Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{208754,
        author = {Aishwarya Kolhe and Geeta Yadav and Dnyaneshwari Chalak and Dr. Shital Ghotekar},
        title = {A Review of Missing Data Handling Techniques in Data Science},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {no},
        pages = {660-665},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=208754},
        abstract = {Incomplete data is one of the most common quality problems encountered in data science. Missing values can occur because of non-response, faulty sensors, data-entry mistakes, transmission failures, privacy restrictions, or differences between data sources. If missing values are handled without considering the reason for missingness, the resulting analysis may be biased and machine-learning models may lose predictive reliability. This paper presents a review of major approaches for handling missing data, beginning with three classical missingness mechanisms: Missing Completely at Random (MCAR), Missing at Random (MAR), and Missing Not at Random (MNAR) [1], [2]. The reviewed methods include deletion, statistical imputation, regression-based imputation, multiple imputation, expectation-maximization, K-nearest-neighbour imputation, random-forest-based methods, and deep-learning approaches [4][10]. These techniques are compared in terms of their assumptions, information retention, computational requirements, scalability, and suitability for different data conditions. The review emphasizes that the effectiveness of an imputation method depends strongly on the mechanism and pattern of missingness. Simple methods may be suitable for small amounts of approximately random missingness, whereas advanced approaches can be useful when variables have strong relationships or the dataset is high-dimensional. Important challenges include uncertainty about the missingness mechanism, computational cost, evaluation when the actual values are unavailable, and the possibility of introducing or amplifying bias. The paper concludes by identifying future directions involving adaptive, uncertainty-aware, scalable, and fairness-conscious missing-data frameworks.},
        keywords = {Missing Data, Data Imputation, MCAR, MAR, MNAR, Data Preprocessing, Machine Learning.},
        month = {September},
        }