A. Creating the Network
- Use the following parameters to create your first FuNN neural network using FuzzyCope/3:
|Number of Inputs
|Number of Members
||(3 Memberhip Functions per input, eg High, Medium, Low)|
|Number of Rules
|Number of Outputs
|Number of Actions
||(2 Memberhip Functions per input, eg High, Low)|
- Press the create button.
- Save the network as IrisFUNN.wgt
- View the network
B. Training the network
In neural net applications the weights are determined so that they minimise the Root Mean Squared errors (RMSE).
- Open the training file that you formatted earlier - Iris.trn
- View the training file
- Set the parameters for your Iris neural network:
||100 then 500 (increase until the terminating error is reached)|
- Press the Train button to commence training
- Save the trained network as TIrisFUNN.wgt
C. GA Training
- If your training error isn't going down try some GA (Genetic Algorithm) training. Use the following parameters:
The training error will result.
- Save the trained network again as TIrisFUNN.wgt
D. Network Evaluation
- Choose the Recall Tab
- Open the recall file that you formatted earlier - Iris.rcl
- Open a new output file for the results and name it IrisFUNN.out
- Press the Recall Button.
- Save the new output file.
- View the Output file.
- Compare your results with the desired outputs
E. Extracting Rules
- Choose the Extract Rules Tab
- Press the label button
- Save the label file (using the Save As Label file option from the Extract Rules menu)
- Open the label file in a text editor and modify the labels
- Return to FuNN and load the modified label file
- Set the Threshold values
- Press the Extract Rules button
- View your rule file (using the View Rule file option from the Extract Rules menu)
F. Additional Work
Try different combinations of:
Rules (Hidden Nodes)
Pattern / Batch Training Mode
This page is maintained by Melanie Middlemiss
Last modified on: 22/12/98.