Models In Cotton Production Forecast In The Comparative Study

This study adopts the national bureau of statistics national cotton area, the total yield and 60 years data, Six major cotton producing provinces of cotton output 16 years data, and the Chinese academy of agricultural sciences cotton institute survey data. First use of Excel to the national 60 years of planting area, yield and cotton output data standardization and makes an overall analysis; The second according to survey data, do V8.0 software application SAS data processing analysis, a revenue and output value model; The third national cotton production and Six major cotton producing provinces do time series analysis. The fourth respectively by comparative analysis and unitary regression model and the multiple regression model, ARIMA model and the combination model of yield predicted advantages and accuracy, to the national cotton output, Eight major cotton producing provinces, and predicted to total yield and actual production trend analysis of data, aiming to dig deeper into the internal connection of implicit deposit, provided the scientific basis for scientific cotton. The main research results are as follows:1, Farmers value modelWith the total cost of production equation, cotton output, seed cotton price impact factor composition equation. SAS analysis results:the value of the Prob> F for less than 0.0001, belongs to the very significant, and fitting degree (R-for 0.9896 Square) that can use this equation is multivariate regression model and results of and significant level. Equation is Y=-1895+0.0004763*x1+5.12993*x2+3700.9362*x3.2, The national cotton output forecast model(1) ARIMA times series, the test of residual sequence of this model can draw the result that Chi-Square all greater than 0.05, indicates sequence data was made full use, by testing the residual sequence, it reaches to a significant level, making use of this forecast the total production from 2010 to 2014, the total cotton production from 2010 to 2014 respectively is 7.42,6.60,7.10,7.05 and 8.12 million tons.(2) Unary regression yield prediction model. The result of SAS shows regression equation is very manifest, the parameters less than 0.0001, reach at a significant level, we can get the model:Tt=21.87025+10.16873t, putt=61,62,63,64,65 into this equation, the result of the forecast of China' total cotton production is6.42,6.52,6.62,67.63 and 6.83 million tons.(3) Regression model, it employs the equation:total production equals to unite production multiply area, establish regression equation:Y=aX1*X2+c (X1 means unit price, X2 means area, a means adjustment coefficient, c means constant), namely regression model:Y= 1.01285X1*X2-3.36129,R-Square=0.9973, Pr> F less than 0.0001, therefore, it can be concluded that it has manifest meaning, the predicting result is 6.77,6.78,6.80,6.82,6.84 million tons.(4) Combination model, using weighted formula, calculated the weight of production function model, exponential smoothing model and multivariate linear regression model, put them into combination model equation, the combination predict model is Y=0.364y1+0.041y2+0.595y3 (y1,y2,y3 are ARIMA model predict result respectively unary regression model prediction result, multiple regression model result, prediction result is 6.99,6.70,6.90,6.90,7.30 million tones)(5) The main conclusion:the next five years in total national cotton trend between 6.42～6.89 million tons, with an average 6.76 million tons, and for the average of the last five years slightly reduced or unbiased 6.95 million tons. Judging from four model, four kinds of model than on a five-year reduction respectively; combination forecast model, reduce the rate of 2.7%;ARIMA model, educe the rate of 3.4%; The multiple regression,reduce the rate of 2.2%; A dollar regression model output to reduce the rate of 4.6%.3, Making use of ARIMA model predicts the cotton production of the major province which product cotton.(1) The forecast equation of cotton production in Xinjiang is:ΔX3t = 330.85595+0.60203Xt-1+(?)t+0.75124(?)t-1+0.2483(?)t-2,(?)t～WN(0,σ2(?))(2)The forecast equation of cotton production in ShanDong is: Xt= 38.24298+0.20002Xt-1-0.26736Xt-2+0.53262Xt-3+(?)t+0.50082(?)t-1, (?)t WN(0,σ2(?))(3) The forecast equation of cotton production in HeBei is (ΔX2(?)=53.26862+0.39796Xt-1+(?)t+0.28533(?)t-1,(?)t～WN(0,σ2(?)).(4)The forecast equation of cotton production in HeNan is Xt=688.50018+0.04804Xt-1-0.17018Xt-2-0.11307Xt-2,(?)t WN(0,σ2(?)).(5)The forecast equation of cotton production in HuBei is Xt=4.20328+0.46822Xt-1+0.53178Xt-2+(?)t-0.09005(?)t-1,(?)t～WN(0,σ2(?)).(6)The forecast equation of cotton production in Anhui is Xt=205.3145+(?)t-0.59361(?)t-1-0.63806(?)t-2+0.37404(?)t-3,(?)t～WN(0,σ2(?)).(7)The forecast equation of cotton production in HuNan is Xt=0.6335-0.755.1Xt-1-0.4004Xt-2,(?)t～WN(0,σ2(?))(8)The forecast equation of cotton production in Jiangsu is Xt=41.63724-0.83734(?)t-0.24745(?)t-1-0.30283(?)t-2+0.68303(?)t-3,(?)t～WN(0,σ2(?))The next few years,Predict the results shov that xinjiang,shandong,hebei,hubei Anhui and Jiangsu cotton output shows ascendant trend,which rose faster is shandong, xinjiang and hebei steadily rising,with a slow speed,hubei and Anhui rise the slowest, Jiangsu cotton output has certain fluctuations；Hunan and Henan decreased.