ML: Review

• in Witten: Chapter 6 instance methods, clustering, EM, Bayesian Networks) 6.5, 6.7, 6.8
• Exam on Thursday: There will be a lecture/workshop after the exam, there will be one takehome problem due in Week 10.

Review for the Exam

• Bayes Rule, Naïve Bayes estimates
• Concepts, instances, attributes
• kmeans clustering (EM is optional)
• Decision trees
• Information gain
• decision rules
• coverage and accuracy
• regression and linear models
• Perceptron and neural networks
• ROC, you don’t need to memorize the formula, but know what it is
• type I and type II errors

Bayesian Networks

• What makes Naïve Base naïve