PatternRecognitiontool
所属分类:matlab编程
开发工具:matlab
文件大小:344KB
下载次数:87
上传日期:2008-04-01 11:37:27
上 传 者:
strongbox
说明: matlab 模式识别工具包 希望能对你们有用
(pattern recognition matlab toolkit hope to be useful to you)
文件列表:
模式识别工具包\prtools3.1.7\normalm.m (1801, 2002-09-18)
模式识别工具包\prtools3.1.7\normm.m (1382, 2002-09-18)
模式识别工具包\prtools3.1.7\nstrcmp.m (520, 2002-09-18)
模式识别工具包\prtools3.1.7\parsc.m (516, 2002-09-18)
模式识别工具包\prtools3.1.7\parzen_map.m (1788, 2002-09-18)
模式识别工具包\prtools3.1.7\parzenc.m (2194, 2002-09-18)
模式识别工具包\prtools3.1.7\parzenm.m (1896, 2002-09-18)
模式识别工具包\prtools3.1.7\parzenml.m (2184, 2002-09-18)
模式识别工具包\prtools3.1.7\pca.m (2747, 2002-09-18)
模式识别工具包\prtools3.1.7\perlc.m (1765, 2002-09-18)
模式识别工具包\prtools3.1.7\persc.m (2048, 2002-09-18)
模式识别工具包\prtools3.1.7\pfsvc.m (778, 2002-09-18)
模式识别工具包\prtools3.1.7\plot2.m (791, 2002-09-18)
模式识别工具包\prtools3.1.7\plotd.m (2874, 2002-09-18)
模式识别工具包\prtools3.1.7\plotdg.m (1647, 2002-09-18)
模式识别工具包\prtools3.1.7\plotf.m (1119, 2002-09-18)
模式识别工具包\prtools3.1.7\plotm.m (1528, 2002-09-18)
模式识别工具包\prtools3.1.7\polyc.m (972, 2002-09-18)
模式识别工具包\prtools3.1.7\prdata.m (960, 2002-09-18)
模式识别工具包\prtools3.1.7\prex1.m (665, 2002-09-18)
模式识别工具包\prtools3.1.7\prex2.m (745, 2002-09-18)
模式识别工具包\prtools3.1.7\prex3.m (659, 2002-09-18)
模式识别工具包\prtools3.1.7\prex4.m (1281, 2002-09-18)
模式识别工具包\prtools3.1.7\prex5.m (1561, 2002-09-18)
模式识别工具包\prtools3.1.7\prex6.m (605, 2002-09-18)
模式识别工具包\prtools3.1.7\prodc.m (1228, 2002-09-18)
模式识别工具包\prtools3.1.7\proxm.m (2593, 2002-09-18)
模式识别工具包\prtools3.1.7\prtools.m (8690, 2002-09-18)
模式识别工具包\prtools3.1.7\qdc.m (1170, 2002-09-18)
模式识别工具包\prtools3.1.7\quadrc.m (1667, 2002-09-18)
模式识别工具包\prtools3.1.7\rbnc.m (1145, 2002-09-18)
模式识别工具包\prtools3.1.7\reducm.m (1072, 2002-09-18)
模式识别工具包\prtools3.1.7\reject.m (721, 2002-09-18)
模式识别工具包\prtools3.1.7\renumlab.m (1904, 2002-09-18)
模式识别工具包\prtools3.1.7\rnnc.m (832, 2002-09-18)
模式识别工具包\prtools3.1.7\roc.m (1334, 2002-09-18)
模式识别工具包\prtools3.1.7\rsubc.m (2060, 2002-09-18)
模式识别工具包\prtools3.1.7\scalem.m (2219, 2002-09-18)
模式识别工具包\prtools3.1.7\scatterd.m (5885, 2002-09-18)
... ...
% PRTools3.0 December 1999
%
% PRTOOLS is a basic set of statistical pattern recognition
% tools under Matlab. Not all commands are entirely tested.
%
% It is heavily upgraded from the previous version, November 1997
%
% The main change is the use of "classes' and 'objects' offered
% by Matlab-5 (Don't confuse them with the pattern recognition
% classes and objects!). This simplifies the use, but may be
% hard to understand for old users. Moreover, it makes PRTools3.0
% completely incompatible with old versions.
%
% There are two of these class-constructors: datasets and mappings,
% the basic elements of pattern recognition. As use is made of
% structures (entirely hidden for the user), they contain all types
% of information that should be transferred otherwise explicitely
% by the user: labels, sizes, weigths, mapping types, etcetera.
