Clearing feedback data for text analytics models
Clearing feedback is essential in the production environment if you have already downloaded and used the feedback to train a new version of the model.
If you fail to clear the old feedback data before you work on the next model version, you retrain the model on the old feedback. As a result, the model becomes biased.
By clearing the feedback data at the end of a model development cycle, you can collect new feedback for the next version of the model. The next time you export feedback data, the data does not include the feedback that you already used to train the previous version of the model. For more information, see Increase the accuracy of text analytics models with imported feedback data.
In the navigation pane of Prediction Studio, click Models.
Click the model for which you want to clear feedback data.
In the header of the model work area, click.
In the Clear feedback dialog box, in the Clear feedback for language list, select the language for which you want to clear feedback, and then click Yes.The system deletes the feedback data for the selected language.