annotate toolboxes/FullBNT-1.0.7/nethelp3.3/demtrain.htm @ 0:e9a9cd732c1e tip

first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
rev   line source
wolffd@0 1 <html>
wolffd@0 2 <head>
wolffd@0 3 <title>
wolffd@0 4 Netlab Reference Manual demtrain
wolffd@0 5 </title>
wolffd@0 6 </head>
wolffd@0 7 <body>
wolffd@0 8 <H1> demtrain
wolffd@0 9 </H1>
wolffd@0 10 <h2>
wolffd@0 11 Purpose
wolffd@0 12 </h2>
wolffd@0 13 Demonstrate training of MLP network.
wolffd@0 14
wolffd@0 15 <p><h2>
wolffd@0 16 Synopsis
wolffd@0 17 </h2>
wolffd@0 18 <PRE>
wolffd@0 19 demtrain</PRE>
wolffd@0 20
wolffd@0 21
wolffd@0 22 <p><h2>
wolffd@0 23 Description
wolffd@0 24 </h2>
wolffd@0 25 <CODE>demtrain</CODE> brings up a simple GUI to show the training of
wolffd@0 26 an MLP network on classification and regression problems. The user
wolffd@0 27 should load in a dataset (which should be in Netlab format: see
wolffd@0 28 <CODE>datread</CODE>), select the output activation function, the
wolffd@0 29 number of cycles and hidden units and then
wolffd@0 30 train the network. The scaled conjugate gradient algorithm is used.
wolffd@0 31 A graph shows the evolution of the error: the value is shown
wolffd@0 32 <CODE>max(ceil(iterations / 50), 5)</CODE> cycles.
wolffd@0 33
wolffd@0 34 <p>Once the network is trained, it is saved to the file <CODE>mlptrain.net</CODE>.
wolffd@0 35 The results can then be viewed as a confusion matrix (for classification
wolffd@0 36 problems) or a plot of output versus target (for regression problems).
wolffd@0 37
wolffd@0 38 <p><h2>
wolffd@0 39 See Also
wolffd@0 40 </h2>
wolffd@0 41 <CODE><a href="confmat.htm">confmat</a></CODE>, <CODE><a href="datread.htm">datread</a></CODE>, <CODE><a href="mlp.htm">mlp</a></CODE>, <CODE><a href="netopt.htm">netopt</a></CODE>, <CODE><a href="scg.htm">scg</a></CODE><hr>
wolffd@0 42 <b>Pages:</b>
wolffd@0 43 <a href="index.htm">Index</a>
wolffd@0 44 <hr>
wolffd@0 45 <p>Copyright (c) Ian T Nabney (1996-9)
wolffd@0 46
wolffd@0 47
wolffd@0 48 </body>
wolffd@0 49 </html>