Trained Model

The Trained Model tab displays statistics for all trained models that were saved. The first item in the list always represents the current model. Every time a model is trained and saved, the current trained model (if it exists) is overwritten with the new model. Many functions on this tab are enabled only for the current model.

Figure 44: Trained Model Tab

On this tab you can perform the following activities:

•

Save a copy of the current trained model. See Saving and Reviewing Trained Models for more information.

•

Test the model using a selected set of record pairs. See Testing a Trained Model for more information.

•

Set query cutoff score to adjust the balance of false positives to false negatives. Depending on the application, it can be better to reduce false positives at the cost of increasing false negatives or vice versa. See Setting the Query Cutoff Score for more information.

•

Export a model to be used for the prediction of different record pairs from real-world data. See Exporting a Model for more information.