The depth is reduced to 4 and the accuracy is

This preview shows page 9 - 14 out of 18 pages.

to 70, the complexity of the tree is reduced to 9 number of nodes and 5 number of terminal nodes. The depth is reduced to 4 and the accuracy is decreased to 57.5% for the training and 52.2% for the testing. The top three importance variables remained at alcohol, sulphates and volatileacidity.3.Now bin the class variable in such a way that data is not so imbalanced with respect to the class variable. Repeat Problem 1 but on the wine data with less number of classes (the binned class variable).The data was binned into two classes. The first class consists of quality classes 3, 4, and 5. The second class consists of classes 6, 7, and 8. The first class makes up 47% of the data while the second class holds 53% of the data so the two classes are balanced.
DSC441- Fall 2018, Assignment 3, Page 10 of 18
DSC441- Fall 2018, Assignment 3, Page 11 of 18
DSC441- Fall 2018, Assignment 3, Page 12 of 18
4.How the performance of the best classification model on the original class variable compares with the accuracy of the best classification model on the binned classification variable?
DSC441- Fall 2018, Assignment 3, Page 13 of 18
Cross-Validation:Holdout Partitioning:

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture