Overall Model Quality

The Overall Model Quality section shows the name of the data set that was used to build the model. It also shows whether the data set was used to train the model, or whether it was used for the validation or the testing of the model quality.

The scale in the Overall Model Quality section shows the overall quality of the model. The overall model quality is based on quality details in terms of accuracy, reliability, and the ability to sort the data.

The overall model quality compares the predictive performance of a model to a trivial model that randomly assigns one of the possible values to the target field according to their occurrences in the data set.
  • An overall model quality value close to 0 indicates a very poor model with respect to predictive accuracy, reliability, and the ability to rank data.
  • An overall model quality value close to 1 indicates an excellent model that correctly classifies the data, ranks the records correctly, and is maximally reliable.
The overall model quality value depends not only on the model itself, but also on the validation data or the test data that was used. Even if the overall model quality value is high, it can only be assumed that model application produces high quality results if the application data is similar to the validation data or the test data that was used.