Ten WEKA sessions: datasets and preprocessing, Apriori and FP-Growth association rules, classification with decision trees, Naive Bayes, k-NN and SVM, regression, and k-means, hierarchical and density-based clustering.
Objectives
Do not copy. Read for understanding and the viva- Handle data mining tasks using a data mining toolkit such as open source WEKA
- Understand the various kinds of algorithms available in WEKA
- Understand the data sets and data preprocessing
- Demonstrate classification, clustering and association rule mining in large data sets
- Exercise the data mining techniques with varied input values for different parameters
- Apply mining techniques for realistic data
Sessions
| Session | Topic | Questions |
|---|---|---|
| Session 1 | WEKA, datasets and ARFF files | Q1 to Q4 |
| Session 2 | Basic preprocessing | Q5 to Q6 |
| Session 3 | Unsupervised filters and Apriori | Q7 to Q8 |
| Session 4 | FP-Growth and Apriori parameter studies | Q9 to Q11 |
| Session 5 | Reading Apriori output | Q12 |
| Session 6 | Apriori on the zoo dataset and regression | Q13 to Q14 |
| Session 7 | Classification: logistic regression, decision tree, Naive Bayes, k-NN, SVM | Q15 to Q16 |
| Session 8 | Decision trees, rules, entropy, kappa and ROC | Q17 to Q18 |
| Session 9 | k-means clustering | Q19 to Q21 |
| Session 10 | Hierarchical and density-based clustering | Q22 to Q23 |
Before the first session
Do not copy. Read for understanding and the viva- Install WEKA 3.8 or later (Java bundled). The sample datasets (weather, iris, glass, contact-lenses, labor, zoo, bank, credit) are in the
datafolder of the installation;student.arffandemployee.arffyou create yourself in Session 1. - For every run, record the exact parameter values, the run information block, and the numbers the formula sheet lets you check by hand (support, confidence, entropy, accuracy, kappa, SSE).
- Screenshots go in the record only when the numbers are also written out; the evaluator marks the interpretation, not the picture.