InfoEvolve: moving from data to knowledge using information theory and genetic algorithms.

Journal Article

InfoEvolve is a unified suite of data mining and empirical modeling tools capable of discovering low-bias and low-variance solutions to complex processes. The method is based on a common set of principles involving information theory and genetic algorithms. InfoEvolve can also discover multiple strategies embedded in complex data sets for achieving a desired target or goal. This latter aspect may prove to be very useful in drug design. The paper analyzes the following: InfoEvolve from a theoretical standpoint; a conceptual overview of InfoEvolve with a short description of the modeling method; the method using the example of homogeneous identification of DNA from an analysis of its melting curve behavior; and key learnings and additional applications of the technology for both drug design and genome analysis.

Full Text

Duke Authors

Cited Authors

  • Vaidyanathan, G

Published Date

  • May 2004

Published In

Volume / Issue

  • 1020 /

Start / End Page

  • 227 - 238

PubMed ID

  • 15208195

International Standard Serial Number (ISSN)

  • 0077-8923

Digital Object Identifier (DOI)

  • 10.1196/annals.1310.019

Language

  • eng

Conference Location

  • United States