readme

国外编的信号识别的程序

This is version 0.3 of the "Discriminant Analysis Toolbox" with major
bug fixes from the first and second versions and with the addition of
logistic discriminant anlsyis and multinomial classification
feed-forward neural-network.

To install, place the class directories starting with '@' somewhere in
your MATLAB path along with the other .m files listed below.

This is a series of .m files implementing linear, quadratic and
logistic discriminant analysis and classification generally following
the text [1]. 

References: 
[1] B. D. Ripley (1996) Pattern Classification and Neural
   Networks. Cambridge.

The following files are included:
   Contents.m          - Contents file
   README              - This file

   Parent classifier object:
   @classifier/classifier.m - Null classifier
   @classifier/classify.m   - Classify with null classifier
   @classifier/disp.m       - Display CLASSIFIER object
   @classifier/display.m    - Display CLASSIFIER object
   @classifier/subsref.m    - Access fields of CLASSIFIER object

   Linear discriminant analysis:
   @lda/classify.m          - Classify with linear discriminants
   @lda/cov.m               - Within-groups covariance matrix
   @lda/cvar.m              - Canonical variates
   @lda/disp.m              - Display LDA object
   @lda/display.m           - Display LDA object
   @lda/lda.m               - Linear discriminant analysis
   @lda/plotcov.m           - Plot covariance ellipsoids
   @lda/shrink.m            - Shrink covariance matrix
   @lda/subsref.m           - Access fields of LDA object

   Logistic discriminant analysis:
   @logda/classify.m        - Classify with logistic discriminants
   @logda/disp.m            - Display LOGDA object
   @logda/display.m         - Display LOGDA object
   @logda/logda.m           - Logistic discriminant analysis
   @logda/subsref.m         - Access fields of LOGDA object

   Quadratic discriminant analysis:
   @qda/classify.m          - Classify with quadratic discriminants
   @qda/cov.m               - Within-groups covariance matrices
   @qda/disp.m              - Display QDA object
   @qda/display.m           - Display QDA object
   @qda/plotcov.m           - Plot covariance ellipsoids
   @qda/qda.m               - Quadratic discriminant analysis
   @qda/shrink.m            - Shrink covariance matrices
   @qda/subsref.m           - Access fields of QDA object

   Multinomial feed-formward neural network:
   @softmax/classify.m      - Classify with logistic discriminants
   @softmax/disp.m          - Display LOGDA object
   @softmax/display.m       - Display LOGDA object
   @softmax/logda.m         - Logistic discriminant analysis
   @softmax/subsref.m       - Access fields of LOGDA object
   
   Other .m files:
   confmat.m                - Confusion matrix
   crossval.m               - Cross-validation classification
   kmeans.m                 - K-means clustering
   mahalanobis.m            - Mahalanobis distance
   mcnemar.m                - McNemar tests
   parseopt.m               - Get options from struct
   plotdr.m                 - Plot decision regions
   plotobs.m                - Observation scatter plot

There are examples of how to use some of these function in the help
for LDA, QDA and LOGDA. Hopefully, this should be enought to get you
started.

The .m files have been tested on MATLAB version 5.3 and definitely
won't work on anything prior to 5.0.

This is a work in progress, so if you have any comments or suggestions
feel free to send me an e-mail.

-- Michael Kiefte, June 3, 1999.

Contact Information:

Michael Kiefte                 e-mail: [e-mail]
Department of Linguistics      tel:    [phone]
4-32 Assiniboia Hall           fax:    [phone]
University of Alberta
Edmonton, Alberta
T6G 2E7 Canada

$Log$