Code and Results for Chapter 7

• treepredict.py
  
  • added the ability to the function mdclassify to be able to classify an instance vector that contains a tuple
  • added the ability to the function buildtree to accept a minimum gain parameter
  • fixed bug in mdclassify on how to aggregate counts when a feature is None
  • fixed bug in buildtree by recursively pass all input arguments into the recursive function calls
• run_all.py (runs and demonstrates all commands from the text)
  
• chap_7_ex_1.py (a function that turns counts into probabilities)
  
• chap_7_ex_2.py (exercises the ability to classify an instance with a tuple)
  
• chap_7_ex_3.py (demonstrates the ability to stop building a classification tree when the entropy reduction is not great enough)
  

Alternative Classification Tree Codes:

• A Matlab implementation of AdaBoost based on
  • The Elements of Statistical Learning by Friedman, Hastie, and Tibshirani