组合预测模型在畜产品价格预测中的应用研究

Application of Combination Forecasting Model on Price Forecast of Animal By-products

频繁波动的畜产品价格已经成为影响我国居民消费价格指数的重要因素之一,如何准确预测价格走势以及时调整生产受到广泛关注。选取Holt-Winter季节乘积模型和ARIMA模型对2002—2015年北京市生猪、肉鸡、鸡蛋3种畜产品的市场周价进行实证研究,建立组合预测模型对价格进行拟合预测。模型评价结果显示,组合模型的预测精度最高,ARIMA模型次之,Holt-Winter季节乘积模型最差。利用组合预测模型对2015年7月之后5周的价格进行预测,结果显示组合预测模型在畜产品短期预测当中具有较好的准确性和可行性。

Frequent fluctuation of animal by-products price has become one of the important factors affecting China＇s consumer price index,and how to accurately predict price movements and adjust production has been widely concerned. Empirical research on Beijing＇s animal by-products＇ including pig,poultry,and eggs weekly market price from 2002 to 2015 was conducted by adopting Holt-Winter Multiplicative Model and ARIMA Model. Based on these two single prediction models,the Combination Forecasting Model was set up to fit and predict the price. The results of the model evaluation indicated that the level of each model＇s prediction accuracy ranked in a descending order from the Combination Forecasting Model to ARIMA model,then to Holt- Winter Multiplicative Model. The Combination Forecasting Model showed higher accuracy and feasibility in the short-term forecasting of the price of animal by-products,which was used to predict the market price of animal by-products in the subsequent five weeks after July 2015.