Affiliations: Applied Computational Intelligence Laboratory, Department of Computer Science, Pontifical Catholic University of Minas Gerais, Minas Gerais, Brazil
Correspondence:
[*]
Corresponding author: Luis E. Zárate, Department of Computer Science, Pontifical Catholic University of Minas Gerais, Minas Gerais, Av. Dom José Gaspar 500, Coração Eucarístico, Belo Horizonte, 30535-610, MG, Brazil. E-mail: zarate@pucminas.br.
Abstract: Missing data is a significant problem found in data mining projects and data analysis. Despite being a common problem, the missing data is dealt in a simplistic way and may lead to inconsistent knowledge discovery. Through literature review, it was possible to observe that the missing data mechanisms are not always considered when methods of treatment or imputation are chosen. This work presents a review about the main treatment methods of missing data that can be considered in a process of knowledge discovery in database. We emphasize that attention should be given to the identification of the absence mechanism, for the choice the most appropriate treatment method.
Keywords: Missing data, missing-data mechanism, data mining
