Decision tree analysis of construction fall accidents involving roofers
作者：
Highlights：
• We demonstrated the utility of the decision tree methods in fall accidents experienced by roofers.
• We used C5.0 and CHAID algorithm.
• The DT models that resulted from the analysis of research data proved to be sufficiently accurate.
• The study showed that fall distance was the most important attribute, followed by fatality cause.
• Decision rules revealed that roofers who received training were safer against fatal falls.
摘要
•We demonstrated the utility of the decision tree methods in fall accidents experienced by roofers.•We used C5.0 and CHAID algorithm.•The DT models that resulted from the analysis of research data proved to be sufficiently accurate.•The study showed that fall distance was the most important attribute, followed by fatality cause.•Decision rules revealed that roofers who received training were safer against fatal falls.
论文关键词：Fall accidents,Data mining,Degree of injury,Decision tree,Predictive power
论文评审过程：Available online 13 October 2014.
论文官网地址：https://doi.org/10.1016/j.eswa.2014.10.009
