A Data-Driven Approach To Detect Remote Work Burnout using Machine Learning Models

Rizky, Muhammad (2026) A Data-Driven Approach To Detect Remote Work Burnout using Machine Learning Models. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026201102-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Perubahan menuju kerja jarak jauh telah mengubah dinamika tempat kerja secara signifikan. Meskipun menawarkan fleksibilitas, sistem ini juga menimbulkan tantangan seperti batasan kerja–hidup yang kabur, peningkatan beban kerja, serta isolasi sosial. Dengan menganalisis data primer yang dikumpulkan melalui survei terstruktur terhadap pekerja jarak jauh di seluruh Asia Tenggara, penelitian ini melanjutkan studi sebelumnya untuk mengidentifikasi faktor-faktor penting yang berkontribusi terhadap tingginya tingkat stres dan burnout.Penelitian ini mengevaluasi kemampuan model dalam mengidentifikasi pekerja yang berisiko tinggi mengalami burnout atau stres berat dengan menggunakan teknik machine learning seperti XGBoost, Support Vector Machines (SVM), Ridge Regression, dan Lasso Regression. Hasil penelitian menunjukkan bahwa keseimbangan kerja–hidup, ergonomi lingkungan kerja jarak jauh, dan tekanan psikologis merupakan prediktor yang signifikan. Lasso Regression memiliki kinerja terbaik dibandingkan model lainnya dalam hal akurasi dan presisi, dengan berhasil meminimalkan false positives dan false negatives. Temuan ini menegaskan potensi organisasi untuk mengimplementasikan pendekatan berbasis data guna secara proaktif mengidentifikasi serta mendukung karyawan yang berisiko. Penelitian selanjutnya dapat diperluas pada sampel yang lebih besar dan beragam serta mengeksplorasi pendekatan ensemble untuk meningkatkan stabilitas prediksi
===================================================================================================================================
Workplace dynamics have been profoundly altered by the move to remote work, which offers flexibility but also produces problems like work-life boundaries issues, heavier workloads, and social isolation. By analyzing primary data gathered from a structured survey of remote workers throughout Southeast Asia, this study builds on earlier research by identifying important factors that contribute to high levels of stress and burnout. This study assesses the models' capacity to identify workers who are at high risk of burnout or high stress by utilizing machine learning techniques such as XGBoost, Support Vector Machines (SVM), Ridge Regression, and Lasso Regression. The results show that work-life balance, ergonomics of remote work, and psychological strain are important predictors. Lasso Regression outperformed the other models in terms of accuracy and precision, successfully reducing false positives and negatives. These results highlight the potential for organizations to implement data-driven approaches to proactively identify and support at-risk employees. Future work could expand to larger and more diverse samples and explore ensemble approaches to further strengthen predictive stability

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Kerja Jarak Jauh, Stres Karyawan, Burnout / Stres Tinggi, Pembelajaran Mesin (Machine Learning), XGBoost, Support Vector Machines (SVM), Regresi Ridge, Regresi Lasso, Deteksi Stres, Asia Tenggara. ======================================================================================================================== Remote Work, Employee Stress, Burnout / High Stress, Machine Learning, XGBoost, Support Vector Machines (SVM), Ridge Regression, Lasso Regression, Stress Detection, Southeast Asia
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Information Technology > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Muhammad Rizky
Date Deposited:	29 Jan 2026 08:44
Last Modified:	29 Jan 2026 08:44
URI:	http://repository.its.ac.id/id/eprint/131291

Actions (login required)

View Item