The following instructions will allow you to quickly try NeuNet Pro.
For more detailed instructions, see the NeuNet Pro Tutorial.
- Run NNP23.EXE to install the program. Follow the on-screen prompts.
- Begin NeuNet Pro by clicking the desktop icon or using Windows Start Button... Programs.
- Once NeuNet Pro has loaded, select Stocks.nnp as a typical Backprop project from the recent file list.
You are delivered to Browse Test Data which shows the results of a previous training session.
Notice the status bar at bottom provides some information about this project and some helpful hints.
Notice how the yellow interactive row allows you to experiment with field values while observing the effect on the prediction.
- Try right clicking on the blue column headers and experiment with the sort and find features.
Notice how the previous column sorts provide secondary and tertiary sorting.
- Try double clicking on a data row to set a checkmark.
Notice how the keyboard down arrow and spacebar can be used as a quick way to set checkmarks.
Use Shift_click or Ctrl_Click to select multiple rows for checkmarking.
Experiment with Mark All, Unmark All, and Toggle All checkmarks.
These checkmarks are used to flag interesting data rows for export.
- Try the available graphic data views.
The Time Series and Scatter Graphs are available for Backprop projects.
Experiment with the Zoom button in the Time Series Graph.
The Confusion Matrix is available for SFAM classification projects.
Notice how double clicking on a certain area of the Scatter Graph or Confusion Matrix
delivers you to a data browse for that region.
- Have a look at the Split Data screen.
Do not change the current settings at this time.
This screen is used to split you data table into two sets: "Training Set" and "Testing Set".
Notice how some data rows may be overlapped to appear on both sets.
Some rows can be omitted from both sets.
Press Cancel to exit from this screen
- Have a look at the Configure Project screen.
Notice how you may choose what field to predict and which fields are to be used as inputs in making the prediction.
Text fields cannot be chosen as inputs.
Press the Advanced button and see the statistical report for every field.
Do not change anything on the Advanced screen at this time.
Just press Cancel to exit.
Now press Cancel to exit from the Configure Project screen.
- Click GO and review the Training Screen.
You are looking at the training results from a previous training session.
The settings like Learn Rate, Momentum, and Vigilance will work quite well at their default settings,
so do not be concerned about these settings.
Try clicking the Reset button to begin a fresh training session; then press GO to begin training.
Notice how the progress of the training is reported periodically as trainng cycles are performed on your Training Set.
When you wish apply this training to your Test Set, Stop training then press the Finish button.
- When you are finished experimenting with the Stocks.nnp project, use the FILE menu to load Cancer.nnp as a typical SFAM project.
Repeat the above experiments with this SFAM project.
Notice how the screens are slightly different for an SFAM project compared to a Backprop project.
A Complete Neural Network Development System
CorMac Technologies Inc.
34 North Cumberland Street ~ Thunder Bay ON P7A 4L3 ~ Canada
E m a i l