基于ARIMA-SVM组合模型的我国农产品价格预测研究  

Study on the Agricultural Prices Prediction in China Based on ARIMA and SVM Hybrid Model

作　　者：陈兆荣[1] 雷勋平[1,2] 王亮[1] 叶松[1] 

机构地区：[1]铜陵学院农村经济与文化研究所,安徽铜陵244000 [2]南京航空航天大学经济与管理学院,江苏南京210016

出　　处：《财经理论研究》2013年第2期103-107,共5页Journal of Finance and Economics Theory

基　　金：教育部人文社科青年基金资助项目(12YJC630082);安徽省高校省级自然科学研究项目(KJ2013Z329);安徽省高校省级自然科学研究项目(KJ2013Z329);安徽省教育厅人文社科重点研究基地项目(SK2012A183)

摘　　要：为准确把握国内农产品价格波动规律,提高农产品价格预测精度,构建农产品价格自回归移动平均与支持向量机(ARIMA-SVM)组合预测模型,以ARIMA模型揭示农产品价格线性变动规律,以SVM模型揭示非线性变动规律,并结合1999-2011年我国农产品价格指数月度数据,使用组合模型和ARIMA、SVM单个模型对农产品价格进行预测。预测结果显示:组合模型比单个ARIMA、SVM模型预测精度高,能够提高农产品价格预测的准确性,是一种有效的农产品价格预测模型。In order to grasp the fluctuation of the domestic agricultural price and to raise the accuracy of price prediction. A new hybrid forecasting model based on ARIMA and SVM was proposed, ARIMA model was used to predict the linear component of the agricultural prices and SVM model was applied to the nonlinear residual compo- nent. The prediction performances of the single models and hybrid model were tested based on the data of agricul- tural prices in 1999 -2011. The results indicate that the hybrid model has better accuracy than the single models. The hybrid model which can raise the accuracy of price prediction is an effective method for the agricultural prices.

关 键 词：农产品价格 组合预测 ARIMA-SVM组合模型 

分 类 号：F304.2[经济管理—产业经济]