📄 lle_pd.m
字号:
% LLE ALGORITHM with pairwise distances (using K nearest neighbors)%% [Y] = lle_pd(X,K,dmax)%% X = pairwise distances as N x N matrix (N = #points)% K = number of neighbors% dmax = max embedding dimensionality% Y = embedding as dmax x N matrix%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%function [Y] = lle(X,K,d)function [Y] = lle_pd(X,K,d)[N,tmp] = size(X);fprintf(1,'LLE running on %d points\n',N);% STEP1: FIND NEIGHBORS fprintf(1,'-->Finding %d nearest neighbours.\n',K);%distance = X;[sorted,index] = sort(X);neighborhood = index(2:(1+K),:);% STEP2: SOLVE FOR RECONSTRUCTION WEIGHTSfprintf(1,'-->Solving for reconstruction weights.\n');tol = 1e-3;W = zeros(K,N);for ii=1:N for j=1:K for k=1:K D(j,k) = X(neighborhood(j,ii), neighborhood(k,ii)); D(j,k) = D(j,k)*D(j,k); end end D0 = sum(sum(D)); for j=1:K for k=1:K %C(j,k) = (sum(D(j,:))/K+sum(D(k,:))/K-D(j,k)-D0/(K*K))/2; C(j,k) = (sum(D(j,:))/N+sum(D(k,:))/N-D(j,k)-D0/(N*N))/2; %C(j,k) = (sum(D(j,:))+sum(D(k,:))-D(j,k)-D0)/2; end end C = C + eye(K,K)*tol*trace(C); W(:,ii) = C\ones(K,1); % solve Cw=1 W(:,ii) = W(:,ii)/sum(W(:,ii)); % enforce sum(w)=1end;% STEP 3: COMPUTE EMBEDDING FROM EIGENVECTS OF COST MATRIX M=(I-W)'(I-W)fprintf(1,'-->Computing embedding.\n');% M=eye(N,N); % use a sparse matrix with storage for 4KN nonzero elementsM = sparse(1:N,1:N,ones(1,N),N,N,4*K*N); for ii=1:N w = W(:,ii); jj = neighborhood(:,ii); M(ii,jj) = M(ii,jj) - w'; M(jj,ii) = M(jj,ii) - w; M(jj,jj) = M(jj,jj) + w*w';end;% CALCULATION OF EMBEDDINGoptions.disp = 0; options.isreal = 1; options.issym = 1; %[Y,eigenvals] = eigs(M,d+1,0,options);[Y,eigenvals] = eig(full(M));Y = Y(:,2:d+1)*sqrt(N); % bottom evect is [1,1,1,1...] with eval 0%DY = allpairdistances(N, Y);%DY = squareform(pdist(Y));DY = euclidean_distance(Y,Y);Y(:,:) = Y(:,:) * (sum(sum(X))/sum(sum(DY)));fprintf(1,'Done.\n');
⌨️ 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 + -