Pattern Recognition and Machine Learning

9

Learning by Classification and Discovery

KEYWORDS

massive data

decision tree

information theory

noisy data

simultaneous presentation

incremental presentation

Bayes

theorem

concept representation space

learning by discovery

problem solving

heuristic search

discovery of attributes

proportionality graph search

suspended node

heuristic knowledge

frame representation

heuristic rule

meta rule

This chapter will explain methods for learning by classification and discovery. The first half of this chapter describes a classification method using decision trees and Bayesian statistics. A decision tree can be generated by a simple algorithm and is an efficient method of finding regularities of massive data. Bayesian statistics can be applied to a learning ...