The evaluation of occupational accident with sequential pattern mining

Abstract

Accidents in manufacturing systems greatly affect productivity and efficiency, which are well known perfor-mance indicaters in practice. Therefore, it is very important to know the sequential patterns among the accidents to avode possible losses decrasing performance of the manufacturing systems. In order to reduce accidents, it is necessary to determine the patterns that cause the accident first. The associations among the causes of the occurrence of accidents is rarely investigated in the literature. To fill this gap, the patterns of causes among the accidents in the manufacturing system are revealed by using sequential pattern mining in this study. The most important contribution of this study is the discovery of sequential patterns formed by accident characteristics of pre-accident, moment of accident and post-accident stages unlike traditional accident investigation methods. Additionally, knowing the patterns of causes among the accidents can help decision makers to prepare a more proactive security program in real life. The CloFast algorithm is performed to go into the details of accidents in manufacturing systems. Accident records induding data between 2013 and 2019 are used to discover the sequential patterns. The results of this study showed that each accidents has its own sequential accident patterns and it is also posible to prevent possible accidents and reduce losses due to accidents considering sequential patterns in real life. Safety engineers and occupational safety specialists should take into account the sequential patterns among the accidents to avoid similar accident in the near future.