fsym_neighbors2.m

来自「空间统计工具箱」· M 代码 · 共 32 行

function [wswnn, wwsnn, wmatnn]=fsym_neighbors2(m, rho)
%
%[wswnn, wwsnn, wmatnn]=fsym_neighbors2(m, rho)
%
%This function takes the individual neighbor matrices and manipulates
%these to form  symmetric, normalized weight matrices based on m nearest
%neighbors.
%
%INPUT:
%
%The scalar m which gives how many neighbors to use
%
%The scalar rho which sets the geometric weighting (1-rho)=decay
%
%The file smats.mat on disk from a previous invocation of fneighbors2.
%
%OUTPUT:
%
%wswnn is a symmetric n by n spatial matrix similar to (same eigenvalues) the asymmetric row-stochastic wwsnn
%
%wwsnn is a asymmetric row-stochastic weight matrix
%
%wmat is the n by n diagonal matrix so that wmat*wmat*s is row-stochastic
%and wmat*s*wmat is a symmetric matrix.
%
%NOTES:
%
%One can optimize the likelihood over neighbors and weighting of neighbors
%as 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,