In order to improve the tradition GA's searching ability in regression problems,enhance the algorithm in convergence rate and precision,propose GEPAdaBoost algorithm. Based on the frame of AdaBoost,the GEPAdaBoost takes GEP as a weak-learner for every iteration by using its power ability of symbolic regression. Then the new fitness function of every iteration can produce good hypothesis based on weight computing and distributing. Finally the algorithm will get optimization result by using voting strategy in multi-hypothe-sis. Experiments show that the new algorithm is more accurate than the traditional GEP algorithms by 16. 7% and GPBoosting algorithms by 40. 8%.
国家科技期刊平台
NSTL
万方数据
维普
