基于异常模式识别的服务过程改进  

Service process improvement based on fault pattern recognition

作　　者：李兵[1] 赵宇[1] 李秀[1] 刘文煌[1] 

机构地区：[1]清华大学自动化系,北京100084

出　　处：《清华大学学报（自然科学版）》2007年第4期494-497,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(70202008);中国博士后基金资助项目(023209020)

摘　　要：持续地改进服务过程,是提高服务过程管理水平的重要方法之一,但在服务过程改进中往往难以识别服务缺陷产生的原因。针对这一问题,提出一种基于关联规则挖掘的服务过程缺陷识别和改进方法,以为有效改进服务过程提供决策依据。该方法在基本的关联规则算法中,引入了缺陷原因和结果互斥约束、规则前件数约束、改进度约束以及新颖性约束。仿真试验表明,该方法与传统关联规则算法相比,降低了生成规则的数目,提高了规则有效性,降低了算法运行时间,取得了更好的效果。The continual improvement of service processes is an important way to improve service process management, but the real causes of service failures are difficult to identify. This paper presents a service fault recognition and improvement approach based on association rules mining. The method provides a decision-making warrant for enterprises to effectively improve service processes. This approach identifies the bug cause and result restriction, the pre-rule restriction, the improvement rate restriction, and the novelty restriction to improve traditional association rules arithmetic. Experiments show that this approach reduces the number of rules and the run-times with improved efficiency compared with the traditional association rules method.

关 键 词：服务过程管理 异常事件 关联分析 模式识别 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]