TCA: Telecom Credit Assessment Assisted by Edge Intelligence

Abstract

With China’s economic transformation into a high-quality development stage, the importance of credit system construction has become increasingly prominent. The problems existing in the current telecom credit system include: (1) insufficient coverage of credit features; (2) traditional credit assessment models are difficult to reflect user credit status objectively, comprehensively and timely; (3) user demand for credit management and credit services are ignored. Due to these deficiencies, a new multi-level credit system is necessary to meet the rapid development of market economy. Telecom operators have large amount of precious data, with the advantages of large-scale, high-precision and data-diversity, which can provide new ideas for the construction of credit system. This work focuses on the current problems and conducts research as follows: design a Telecom Credit Assessment Model based on Boosting and Stacking ensemble techniques, called TCAMBS, to improve the evaluation accuracy, and to select the best model according to the experimental results. On the one hand, this work can promote the innovation of telecom credit assessment models and provide new ideas for the construction of the credit system. On the other hand, this work will also help telecom operations to improve the quality of telecom credit services.