The missing data is one of the common problems of data quality. Most of the real datasets have missing values. Imputing the missing values makes the analysis easier by creating a complete dataset as it eliminates the problem of handling complex patterns of missingness. The conventional methods for imputations are easy to implement but introduce biasness in the data. Under certain assumptions modern and hybrid methods perform better. For missing data compensation, a variety of methods have been developed. This paper gives the comprehensive overview of the various imputation methods for missing data compensation. Also various hybrid techniques for missing value imputation along their benefits and limitations have been discussed. The objective of this review is to draw special attention on possible improvement of existing methods and to give readers better understanding of trends for imputation methods.
Volume 11 | 07-Special Issue
Pages: 312-318