Service Process Improvement Based on Exceptional Pattern Analysis

The continual improvement of service process is one of the most important factors for improving service process management level. But it is very difficult to identify the real causes of service fault executing service process improvement. Aiming at the problem, this paper proposed a service fault recognition & improvement method based on association rules mining. The method provides a new approach for enterprises to improve effectively service process fault. 1. Service Fault Analysis Based on Association Rules In the present complicated and fierce competitive circumstance, it is increasingly important to improve the management of service process so as to improve enterprise’s service effectively and quickly satisfy the customers’ needs. Therefore, many scholars, domestic or foreign, have done lots of research in this area and attained abundant achievements. For instance, many companies and scholars have studied the improvement of service process in terms of 6σ management [1,2,3] and then use method DMAIC to achieve the optimization and improvement of service process. However, these methods need to be measured, analyzed and improved by the professional statistical means and specialists, which is comparatively difficult because of professional obstacles. Moreover, sometimes the real reasons for faults generated in service are too complicated to be improved, or even, be identified by these methods related above. In order to realize the dynamic optimization of service process, Yushun Fan and Fabio Casati[4,5] have proposed the framework of a real-time business process performance management(RTBPPM) and a toolbox of business process intelligence(BPI). Both of them have the common features as the real-time monitor of service process, which is efficient and convenient as well. J.Leon Zhao and Edward Stohr[6] have proposed a method of service time assessment based on workflow management and have put forward control strategy of service process. Besides, Cheung.C.F[7,8] has proposed multi-perspective knowledge-based system(MPKBS), which integrates case inference technique and time sequence model, predicts service quality, navigates automatically, and assesses service performance by analyzing customers’ rights or conditions of customers and staffs. Although these service-oriented implementations and optimizations are effective, we still lack a good method to improve the service. It is undoubtedly beneficial for enterprise’s management to find a better method to improve services based on the research mentioned above. Considering this demand, this paper proposes the method of service fault recognition and improvement which utilizes the historical records of service process, analyzes exceptional past events, discovers and fixes the imperfect links, and elevates the management level of service process. Service faults are always the effects of several causes. The way to analyze service faults is to scan mass historical logs about service faults. Owing to masses of historical information generated in enterprise service process, one discovers that it is difficult to find out the causes of faults both by traditional cause-effect method and 6σ management. Therefore, we will use association analysis belonging to data mining technique to analyze the association between the service faults and their causes.