home *** CD-ROM | disk | FTP | other *** search
- 1. Functions of the Network Structure Estimation Program
- a. Reads a training data file
- b. Requests a minimum acceptable training error
- from the user,
- c. Requests a maximum number of iterations from the user.
- A typical number is 10.
- d. Estimates the required network structure for an MLP to attain
- the desired classification error percentage
- e. Notifies user if the network will generalize or memorize
-
- 2. Example Run of the Network Structure Estimation Program
- a. Go to the "Batch Processing" option and press <ret>
- b. Observe the parameter file with commented keyboard responses;
-
- Gls ! input training data filename
- 4 ! number of inputs per pattern
- 1 ! number of outputs per pattern
- .001 ! maximum acceptable mean-square training error
- 8 ! maximum allowable number of iterations
-
- Here, we will estimate the required size of an MLP for
- predicting chaotic time series created by the Mackey-Glass
- delay-difference equation,
- (Ref. Lapedes, A. & Farber, R.1987 Nonlinear Signal processing using
- Networks : Prediction & System modelling, Tech. Rep. LA-UR-87-2662,
- Los Alamos National Laboratory, Los Alamos, NM.)
- c. Exit the DOS editor and observe the program running
- d. Go to the "Examine Program Output" option and press <ret>
- e. The program predicts that a network with 9 hidden units can be
- trained with a MSE of .000987, and that networks with the proposed
- structure should successfully generalize
- f. You can run this program on your own data, simply by editing the
- parameter file in the "batch Run" option.