|
wolffd@0
|
1 <html>
|
|
wolffd@0
|
2 <head>
|
|
wolffd@0
|
3 <title>
|
|
wolffd@0
|
4 Netlab Reference Manual gpfwd
|
|
wolffd@0
|
5 </title>
|
|
wolffd@0
|
6 </head>
|
|
wolffd@0
|
7 <body>
|
|
wolffd@0
|
8 <H1> gpfwd
|
|
wolffd@0
|
9 </H1>
|
|
wolffd@0
|
10 <h2>
|
|
wolffd@0
|
11 Purpose
|
|
wolffd@0
|
12 </h2>
|
|
wolffd@0
|
13 Forward propagation through Gaussian Process.
|
|
wolffd@0
|
14
|
|
wolffd@0
|
15 <p><h2>
|
|
wolffd@0
|
16 Synopsis
|
|
wolffd@0
|
17 </h2>
|
|
wolffd@0
|
18 <PRE>
|
|
wolffd@0
|
19 y = gpfwd(net, x)
|
|
wolffd@0
|
20 [y, sigsq] = gpfwd(net, x)
|
|
wolffd@0
|
21 [y, sigsq] = gpfwd(net, x, cninv)
|
|
wolffd@0
|
22 </PRE>
|
|
wolffd@0
|
23
|
|
wolffd@0
|
24
|
|
wolffd@0
|
25 <p><h2>
|
|
wolffd@0
|
26 Description
|
|
wolffd@0
|
27 </h2>
|
|
wolffd@0
|
28 <CODE>y = gpfwd(net, x)</CODE> takes a Gaussian Process data structure <CODE>net</CODE>
|
|
wolffd@0
|
29 together
|
|
wolffd@0
|
30 with a matrix <CODE>x</CODE> of input vectors, and forward propagates the inputs
|
|
wolffd@0
|
31 through the model to generate a matrix <CODE>y</CODE> of output
|
|
wolffd@0
|
32 vectors. Each row of <CODE>x</CODE> corresponds to one input vector and each
|
|
wolffd@0
|
33 row of <CODE>y</CODE> corresponds to one output vector. This assumes that the
|
|
wolffd@0
|
34 training data (both inputs and targets) has been stored in <CODE>net</CODE> by
|
|
wolffd@0
|
35 a call to <CODE>gpinit</CODE>; these are needed to compute the training
|
|
wolffd@0
|
36 data covariance matrix.
|
|
wolffd@0
|
37
|
|
wolffd@0
|
38 <p><CODE>[y, sigsq] = gpfwd(net, x)</CODE> also generates a column vector <CODE>sigsq</CODE> of
|
|
wolffd@0
|
39 conditional variances (or squared error bars) where each value corresponds to a pattern.
|
|
wolffd@0
|
40
|
|
wolffd@0
|
41 <p><CODE>[y, sigsq] = gpfwd(net, x, cninv)</CODE> uses the pre-computed inverse covariance
|
|
wolffd@0
|
42 matrix <CODE>cninv</CODE> in the forward propagation. This increases efficiency if
|
|
wolffd@0
|
43 several calls to <CODE>gpfwd</CODE> are made.
|
|
wolffd@0
|
44
|
|
wolffd@0
|
45 <p><h2>
|
|
wolffd@0
|
46 Example
|
|
wolffd@0
|
47 </h2>
|
|
wolffd@0
|
48 The following code creates a Gaussian Process, trains it, and then plots the
|
|
wolffd@0
|
49 predictions on a test set with one standard deviation error bars:
|
|
wolffd@0
|
50 <PRE>
|
|
wolffd@0
|
51
|
|
wolffd@0
|
52 net = gp(1, 'sqexp');
|
|
wolffd@0
|
53 net = gpinit(net, x, t);
|
|
wolffd@0
|
54 net = netopt(net, options, x, t, 'scg');
|
|
wolffd@0
|
55 [pred, sigsq] = gpfwd(net, xtest);
|
|
wolffd@0
|
56 plot(xtest, pred, '-k');
|
|
wolffd@0
|
57 hold on
|
|
wolffd@0
|
58 plot(xtest, pred+sqrt(sigsq), '-b', xtest, pred-sqrt(sigsq), '-b');
|
|
wolffd@0
|
59 </PRE>
|
|
wolffd@0
|
60
|
|
wolffd@0
|
61
|
|
wolffd@0
|
62 <p><h2>
|
|
wolffd@0
|
63 See Also
|
|
wolffd@0
|
64 </h2>
|
|
wolffd@0
|
65 <CODE><a href="gp.htm">gp</a></CODE>, <CODE><a href="demgp.htm">demgp</a></CODE>, <CODE><a href="gpinit.htm">gpinit</a></CODE><hr>
|
|
wolffd@0
|
66 <b>Pages:</b>
|
|
wolffd@0
|
67 <a href="index.htm">Index</a>
|
|
wolffd@0
|
68 <hr>
|
|
wolffd@0
|
69 <p>Copyright (c) Ian T Nabney (1996-9)
|
|
wolffd@0
|
70
|
|
wolffd@0
|
71
|
|
wolffd@0
|
72 </body>
|
|
wolffd@0
|
73 </html> |
