What Is It? The Cross-Validation Wizard helps you understand the accuracy of your data mining models and whether they are ...

What Is It? 
 The Cross-Validation Wizard helps you understand the accuracy of your data mining models and whether they are generally applicable to your data. Cross-validation partitions your data into a number of "folds" and tests the accuracy and error of the model against each fold in turn. If the results for each fold are good, and if each set of results falls in a similar range, then you know that your model generalizes well for all your data. However, if there is a lot of variance between the results for some folds, then you might infer that your model is not applicable to all of your data, and that it requires fine tuning. 

  What Does It Do? 
 The Cross-Validation Wizard enables you to specify the models you want to test. You can also specify the number of folds into which to partition the data. Based on these selections, the application creates and trains as many models as there are folds. It then returns a set of accuracy metrics for each fold in each model, and aggregates the results for the model. The results include the numbers of True and False positive results predicted by the model, along with the Root Mean Square Error and Log Likelihood. You can find more information about these results in our help files.