Comparing classifiers: use WEKA Explorer for this project: Solution
1. Load dataset sattrn.arff, loaded in the folder "Assignments".
2. Create the following classifiers using the "supplied test set" option using the test set sattst.arff, also loaded in the folder "Assignments":
a) A Decision Tree using J48
b) A Naive Bayesian classifier (listed as "NaiveBayes" under "bayes")
c) A Bayesian network classifier (listed as "BayesNet" under "bayes") with default options
d) A KNN classifier (listed as "IBk" under "lazy"). Build three classifiers with 3, 5, and 7 as the
number of neighbors
Note: All Classifiers are under the ‘Classify’ Tab.
3. Compare the confusion matrix of each of the classifiers created above and identify the best classifier. Justify your choice.
4. For the J48 classifier, answer the following questions: a) What are the values for TP, TN, FP, and FN for class value = 1? b) What are the sensitivity, specificity, and FPR for class value = 1? Show your calculations.
5. Include the screenshots of Explorer for each classifier generated above.