Combination of Boruta's Algorithm and Lasso Logistic Regression in Establishing Blood-Based Biomarkers for Alzheimer's Disease

• Published in: 2024 5th International Conference on Artificial Intelligence and Data Sciences, AiDAS 2024 - Proceedings
• Main Author: Abdullah M.N.; Wah Y.B.
• Format: Conference paper
• Language: English
• Published: Institute of Electrical and Electronics Engineers Inc. 2024
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85209669778&doi=10.1109%2fAiDAS63860.2024.10730261&partnerID=40&md5=8356edf1dbb33e5a9e143d93090cf436

Alzheimer's disease (AD) is the leading neurodegenerative disorder worldwide, significantly impacting morbidity, disability, and mortality rates. This study aims to enhance the early diagnosis of AD by identifying blood-based biomarkers using Boruta's algorithm and lasso logistic regression. The Gene Expression Omnibus (GEO) dataset GSE140829 identified 47,008 potential biomarkers, which preprocessing reduced to 22,372. Boruta's algorithm further narrowed these to 21 significant biomarkers. Lasso logistic regression was applied to this balanced dataset, resulting in six key biomarkers: CST7, DPAGT1, DZIP1, MOSC1, TACC3, and YDJC. The final model, validated with a C-index of 0.763, demonstrated good discriminative power and clinical utility. This research offers a robust method for predicting AD, potentially improving early diagnosis and guiding treatment decisions. © 2024 IEEE.