Random Forest Specifications - Stopping Condition Tab

Select the Stopping Condition tab of the Random Forest Specifications dialog box to access options that control the end of training. Specifically, there are stopping conditions to determine if training should be terminated before the full number of trees are added to the Random Forest model.

Enable advanced stopping condition

Select this check box to enable early stopping of the Random Forest training algorithm, (stop adding trees before the full number of trees are added to the model). Selecting this option will make the rest of the controls on this tab available.

Cycles to calculate mean error

Specifies the number of cycles across which improvement is measured. The Random Forest algorithm may demonstrate noise on the training and test errors. Therefore, it is not usually a good idea to halt training on the basis of a failure to achieve the desired improvement in error rate over a single cycle. This field specifies a number of cycles over which the error rates are monitored for improvement. Training is only halted if the error fails to improve for that many cycles.

Percentage decrease in training error

Specify a minimum percentage of improvement (drop) in error that must be made; if the rate of improvement drops below this level, training is terminated.