|
wolffd@0
|
1 <html>
|
|
wolffd@0
|
2 <head>
|
|
wolffd@0
|
3 <title>
|
|
wolffd@0
|
4 Netlab Reference Manual somtrain
|
|
wolffd@0
|
5 </title>
|
|
wolffd@0
|
6 </head>
|
|
wolffd@0
|
7 <body>
|
|
wolffd@0
|
8 <H1> somtrain
|
|
wolffd@0
|
9 </H1>
|
|
wolffd@0
|
10 <h2>
|
|
wolffd@0
|
11 Purpose
|
|
wolffd@0
|
12 </h2>
|
|
wolffd@0
|
13 Kohonen training algorithm for SOM.
|
|
wolffd@0
|
14
|
|
wolffd@0
|
15 <p><h2>
|
|
wolffd@0
|
16 Synopsis
|
|
wolffd@0
|
17 </h2>
|
|
wolffd@0
|
18 <PRE>
|
|
wolffd@0
|
19
|
|
wolffd@0
|
20 net = somtrain{net, options, x)
|
|
wolffd@0
|
21 </PRE>
|
|
wolffd@0
|
22
|
|
wolffd@0
|
23
|
|
wolffd@0
|
24 <p><h2>
|
|
wolffd@0
|
25 Description
|
|
wolffd@0
|
26 </h2>
|
|
wolffd@0
|
27 <CODE>net = somtrain{net, options, x)</CODE> uses Kohonen's algorithm to
|
|
wolffd@0
|
28 train a SOM. Both on-line and batch algorithms are implemented.
|
|
wolffd@0
|
29 The learning rate (for on-line) and neighbourhood size decay linearly.
|
|
wolffd@0
|
30 There is no error function minimised during training (so there is
|
|
wolffd@0
|
31 no termination criterion other than the number of epochs), but the
|
|
wolffd@0
|
32 sum-of-squares is computed and returned in <CODE>options(8)</CODE>.
|
|
wolffd@0
|
33
|
|
wolffd@0
|
34 <p>The optional parameters have the following interpretations.
|
|
wolffd@0
|
35
|
|
wolffd@0
|
36 <p><CODE>options(1)</CODE> is set to 1 to display error values; also logs learning
|
|
wolffd@0
|
37 rate <CODE>alpha</CODE> and neighbourhood size <CODE>nsize</CODE>.
|
|
wolffd@0
|
38 Otherwise nothing is displayed.
|
|
wolffd@0
|
39
|
|
wolffd@0
|
40 <p><CODE>options(5)</CODE> determines whether the patterns are sampled randomly
|
|
wolffd@0
|
41 with replacement. If it is 0 (the default), then patterns are sampled
|
|
wolffd@0
|
42 in order. This is only relevant to the on-line algorithm.
|
|
wolffd@0
|
43
|
|
wolffd@0
|
44 <p><CODE>options(6)</CODE> determines if the on-line or batch algorithm is
|
|
wolffd@0
|
45 used. If it is 1
|
|
wolffd@0
|
46 then the batch algorithm is used. If it is 0
|
|
wolffd@0
|
47 (the default) then the on-line algorithm is used.
|
|
wolffd@0
|
48
|
|
wolffd@0
|
49 <p><CODE>options(14)</CODE> is the maximum number of iterations (passes through
|
|
wolffd@0
|
50 the complete pattern set); default 100.
|
|
wolffd@0
|
51
|
|
wolffd@0
|
52 <p><CODE>options(15)</CODE> is the final neighbourhood size; default value is the
|
|
wolffd@0
|
53 same as the initial neighbourhood size.
|
|
wolffd@0
|
54
|
|
wolffd@0
|
55 <p><CODE>options(16)</CODE> is the final learning rate; default value is the same
|
|
wolffd@0
|
56 as the initial learning rate.
|
|
wolffd@0
|
57
|
|
wolffd@0
|
58 <p><CODE>options(17)</CODE> is the initial neighbourhood size; default 0.5*maximum
|
|
wolffd@0
|
59 map size.
|
|
wolffd@0
|
60
|
|
wolffd@0
|
61 <p><CODE>options(18)</CODE> is the initial learning rate; default 0.9. This parameter
|
|
wolffd@0
|
62 must be positive.
|
|
wolffd@0
|
63
|
|
wolffd@0
|
64 <p><h2>
|
|
wolffd@0
|
65 Examples
|
|
wolffd@0
|
66 </h2>
|
|
wolffd@0
|
67 The following example performs on-line training on a SOM in two stages:
|
|
wolffd@0
|
68 ordering and convergence.
|
|
wolffd@0
|
69 <PRE>
|
|
wolffd@0
|
70
|
|
wolffd@0
|
71 net = som(nin, [8, 7]);
|
|
wolffd@0
|
72 options = foptions;
|
|
wolffd@0
|
73
|
|
wolffd@0
|
74 <p>% Ordering phase
|
|
wolffd@0
|
75 options(1) = 1;
|
|
wolffd@0
|
76 options(14) = 50;
|
|
wolffd@0
|
77 options(18) = 0.9; % Initial learning rate
|
|
wolffd@0
|
78 options(16) = 0.05; % Final learning rate
|
|
wolffd@0
|
79 options(17) = 8; % Initial neighbourhood size
|
|
wolffd@0
|
80 options(15) = 1; % Final neighbourhood size
|
|
wolffd@0
|
81 net2 = somtrain(net, options, x);
|
|
wolffd@0
|
82
|
|
wolffd@0
|
83 <p>% Convergence phase
|
|
wolffd@0
|
84 options(14) = 400;
|
|
wolffd@0
|
85 options(18) = 0.05;
|
|
wolffd@0
|
86 options(16) = 0.01;
|
|
wolffd@0
|
87 options(17) = 0;
|
|
wolffd@0
|
88 options(15) = 0;
|
|
wolffd@0
|
89 net3 = somtrain(net2, options, x);
|
|
wolffd@0
|
90 </PRE>
|
|
wolffd@0
|
91
|
|
wolffd@0
|
92
|
|
wolffd@0
|
93 <p><h2>
|
|
wolffd@0
|
94 See Also
|
|
wolffd@0
|
95 </h2>
|
|
wolffd@0
|
96 <CODE><a href="kmeans.htm">kmeans</a></CODE>, <CODE><a href="som.htm">som</a></CODE>, <CODE><a href="somfwd.htm">somfwd</a></CODE><hr>
|
|
wolffd@0
|
97 <b>Pages:</b>
|
|
wolffd@0
|
98 <a href="index.htm">Index</a>
|
|
wolffd@0
|
99 <hr>
|
|
wolffd@0
|
100 <p>Copyright (c) Ian T Nabney (1996-9)
|
|
wolffd@0
|
101
|
|
wolffd@0
|
102
|
|
wolffd@0
|
103 </body>
|
|
wolffd@0
|
104 </html> |
