粗糙集理论在海事中的应用研究

Application Study of Rough Set Theory in Maritime Affairs

针对粗糙集理论在海事领域应用研究的不足,运用RSES粗糙集数据处理系统中的遗传算法,对由6个因素组成的海船船员适任决策表进行了数据挖掘,挖掘出了5个长度为4的属性约简集和241条决策规则。研究表明,在所选取的6个因素中,约简后各因素出现的频度分别为2,5,4,3,3,4,学历因素的重要度最高,而年龄因素的重要度最低。各因素对一般及以下等级海事发生的平均可信度分别为0.71,0.68,0.68,0.74,0.73,0.77。各因素对重大及以上等级海事发生的平均可信度分别为0.29,0.32,0.32,0.26,0.27,0.23。各因素主要引起一般及以下等级程度海事。

Aiming at the fact that the rough set theory has not been well applied in maritime field, based on the Genetic Algorithm of Rough Set Exploration System, a six-factor seafarers＇ competency decision table is mined, five reduced attribute sets and 241 decision rules are obtained. Study shows that the frequencies of all attributes are 2,5, 4,3,3,4, respectively among the six faetors, and the Education factor has highest importance, while the Age factor has lowest one. For gentle grade of maritime accidents and below, the average confidence factors are 0. 71, 0.68, 0.68, 0.74, 0.73 and 0.77, respectively for the six factors. However, for those sever maritime accidents, the average CFS are 0.29, 0.32, 0.32, 0.26, 0.27 and 0.23. So, every attribute mainly contributes to small maritime accidents.