首页>
外国专利>
METHOD FOR PARALLEL MINING OF TEMPORAL RELATIONS IN LARGE EVENT FILE
METHOD FOR PARALLEL MINING OF TEMPORAL RELATIONS IN LARGE EVENT FILE
展开▼
机译:大型事件文件中时间关系的并行挖掘方法
展开▼
页面导航
摘要
著录项
相似文献
摘要
Disclosed herein is a method for parallel mining of temporal relations in a large event file using a MapReduce model. In the method for parallel mining of temporal relations in a large even file according to the present invention, an event file is sorted based on customer identification (ID) and event time at which each event has occurred. A set of large event types satisfying a preset support or more is generated from the event file. The event file is converted into a large event sequence including the large event type set. The large event sequence is summarized and then a time interval data file is created. Candidate temporal relations are generated from the time interval data file, and frequent temporal relations satisfying a preset support or more are derived from the candidate temporal relations. A temporal relation rule is generated from the derived frequent temporal relations.
展开▼