|
Daniel@0
|
1 <html>
|
|
Daniel@0
|
2 <head>
|
|
Daniel@0
|
3 <title>
|
|
Daniel@0
|
4 Netlab Reference Manual gpinit
|
|
Daniel@0
|
5 </title>
|
|
Daniel@0
|
6 </head>
|
|
Daniel@0
|
7 <body>
|
|
Daniel@0
|
8 <H1> gpinit
|
|
Daniel@0
|
9 </H1>
|
|
Daniel@0
|
10 <h2>
|
|
Daniel@0
|
11 Purpose
|
|
Daniel@0
|
12 </h2>
|
|
Daniel@0
|
13 Initialise Gaussian Process model.
|
|
Daniel@0
|
14
|
|
Daniel@0
|
15 <p><h2>
|
|
Daniel@0
|
16 Synopsis
|
|
Daniel@0
|
17 </h2>
|
|
Daniel@0
|
18 <PRE>
|
|
Daniel@0
|
19 net = gpinit(net, trin, trtargets, prior)
|
|
Daniel@0
|
20 net = gpinit(net, trin, trtargets, prior)
|
|
Daniel@0
|
21 </PRE>
|
|
Daniel@0
|
22
|
|
Daniel@0
|
23
|
|
Daniel@0
|
24 <p><h2>
|
|
Daniel@0
|
25 Description
|
|
Daniel@0
|
26 </h2>
|
|
Daniel@0
|
27 <CODE>net = gpinit(net, trin, trtargets)</CODE> takes a Gaussian Process data structure <CODE>net</CODE>
|
|
Daniel@0
|
28 together
|
|
Daniel@0
|
29 with a matrix <CODE>trin</CODE> of training input vectors and a matrix <CODE>trtargets</CODE> of
|
|
Daniel@0
|
30 training target
|
|
Daniel@0
|
31 vectors, and stores them in <CODE>net</CODE>. These datasets are required if
|
|
Daniel@0
|
32 the corresponding inverse covariance matrix is not supplied to <CODE>gpfwd</CODE>.
|
|
Daniel@0
|
33 This is important if the data structure is saved and then reloaded before
|
|
Daniel@0
|
34 calling <CODE>gpfwd</CODE>.
|
|
Daniel@0
|
35 Each row
|
|
Daniel@0
|
36 of <CODE>trin</CODE> corresponds to one input vector and each row of <CODE>trtargets</CODE>
|
|
Daniel@0
|
37 corresponds to one target vector.
|
|
Daniel@0
|
38
|
|
Daniel@0
|
39 <p><CODE>net = gpinit(net, trin, trtargets, prior)</CODE> additionally initialises the
|
|
Daniel@0
|
40 parameters in <CODE>net</CODE> from the <CODE>prior</CODE> data structure which contains the
|
|
Daniel@0
|
41 mean and variance of the Gaussian distribution which is sampled from.
|
|
Daniel@0
|
42
|
|
Daniel@0
|
43 <p><h2>
|
|
Daniel@0
|
44 Example
|
|
Daniel@0
|
45 </h2>
|
|
Daniel@0
|
46 Suppose that a Gaussian Process model is created and trained with input data <CODE>x</CODE>
|
|
Daniel@0
|
47 and targets <CODE>t</CODE>:
|
|
Daniel@0
|
48 <PRE>
|
|
Daniel@0
|
49
|
|
Daniel@0
|
50 net = gp(2, 'sqexp');
|
|
Daniel@0
|
51 net = gpinit(net, x, t);
|
|
Daniel@0
|
52 % Train the network
|
|
Daniel@0
|
53 save 'gp.net' net;
|
|
Daniel@0
|
54 </PRE>
|
|
Daniel@0
|
55
|
|
Daniel@0
|
56 Another Matlab program can now read in the network and make predictions on a data set
|
|
Daniel@0
|
57 <CODE>testin</CODE>:
|
|
Daniel@0
|
58 <PRE>
|
|
Daniel@0
|
59
|
|
Daniel@0
|
60 load 'gp.net';
|
|
Daniel@0
|
61 pred = gpfwd(net, testin);
|
|
Daniel@0
|
62 </PRE>
|
|
Daniel@0
|
63
|
|
Daniel@0
|
64
|
|
Daniel@0
|
65 <p><h2>
|
|
Daniel@0
|
66 See Also
|
|
Daniel@0
|
67 </h2>
|
|
Daniel@0
|
68 <CODE><a href="gp.htm">gp</a></CODE>, <CODE><a href="gpfwd.htm">gpfwd</a></CODE><hr>
|
|
Daniel@0
|
69 <b>Pages:</b>
|
|
Daniel@0
|
70 <a href="index.htm">Index</a>
|
|
Daniel@0
|
71 <hr>
|
|
Daniel@0
|
72 <p>Copyright (c) Ian T Nabney (1996-9)
|
|
Daniel@0
|
73
|
|
Daniel@0
|
74
|
|
Daniel@0
|
75 </body>
|
|
Daniel@0
|
76 </html> |
