Togaware DATA MINING
Desktop Survival Guide
by Graham Williams
Google

E

E-Commerce: Conduct of financial transactions by electronic means, usually in the form of on-line business and trade, and often involving the buying and selling of goods over the Internet.

Ensemble Learning: A generic description for a learning system that builds multiple knowledge structures and combines the outcomes in some way. For example, a voting stratgey might be used where we have generated multiple Decision Trees for the same task. Each decision tree will be applied to an unseen record and the decisions from each tree will be combined in some way. One of the earliest examples was developed by [#!williams:acj.19.2.1987!#] (also see [#!williams:aida.1988!#] and [#!williams:phd.anu.1991!#]) where multiple decision trees were combined using a majority rules voting scheme. The idea of a committee of models arose from earlier work in expert systems and then in machine learning with MIL (Multiple Inductive Learning).

Evolutionary Computation: Motivated by Darwinian Evolution heuristic rules are employed to modify a population of solutions in such a way that each generation of the population tends to be on average better than its predecessor. These algorithms generally have three components (analogous to the theory of evolution): A fitness function which is used to guide in the selection of the fitest individual; Reproduction operators to generate new structures from other structures; Genetic operators to guide the combination of the structures in reproduction.

Copyright © 2004-2006 [email protected]
Support further development through the purchase of the PDF version of the book.
Brought to you by Togaware.