Study On Forecasting And Early Warning Of Cotton Production In Hebei Province

Hebei province is one of four major cotton-growing areas including Hebei, Shandong, Henan and Xinjiang, of which cotton planting and spinning industries not only are two economic mainstays of the province, but also play an important role in the nation. The situation of cotton production straightly influences cotton farmers' life and the development of agricultural production and national economy. During recent years, as affected by factors with polices, climate and international market and so on, cotton production in Hebei has fluctuated frequently, which has brought disadvantageous influence to farmers, sellers and the national finance. Therefore, the build-up of a scientific and reasonable system of forecasting and early-warning cotton production in order to take corresponding measures in time and abate its fluctuation is of great significance to the smooth development of cotton industry in Hebei province.According to theories of economic fluctuation, forecasting and early warning and based on the research of cotton production history of Hebei, this paper firstly employs trend-decomposition method to analyze fluctuations of total yield, unit yield and planting areas of cotton and gets their own cv—coefficient of variation respectively, in which the cv of total yield is the highest and planting areas and unit yield in order. Meanwhile, by regression analysis method, it analyzes the degree that unit yield and planting area influence total yield, which shows that after cotton price being decided by market in 1999, the difference between unit yield and planting influencing total yield has appeared and the former is obviously higher than the latter. Secondly, after analyzing the fluctuation of cotton, the paper uses grey model, autoregressive moving-average model and their compositive model to fit yearly total yield between 1990 and 2006, and moreover forecast yearly total yield between 2007 and 2010. Their accurate degrees of forecasting are ranked from high to low as follows: compositive model, autoregressive moving-average model and grey model. Finally, on the basis of theories about agricultural economy and cotton production, the paper chooses 37 indicators affecting cotton production and classifies them by time-difference correlation analysis, on which it establishes early-warning models of signal lights and partial least-squares regression from different angles, estimates situations of cotton production in next several years combining with forecasted results and puts forward corresponding Countermeasures. Furthermore, the paper develops early-warning software of cotton production based on the foundation of the early-warning models and explains its functions, working flow and the whole designing thought at length.