The final step of AgielNN is forecasting. This should be the easiest process in AgielNN. After finished douing some training and checking; and going with best match to actual data; forecasting can be made using the model by taking a step to “Forecasting” tab in AgielNN.Before that, make sure you are not bypassing the data preparation for forecasting as explained in the first tutorial of data preparation.
There are 2 choices we have to do the process.
Case 1: Use the matched model.
- In this case, you can simple select the “Use latest trained ANN” (selected by default).
- Select the data to be used in the combo box of “Data used”
- Select a number of column as input. Be sure that the number of column must be equal to the number of column used during training and checking.
- Click “Do Forecast”
Case 2: Use saved AgielNN model.
- In this case, select the “Use other ANN file”.
- Browse the ANN model saved from previous training and checking. The ANN file model must be a model exported from AgielNN and should have *.ann extension. It is not compatible with another ANN file generated by another neural network software. The file can be exported thru menu File — Export — Artificial Neural Network — ANN Trained Model.
- Select the data to be used in the combo box of “Data used”
- Select a number of column as input. Be sure that the number of column must be equal to the number of column used during training and checking.
- Click “Do Forecast”
The screenshot of the software should be like this.
—————————————————————-
This is the end of tutorial. Should anybody have question, please mail me. Thanks.
Written by: Sugiyanto
28 June 2013.