The wizard will split your data in the mining structure into a training set and a testing set. By default, the data mining ...

The wizard will split your data in the mining structure into a training set and a testing set. By default, the data mining engine will use the training set to train the mining models, and the testing set to test the accuracy of the models. Use the options on this page to specify how much of the input data should be held out for testing. If you set both options, the wizard will use both limits. For example, if the 'Maximum number of rows' is less than the 'Percentage of data for testing', then the 'Maximum number of rows' will be used for testing. If you set the 'Maximum number of rows' to 0, this limit will not be used. The partitions created by this wizard are random, which helps to ensure representative training and testing sets. Because the partitions are created in your mining structure, your source data is not affected.