Reference no: EM132491403
Decision Trees
For the following two projects, use WEKA to build a Decision Tree. Choose the Trees >> J48 model from the Classify tab.
Use Cross-validation with 10 folds. This builds 10 trees, using a different random 10% slice of the whole dataset as the TEST set. Then averages the 10 resulting accuracies together as a best-guess real-world performance rating.
1) Use WEKA to build a Decision Tree on the Iris data set
Append the cross-validation tree's results to the TEST columns of the Iris model-comparison table you have been building all semester. Train a separate model on the whole dataset as a Training Set and append those results in your TRAIN columns.
Display the complete table and discuss how this model's results compare to the rest.
2) Use WEKA to build a J48 Decision Tree on your favorite classification-style dataset you select and download from the UCI Machine Learning Repository
For both trees, paste a copy of WEKA's complete text-output page for your final J48 model into your report.
Also paste a screenshot of WEKA's tree visualisation.
Interpret these results for me with as much detail as possible.
Attachment:- Decision Trees.rar