Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Media Statistika
Titin Siswantining
Department of Mathematics, Universitas Indonesia, Jakarta, Indonesia
Decision Sciences, Operations Research & Management

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
MACHINE LEARNING TO HANDLE MISSING VALUES, PREDICTION AND DETECTION OF RISK FACTOR FOR ATRIAL FIBRILLATION, DIABETES MELLITUS TYPE 2, AND PREDIABETES IN OBSTRUCTIVE SLEEP APNEA (OSA) PATIENTS Kurnia Susvitasari; Yuni Rosita Dewi; Titin Siswantining; Diana Nur Vitasari; Felicia Erinna Puspitaningtyas; Telly Kamelia
MEDIA STATISTIKA Vol 18, No 2 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.2.1-12

Abstract

Machine learning is a discipline of artificialintelligence. It enables computational systems to exhibitbehaviors derived from empirical data, without explicitprogramming for the task. Algorithms are systematicallyapplied to various tasks in the biomedical and healthcare sectors,including classification, regression, clustering, dimensionalityreduction, model selection, and data preprocessing. This clinicalstudy employs machine learning techniques to resolve missingdata anomalies and to create predictive models for the risk ofAtrial Fibrillation (AF), Type 2 Diabetes Mellitus (DM), andPrediabetes in patients with Obstructive Sleep Apnea, in a casestudy at dr Cipto Mangunkusomo Hospital. The assessment ofvarious imputation and deletion methodologies (listwisedeletion, zero-imputation, central tendency substitutions, andfuzzy c-means clustering) was conducted to address thestructural incompleteness of the dataset. The disease risk wascategorized using several algorithms: Support Vector Machines,Random Forests, Extreme Gradient Boosting, and MultilayerPerceptrons. A missing value proportion of 25% or less wasobserved during the preliminary diagnostic. The empiricalfindings show that the classification accuracies for detecting AFwere 0.908, for Type 2 DM were 0.9245, and for Prediabeteswere 0.7924. The results indicate that the optimal machinelearning architecture and data imputation strategy varydepending on the specific clinical targets.
Co-Authors Diana Nur Vitasari Felicia Erinna Puspitaningtyas Kurnia Susvitasari Telly Kamelia Yuni Rosita Dewi