머신러닝 기법을 활용한 한국 노동시장 탐색적 분석

초록

This study aims to derive multi-layered typologies of the Korean labor market based on “Industry × Occupation” units to capture its structural and experiential heterogeneity with high granularity. Using data from the 23rd wave of the Korean Labor and Income Panel Study (KLIPS), we analyzed key indicators including wages, working hours, employment type, welfare/social insurance, job satisfaction, and organizational commitment. K-means cluster analysis was conducted on 544 Industry-Occupation combined units. The analysis identified four distinct groups: (1) Stable Employment Type, (2) Flexible-Precarious Type, (3) Overwork-Intensive Type, and (4) Vulnerable-Peripheral Type. These four worlds demonstrated significant differences in both structural conditions and subjective experiences. These findings suggest that the Korean labor market is not merely a simple dual structure but a multi-layered system where multiple employment ecosystems coexist. By extending existing segmentation theories to clarify detailed heterogeneity and presenting a typological framework applicable to HRM and policy design, this study offers significant academic and practical implications.

키워드