Decision Tree - Output Interpretaion




I built Decision Tree in R. I has 7 leaf nodes with Probability %.

Now I want to know , what are the main statistics we have look in Decision Tree Output.

How to validate Decision Tree whether it is classifying correctly or whether it is predicting correctly.

As in Logistic Regression as per i know we have different Statistics like %Concordance, C Value, HL Test like that…in Decision Tree what are the main statistics to consider…




@kalyan - in decision tree in R the rpart package help to pictorize the tree and at each node there is value of classification percentage by which you can understand how tree is classifying at each node.

Hope this helps!


Yes i can see the values. Is there any tests/values to know the performance of tree?.



You can directly compare the predictions using the confusion matrix, the common metrics are:

Accuracy =Number of correct predictions/Total number of predictions


Error rate =Number of wrong predictions/Total number of predictions

You can also use precision or recall if the observations are skewed.

Alternatively, you should go through this:

Hope this helps!