﻿ 组合时间序列模型在江西省农业总产值预测中的应用 Application of the Combination Time Series Model in Forecasting the Gross Agriculture Output of Jiangxi Province

Statistical and Application
Vol.04 No.04(2015), Article ID:16535,7 pages
10.12677/SA.2015.44025

Application of the Combination Time Series Model in Forecasting the Gross Agriculture Output of Jiangxi Province

Xiaofa Tian

School of Mathematics and Statistics, Yunnan University of Finance and Economics, Kunming Yunnan

Received: Nov. 24th, 2015; accepted: Dec. 13th, 2015; published: Dec. 16th, 2015

ABSTRACT

This article selects Jiangxi province’s 1979-2014 agricultural output [1] , using the Jiangxi agricultural gross output value of 1979-2009 data, respectively to set up the GM(1,1) model, the exponential curve model and mixed time series model. The data from 2010-2012 were used for the model test, and then on the basis of the three models, the combination of time series model was established to predict agricultural output of the Jiangxi province in recent years. Combined model prediction with a single model predicted results comparison shows that the results error of combination forecasting model is superior to the three model predictions respectively, and has more advantages in prediction of time series data. The prediction of 2015 Jiangxi province’s agricultural output was 3385.07 billion Yuan, but it totally reached 3800 billion Yuan in 2015. For the wide gap with the expection, we should increase investment in agriculture.

Keywords:GM(1,1) Model, Exponent Model, Mixed-Time Series Model, Combination Prediction Model, Gross Agriculture Output

1. 引言

2. 组合时间序列模型

“组合预测”思想于1969年首次提出，近年来引起我国学者的重视，其应用范围也逐渐扩大。组合预测理论的基本原理是：通过个体预测值的加权算术平均而得到其组合预测值，在确定加权权重时，以组合预测误差方差最小为原则。

GM(1,1)灰色预测模型要求负荷数据少、不考虑分布规律、不考虑变化趋势、运算方便、短期预测精度高。指数曲线趋势外推法简单方便，适用于趋势明显的时间序列的预测。ARIMA模型较为复杂，适用于数据随机性较强不易提取确定性因素的时间序列。结合这三种时间序列模型的优势，构建组合模型，可以提高预测精度。

,

(1)

3. 实证分析

3.1. GM(1,1)预测模型

, ,

(2)

GM(1,1)模型检验

3.2. 指数曲线趋势外推法

3.3. 混合时间序列模型

3.3.1. 建立回归模型

3.3.2. 混合时间序列模型

3.4. 组合模型

Figure 1. After taking logarithm of agricultural output linear regression model test batches

Figure 2. Hybrid model parametric test

Figure 3. Hybrid model residual error sequence of autocorrelation function and partial correlation function test

Table 1. Three kinds of forecasting methods of prediction results

Table 2. Jiangxi agricultural output combination forecast result table

4. 结论

Application of the Combination Time Series Model in Forecasting the Gross Agriculture Output of Jiangxi Province[J]. 统计学与应用, 2015, 04(04): 226-232. http://dx.doi.org/10.12677/SA.2015.44025

1. 1. 中华人民共和国国家统计局. 中国统计年鉴[M]. 北京: 中国统计出版社, 2013.

2. 2. 陈显周, 区晶莹, 俞守华, 杨春. 组合时间序列模型及其在我国农业总产值预测中的应用[J]. 广东农业科学, 2012(13): 194-196.

3. 3. 陈华友. 组合预测方法有效性理论及其应用[M]. 北京: 科学出版社, 2008.

4. 4. 徐国祥. 统计预测与决策[M]. 上海: 上海财经大学出版社, 2008.

5. 5. 何晓群. 实用回归分析[M]. 北京: 高等教育出版社, 2008.

6. 6. 王燕. 应用时间序列分析[M]. 北京: 中国人民大学出版社, 2008.