摘要
利用逐级逼近思想建立起一种基于DE-SVR的逼近算法。通过运用该算法对居民消费价格指数(CPI)进行预测的实验证实,该算法有着很高的预测精度,是一种进行回归拟合的新算法。
According to the theory of gradual approach,an approximation algorithm based on the differential evolution algorithm(DE) and support vector machine for regression(SVR) was proposed and was applied to CPI forecast.The experimental results show that the new algorithm has obviously superior forecasting accuracy in regression.
出处
《科技通报》
北大核心
2012年第12期39-41,共3页
Bulletin of Science and Technology
关键词
逼近算法
微分进化算法
支持向量机
居民消费价格指数
approximation algorithm
differential evolution algorithm(DE)
support vector machine(SVM)
consumer price index(CPI)