Section 1 : Decision Tree Learning

Commentary

Section Goals

• To introduce some basic concepts of machine learning.
• To discuss the decision tree learning method and its algorithms.

Learning Objectives

Learning Objective 1

• Outline the three different types of learning methods.
• Explain the general principles of inductive learning.
• Describe decision trees, and the methods for learning decision trees.
• Exemplify a decision tree, and illustrate the construction process.
• Discuss the issues that have emerged from decision tree learning, such as attribute choosing, noise and overfitting, and performance assessing.
• Explain the following concepts or terms:
  • Supervised learning
  • Unsupervised learning
  • Reinforcement learning
  • Inductive learning
  • Hypothesis space
  • Decision tree
  • Training set and test set
  • Information gain
  • Learning curve
  • Decision tree pruning
  • Overfitting

Objective Readings

Required readings:

Reading topics:

Inductive Learning, Forms of Learning, Learning Decision Trees (see Chapter 18.1-18.4 of AIMA3ed)

Objective Questions

• What are the main differences between supervised, unsupervised, and reinforcement learning?
• How should you choose the attributes during decision tree learning to improve the efficiency of the learning process?
• How should you avoid overfitting?

Objective Activities

• Explore some application scenarios of decision tree learning applications in business, and discuss them in the course conference.
• Explore the following program code for the decision tree learning algorithm from the textbook's website.
  • Decision-Tree-Learning
• Complete Exercise 18.3 of AIMA3ed.
• Complete Exercise 18.4 of AIMA3ed.