SEG 4630 ECommerce Data Mining Midterm Review 392021

SEG 4630 E-Commerce Data Mining — Midterm Review — 3/9/2021 E-Commerce Data Mining 1

Midterm Time/Location n Time: 10: 30 -12: 15 am Oct. 29, Thursday n Location: LT 2 of LSB n Coverage: Chaps 1, 2, 4, 5, 6 (to slide # 51) n You can bring an A 4 size, double-sided cheat sheet n You may bring a calculator, but cannot use your cell phone 3/9/2021 E-Commerce Data Mining 2

Chapter 2 (1) n Types of attributes n Calculate data distribution n n Calculate distance between data objects n n n 3/9/2021 Mean, median, variance and standard deviation Minkowski distance between binary variables n Simple matching and Jaccard coefficients Cosine similarity E-Commerce Data Mining 3

Chapter 2 (2) n Data normalization n n Data reduction n 3/9/2021 Min-max normalization Z-score normalization Decimal scaling Dimensionality reduction methods Feature subset selection Sampling E-Commerce Data Mining 4

Chapters 4 -5 (1) n Decision tree n n n Bayes theorem and Naïve Bayesian n n 3/9/2021 Calculate probabilities from training datasets Lazy classifier and k-nearest neighbor n n Calculate information gain, gini index, gain ratio Tree construction/induction and pruning Calculate based on different k values and different distance measures Differences between eager and lazy classifiers E-Commerce Data Mining 5

Chapters 4 -5 (2) n Accuracy and error measures n n ROC curve n n 3/9/2021 True positive rate (TPR) and false positive rate (FPR) Area under curve (AUC) Evaluation methods n n Training error vs. validation error, overfitting Confusion matrix Precision, recall, F-measure Hold out Cross validation Ensemble, bagging E-Commerce Data Mining 6

Chapter 6 (1) n Frequent patterns and association rules n n n Apriori algorithm n n 3/9/2021 Support, confidence Generate association rules from frequent itemsets Candidate generation and test Self joining Pruning Database scan E-Commerce Data Mining 7

Chapter 6 (2) n FPgrowth algorithm n n Build FP-tree Extract conditional DB n Closed itemsets and maximal itemsets n Factors affecting complexity 3/9/2021 E-Commerce Data Mining 8

Questions? 3/9/2021 E-Commerce Data Mining 9