Masinõpe II 2018/19 kevad

II. Decision trees and association rules

Given by Sven Laur

• Home exercises to the II session
• Materials for the exercise session

Video Tutorial

Here is the set of videos to acquaint you with association rules and decision trees. Watch them in correct order.

Exercise 1

• 01 Rules - Support, Confidence, Coverage
• 02 Rules - Rule-based prediction in R
• 03 Rules - Confusion matrix, Precision, Recall
• Now start by solving exercise 1 from the exercise sheet

Exercise 2 and 3

• 04 What is a Decision Tree Algorithm
• 05 How to Compute Probability
• 06 Evaluating splits - Entropy of a Target Variable
• 07 How to Compute Entropy
• 08 Evaluating splits - Information Gain
• 09 Enumerate all Possible Attributes
• 10 Split According to an Attribute
• 11 Indexing in GRU R
• 12 Subsetting a Dataset in R
• 13 Lists in GNU R
• 14 Enumerate All Possible Splits
• 15 Enumerate All Possible Splits in R
• 16 Idea of the Recursive ID3 Algorithm
• 17 Recursive Splitter in R
• 18 When to Stop Splitting
• Now you are set to attack exercises 2 and 3

Exercise 4

• 19 Convert Decision Tree Into a Set of Rules
• This should help you with the 4th exercise

Remaining exercises

• Proceed with more open-ended exercises 5, 6 and 7

Attach:boo Δ

Deadline: 1st of March 16:15 EET