Missing value imputation using decision trees and decision forests by splitting and merging records: Two novel techniques

• Published in: Knowledge-based systems Vol. 53; pp. 51 - 65
• Main Authors: Rahman, Md. Geaur, Islam, Md Zahidul
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.11.2013
• Subjects: Confidence intervals; Correlation; Data cleansing; Data pre-processing; Decision forest algorithm; Decision tree algorithm; Decision trees; EM algorithm; Forests; Knowledge base; Merging; Missing value imputation; Segments; Similarity; Decision forest algorithm; Missing value imputation; Decision tree algorithm; Data pre-processing; Data cleansing; EM algorithm
• ISSN: 0950-7051; 1872-7409
• DOI: 10.1016/j.knosys.2013.08.023

•Two novel missing value imputation techniques.•Justification of the basic concepts of the techniques through some empirical analyses.•Experimentation on nine data sets, four evaluation criteria.•Comparison with two existing techniques.•A complexity analysis of all techniques. We present two novel techniques for the imputation of both categorical and numerical missing values. The techniques use decision trees and forests to identify horizontal segments of a data set where the records belonging to a segment have higher similarity and attribute correlations. Using the similarity and correlations, missing values are then imputed. To achieve a higher quality of imputation some segments are merged together using a novel approach. We use nine publicly available data sets to experimentally compare our techniques with a few existing ones in terms of four commonly used evaluation criteria. The experimental results indicate a clear superiority of our techniques based on statistical analyses such as confidence interval.