基于指数平滑法与回归分析相结合的GDP预测

Forecasting of GDP Based on Exponential Smoothing and Regression Analysis

指数平滑法与回归分析分别是对移动平均法和相关分析的继续及深化,为实际生活中对时间序列历史近期数据短期的预测提供了依据。从指数平滑法与回归分析相关理论出发构建两者的预测模型,以指数平滑法的初始平滑值与平滑常数的筛选为准则,并以此为基础,通过二次指数平滑法对时间序列历史近期数据进行分析,再运用回归分析进行再次预测。与以往的不同之处在于运用了二者的算术平均值作为最终短期预测的目标,来弥补两种预测可能出现的误差。最后,运用两者相结合的方法,根据2015年已有的统计数据,对2017年我国的GDP进行预测。通过实例,验证该方法在进行数据短中期预测时的可行性与适用性。

Exponential smoothing and regression analysis respectively is the continuation and deepening of moving average and correlation analysis, for real life history of recent data for time series short-term prediction to provide the basis.This article embarks from the related theories of exponential smoothing method and regression analysis, in which prediction model was established with exponential smoothing initial value and the selection of smoothing coefficient for the principle.On the basis, through the secondary exponential smoothing analyze the time sequence of recent history data, and then use regression analysis to predict again.In this paper, the difference is to use the arith- metic mean of both short-term prediction as the final goal, to compensate the occurrence for the two kinds of forecast error.Finally, the article uses the method of the combination, according to the existing statistical data in 2015, to predict China＇s GDP in 2017.Through exampies, the method is verified in the middle of the short data to predict the feasibility and applicability.