Dynamic Imputation Techniques for Enhancing Predictive Accuracy in Healthcare Data

Big data is the representation of large datasets, while analytics is the use of mathematical tools. Together, these two fields comprise "big data analytics," which uses cutting-edge digital approaches to find hidden patterns in datasets. Missing data poses a significant challenge in healthcare analytics, impacting the accuracy and reliability of predictive models. In this paper, we propose a dynamic missing value imputation module tailored to structured healthcare data. Our approach adapts imputation methods based on identified missingness mechanisms and percentages, ranging from simple techniques like mean imputation to advanced methods such as Multiple Imputation by Chained Equations (MICE) and predictive modeling using machine learning algorithms. Through comprehensive evaluations across diverse datasets, including complaint classification and life expectancy prediction, we demonstrate the module's effectiveness in improving dataset quality and predictive performance.