The study of regression based on grey system theory

The traditional methods for the regression of alternate data, methods such as linear regression, polynomial regression and exponential regression, are used widely. When we compared their results with the original data, we found some unacceptably large errors. At this point view we present the GM(1,1) model to decrease the error. In this paper, first, we investigate the traditional and the GM(1,1) regression models. Second, we give an alternate data set as an example and to find their regression results. The third step, we compare the errors of traditional method with the error of GM(1,1) model, and list their rank to improve that the GM(1,1) model is a more suitable method in the regression.