blob: 439233771aefd8d64bc867b7a380c01e94553a8c (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
|
<html>
<head>
<title>
Netlab Reference Manual demard
</title>
</head>
<body>
<H1> demard
</H1>
<h2>
Purpose
</h2>
Automatic relevance determination using the MLP.
<p><h2>
Synopsis
</h2>
<PRE>
demmlp1</PRE>
<p><h2>
Description
</h2>
This script demonstrates the technique of automatic relevance
determination (ARD) using a synthetic problem having three input
variables: <CODE>x1</CODE> is sampled uniformly from the range (0,1) and has
a low level of added Gaussian noise, <CODE>x2</CODE> is a copy of <CODE>x1</CODE>
with a higher level of added noise, and <CODE>x3</CODE> is sampled randomly
from a Gaussian distribution. The single target variable is determined
by <CODE>sin(2*pi*x1)</CODE> with additive Gaussian noise. Thus <CODE>x1</CODE> is
very relevant for determining the target value, <CODE>x2</CODE> is of some
relevance, while <CODE>x3</CODE> is irrelevant. The prior over weights is
given by the ARD Gaussian prior with a separate hyper-parameter for
the group of weights associated with each input. A multi-layer
perceptron is trained on this data, with re-estimation of the
hyper-parameters using <CODE>evidence</CODE>. The final values for the
hyper-parameters reflect the relative importance of the three inputs.
<p><h2>
See Also
</h2>
<CODE><a href="demmlp1.htm">demmlp1</a></CODE>, <CODE><a href="demev1.htm">demev1</a></CODE>, <CODE><a href="mlp.htm">mlp</a></CODE>, <CODE><a href="evidence.htm">evidence</a></CODE><hr>
<b>Pages:</b>
<a href="index.htm">Index</a>
<hr>
<p>Copyright (c) Ian T Nabney (1996-9)
</body>
</html>
|
