Abstract:

Telecom operators innovate continually in data management modes and combine data marketing with social red envelope activities,consequently launching data red envelope activities,which stimulates customers’ interest of data using.Faced with the research pain points of customer characteristics and propagation modes of socialized traffic red envelopes and the current status of the single community modeling algorithm,the detailed research and comparison of the application effects of the six community modeling algorithms were conducted to select the optimal traffic envelope algorithm,and the characteristics of the core value customer groups were located and analyzed.Results show that Multi-Level algorithm performs better in data red-envelope scenario.Based on this algorithm,four characteristic customer groups were mined,namely seed customer,high-value customer,low-value customer and silent customer.The social network clustering conclusion provided effective guidance for operators’ precise marketing,precise push and other marketing activities,as well as other customer operation management such as silent customer activation,lost customer recovery and so on.

Key words: social network analysis, data red envelope, customer group division