This function is only available in the Enterprise Solution.
Your categories are listed in the overview, as well as the Precision, Recall, F1-mass and the amount of uploaded examples per category. These values are connected to the Confusion Matrix. In the following, the meaning of these values is explained:
Precision describes the ratio of correctly classified examples in relation to all examples classified in this category, correct as well as incorrect.
For example: The training data contains 100 examples that are part of the category «Credit Card». The classifier predicts the credit card category based on 88 examples, but only 53 are correct, which means that the precision is 60% (53 / 88 = 0.6 = 60%).
Recall indicates the ratio of the correctly classified examples in relation to all examples of this category, as well as the incorrect examples (meaning examples that were incorrectly assigned to another category).
For example: There are 100 examples inside your training data that were predicted as belonging to the category «Credit Card». Only 53 of these 100 examples are actually assigned to the Credit Card category by the classifier. This leads to a recall of 53% (53 / 100 = 0.53 = 53%).
F1-Mass is the weighted mean between precision and recall within one category.
Precision, Recall and F1-Mass should be as high as possible. The smallest acceptable value is 0 and the highest is 1. The ideal value for Precision, Recall and F1-Mass therefore is = 1. Realistically, however, this value will rarely be reached.