Exploring Imputation Techniques On Single and Ensemble Medical Classification

This study investigates the efficacy of four imputation techniques K-Nearest Neighbors (KNN-I), Multilayer Perceptron (MLP-I), Support Vector Machine (SVM-I), and Decision Trees (DT-I) on six single/ensemble classifiers (MLP, KNN, SVM, XGB, RF, BAGGED SVM) across seven distinct medical da-tasets. The analysis revealed that the best classification performance was achieved using MLP-I, across single and ensemble classifiers. Further, En-sembles enhanced classification accuracy, demonstrating superiority over single classifiers irrespective of the imputation technique used. The findings also underscored the importance of using imputation techniques in optimiz-ing medical data analysis and improving diagnostic accuracy.