Demonstration 2b Testing a Trained NNet By Entering a Test Data Set - adam-nnl/Erudite GitHub Wiki

For this demonstration we will test a trained neural network be manually entering a input data set.

  1. Go to the ‘Input Settings‘ section and click the ‘Input Data‘ button

  2. For “How many examples in this processing run” enter ‘4‘ and click ‘OK‘

  3. For “Enter the name of this run example” enter ‘ex1-0‘ and click ‘OK‘

  4. For “Desired Input for IN001 (input 1)” enter ‘1‘ and click ‘OK‘

  5. For “Desired Input for IN002 (input 2)” enter ‘0‘ and click ‘OK‘

  6. For “Enter the name of this run example” enter ‘ex0-1‘ and click ‘OK‘

  7. For “Desired Input for IN001 (input 1)” enter ‘0‘ and click ‘OK‘

  8. For “Desired Input for IN002 (input 2)” enter ‘1‘ and click ‘OK‘

  9. For “Enter the name of this run example” enter ‘ex0-0‘ and click ‘OK‘

  10. For “Desired Input for IN001 (input 1)” enter ‘0‘ and click ‘OK‘

  11. For “Desired Input for IN002 (input 2)” enter ‘0‘ and click ‘OK‘

  12. For “Enter the name of this run example” enter ‘ex1-1‘ and click ‘OK‘

  13. For “Desired Input for IN001 (input 1)” enter ‘1‘ and click ‘OK‘

  14. For “Desired Input for IN002 (input 2)” enter ‘1‘ and click ‘OK‘

  15. Click the ‘Run NNet‘ button, the data set you just entered will be processed through the neural network and the output shown. With successful training you will see the outputs match the training set desired outputs closely.