A genetic algorithm approach to recognition and data mining
- Michigan State Univ., East Lansing, MI (United States); and others
We review here our use of genetic algorithm (GA) and genetic programming (GP) techniques to perform {open_quotes}data mining,{close_quotes} the discovery of particular/important data within large datasets, by finding optimal data classifications using known examples. Our first experiments concentrated on the use of a K-nearest neighbor algorithm in combination with a GA. The GA selected weights for each feature so as to optimize knn classification based on a linear combination of features. This combined GA-knn approach was successfully applied to both generated and real-world data. We later extended this work by substituting a GP for the GA. The GP-knn could not only optimize data classification via linear combinations of features but also determine functional relationships among the features. This allowed for improved performance and new information on important relationships among features. We review the effectiveness of the overall approach on examples from biology and compare the effectiveness of the GA and GP.
- OSTI ID:
- 466461
- Report Number(s):
- CONF-9610138-; TRN: 97:001309-0041
- Resource Relation:
- Conference: International multi-disciplinary conference on intelligent systems: a semiotic perspective, Gaithersburg, MD (United States), 21-23 Oct 1996; Other Information: PBD: 1996; Related Information: Is Part Of Intelligent systems: A semiotic perspective. Volume I: Theoretical semiotics; Albus, J.; Meystel, A.; Quintero, R.; PB: 303 p.
- Country of Publication:
- United States
- Language:
- English
Similar Records
Analytical optimal pulse shapes obtained with the aid of genetic algorithms
Using genetic algorithms to select and create features for pattern classification. Technical report