Short-Term Price Analysis And Forecasting Methods Selection Of Agricultural Products

The fluctuation of farm produce price concerns the stable development of agriculture and farm produce's effective supply. It's the basis of government's decision-making, and it has a relation with the increase of farmers'income. The frequent variation of farm produce price has been becoming the main obstacle of increase of both of agricultural yield and farmer's income. With the deeply advance of globalization, the world farm produce price has become the inevitable factor influcing the price on our national market. However, the study of farm produce price in china most focus on present situation rather than the price formulation mechanism; and has more research concerning the long-term trend than short-term forecasting. Under this situation, it's siganificante to study the short-term farm produce price. First, it will strengthen the government's regulating capability. In addition, it will be helpful for farmers to adjust production with sound judgement to the market signal. In the meanwhile, the law of price fluctuation is one of the key reaserch fields.The purposes of this paper lie in comparing the current forecasting theories and methods, exploring modeling of farm produce short-term price arcording the period length, verifying the model and making improvement to it.This paper will modele and forecast the Fuji apple price on Beijing Xin Fadi wholesome market in accordance with the different characteristics of daily, weekly, ten-days', and monthly price. ARIMA (1,0,1) and GARCH (1,1) model will be applied to fit and forecast the dayly price from Jan, 1st 2004 to Mar, 31st 2009; ARIMA (3,1,3), (3,1,14) model and the VAR model are both used to fit and forecast the weekly price from 1st 2004to Mar 26th 2009; ARIMA (13,1,13), (2,1,13), (3, 1,13) model will be used to predict ten-days price; WLS, ARIMA (1,1,10), (10,1,10), and VAR models are all applied to fit and forecast the month price from Jan,2006 to Mar, 2009.In sum, all the models applied are suitable for the price research about particular variety and specific market.Through quantitative research and qualitative analysis, the paper draws following conclusions: GARCH model is fit for the day price fitting and forecast; ARIMA model is superior to VAR model for weekly price fitting and forecast; ARIMA model is better from the fitting aspects, while the WLS model is proper from the predicting sides.