贝叶斯动态线性模型(BDLM)在水产品价格监测中的应用  

The Application of the Bayes' Dynamic Liner Model in the Price Forecast of Fishery

作　　者：方平[1] 宋瑞凤[1] 

机构地区：[1]华南农业大学理学院,广东广州510642

出　　处：《江西科学》2009年第1期50-54,共5页Jiangxi Science

基　　金：国家自然基金项目(10726070);华南农业大学校长基金资助项目(4900-K07277)

摘　　要：应用简单差分方程模型、ARMA模型和BDLM分析了广州市水产品草鱼的价格数据的时间序列特征。借助Matlab软件进行编程,对上述3种模型进行分析比较和拟合优度检验。结果表明,对于水产品草鱼的价格预测,BDLM模型较ARMA模型更适合。此外,还得出广州市草鱼价格自2003年以来不仅随季节和供求发生变化,还存在明显上涨趋势的结论。Applying the simple difference equation model, the ARMA model and the bayes＇dynamic linear models （BDLM） ,the time series characterizations about the fish＇s price data of Guangzhou are analyzed. By means of the programming language MATLAB, with considering and proceeding a test of goodness of fit for three models above,we find that the fish＇s price of Guangzhou not only changes with the seasons and the relation of supply and demand,but also has the trend of rising in price evidently.

关 键 词：BDLM ARMA 先验信息 预测误差 拟合优度检验 

分 类 号：O213.9[理学—概率论与数理统计]