一种基于事件的海洋关联规则挖掘方法

Marine Association Rule Mining Based on Events

在Apriori算法的递归链接-剪枝概念上,设计了面向海洋异常事件的关联规则挖掘算法.首先给出事件的相关概念与定义、事件的规则表达及评价指标.根据事件的定义和支持度阈值,生成事件频繁1-项集,并设计面向事件的链接-剪枝算法,实现频繁k-项集到(k+1)-项集的产生.根据事件强关联规则评价指标,提取海洋事件强关联规则.通过太平洋海洋异常事件的关联规则挖掘和典型异常事件间的关联规则分析,验证了该方法的正确性和可行性.

An association rule mining algorithm is designed for anomalous oceanm events based on recursive ＂link-prune＂ of a priori algorithm. Concepts, definitions, rule expression and evaluation indicator related to events are introduced first. Based on a threshold of support and definition of event, event frequent 1-item set and designed an event-oriented link-prune algorithm is generated for frequent （k ＋ 1）-item set from frequent k-item set. Marine events＇ strong association rule is then extracted based on events＇ strong association rule evaluation indicator. A case study on the association rule mining of Pacific marine abnormal events and association rule analysis on typical abnormal events are used to show correctness and feasibility of the proposed method.