基于混合模型的模糊时间序列预测的研究  

Research on Fuzzy Time Series Prediction Based on Hybrid Model

作　　者：秦田田[1] 王涛[1] 郭小蕾[1] 范秋枫 

机构地区：[1]辽宁工业大学理学院,辽宁锦州121001

出　　处：《辽宁工业大学学报（自然科学版）》2015年第3期200-205,共6页Journal of Liaoning University of Technology(Natural Science Edition)

基　　金：辽宁省高人才资助项目(2012921055)

摘　　要：将移动自回归模型,减法聚类,神经网络,模糊逻辑系统相结合,设计基于模糊时间序列预测的混合系统模型,基于所设计的混合模型研究居民消费价格指数(CPI)的模糊时间序列预测问题,应用BP算法调整系统参数,并给出仿真研究。仿真结果表明,所设计的混合模型应用于CPI的模糊时间序列预测是可行的与有效的。Autoregressive integrated moving average model, subtractive clustering, neural network, and fuzzy logic system were combined, a kind of hybrid system model based on time series prediction was designed, fuzzy time series prediction problem of CPI was studied based on the hybrid model and BP algorithm was applied to adjusting the system parameters, and the simulation research was given. The simulation results show that the design of the hybrid model which was applied to fuzzy time series prediction of CPI is feasible and effective.

关 键 词：时间序列 移动自回归模型 模糊逻辑系统 神经网络 BP算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]