home *** CD-ROM | disk | FTP | other *** search
- Info on Network Modeling and Pruning Program
-
- 1. Given a trained MLP network, the program;
- a. Models the net
- b. Measures its performance as a function of
- the number of hidden units
-
- 2. The non-demo version of the program;
- a. Prunes the net to a user-controlled size and saves a pruned
- net to disk
- b. Determines the effective degree (amount of nonlinearity) of the net
-
- 3. Example Run of Network Modeling and Pruning Program
- a. Go to the "Batch Processing" option and press <ret>
- b. Observe the parameter file with commented keyboard responses;
-
- grng.top ! network structure file
- 2 ! 1 if the outputs are coded, 2 for uncoded output format
- grng ! data file
- 0 ! number of training patterns (0 for all data)
- 0, 0 ! Enter numbers of 1st and last patterns to examine (0 0 for none)
- grng1.wts ! weight file of trained network
- 7.0 ! Maximum acceptable classification error percentage in pruned network
-
-
- The program will prune the network from the fast training program,
- which has network structure file grng.top and weight file grng1.wts,
- so that it has 7 % classification error or less.
- c. Exit the DOS editor and observe the program running
- d. Go to the "Examine Program Output" option and press <ret>
- e. The program has created a new network, having 10 hidden units,
- which performs about as well as the original network.
- f. You can run this program on your own data, simply by editing the
- parameter file in the "batch Run" option. Realistically, it is
- more productive to use the "Manual Run" with this program.
-
-