The Least Squares Method for Solving Autoregressive Model Parameter
In the calculation of autoregressive model parameters,this paper proposes the least squares method for AR model parameter estimation to address the issue of different corrections for the same observation at different positions in the augmented matrix.Firstly,AR model is transformed equivalently,and a new function model is formed.Then,the error matrix and unknown parameters are solved by the successive least squares method.Finally,the first-order approximate estimation formula of the parameter cofactor matrix is giv-en.The results of simulated examples show that the method is feasible in solving AR model paremeters.By using this method,the so-lution of the total least squares problem is transformed into the least squares problem,which is simple to calculate and easy to pro-gram.At the same time,this method can effectively reduce the amount of computation,thus reducing the complexity of the solution.
autoregressive modelleast squarestotal least squaresparameter estimation
NETL
NSTL
万方数据
