• 大小: 167KB
    文件类型: .rar
    金币: 1
    下载: 0 次
    发布日期: 2021-06-11
  • 语言: Matlab
  • 标签: LPP,MATLAB  

资源简介

LPP是流形学习的一种算法,该资源为LPP局部保持投影的MATLAB源代码。

资源截图

代码片段和文件信息

function [eigvector eigvalue elapse] = LGE(W D options data)
% LGE: Linear Graph embedding
%
%       [eigvector eigvalue] = LGE(W D options data)

%             Input:
%               data       - data matrix. Each row vector of data is a
%                         sample vector. 
%               W       - Affinity graph matrix. 
%               D       - Constraint graph matrix. 
%                         LGE solves the optimization problem of 
%                         a* = argmax (a‘data‘WXa)/(a‘data‘DXa) 
%                         Default: D = I 
%
%               options - Struct value in Matlab. The fields in options
%                         that can be set:
%
%                     ReducedDim   -  The dimensionality of the reduced
%                                     subspace. If 0 all the dimensions
%                                     will be kept. Default is 30. 
%
%                            Regu  -  1: regularized solution 
%                                        a* = argmax (a‘X‘WXa)/(a‘X‘DXa+ReguAlpha*I) 
%                                     0: solve the sinularity problem by SVD (PCA) 
%                                     Default: 0 
%
%                        ReguAlpha -  The regularization parameter. Valid
%                                     when Regu==1. Default value is 0.1. 
%
%                            ReguType  -  ‘Ridge‘: Tikhonov regularization
%                                         ‘Custom‘: User provided
%                                                   regularization matrix
%                                          Default: ‘Ridge‘ 
%                        regularizerR  -   (nFea x nFea) regularization
%                                          matrix which should be provided
%                                          if ReguType is ‘Custom‘. nFea is
%                                          the feature number of data
%                                          matrix
%
%                            PCARatio     -  The percentage of principal
%                                            component kept in the PCA
%                                            step. The percentage is
%                                            calculated based on the
%                                            eigenvalue. Default is 1
%                                            (100% all the non-zero
%                                            eigenvalues will be kept.
%                                            If PCARatio > 1 the PCA step
%                                            will keep exactly PCARatio principle
%                                            components (does not exceed the
%                                            exact number of non-zero components).  
%
%
%             Output:
%               eigvector - Each column is an embedding function for a new
%                           sample vector (row vector) x  y = x*eigvector

 属性            大小     日期    时间   名称
----------- ---------  ---------- -----  ----

     文件     200194  2009-12-02 10:20  Locality Preserving Projections.pdf

     文件      10940  2009-12-02 10:43  LGE.m

     文件       5044  2009-12-02 10:43  LPP.m

----------- ---------  ---------- -----  ----

               216178                    3


评论

共有 条评论