Code and Results for Chapter 10 (Boosting and Additive Trees):

• gen_eq_10_2_data.R (generates classification data according to equation 10.2)
  
• dup_fig_10_2.R (duplicates figure 10.2 (boosting applied to equation 10.2))
  
• dup_fig_10_4.R (duplicates figure 10.4 (loss functions for two-class classification))
  
• dup_spam_classification.R (duplicates the SPAM classification example)
  
• dup_fig_10_9.R (duplicates figure 10.9 (boosting with different sized trees))
  

• Exercise 10.4 (A MATLAB implementation of AdaBoost)
• ex_10_4_Script.m (a script to replicate Figure 10.2)
  
• gen_data_pt_b.m (generates the suggested training data for parts A and B)
  
• AdaBoost Code (developed by Kelly Wallenstein and John Weatherwax):
  • k_adaboost_training.m (boosts binary classification trees by selecting difficult patterns)
    
  • k_adaboost_evaluate.m (evaluates the binary classification tree using adaboost voting)
    
  • k_classify_w_tree.m (uses a binary tree to classify input patterns)
    
  • k_find_split.m (find the variable and value which splits a set of data producing the greatest decrease in impurity)
    
  • k_make_tree.m (recursively builds a classification tree)
    
  • k_make_tree_n_test.m (builds a binary tree using training data and computes in and out of sample classification error)
    
• From Kevin Murphy's Bayes Net Toolbox:
  • sample_discrete.m (sample discrete random variables according to a specified probability mass function)