Machine Learning for Return-to-Work Prediction and Welfare Policy Prioritization among Cargo Workers
- 주제(키워드) return-to-work , industrial accident , machine learning , Welfare Policy , AHP , Critical Factors , Job Safety
- 주제(DDC) 658
- 발행기관 아주대학교 일반대학원
- 지도교수 Min Jae Park
- 발행년도 2026
- 학위수여년월 2026. 2
- 학위명 박사
- 학과 및 전공 일반대학원 비즈니스애널리틱스학과
- 실제URI http://www.dcollection.net/handler/ajou/000000035763
- 본문언어 영어
- 저작권 아주대학교 논문은 저작권에 의해 보호받습니다.
초록/요약
The above study investigates the key predictors of industrial accident victims Return-to-Work (RTW) to their original workplaces using national panel data and machine learning-based analysis approaches from South Korea. Moreover, the number of industrial accident accidents among transport workers is increasing in the industry, which examines the relationship between them along with the welfare policies of transport workers. Based on 3,423 cases from the2023 Industrial Insurance Panel Study (PSWCI), we applied logistic regression, decision tree, and random forest models to examine how individual characteristics, job conditions, and organizational support affect the outcome of return to work. The dependent variable was defined as successful return to work and non-return to work before injury. In all models, length of service at work, employment status, employer involvement during recovery, and self-evaluated functional ability were consistently shown as strong predictors of return to work. This suggests that return to work is not simply a clinical or administrative outcome but a socially embedded process shaped by employment relationships, economic stability, and workplace practices. It provides evidence on the importance of early intervention, structured employer-worker communication, and integrated support for vulnerable groups. Machine learning methods improve the ability to identify high-risk individuals and provide scalable tools for personalized rehabilitation plans and labor policy development. Meanwhile, it is playing a pivotal role in the development of the national economy and the welfare of transport workers is essential for the sustainable growth of the logistics industry. Although the number of transport workers is gradually increasing due to industrial development, interest and government welfare policies are poor and taking risks. Based on overseas cases and existing studies, we looked at the welfare policies of transport workers, looked at the priorities of welfare policies mentioned in previous studies, drew priorities using analytic stratification techniques (AHP), analyzed them as financial and policy factors, and analyzed the introduction of advanced technologies or psychological factors of elderly workers as technical and psychological factors to analyze problems that will arise in the future. By classifying it as a wide range of industrial accident victims and analyzing the factors and causes of the return-to-work (RTW), it will reflect occupational safety and government policies for transport workers, which are comprehensive meanings of freight transport, and further help find causes that contribute to industrial development. Keywords: return-to-work, industrial accident, machine learning, Welfare Policy, AHP, Critical Factors, Job Safety
more목차
I. Introduction 1
II. Literature Review 3
2.1 Utilize Machine Learning (ML) Return to Work Predictors 4
2.2. A Study on the Priority of Welfare Policy of Freight Carriers 5
2.3 Purpose and Priorities of the Study 5
III. Study 1 - A Machine Learning Approach to Return-to-Work Predictors Among Industrial Accident Victims 7
3.1 Introduction 7
3.2 Literature Review 9
3.2.1 Return to Work and Original Workplace Reintegration 9
3.2.2 Toward Predictive Modeling: Machine Learning in Occupational Rehabilitation 10
3.2.3 Gaps and Research Contribution 11
3.3 Research Hypotheses 12
3.3.1 Human Capital and Socioeconomic Status 12
3.3.2 Organizational and Job Characteristics 13
3.3.3 Functional Capacity and Recovery Support 13
3.3.4 Psychosocial and Environmental Factors 13
3.4. Methodology 14
3.4.1 Research Design 14
3.4.2 Data Source 14
3.4.3 Variables 16
3.4.4 Analytical Strategy 18
3.4.5 Ethical Considerations 19
3.5. Results 19
3.5.1 Logistic Regression Analysis 20
3.5.2 Decision Tree Classification 22
3.5.3 Random Forest Classification 22
3.5.4 Model Evaluation Metrics 25
3.6 Discussion 27
3.7 Conclusion 29
IV. Study 2 - Establishing a welfare policy program for cargo drivers in the Rep. of Korea: A comparative analysis of cargo driver employees 31
4.1 Introduction 31
4.2 Theoretical Framework 32
4.3 Research Model 34
4.3.1 Financial factors 36
4.3.2 Psychological factors 38
4.3.3 Technical factors 40
4.3.4 Policy factors 43
4.4 Methods. 45
4.4.1 The AHP Method 45
4.4.2 Data Source and Collection. 46
4.4.3 Analysis Results 48
4.5 Conclusion 53
V. Conclusion 55
5.1 Summary of Findings 55
5.2 Policy Implications and Priority Setting of Research 55
5.3 Limitations and Suggestions for Future Research 56
References 57
