source: distools/distools_long.m @ 33

Last change on this file since 33 was 18, checked in by bduin, 13 years ago

clevald, parzenddc, parzend_map and kem added
File size: 6.9 KB
RevLine 
[18]1%%   DisTools Table of Contents
2%%  23-Nov-2009_17:27
3%%
4%%  This Matlab toolbox for the analysis of dissimilarity data works only
5%%  if also the pattern recognition toolbox PRTools is available.
6%%  See http://prtools.org
7%%
8%%    E. Pekalska,  ela.pekalska@googlemail.com, University of Manchester
9%%    R.P.W. Duin,  r.duin@ieee.org, Delft University of Technology
10%%
11%% Characterization of dissimilarity matrices
12%% ------------------------------------------
13%% CHECKEUCL    Check whether a square dissimilarity matrix has a Euclidean behavior
14%% CHECKTR      Check whether a square dissimilarity matrix obeys triangle inequality
15%% CHARDMAT     Fiand several characteristic of (dis)similarity data
16%% CORRTR       Correct a square dissimilarity matrix to obey the triangle inequality
17%% DISCHECK     Dissimilarity matrix check
18%% DISNORM      Normalization of a dissimilarity matrix
19%% DISSTAT      Basic statistics of the dissimilarity matrix
20%  GOFCL        Goodness of clusters/classes separability vs compactness for dissimilarity data
21%  INTRDIM      Estimate Intrinsic dimension from dissimilarity data
22%% ISSQUARE     Check whether a matrix is square
23%% ISSYM        Check whether a matrix is symmetric
24%% ASYMMETRY    Compute asymmetry of dissimilarity matrix
25%% NEF          Negative Eigen Fraction
26%% NNE          Leave-one-out Nearest Neighbor error on a dissimilarity matrix
27%% NNERR        Exact expected NN error from a dissimilarity matrix
28%  NNERROR      Exact expected NN error from a dissimilarity matrix (2)
29%  VAT          Visual Assessment of cluster Tendency for dissimilarity matrices       
30%
31%%
32%% Dissimilarity Measures
33%% -----------------------------------------------
34%  BINDISTM     Dissimilarity matrix between binary vectors
35%  BLURDISTM    Blurred Euclidean distance matrix between blobs
36%   BLOBBOX     Find box around a binary blob and resample
37%  CORRDISTM    Distance matrix based on correlations
38%% COSDISTM     Distance matrix based on inner products
39%  DPROCRUSTDM  Distance matrix between datasets based on extended Procrustes problem
40%  EDGEDISTM    Distance matrix between images based on their edges
41%  EDITDISTM    Edit distance matrix between strings
42%% EUDISTM      Euclidean distance matrix
43%  EXPDISTM     Exponential-type of distance matrix
44%  FLPDISTM     Fast computation of the lp (p > 0)  distance matrix
45%% HAMDISTM     Hamming distance matrix between binary vectors
46%% HAUSDM       Hausdorff and modified Hausdorff distance between datasets of image blobs
47%  JACSIMDISTM  Jaccard-like distance matrix based on similarities
48%  LPDISTM      l_p (p > 0) distance matrix
49%  QDISTM       Distance matrix for quantitative variables
50%  RANKDISTM    Distance matrix between two data sets based on ranking
51%  SAMDISTM     Distance matrix based on Spectral Angular Mapper (SAM)
52%  STRKERM      String Kernel Matrix by Lodhi et al
53%
54%%
55%% Transformations and projections
56%% -----------------------------------------------
57%% DISSIMT      Fixed DISsimilarity-SIMilarity transformation
58%  KCENTERM     Kernel weighted centering mapping (also for a similarity matrix)
59%% MAKESYM      Make a matrix symmetric
60%  PROXXM       Proximity mapping
61%  SIGMOID      Element-wise sigmoid tranformation of a matrix
62%  FASTMAPD     FastMap; inear projection of Euclidean distances
63%% PE_EM        Pseudo-Euclidean embedding (includes Classical Scaling as a special case)       
64%  SPHEM        Spherical Embedding
65%
66%%
67%% Classification in Pseudo-Euclidean Space and indefinite kernels
68%% -----------------------------------------------
69%  PE_AFFINE
70%% SETSIG       Set PE signature for mappings or datasets
71%% GETSIG       Set PE signature for mappings or datasets
72%% ISPE_DATASET Test dataset for PE signature setting
73%% ISPE_EM      Test mapping for PE signature setting
74%% PE_DISTM     Square pseudo-Euclidean distance between two datasets
75%% PE_KERNELM   Compute kernel in PE space
76%  PE_LIBSVC    Libsvc for PE spaces
77%% PE_MTIMES    Matrix multiplication (inner product) in PE space
78%% PE_PARZENC   Parzen classifier in PE space
79%% PE_KNNC      KNN classifier in PE space
80%% PE_NMC       Nearest mean classifier in PE space
81%% PE_EM        Pseudo-Euclidean linear embedding of dissimilarities
82%  KEM          Kernel embedding
83%% PLOTSPECTRUM Plot spectrum of eigenvalues
84%  PSPCA        Pseudo-Euclidean Principal Component Analysis
85%
86%  Indefinte kernel routines
87%  -------------------------
88%  IKFD         Indefinite Kernel Fisher discriminant
89%  IKPCA        Indefinite Kernel PCA
90%
91%%
92%% Routines supporting in learning from dissimilarity matrices
93%% -----------------------------------------------------------------------
94%% CROSSVALD    Cross-validation error for dissimilarity representations
95%  CLEVALD      Classifier evaluation (Learning curve)
96%% DISSPACES    Compute various spaces out of a dissimilarity matrix
97%% GENDDAT      Generate random training and test sets for dissimilarity data
98%% GENREP       Generate a representation set
99%% GENREPI      Generate indices for representation, learning and testing sets
100%% SELCDAT      Select Class Subset from a Square Dissimilarity Dataset
101%% PROTSELFD    Forward prototype selection   
102%  AUCDLPC      AUC-LP classifier on dissimilarity data
103%% DLPC         LP-classifier on dissimilarity (proximity) data
104%  DRSSCC       Dissimilarity-based random subspace combining classifier
105%% KNNDC        K-Nearest Neighbor classifier for dissimilarity matrices
106%% PARZENDDC    Parzen classifier for dissimilarity matrices
107%  KFD          Kernel Fisher Discriminant
108%  KSVC         Kernel Support Vector classifier on a kernel matrix
109%       KSVO        Kernel Support Vector Optimizer
110%  KSVC_NU      Kernel Support Vector classifier on a kernel matrix; nu-version
111%       KSVO_NU     Kernel Support Vector Optimizer; nu-version
112%  MCLASSDC     Multi-Class Dissimilarity-based Classifier from Two-Class Discriminants
113%% TESTKD       Test k-NN classifier for dissimilarity data
114%% TESTPD       Test Parzen classifier for dissimilarity data
115%  TQDC         Trade-off Quadratic Discriminant (Regularized Bayes Normal Classifier)
116%
117%
118%  Graphs and distances
119%  -----------------------------------------------
120%  DISTGRAPH    Computes distances in a graph
121%  DMSTSPM      Finds the shortest paths along K minimum spanning trees
122%  DSPATH       Single shortest path in a (dissimilarity) graph
123%  DSPATHS      All shortest paths in a (dissimilarity) Graph
124%  GRAPHPATH    Compute shortest paths in a graph
125%  KMST         Finds K minimum spanning trees based on a distance matrix
126%  MSTPLOT      Plot minimum spanning trees
127%  NHGRAPH      Find a neighborhood graph and its shortest paths
128%  PLOTGRAPH    Plot a 2D graph
129%%
130%% EXAMPLES
131%% --------
132%% CROSSVALD_EX Crossvalidation of several classifiers
133
134%  Superfluous / outdated but still available
135%  -------------------------------------------------
136%  KPCA, AUGPSEM, PSEM
137% 
Note: See TracBrowser for help on using the repository browser.