基于模糊关联规则的交通事故分析应用研究

Application Research of Traffic Accident Analysis Based on Fuzzy Association Rule

研究发现交通事故潜在规律,预测交通事故的发生,针对关联规则方法用于交通事故分析,对交通的数值型属性无法给出有效地划分,为提高安全管理决策,提出引入模糊聚类,用改进FCM(Fuzzy c-Means)方法对数值属性进行聚类,可用取值的范围对分类属性进行聚类,采用模糊关联规则挖掘导致交通事故的原因和规律。模糊关联规则首先对FCM算法进行了改进,包括隶属度、权值和中心点的计算和修正方法,利用模糊关联规则方法进行挖掘,最后对算法进行了仿真和可视化显示,结果表明模糊关联规则方法挖掘出的规则符合现实情况,为交通管理提供有效的方法。

In order to search the rules of traffic accidents and prevent their occurrences, the fuzzy clustering was introduced into this paper because traditional association rule is difficult to gain partition for numerical attributes while it is applied in analyzing traffic accident. This paper uses Fuzzy c-Means（FCM） to classify numerical data and then employs fuzzy association rule to gain the reasons and potential rules of traffic accident. This paper improves the FCM algorithm, including the methods for computing the membership degree and modifying weights and centers of clusters, and then makes use of fuzzy association rule to mine the rules of traffic accidents. Finally, we implement the algo- rithm and its visualization, and results show that it is feasible.