Note about Bias in Bayesian Genetic Algorithms for Discrete Missing Values Imputation

A Genetic Algorithm is a sophisticated searching technique that finds the best possible solution in the solution space. The search process is driven by a fitness function which measures the fitness level of a candidate solution. The chosen fitness function varies according to the problem being solve...

Full description

Saved in:
Bibliographic Details
Published in2019 International Arab Conference on Information Technology (ACIT) pp. 87 - 90
Main Authors Migdady, Hazem, Alrabaiah, Hussam, Al-Talib, Mohammad
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.12.2019
Subjects
Bayes Theorem
Data Mining
Genetic Algorithms
Missing Values Imputation
Online AccessGet full text
DOI10.1109/ACIT47987.2019.8991024

Cover

More Information
Summary:A Genetic Algorithm is a sophisticated searching technique that finds the best possible solution in the solution space. The search process is driven by a fitness function which measures the fitness level of a candidate solution. The chosen fitness function varies according to the problem being solved. However, any fitness function should satisfy some conditions. In the Bayesian genetic algorithm for missing values imputation we noted that the used fitness function performs poorly when be applied with datasets that contain missing values, since such datasets hold a level of bias which would adversely affect the efficiency of the entire searching process. In this paper we mentioned that some assumptions should hold in order to apply the technique of Bayesian genetic algorithm efficiently.
DOI:10.1109/ACIT47987.2019.8991024