%
% What we can do now are defining datasets by
% A = dataset(a,labels);
% use it for training by defining a mapping by
% W = ldc(A);
% which may also be written as
% W = A*ldc;
% and use it for testing by:
% e = A*W*testd;
% This can be read as: map the dataset A on W (i.e. a space constructed
% by class memberships) and test it (count the errors).
% Mappings can be combined sequentially (W = W1*W2), by stacking
% ([W = [W1 W2 W3]) and in parallel (W = [W1; W2; W3]). This facilitates
% the construction of combined classifiers, e.q. W = maxc([W1 W2 W3]).
%
% A number of routines has been renamed, e.g.:
% nlc : ldc
% nqc : qdc
% mlslc : fisherc
% nmlc : nmc
% Others are combined or have been moved outside the user sight into the
% private directory. Many routines have been added. See the Contents file.
%
% PRTools3.1, January 2000
%
% For affine mapping and normalizing mappings the multiplicative output
% constant w.v is now integrated in the mapping coefficients. w.v is set
% to one.
%
% Untrained mappings are no longer empty, so now isempty(fisherc) is 0.
% These mappings can now be detected by the new routine 'istrained', so
% istrained(fisherc) is 0.
%
% The display command for datasets has been changed such that it returns
% the actual number of classes available in the dataset.
%
% New commands:
% seldat: Select classes / features / objects from dataset
% classim: Classify image using a given classifier
% clevalf: Classifier evaluation (feature size curve)
% cnormc: Renamed copy of normc
% emclust: Expectation - Maximization Clustering
% spatm: Augment image dataset with spatial label information
% nlfisherm:Renamed copy of nlklm
%
% PRTools3.1.2, January 2001
%
% The dataset structure changed in order to speed up the handling of labels.
% This should not effect the calls as it is intended to be upwards compatible.
% Newly created datasets, however, cannot be used under previous versions
% of PRTools.
%
% The confmat command changed slightly to enable a more general use. As a
% result the number of errors for the specific useage in which the two
% labellists are identical is not returned anymore.
%
% A new command matchlab is added to rotate labels for optimal match.
%
% PRTools3.1.3, July 2001
%
% New commands
% pca: Principal component analysis (replaces overloaded procedure by klm)
% maf: Maximum autocorrelation mapping, (pca for multi-band images)
%
% Consequently, some of the old, overloaded and confusing possiblities in klm
% are removed.
%
% PRTools3.1.4, August 2001
%
% some bugs are removed in @dataset/subsasgn.m and @dataset/dataset.m
% all relating to the erroneous working of expressions like
% A([1 2 3],:) = []; A(:,[1 2 3]);
% if A is a dataset.
%
% kljlc: bug removed
% gauss: label generation improved
% gendats: labels changed from numeric to character
% featsel* redundant output suppressed
%
% PRTools3.1.5, August 2001
%
% knnc: Makes no use anymore of prior probabilities.
%
% New command
% getprob: Retrieves dataset prior probabilities
%
% PRTools3.1.6, September 2001
%
% normal_map: now generates scaled densities
% qdc: now also in 2-class problems based on densities
% plotm: internal scaling improved. Parameter added for
% selecting contour.
% scatterd: extended, a.o. with 3d plot
% distm: now returns a dataset
% gendat: relative class sizes supported
%
% New commands:
%
% subsc: Subspace Classifier
% quadrc: Quadratic classifier (original 2-class qdc)
% mclassc: Multi-class classifier by 2-class discriminants
% classsizes: Returns sizes of classes in a dataset
% getlablist: Returns label list of a dataset
% See 'help datasets' for all means to retrieve data stored
% in a dataset.
% mds: Non-linear mapping by multi-dimensional scaling (Sammon)
% classs: Multi-dimensional mapping by classical scaling
%
% *************************************************************
%
% More information can be found by
% help prtools
% help datasets
% help mappings
% or in the manual (http://www.ph.tn.tudelft.nl/~bob/postscript/PRTools3.ps)
%
% This software can be used freely for inspection and academic research.
% Please refer to
%
% R.P.W. Duin, PRTools 3, A Matlab Toolbox for Pattern Recognition, Delft
% University of Technology, January 2000
%
% if it has been useful to you. If you like to use the toolbox for commercial
% purposes, please contact me.
%
% Bob Duin
%
% Copyright: R.P.W. Duin, duin@ph.tn.tudelft.nl
% Faculty of Applied Physics, Delft University of Technology
% P.O. Box 5046, 2600 GA Delft, The Netherlands
近期下载者:
相关文件:
收藏者: