A Bayesian imputation method for a clustering genetic algorithm

Issue title: Advances in Intelligent Systems

Guest editors: Vassilis Kodogiannisx and Ilias Petrouniasy

Article type: Research Article

Authors: Hruschka, Estevam R.^{a; *} | Hruschka, Eduardo R.^b | Ebecken, Nelson F. F.^c

Affiliations: [a] Federal University of São Carlos, CP 676, 13.565-905 São Carlos, SP, Brazil | [b] Catholic University of Santos, R. Carvalho de Mendonça, 144, 11.070-906, Santos, SP, Brazil | [c] Federal University of Rio de Janeiro, CP 68506, 21945-970, Rio de Janeiro, RJ, Brazil | [x] University of Westminster, Westminster, UK | [y] The University of Manchester, Manchester, UK

Correspondence: [*] Corresponding author. E-mail: estevam@dc.ufscar.br.

Abstract: Missing values are a critical problem in data mining applications. The substitution of these values, also called imputation, can be performed by several methods. This work describes the application of an optimized version of the Bayesian Algorithm K2 as an imputation tool for a clustering genetic algorithm. The resulting hybrid system is assessed by means of simulations in five benchmark datasets. The obtained results indicate that the proposed imputation method is a suitable data preparation tool for the employed clustering genetic algorithm.

Keywords: Missing values, bayesian networks, clustering, genetic algorithms

DOI: 10.3233/JCM-2011-0362

Journal: Journal of Computational Methods in Sciences and Engineering, vol. 11, no. 4, pp. 173-183, 2011