Reviewing Results of a Trained Model
The Trained model tab displays the results of any trained model that was saved.
Figure 50: Review Results
Displaying Results for a Specific Dataset
After selecting a trained model in the list, you can display results for one of the following:
| • | The validation dataset saved with the selected model. |
| • | The training dataset saved with the selected model |
| • | The latest manual test that used a specific dataset from a selected project and model. See Testing a Trained Model for details. |
Testing Results
This section displays statistics of model predictions for the selected dataset:
| • | Error Rate |
| • | False Positive Rate |
| • | False Negative Rate |
| • | Percentage of No Confidence Predictions |
| • | Average Confidence |
Refer to Results for Validation Dataset, for more information.
Links provided can be used to review a filtered list of pairs that correspond to the specific item in the “Testing Results” section.
Deleting a Model
To delete a saved model together with features and datasets associated with it, do the following:
| 1. | Select the saved model in the grid. |
| 2. | Click Delete. |
| 3. | Click Yes to confirm your decision. |
The selected saved model and datasets related to it are deleted.
Note: If the current model is selected, this button only deletes the trained model and model results, but features and datasets are preserved. Typically this significantly reduces the size of the project directory.