📄 mds.m
字号:
% MDS based on distance matrix ALGORITHM%% function [Y,eigenvals] = dist_pca(D, dim)%% D = distance as N x N matrix ( N = #points)% dim = embedding dimensionality% y = embedding as N x dim matrixfunction [y,eigenvals] = mds(D,dim) S = D.^2; %the squared distance N = length(S); g = gramMatrix(S); [mu, lam] = eig(g); eigenvals = dot(lam,eye(N)); rep_lam = repmat(eigenvals(N-dim+1:N),N,1); new_mu = mu(:,N-dim+1:N); % last d col of mu y = sqrt(rep_lam).*new_mu; function g = gramMatrix(s)%s is the squared distance matrix N = length(s); over_i = sum(s); total = sum(over_i); Gjj = (over_i - total/N/2) / N; rep_Gjj = repmat(Gjj,N,1); rep_Gii = rep_Gjj'; g = (rep_Gjj+rep_Gii-s)./2;
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -