Um Ambiente Orientado a Agentes para Experimentação em Imputação Seqüencial

Rafael CastanedaRonaldo GoldschmidtRicardo Choren

Missing data frequently complicates data analysis for scientific investigations. The development of methods to address missing data has been an active area of research. Imputation denotes a procedure that replaces the missing values in a data set by some plausible values. However, imputation methods usually deal with monotone missingness. Sequential imputation is a general purpose method for analyzing data sets with multiple-variable missing data. In this paper we propose a sequential imputation process, which allows imputed value reuse, and we describe a multi-agent environment that implements this process. This paper also shows some experiments, to illustrate the use of the environment, and a brief discussion about the imputation results.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: