Rattle R package: What is the best metric to evaluate model performance of logistic regression model and decision tree model?

Question

I have a data set asked to be evaluate in two models : logistic regression and decision tree. What is the best metric to evaluate these two model performance?

This should be on Cross Validated as it's not a programming question. The answer also hinges on what your goal is. Probably prediction, since you're using a decision tree, but you should state your goal to be clear. — Gregor Thomas, Mar 09 '15 at 16:14

score 0 · Answer 1 · answered Mar 12 '15 at 01:08

There are many ways to measure the performance of a classification method. Some of them are Accuracy, ROC curve, AUC (area under ROC curve), precision, recall, F1 and so on. ROC curve and the AUC are not dependent to any specific cut point. Rattle has the ability to generate ROC curve and calculate AUC.

Rattle R package: What is the best metric to evaluate model performance of logistic regression model and decision tree model?

1 Answers1