Abstract
Cause analysis of accidents associated with industrial machines and devices is essential to improve the effectiveness and the efficiency of industrial safety system in Korea. This study focuses on cause analysis of accidents associated with industrial machines and devices. In particular, analysis of abstracts of accidents which are written in descriptive format and, therefore, inherently unstructured and exhibits characteristics of big data is suggested and tested. Automatic analysis of such big data performed in this study reveals the consistent results with the manual analysis results in previous studies. Analysis results also suggest that incorporating transition from the current user-oriented indirect regulations to more manufacturer and user balanced direct regulations will guarantee more effective prevention of industrial accidents at the early stage of generation of danger.