Uncertain Data Analysis using Gray Set based Multiple Imputation with Penalized Optimization Algorithm

Authors

  • G. V. Suresh, Dr. E. Sreenivasa Reddy

Abstract

Many real-world applications face an uncanny predicament of uncertainty plaguing the available knowledge.In the normal run of things, some uncertainties arise due to the prevalence of incorrect measurements and inaccurate decision-making, resulting in unreliable data transmission and data storage.Furthermore, the inevitable randomization that occurs during the physical data generation and gathering process contributes to the aforementioned issue.However, data imperfection issues create lots of trouble for people while working on real-world data wherein data insufficiency emerges as the primary source of trouble. Invalid approaches to handle missing data results in biased outcomes, over-confident intervals, and inaccurate inferences. In recent times, multiple imputation has emerged as a potential alternative to the traditional approaches in missing data analysis. This paper proposes a missing data imputation method i.e., Multiple Imputation–Penalized Optimization Algorithm(MIPOA). By removing the missing-valued items, it first divides the non-missing data instances into various clusters. The entropy of the proximal category is then used to calculate the similarity metric for each incomplete instance using gray relational analysis.Our algorithm is superior to previous approaches in terms of validity, according to experiments on UCI datasets.

Downloads

Published

2022-06-09

How to Cite

G. V. Suresh, Dr. E. Sreenivasa Reddy. (2022). Uncertain Data Analysis using Gray Set based Multiple Imputation with Penalized Optimization Algorithm. Mathematical Statistician and Engineering Applications, 71(3), 252 –. Retrieved from https://philstat.org.ph/index.php/MSEA/article/view/163

Issue

Section

Articles