A Systematic Review of Data Analytics Techniques for Central Tendency–Based Missing Value Replacement (2001–2025)

Author: Sneh B. Patel1, Dr. Darshanaben Dipakkumar Pandya2
Published Online: January 5, 2026
Abstract
References

Incomplete or partially observed data continue to pose a critical obstacle in contemporary analytics and research, often diminishing model accuracy, interpretability, and reliability. The process of compensating for such data gaps—known as imputation—remains essential to maintaining data quality and analytical consistency. Among the broad range of existing strategies, methods grounded in measures of central tendency, such as mean, median, and mode, persist as reliable and computationally economical options. Their enduring popularity arises from their simplicity, transparency, and the ease with which they integrate into larger analytical frameworks. This systematic review consolidates research published between 2001 and 2025 that employs central-tendency-based techniques for missing value replacement across diverse fields, including healthcare, finance, environmental modeling, and social analytics. It compares the empirical performance of these methods, summarizes their practical strengths and weaknesses, and identifies conditions under which each performs optimally. The review reveals that median-based imputation consistently outperforms mean-based approaches in datasets exhibiting skewness or outliers, while mode-based substitution demonstrates superior stability and accuracy for categorical variables. The findings suggest that median-based imputation is particularly effective for skewed or non-normal data, whereas mode-based substitution performs best with categorical attributes. Despite certain limitations—such as reduced variance representation, these methods remain vital reference points for benchmarking more sophisticated approaches. The paper concludes with domain-oriented best practices and prospective directions for enhancing the interpretability, efficiency, and adaptability of imputation processes in future data-driven applications.

Keywords: Central Tendency Measures, Data Analytics, Missing Data, Missing Value Replacement, Statistical Techniques, Systematic Review.
Download PDF Pages ( 198-209 ) Download Full Article (PDF)
←Previous Next →