Customer Churn Analysis System For Broadcast Television Network Design

With the continuous development of "three networks convergence", the radio and television market is growing rapidly. The major domestic carriers are entering or preparing to enter the market of radio and television programs, which is just opening up for competition. This gives unprecedented pressure to the original monopoly in this area. In order to survive in the heated market competition, radio and television network enterprises now put a lot of efforts on increasing the bandwidth and improving the network quality so that they can catch up with the original network operators in short period. However their experience in the environment of enterprise corporatization and mode of operation are far less comparing to those well-established network operators who are grown up in the intensive competitions. It is impossible for the new operators in the market to catch up with the existing operators overnight. It requires exploration and experience accumulation. The market share lost analysis system described in this paper is part of the enterprise operation soft environment.This topic combines computer technology with customer retention strategies. Firstly it introduces the necessity analysis of customer retention, the customer loss and its reasons. After that, it discusses how to use data mining technology, combined with the structure of radio and television network, customer information and service system, to collect all existing data and create a customer retention analysis system. This system is capable of gathering distributed data of customer statistics, sales data, service information, etc. Useful information will be automatically generated based on them and a prediction model of customer lost will be created. Hence we can analyze the key reasons of customer lost and to take appropriate measures to effectively retain the customers that may be lost. It also provides scientific data to support marketing. At last, the paper describes the model-based implementation of the system.Two core technologies are discussed in this paper. The first one is the technology to resolve the problem of data selection used in building model. The modeling algorithm is cover as well. The type of data used to build the model is determined by analyzing the role and scope of data. Based on the various features of data mining algorithms, combining with customer retention, the appropriate algorithm to build the model are chosen and details of modeling computation are described in this paper. The second part of this paper focuses on resolving the problem raised in data reading process. By using the data extracted from PowerBuild's dynamic data pipeline from relevant database system, and the features of dynamic data pipeline technology, data processing like data transfer and data copy can be carried out smoothly.Currently, this system is still in the stage of development and improvement, combining radio and television network. After conducting some data testing, we have achieved desired result on the market predication. This system is believed to play a more important role in the future market.