Grey Multivariate Linear Regression and Application in Search Engine

The classical multivariate linear regression contains two shortages: it cannot run after response variables,the abnormal data which the multivariate linear regression produces affect the simulation.To cope with the problem,based on the time sequence characteristic of the gray system,a new model,the grey multivariate linear regression model,which combines the gray system and multivariate linear regression analysis model is presented.The result turns out to be that the new model can make up the shortages of the classical multivariate linear regression analysis model.First,it can exactly present the expected value of the response variable and the change trend of the response variable.At the same time,it can filter out the abnormal data which affect the simulation.So it is an effective method.At last,we apply the grey multivariate linear regression to the search engine of network station.Through retreating the accessing flux of the network station and the searching frequency of the different in some times,we forecast the accessing flux of the network station in the next time.The result of the forecast would provide the theory of DSS(Decision Support System) to network station's super administrator.