Design And Implementation Of Prediction System For Inventory User Indexes Based On Big Data Technology

In recent years,telecom suppliers have come to know that the pivotal of market competition has shifted to existing business market from the previous increment market.Incremental and existing clients are held at equal state as of now and market operation is giving increasing more attention to the simultaneous promotion of measures for both types of clients.In the last three years,China Unicom has witness a transformation of its existing clients management from the client preservation and management,to value growth and client operation.Amid the changes,to predict the number of clients and revenue from existing clients through existing clients business analysis by using system dynamics become immensely important,which can be a critical,verifiable and fundamental basis for decision makers to spot problems out of business development and existing client operation under the current competing environment,and provide support for decision to adjust marketing strategy and improve business management.The paper,representing the business need of the department for existing business operation,takes the original 36-months revenue data from a branch company of China Unicom in a certain province,builds a local big data platform using the home-designed big data system,and succeeds in establishing a existing clients operation surveillance system through application of data mining algorithm.The surveillance system can effectively predict the number and income status of existing clients in billing months,and predict the number and preservation.The four-tier application of the surveillance system in provincial branch company,city branch company,country branch company and marketing grids demonstrate an accurate match between the prediction of operating indicator and the actual result that helps management personnel on all levels monitor the major existing business indictors at all time,find the cause of problem,send out forewarning and interfere in advance.Hence the effectiveness and operating value of the system are proven and validated.