摘要本文采用灰色模型和ARIMA时间序列模型对杭州市的GDP进行预测。文章完整的介绍了这两个模型的运算过程，分别使用了R语言和SAS软件进行编程拟合。灰色模型部分，需要先对原始GDP序列进行累加，生成均值序列后通过最小二乘法估计得出系数，代入模型后得到预测序列。使用ARIMA模型时，要先对原始GDP序列做差分计算，通过自相关系数和偏自相关系数检验定阶得到最恰当的拟合模型，随后对模型中的参数进行检验，最终得到一个稳定的白噪声序列，最终得到拟合模型代入数据得出相应的预测结果。两种模型各有优缺点，灰色模型模型可在样本较小的情况下进行预测，但计算过程较为复杂；ARIMA模型对样本数量要求较高但拟合效果更好。运用这两个模型得出的各自的预测值，与实际数值比对分析，得出在本次GDP预测中，ARIMA模型的预测结果相对更准确。88309

    In this paper, grey model and ARIMA time series model are used to forecast the GDP of Hangzhou。We introduce the operation process of the two models, and uses the R language and SAS software to fit the program。 In the part of grey model, we need to accumulate the original GDP sequence firstly, then generate the mean sequence,estimate the coefficient by the least square method。After the model is selected, the prediction sequence is obtained。When using ARIMA model, firstly the differential operation of the original GDP sequence is needed, then we have to test the correlation coefficient and partial correlation coefficient in order to find the most appropriate mode, next test the parameters in the model, finally we can obtain a stable white noise sequence, finally we can forecast the corresponding results。The two models have their own advantages and disadvantages。 The grey model can be used in the case of small sample size, but it's calculation process is more complex。 The ARIMA model requires a higher number of samples, but the result is much better。In this paper, we use the two models to obtain the respective prediction values。 Compared with the actual numerical analysis, we can conclude that in this text, the ARIMA model is more accurate in the prediction of GDP。

毕业论文关键词：灰色模型；ARIMA模型；GDP预测； 时间序列；

Keyword: Grey Model； ARIMA Model;  Prediction of GDP；Time Series;

目录

一、引言 4

二、杭州市GDP的预测 5

2。1模型选取 5

2。1。1灰色模型概述 5

2。1。2 ARIMA模型概述 5

2。2数据收集 6

2。3灰色模型的预测过程 6

2。3。1灰色模型原始时间序列 6

2。3。2灰色模型源-于,优W尔Y论L文.网wwW.youeRw.com 原文+QQ75201,8766生成累加数据 6

2。3。3灰色模型生成均值序列 7

2。3。4灰色预测模型 9

2。3。5灰色模型数据检验 9

2。4 ARIMA模型的预测过程 12

2。4。1 ARIMA模型时间序列 12

2。4。2 ARIMA模型差分 12

2。4。3 ARIMA模型拟合 14

2。4。4 ARIMA模型检验 15

2。4。5 ARIMA模型预测 15

2。5两种模型的比较