Alternating Iterative Projection Algorithm of Multivariate Time Series Mixed Models

摘要

In this paper, a new parameters estimation method is proposed based on the ill-condition separation method and the alternating iterative projection algorithm for the problem of column multi-collinearity of design matrix in the multiple regression and multivariate time series mixed models. Furthermore, the parameters estimation is improved according to the character of model. Therefore, the improved estimated parameters have better character.