Mercurial > hg > camir-aes2014

comparison toolboxes/FullBNT-1.0.7/nethelp3.3/gtmem.htm @ 0:e9a9cd732c1e tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:e9a9cd732c1e
1 <html>
2 <head>
3 <title>
4 Netlab Reference Manual gtmem
5 </title>
6 </head>
7 <body>
8 <H1> gtmem
9 </H1>
10 <h2>
11 Purpose
12 </h2>
13 EM algorithm for Generative Topographic Mapping.
14
15 <p><h2>
16 Synopsis
17 </h2>
18 <PRE>
19
20 [net, options, errlog] = gtmem(net, t, options)
21 </PRE>
22
23
24 <p><h2>
25 Description
26 </h2>
27 <CODE>[net, options, errlog] = gtmem(net, t, options)</CODE> uses the Expectation
28 Maximization algorithm to estimate the parameters of
29 a GTM defined by a data structure <CODE>net</CODE>.
30 The matrix <CODE>t</CODE> represents the data whose expectation
31 is maximized, with each row corresponding to a vector. It is assumed
32 that the latent data <CODE>net.X</CODE> has been set following a call to
33 <CODE>gtminit</CODE>, for example.
34
35 The optional parameters have the following interpretations.
36
37 <p><CODE>options(1)</CODE> is set to 1 to display error values; also logs error
38 values in the return argument <CODE>errlog</CODE>.
39 If <CODE>options(1)</CODE> is set to 0,
40 then only warning messages are displayed. If <CODE>options(1)</CODE> is -1,
41 then nothing is displayed.
42
43 <p><CODE>options(3)</CODE> is a measure of the absolute precision required of the error
44 function at the solution. If the change in log likelihood between two steps of
45 the EM algorithm is less than this value, then the function terminates.
46
47 <p><CODE>options(14)</CODE> is the maximum number of iterations; default 100.
48
49 <p>The optional return value <CODE>options</CODE> contains the final error value
50 (i.e. data log likelihood) in
51 <CODE>options(8)</CODE>.
52
53 <p><h2>
54 Examples
55 </h2>
56 The following code fragment sets up a GTM, initialises
57 the latent data sample and RBF
58 parameters from the data, sets the options and trains the model.
59 <PRE>
60
61 % Create and initialise GTM model
62 net = gtm(latentdim, nlatent, datadim, numrbfcentres, ...
63 'gaussian', 0.1);
64
65 <p>options = foptions;
66 options(1) = -1;
67 options(7) = 1; % Set width factor of RBF
68 net = gtminit(net, options, data, 'regular', latentshape, [4 4]);
69
70 <p>options = foptions;
71 options(14) = 30;
72 options(1) = 1;
73 [net, options] = gtmem(net, data, options);
74 </PRE>
75
76
77 <p><h2>
78 See Also
79 </h2>
80 <CODE><a href="gtm.htm">gtm</a></CODE>, <CODE><a href="gtminit.htm">gtminit</a></CODE><hr>
81 <b>Pages:</b>
82 <a href="index.htm">Index</a>
83 <hr>
84 <p>Copyright (c) Ian T Nabney (1996-9)
85
86
87 </body>
88 </html>