踪迹聚类下组织实体的重要度排序方法  

Importance sorting method of organizational enities based on trace clustering

作　　者：徐涛[1,2] 孟野[1] 

机构地区：[1]中国民航大学计算机科学与技术学院,天津300300 [2]中国民航大学中国民航信息技术科研基地,天津300300

出　　处：《计算机应用》2016年第5期1284-1289,共6页journal of Computer Applications

基　　金：国家自然科学基金资助项目(61502499);中国民航科技创新引导资金项目重大专项(MHRD20140105);中央高校科研业务费专项资金资助项目(3122015D015)~~

摘　　要：针对简单套用交接网络等社会网络分析方式不能很好地反映踪迹聚类生成的一系列流程的组织实体的重要度的问题,提出了一种踪迹聚类下组织实体的重要度排序方法。首先,对于参与踪迹聚类生成的一系列流程的组织实体构建踪迹聚类与组织实体关系网络;其次,定义基于踪迹聚类与组织实体关系网络的节点重要度评估方法;最后,对踪迹聚类下的各个组织实体节点计算其在关系网络中的重要度评分并排序。实验结果表明,所提方法构建的关系网络相比踪迹聚类下的交接网络能够更准确地反映组织实体的实际重要度;与基于拓扑势的网络社区节点重要度排序算法相比,所提方法的节点重要度排序结果更符合实际业务流程,能更好地区分关系网络中重要度不同的节点。Aiming at the issue that the social network analysis method like hand-over network cannot express the importance of organizational entities precisely,a method to sort the quantified importance of organizational entities organized under the trace clusters was proposed. Firstly,a relation network was constructed to describe the relationship between trace clusters and organizational entities; secondly,a quantitative assessment of the nodes＇ importance of this network was defined;finally,all these nodes were sorted respectively according to their quantified importance. The experimental results show that this relation network can express the actual importance of organizational entities more precisely than the hand-over network generated by trace clustering. Compared to the importance sorting algorithm of network community nodes based on topological potential,the proposed method is more suitable for the actual business processes,meanwhile it can distinguish distinct organizational entities better than the importance-sorting algorithm based on topological potential.

关 键 词：流程挖掘 组织挖掘 重要度排序 社会网络 复杂网络 

分 类 号：TP391.72[自动化与计算机技术—计算机应用技术]