Pattern-based Comparative Analysis of Techniques for Missing Value Imputation

摘要

Missing data is common in real-world problems and can affect the any statistical analysis significantly. A common way of dealing with this problem is to fill in the missing value. The technique used to impute data should be chosen very carefully to avoid incorrect inference about the data. This paper presents a comparative analysis of various techniques that can be used for imputing missing values in cross-sectional data under different scenarios. We use several algorithms on a variety of datasets to evaluate their performance in imputing missing values. Finally, we observe that a carefully chosen imputation technique helps in minimal distortion of analysis results.