Economic Forecast For The Research And Application Of Statistical Models

Contents of this dissertation are a portion of question for study that my supervisor rofessor Teng suzhen is researching .The project mainly studies the Input-Output analysis and methods which can optimize products of the Polyester Complex of Liaoyang Petroleum Chemical Fiber Co.. The main results obtained from this dissertation may be submitted as follows: LThis paper extends the principal component estimate method from one-to-many to many-to-many in regression analysis, and discusses some properties of the principal component estimate, and obtains the similar results as one-to-one principal component estimate. 2.It makes use of the new method to prove the stationary domain of AR(2)-model. 3.The principal component estimate method is used to analyze the real data of the Polyester Complex, which shows that mean square error of principal component estimate is smaller that one of the least square estimate. 4.This dissertation obtains some satisfying results by making use of time series method to predict sales volume of short silk and long silk about Dacron in the Polyester Complex. 5.Combined efficiently the time series analysis and regression analysis, Input-Output analysis about the device Of short silk and one of long silk is constructed, what's more, input volume of raw material in a few of future months is estimated and predicted, which provides scientific foundation for plan produce of companies. 6.The correlation relation model polyester melt and polyester section, the correlation relation model polyester melt, polyester section and long silk, the correlation relation model polyester melt and short silk and time series predicting model of short silk and long silk are constructed.