Statistical Learning from a Regression Perspective by Richard A. Berk

Author:Richard A. Berk
Language: eng
Format: epub, pdf
Publisher: Springer International Publishing

2.Drop the evaluation data down each tree, and compute the fitted values. For classification trees, then construct a evaluation data confusion table. From those tables, performance measures can be obtained.

3.Select a “best” classification tree.

4.Using the selected tree, drop the test data down the selected tree and compute the fitted values.

5.Cross-tabulate the predicted outcome classes in the test data by the actual test data response classes. Construct a confusion table from that cross-tabulation. That confusion table provides an asymptotically unbiased estimate of the population confusion table approximation if the population realizations were dropped down the selected tree. The overall misclassification rate provides an estimate of generalization error. If misclassification costs are asymmetric, the different classification errors should be weighted by their relative costs if an overall measure of performance is desired.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.