| 本文已被：浏览 6177次 下载 7825次 |
码上扫一扫！ |
|
|
| 一种从无“aba”模式的日志中挖掘2度循环的方法 |
|
林雷蕾1,2, 周华3, 代飞3,2, 朱锐1,2, 李彤1,2
|
|
1.云南大学 软件学院, 云南 昆明 650091;2.云南省软件工程重点实验室(云南大学), 云南 昆明 650091;3.西南林业大学 大数据与智能工程学院, 云南 昆明 650224
|
|
| 摘要: |
| 现有的过程挖掘算法依赖于"aba"模式来挖掘2度循环，而满足局部完备性的日志文件中不一定出现该模式.为此，扩展了经典Alpha算法，提出了αL+算法，用于从没有"aba"模式的日志文件中挖掘出2度循环.首先建立任务间的次序向量矩阵，用于抽象2度循环结构的变体结构；然后从全局视角，根据事件的出现次数及位置来区分2度循环和并发关系；最后提出紧邻度和回路抽象，以排除并发分支上同类型循环带来的干扰.实验结果表明，与现有的挖掘算法相比，αL+算法能够从具有"aba"模式或不具有"aba"模式的日志文件中挖掘出2度循环.此外，该算法实现且集成在开源框架ProM中. |
| 关键词: 日志抽象 向量矩阵 过程挖掘 局部完备性 Petri网 |
| DOI：10.13328/j.cnki.jos.005478 |
| 分类号: |
| 基金项目:国家自然科学基金（61462095，61702442，61662085）；云南省自然科学基金（2016FB102）；云南省中青年学术和技术带头人后备人才培养项目（C6143002）；云南省软件工程重点实验室开放基金（2017SE201，2016SE202）；云南省教育厅科学研究基金（2017YJS107，2017ZZX227）；云南大学研究生创新项目（YDY17095） |
|
| Approach to Mining Length-Two Loops From the Log Without “aba” Pattern |
|
LIN Lei-Lei1,2, ZHOU Hua3, DAI Fei3,2, ZHU Rui1,2, LI Tong1,2
|
|
1.School of Software, Yunnan University, Kunming 650091, China;2.Key Laboratory for Software Engineering of Yunnan Province, Yunnan University, Kunming 650091, China;3.School of Big Data and Intelligence Engineering, Southwest Forestry University, Kunming 650224, China
|
| Abstract: |
| The current research in mining length-two loops depends on "aba" pattern. However, the pattern does not necessarily appear in the logs that satisfies local completeness. This research aims at finding ways to mine length-two loops without the pattern. It results in a new algorithm (αL+-algorithm) that is based on the α-algorithm. First, an order vector matrix is established by tasks in logs to abstract variant structures of length-two loops. Then, distinction between loops and concurrency structure is obtained by event's frequency and location in traces. Finally, proximity and circuit abstraction are used to eliminate the interference caused by the concurrent branches. The experimental results show that the αL+-algorithm can handle length-two loops with or without "aba" pattern. In addition, the αL+-algorithm is implemented in the ProM tool. |
| Key words: log abstraction vector matrix process mining local completeness Petri net |
