Searching for novel classifiers

Published in Conferences Papers
  1. Mario Garc\'ıa-Valdez, Juan J Merelo, Leonardo Trujillo, Francisco Fernández-de-Vega, José C Romero and Alejandra Mancilla. EvoSpace-i: A Framework for Interactive Evolutionary Algorithms. In Proceedings of the 15th Annual Conference Companion on Genetic and Evolutionary Computation. 2013, 1301–1308. URL, DOI BibTeX

    @inproceedings{Garcia-Valdez:2013:EFI:2464576.2482709,
    	author = "Garc\'{\i}a-Valdez, Mario and Merelo, Juan J. and Trujillo, Leonardo and Fern\'{a}ndez-de-Vega, Francisco and Romero, Jos{\'e} C. and Mancilla, Alejandra",
    	title = "EvoSpace-i: A Framework for Interactive Evolutionary Algorithms",
    	booktitle = "Proceedings of the 15th Annual Conference Companion on Genetic and Evolutionary Computation",
    	series = "GECCO '13 Companion",
    	year = 2013,
    	isbn = "978-1-4503-1964-5",
    	location = "Amsterdam, The Netherlands",
    	pages = "1301--1308",
    	numpages = 8,
    	url = "http://doi.acm.org/10.1145/2464576.2482709",
    	doi = "10.1145/2464576.2482709",
    	acmid = 2482709,
    	publisher = "ACM",
    	address = "New York, NY, USA",
    	keywords = "cloud-based platforms, interactive evolutionary computation, interactive systems"
    }
    
Abstract

Natural evolution is an open-ended search process without an a priori fitness function that needs to be optimized. On the other hand, evolutionary algorithms (EAs) rely on a clear and quantitative objective. The Novelty Search algorithm (NS) substitutes fitness-based selection with a novelty criteria; i.e., individuals are chosen based on their uniqueness. To do so, individuals are described by the behaviors they exhibit, instead of their phenotype or genetic content. NS has mostly been used in evolutionary robotics, where the concept of behavioral space can be clearly defined. Instead, this work applies NS to a more general problem domain, classification. To this end, two behavioral descriptors are proposed, each describing a classifier’s performance from two different perspectives. Experimental results show that NS-based search can be used to derive effective classifiers. In particular, NS is best suited to solve difficult problems, where exploration needs to be encouraged and maintained.

Published in
Proceedings of the 16Th European Conference on Genetic Programming (EvoGP'13)
Volume 7831
Pages 145-156
http://link.springer.com/chapter/10.1007%2F978-3-642-37207-0_13
Date of conference
03-05 Abril 2013
ISSN
0302-9743
ISBN
978-3-642-37207-0
Feedback