The Design And Application Of KPI Based On The Complaint Service Flow

With the popularity of3G networks and promotion of intelligent terminals, competition the telecom operators becomes increasingly fierce. At the same time, the complaint data presents diversification and magnanimity; operators who can promptly solve complaints and curb complaints expansion will be able to occupy the initiative, to seize the market and the customers. A sound complaint handling system is imminent. In order to simplify the processing steps and shorten time, the closed-loop control system of handling complaints service emerges as the times require,this system can put the complaint organically, complement each other, push each other, constantly improve itself, and effective realize the coordination of various departments, and as a result it improves the processing efficiency of complaints.But the closed-loop control system of handling complaints can only hold the overall situation, it is difficult to take into account the specific problems which each sector are concerned about.For example, network department may only be concerned with some kind of network performance anomaly caused complaints among the mass data, want to put these data mining and analysis them multidimensional, and find the relevant network properties associated with these complaints; whereas the marketing department may only want the complaints during a double promotion. According to such demand and so on, how to solve them? KPI is such a kind of this system, from all angles, to inspect multiple aspects personalized system model of the same complaint data. We can define an index, and choose some sample data belong to the index from the complaints text,then, with the help of the data mining knowledge, we can analysis the characteristic words of this index,choose a large number of complaints data of this index from all records of complaints, and after the subsequent statistical analysis, results can be multi-dimensionally displayed by graphs, so that the related personnel analyze the reasons of these complaints and give supportive decisions, can also push the same kind of mining complaint information to relevant departments for processing and so on, satisfying the diversified needs.This thesis mainly discusses the following aspects:(1) In order to ensure the flexibility of the system, this thesis analyzes the demand deeply, and establishes the related model using the knowledge of data mining.(2) Machine learning techniques are used for interested categories firstly,then the thesis classifies the complaint data based on different topics and using related classification method, finds out interested complaint data and displays them by the graphs from different dimension.(3) Finally practical applications of this system in the complaint service are presented.