Fitness function comparison for GA-based feature construction

Leila S. Shafti, Eduardo Pérez

Research output: Chapter in Book/Report/Conference proceedingConference contribution book

6 Citations (Scopus)

Abstract

When primitive data representation yields attribute interactions, learning requires feature construction. MFE2/GA, a GA-based feature construction has been shown to learn more accurately than others when there exist several complex attribute interactions. A new fitness function, based on the principle of Minimum Description Length (MDL), is proposed and implemented as part of the MFE3/GA system. Since the individuals of the GA population are collections of new features constructed to change the representation of data, an MDL-based fitness considers not only the part of data left unexplained by the constructed features (errors), but also the complexity of the constructed features as a new representation (theory). An empirical study shows the advantage of the new fitness over other fitness not based on MDL, and both are compared to the performance baselines provided by relevant systems.
Original languageEnglish
Title of host publicationCurrent Topics in Artificial Intelligence
Subtitle of host publication12th Conference of the Spanish Association for Artificial Intelligence, CAEPIA 2007, Salamanca, Spain, November 12-16, 2007, Selected Papers
EditorsDaniel Borrajo, Luis Castillo, Juan Manuel Corchado
Place of PublicationCham
PublisherSpringer
Pages249-258
Number of pages10
ISBN (Electronic)9783540752714
ISBN (Print)9783540752707
DOIs
Publication statusPublished - 7 Nov 2007
Event12th Conference of the Spanish Association for Artificial Intelligence - Salamanca, Spain
Duration: 12 Nov 200716 Nov 2007

Publication series

NameLecture Notes in Computer Science
Volume4788
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference12th Conference of the Spanish Association for Artificial Intelligence
Abbreviated titleCAEPIA 2007
Country/TerritorySpain
CitySalamanca
Period12/11/0716/11/07

Keywords

  • attribute interaction
  • entropy
  • feature construction
  • feature selection
  • genetic algorithms
  • machine learning
  • MDL principle

Fingerprint

Dive into the research topics of 'Fitness function comparison for GA-based feature construction'. Together they form a unique fingerprint.

Cite this