frecursive2.m

空间统计工具箱

function [erecur,sset,brecur]=frecursive2(xt,yt,start)
%
%[erecur,sset,brecur]=frecursive2(xt,yt,start)
%
%This function estimates recursive residuals, sum-of-squared errors,
%regression parameters.
%
%INPUT:
%
%xt represents n observations on k independent variables
%
%yt represents n observations on the dependent variable and its spatial
%lag. 
%
%start gives the minimum number of observations to use before beginning
%recursion
%
%OUTPUT:
%
%erecur represents nsub recursive residuals from the prediction of yt.
%Hence, erecur is a n by 1 vector.
%
%sset is a sequence of sum-of-squared residuals from y.
%
%brecur is the n by k recursive regression parameter estimates from
%estimating using y.
%
%NOTES:
%
%The values of erecur, sset, and brecur are zeros before observation start. The variable start must be greater than
%the number of variables. Recursive least squares is not the most numerically robust algorithm in the world, so be careful
%when using it with many correlated independent variables.
%
%This function was used in:
%
%Pace, R. Kelley and Ronald Barry, O.W. Gilley, C.F. Sirmans, 
%揂 Method for Spatial-temporal Forecasting with an Application to Real Estate Prices,