건설현장 재해요인의 연관성 도출을 위한 연관규칙분석 적용 연구

Abstract

The purpose of safety management is to identify and control risk factors to prevent accidents. It is very important to analyze the causes of accidents based on past accident cases. The current state of construction accident analysis is often classified simply based on accident statistics. In addition, Analysis of accidents using simple statistics focuses on high frequency factors. However, since accidents are caused by a combination of factors, the relationships of accident factors must be considered. In this study, the association rule mining(ARM) is applied to extracting relationships of accident factors in construction sites. ARM of data mining techniques is suitable for scientific and probabilistic analysis of the associations between various items. This study collects construction accident data from 2010 to 2014 and extracts relationships of construction accident factors that occur simultaneously, regardless of frequency, based on probability. The purpose of this study is to identify the relationships of factors causing construction accidents according to the accident type, construction size, and process rate. It is expected to reduce the selection time of variable in the processing of information for accident analysis and be helpful for the decision making or the analyst.