两种时间序列模型在客流量预测上的比较  

Comparison between two kinds of time series models for forecasting passenger flow

作　　者：刘建军[1,2] 廖闻剑 彭艳兵 

机构地区：[1]武汉邮电科学研究院通信与信息系统系,武汉430074 [2]南京烽火星空通信发展有限公司研发部,南京210019

出　　处：《计算机工程与应用》2016年第9期228-232,270,共6页Computer Engineering and Applications

基　　金：江苏省科技支撑计划(No.BE2011173)

摘　　要：模糊时间序列模型和季节模型都是基于时间序列的模型,为了探讨在时间序列表现出一定的周期性时,哪种模型的预测效果会更好,分别利用模糊时间序列模型和季节模型对南京某商场的客流量进行预测,计算并比较两种方法下的相对误差值和RMSE(Root Mean Square Error)值,发现季节模型的相对误差值图形的平滑度要优于模糊时间序列模型,季节模型的RMSE值小于模糊时间序列模型,这表明考虑到数据特征的模型有更好的预测结果。Fuzzy time series model and seasonal model are both based on time series. In order to investigate which one is better in forecasting when time series is periodic, forecasting the passenger flow of a shopping mall in Nanjing uses fuzzy time series model and seasonal model respectively. By comparing the value of relative error and the value of RMSE after computing them which are produced by the two different methods, it finds out that the relative error graphics of seasonal model is smoother than fuzzy time series model＇s. And the value of RMSE of seasonal model is also smaller than fuzzy time series model＇s. It indicates that it will get a better prediction when the character of data is concerned.

关 键 词：模糊时间序列模型 季节模型 相对误差 均方根误差 

分 类 号：TP273.4[自动化与计算机技术—检测技术与自动化装置]