Artificial Intelligence: A Modern Approach

This book is considered a classic in the field of artificial intelligence and is the first book that most undergraduates visit when they are first learning this material. Readers unfamiliar with this book and who want to see what others have said can do so here. To make sure that I understood this material as well as possible these are some notes and problem solutions that I came up with as I worked through the book.

• Here
  

• Chapter 3 (Solving Problems by Searching)
  • Finds the sequence of pours needed to get a desired amount
    
  • Solves the eight puzzle using recursive depth first search
    
  • Solves the missionaries and the cannibals problem
    
• Chapter 18 (Learning From Observations)
  • Generate data for the restaurant problem
    
  • Implements table lookup learning (Problem 18.6)
    
  • Implements DECISION-TREE-LEARNING in R
    
    • Implements code to handle missing attributes
    • Implements code to do chi2 pruning
    • Implements code to use the gain ratio (rather than the information gain) to select the attribute to split on
  • Implements Adaboost with a decision tree as the base learner
    
  • Duplicates the books figure 18.7 (the decision tree learning curve on the restaurant problem)
    
  • Duplicates the books figure 18.11 (Adaboost with decision stumps on the restaurant problem)
    
  • Problem 11: Decision tree learning with chi^2 pruning
    
  • Problem 12: Decision tree learning with missing attribute values
    
  • Problem 14: Evaluating error rates for ensemble classifiers
    
• Chapter 20 (Learning Probabilistic Models)
  • Naive Bayes learning
    
  • Ada-boosting with a naive Bayes as the base learner
    
  • Code to generate data from the majority function
    
  • Code to implement PERCEPTRON-LEARNING
    
  • Runs Naive Bayes learning to compute its learning rate (duplicates the books figure 20.3)
    
  • Duplicates the numbers in the book's Bayes network EM example
    
  • Uses decision trees and a perceptron to learn the majority function
    
  • Uses decision trees and a perceptron to learn the restaurant problem mapping
    
  • Problems 1 & 2: Bayesian learning of the type of candy bag
    
  • Problem 5: Ada-boosting with naive Bayes
    
  • Problem 8: Plots the beta(epsilon,epsilon) distribution