Tutorial
Cancer Diagnosis
Stock Market Prediction
NeuNet Pro SFAM Tutorial
"Cancer Diagnosis"
- Using the NeuNet Pro FILE menu, select Create New SFAM Project. Name this project "Cancer2".
- You are delivered to the Project Configuration Screen. Using the Data Source Browse button, select "Cancer.nns" as the data source for this project. This database contains only one table "Breast" and it will automatically be selected for you.
- The target field to be predicted is "Class". Select this field in the list of available fields and pull it leftward into the Prediction Field by using the < button.
- Use the >> button to move all of the remaining available fields into Input Fields. The field named "Sample" is not useful as input, so use the < button to move this field back into Available Fields.
- Press the Advanced button and have a look at the Advanced Configuration screen. It is not necessary to change anything on this screen. Press OK to return to Project Configuration.
- Complete the Project Configuration by pressing OK. You are delivered to the Data Split screen.
- Split your data so rows 1 to 400 are used for training set and all 457 rows are used for testing set. Simply drag the split handles up and down as necessary. Hint: By double clicking on the number box, you can type in the split points. Press OK when your split is complete. You are delivered to the SFAM Training screen.
- Press GO to begin the first SFAM training cycle. This first training cycle results in 31 nodes created with overall error of 2%. This means that 98% of the training set has been correctly predicted.
- Press GO to initiate a second training cycle. This cycle creates 3 additional nodes with overall error of 0.75%.
- Press GO again to initiate the third training cycle. This cycle creates one additional node with an overall error of 0.50%.
- Press Finish now. A test cycle is automatically performed on your testing set, and you are delivered to the Confusion Matrix to see the test results. Notice that all of the benign cases have been successfully predicted, but six of the malignant cases have been predicted as benign. Try double clicking on any cell of the confusion matrix to browse the data that contributed to that cell.
- Now experiment with the four browse views: Browse All; Browse Test Results; Browse Incorrect Rows only; and Browse Correct Rows Only. Notice how you can right-click on any column header to sort and find within that column. Notice how double-click or spacebar can be used to set a checkmark for any row.
- Notice the yellow "interactive" row at the bottom of Browse Test Results. This row allows you to experiment with field values and see how they affect the prediction.
- Suppose you wish to export the six cases which may have been incorrectly diagnosed as benign. Select the view: Browse Incorrect Rows Only and click Mark All Rows. Then use File Menu... Export Marked Rows to create a text file of these six cases. This text file can now be imported into your word processor, spreadsheet or database program.
- Congratulations on completing this tutorial. For additional practice, go back to Configure This Project and try selecting some other field as the target prediction. Then repeat the SFAM training process.