CAJ | 학술논문

为克服FCM M iner算法在挖掘频繁闭情节时存在的不足，基于最小且非重叠发生的支持度定义，提出一种事件序列上频繁闭情节挖掘算法FCM＋＋。定义两种特殊的数据结构：频繁情节树（FET ）、层头表（L H ），采用广度优先搜索策略进行层扩展操作，扩展时将挖掘的频繁情节逐层压缩到FET和L H结点链中。通过动态维护FET及闭合性检查过程挖掘所有的频繁闭情节。实验结果表明， FCM ＋＋算法较 FCMMiner算法有更高的挖掘效率，能有效地挖掘所有的频繁闭情节。
위극복FCM M iner산법재알굴빈번폐정절시존재적불족，기우최소차비중첩발생적지지도정의，제출일충사건서렬상빈번폐정절알굴산법FCM＋＋。정의량충특수적수거결구：빈번정절수（FET ）、층두표（L H ），채용엄도우선수색책략진행층확전조작，확전시장알굴적빈번정절축층압축도FET화L H결점련중。통과동태유호FET급폐합성검사과정알굴소유적빈번폐정절。실험결과표명， FCM ＋＋산법교 FCMMiner산법유경고적알굴효솔，능유효지알굴소유적빈번폐정절。
To overcome the defect of FCMMiner algorithm in mining frequent closed episodes ,an algorithm called FCM+ + was proposed which defined support threshold based on min‐and‐non‐overlapped occurrence .Two special data structures were de‐fined:frequent episode tree (FET) and lay header (LH) ,by using the breadth‐first search strategy and layer extended opera‐tion ,the frequent episodes were compressed to the FET and LH chains .All frequent closed episodes were discovered by dynami‐cally maintaining the FET and closed inspection procedure .Experimental results show that ,compared with FCMMiner algo‐rithm ,FCM+ + algorithm has higher mining efficiency ,can effectively mining all frequent closed episodes .