Matlab - need help urgently-based on LDA

Author

need help urgently-based on LDA

maheema

2006-09-30, 8:05 am

hi,

i'm doing project.plz help me about how to implement
this first step in matlap.

1. Formulate the data sets and the test sets, which are to be
classified in the original space. The
given data sets and the test vectors are formulated, a graphical plot
of the data sets and test
vectors for the example considered in original space is shown in
Figure 1. For ease of
understanding let us represent the data sets as a matrix consisting
of features.

plz send me the matlap code,
-bye

Predictor

2006-09-30, 8:05 am

The software is called MATLAB, not "matlap". Anway, while I will not
do your homework for you, here is a basic explanation of what you need
to do:

1. Training data needs to be available. Individual training cases may
be represented as rows in a matrix (with descriptors in columns) and a
vector reprenting the true output (in your case, a known class).
Sometimes these are mashed into a single matrix (typically with the
target variable in the last column).

2. Some function is needed to model the data. This function will
accept the training data and possibly some parameters as input, and
will yield a model and possibly some diagnostics as output.

Given the subject line of your original message, I assume that you are
using linear discriminant analysis. Be aware that there are several
different linear discriminant procedures (which will not neccessarily
give the same results). If you have access to the Statistics Toolbox,
then investigate the 'classify' function. If you do not, and thus need
to code LDA from scratch, I suggest any of the following on-line
resources:

http://www.cavs.msstate.edu/hse/ies.../lda_theory.pdf
http://www.ics.uci.edu/~welling/cla.../Fisher-LDA.pdf
http://www.answers.com/main/ntq-tna...ysis-fts_start-
http://people.revoledu.com/kardi/tutorial/LDA/LDA.html

....especially this one, which includes MATLAB code:

http://white.stanford.edu/~dicarlo/...algorithms.html

....or a book such as:

"Multivariate Statistical Methods: A Primer", by Manly (my copy has
ISBN 0-412-28620-3, but there is at least one newer edition).

-Will Dwinnell
http://will.dinnell.com

maheema wrote:
> i'm doing project.plz help me about how to implement
> this first step in matlap.
>
> 1. Formulate the data sets and the test sets, which are to be
> classified in the original space. The
> given data sets and the test vectors are formulated, a graphical plot
> of the data sets and test
> vectors for the example considered in original space is shown in
> Figure 1. For ease of
> understanding let us represent the data sets as a matrix consisting
> of features.
>
> plz send me the matlap code,

Sponsored Links