introduced into immune programming algorithm. By this means the improved algorithm can not only achieve global search performance and adaptive capacity, but also improve the ability of local search. The improved algorithm was applied in multiple model modeling to carry out the function of data clustering, sub-space division, and ultimately generate model set. Simulation results show that compared with traditional clustering algorithm, the improved algorithm has better accuracy and shorter calculating time.