Evaluating a Decision Tree Model

You can evaluate the results of a decision tree model. The purpose of valuation is to verify the validity or accuracy of the training result using historic data. You can do this by using the tree to classify a separate set of data whose outcomes are already known. If you compare the predicted outcome with the known outcome, you can easily discover the number of correct predictions and ones that were not correctly predicted.  This information can then be displayed in the form of a matrix, called the Error Matrix.You use this matrix to know which outcome values the tree predicts well and the values that the tree doesn't predict properly.

• The model must be fully created and activated
• The analysis process must be activated

To evaluate a model, you must follow the steps described inExecuting the Prediction. The only difference is that you need to check the option Run in Evaluation Mode.