Data mining (WEKA)

Q1) A University needs to decide which students to accept. Its concern is to accept students who do not make problems and have academic attitude. From the history of the university, we collected the following sample.

• Student 1 had GPA 88% at school, male, is member of many social groups, and this student was good.
• Student2 had GPA 80% at school, female, is not a member of many social groups, and this student was good.
• Student 3 had GPA 62% at school, female, is not member of many social groups, and this student was good.
• Student 4 had GPA 65% at school, male, is not member of many social groups, and this student was bad.
• Student 5 had GPA 67% at school, male, is member of many social groups, and this student was good.
• Student 6 had GPA 70% at school, male, is not member of many social groups, and this student was bad.
• Student 7 had GPA 67% at school, male, is not member of many social groups, and this student was bad.
• Student 8 had GPA 80% at school, male, is member of many social groups, and this student was good.

Six new students applied to the university. Their information was:

• Student2 had GPA 80% at school, male, is not a member of many social groups
• Student 3 had GPA 62% at school, male, is not member of many social groups.
• Student 4 had GPA 55% at school, male, is not a member of many social groups
• Student5 had GPA 84% at school, male, is a member of many social groups
• Student 6 had GPA 52% at school, female, is not member of many social groups.

1. Use Weka to predict which students would be good/bad. Use SVM classifier(SMO) under functions folder.

2. Assume that the correct decision set for the testing sample is {good, good, bad, bad, good, good}, find the accuracy of SMO classifier.

3. Find the accuracy of the J48 classifier under the trees folder.

4. Print the decision tree.