Prediction Of Raw Coal Production Based On Seasonal Adjustment Model

Because of the fact that raw coal has occupied the dominant position of China's energy structure,which also has been the basic industry of national economy,it is extremely important to study the dynamic evolution of raw coal production.The basic theory of time series is analyzed deeply in this dissertation,which includes the modeling,identification,parameter estimation,diagnosis and prediction of random sequence.Traditional grey model and seasonal adjustment model are respectively set up to predict coal production in China,after which compared result is given.Firstly,the longitudinal sequence is divided into three parts:trend,season and random according to the scatter plot by seasonal adjustment method,after which the corresponding model for each of these three parts should be established.The establishment of the trend model can be performed by the principle of least squares,after which prediction of the trend part can be directly predicted.Then the establishment of multiplication model and separation of season part can be gained after trend part is removed and the average of actual seasonal index,which is called the seasonal factor,can be calculated by adjusted season order.For the separation of the last part,firstly this dissertation verifies whether the random part remaining after removing trend part and season part is stable.If the noise sequence is not stable,relevant differential processing must be performed until the sequence becomes stable.By analyzing the autocorrelation and partial autocorrelation of the differential sequence,four models are respectively set up in this dissertation,Arma(1,1),Arma(2,1),Arma(3,1)and Arma(4,1),to simulate the noise sequence.The second model is proved to be the best,which can be used to predict the raw coal production,and the trend part is multiplied by the seasonal factor and pluses the random part,after which it is found that the prediction of coal production whose errors are-0.93%and 0.53%is finally obtained,28919.55 and 29473.43.Finally,this dissertation concludes with a contrastive analysis,which shows that the error of the seasonal adjustment model is lower than grey model,and compared with the grey method,seasonal adjustment model can be improved effectively.Therefore,the seasonal adjustment model can be used as a favorable tool of raw coal production analysis and prediction,which can also provide services for policy decision-making and management of national non-renewable energy in the future.