Analysis And Design Of The Closed-Loop Management And Control System Of The Mobile Complaints Business

Closed-loop management and control system is constituted by the forward treatment process of the complaint data and output feedback. Closed-loop management and control system is a self-stable and continuous improvement system, and can make itself in a suitable condition through modifying control constantly based on the feedback results.With the popularization of3G network and the promotion of intelligent terminals, mobile service is more and more complicated, the competition among the telecom companies is becoming fiercer, and it also appears diversified and mass for the complaint data. The one who can quickly finds the complaint issue from the complaint data and solve it in time, to stop the expansion of the complaint and improve the customer satisfaction, will be able to occupy the initiative, control the market and seize the customers. The system relies on natural language processing, data mining, artificial intelligence algorithm and other technology to make group events mining, automatic classification and trend analysis from the complaint data, and localize group events, hot complaints, system failures and other problems quickly; it utilizes the automatic early-warning mechanism and automatic summarization technology to deliver event briefing to the appropriate leaders and departments automatically, to guarantee the accuracy of complaint business process and ensure maximizing the utilization of support resources, thus making the complaints problems solved quickly and providing strong support for improving the quality of the service. This system can locate the problem from the complaints data rapidly, trigger processing procedure, then return to the corresponding departments and leaders to resolve complaints problems, and eventually reduce the occurrence of customer complaints, thus forming a closed-loop management and control system.The paper comes from a project cooperated with one telecom operator, and the main work is as follows:(1) Analyzing the closed-loop management and control system of the complaint business, key technology and faced difficulties;(2) Designing closed-loop management and control system of the complaints business, including the design of group event mining closed loop, automatic classification closed loop, trend analysis, automatic warning, automatic delivery and KPI evaluation system, etc;(3)Based on the characteristics of the complaint business data, using the reverse frequent itemset of apriori with cosine similarity to design group event mining algorithm, and automatic classification algorithm which combines the feature vector with fulltext retrieval together;(4) Introducing platform and modules of the system, and making detailed instructions and display for the core function.