基于优化模糊神经网络的CPI趋向预测  

Consumer price index's trend forecasting based on optimized fuzzy neural network

作　　者：黄胜忠[1] 黄天开[1,2] 

机构地区：[1]柳州师范高等专科学校数学与计算机科学系,广西柳州545004 [2]广西师范大学计算机科学系,广西桂林541004

出　　处：《计算机工程与应用》2012年第4期176-178,共3页Computer Engineering and Applications

基　　金：新世纪广西高等教改科研工程项目(No.2010JGB135);高校教师教学反思的形式与效果的实践研究(No.LB2009A051)

摘　　要：居民消费价格指数(CPI),是管理层制定居民消费价格政策、工资政策、国民经济发展战略的重要科学依据。已有很多研究学者对CPI进行了大量的研究,并取得了许多成果。但是,过去的预测方法多数是用简单的神经网络方法进行的,很难获得完满的预测结果。针对存在的问题提出了用一种模糊规则自适应的学习算法,可根据具体问题的需要生成不同数目的模糊规则,并根据模糊规则来确定网络结构。在网络结构的确定上,由于不是随机生成,也无需反复实验,所以更具科学性。仿真实验结果表明:用该方法对居民消费价格指数的预测,具有更好的预测效果。The consumer price index is an important basis of consumer price index policy, salary policy, and national economy develop ment strategies. Many researchers have carried out a lot of researches for CPI, and a lot of achievements have been got, however, past predicting method applies the simple neutral network method, and the good predicting results are difficult to be obtained. The paper puts forward a fuzzy rule self adaptive training algorithm according to the existing problems, and the fuzzy rules of different numbers are produced according to the actual problems, and the network structure is confirmed according to fuzzy rule. For the confirmation of the network structure, the test needs not be carried again and again because it can not be formed randomly. Therefore it is responsible. Simulation results show that this method gets the good effect for predicting consumer price index.

关 键 词：消费价格指数 优化模糊神经网络 趋向预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]