No BSD License
Highlights from
discrim
-
confmat(c, d)
CONFMAT Confusion matrix.
-
crossval(rule, X, k, v)
CROSSVAL K-fold cross validation.
-
mahalanobis(X, Mu, C)
MAHALANOBIS Mahalanobis distance.
-
mcnemar(k, a, b)
MCNEMAR McNemar's test for comparison of error rates.
-
parseopt(opt, varargin)
PARSEOPT Get values from fields of option struct.
-
plotdr(f, varargin)
PLOTDR Plot decision regions for classifier object.
-
plotobs(X, k)
PLOTOBS Scatter plot of observation classes
-
classifier(X, k, prior)
CLASSIFIER Generic discriminant analysis object.
-
lda(X, k, prior, est, nu)
LDA Linear Discriminant Analysis.
-
logda(X, k, prior, maxit, est)
LOGDA Logistic Discriminant Analysis.
-
qda(X, k, prior, est, nu)
QDA Quadratic Descriminant Analysis.
-
softmax(X, k, prior, varargin)
SOFTMAX Multinomial feed-forward neural-network
-
Contents.m
-
View all files
from
discrim
by Michael Kiefte
This is version 0.3 of the Discriminant Analysis Toolbox with major bug fixes.
|
| Contents.m |
% Discriminant analysis toolbox.
% Version 0.3 99/04/30
% Copyright (C) 1999 Michael Kiefte.
% See the accompanying README file for information.
%
% Discriminant Analysis.
% lda - Linear discriminant analysis
% logda - Logistic discriminant analysis
% qda - Quadratic discriminant analysis
%
% Feed-forward Neural Network.
% softmax - Multinomial feed-forward neural network
%
% Classification.
% classify - Classify new observations
% crossval - Cross-validation classification
%
% Performance Assessment.
% confmat - Confusion matrix
% mcnemar - McNemar tests for comparison of classifiers
%
% Plotting Functions.
% plotcov - Plot covariance matrices of 2 feature LDA or QDA objects
% plotdr - Plot decision regions for 2 feature classifier
%
% Helper Functions.
% classifier - Parent class to all classifier objects
% parseopt - Chreck structure for options and return field values
%
% Other Functions.
% cov - Within groups covariance matrices
% mahalanobis - Mahalanobis distance
% shrink - Shrink covariance matrices of LDA or QDA object
|
|
Contact us at files@mathworks.com