基于VC的时间序列分析法的应用研究  

Application Research of Time Series Analysis Method Based on VC

作　　者：李洪雷[1] 管群[1] 

机构地区：[1]四川大学计算机学院,四川成都610065

出　　处：《计算机仿真》2009年第7期113-117,共5页Computer Simulation

摘　　要：时间序列分析是概率统计学的一个重要组成部分,现在从经济到工程技术,为了得到时间序列分析预测,对时间序列分析、预测、比较都具有重要的理论意义和实际应用价值。选取两种有代表性的数据,利用VisualC++实现了三种时间序列预测方法:移动平均法、指数平滑法和最小二乘法,并通过图表以及相关的误差分析,详细地比较了三种方法的特点和适用范围,从而既可以从直观上了解三种方法的共同点以及不同点,又可以从误差分析的数据中了解各种方法适合于分析预测哪种类型的数据。Time series analysis forecasting is an important part of Statistics ＆ Probability, Now, from economy to engineering, time series analysis forecasting is met in almost all kinds of fields. To compare and analyze the various forecasting and analysis methods in time series analysis has important theoretical and practical value. Two kinds of representative data were selected. Three kinds of time series forecasting methods ： the method of moving average, the method of exponential smoothness and the method of least squares were realized by using Visual C ＋ ＋. The charac- teristic and the application scope of three kinds of methods were compared in detail with charts and the related errors analysis. Thus not only the common points and the different points of three methods could be understood intuitively, but also various methods suitable for analyzing and forecasting corresponding type of data could be known from the data of error analysis.

关 键 词：时间序列 预测 移动平均法 指数平滑法 最小二乘法 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]