BigCredit: A Novel Framework for Big Social Media Data Enhanced Online Credit Scoring

Project: Research

View graph of relations


In his 2015-16 budget speech, HK SAR financial secretary John Tsang reported that he had askedthe Financial Services and the Treasury Bureau to set up a steering group to study how to developthe city into a financial technology (FinTech) hub.1In fact, senior officials from the HK SARgovernment have repeatedly stressed that FinTech is a key enabler for developing Hong Kong as aninternational financial hub.2On the other hand, our premier Li Keqiang has highlighted the“Internet Plus” strategy in his Government Work Report presented in the National People'sCongress early this year; “Internet Finance” is one of the essential components of the InternetPlus strategy which will drive the economic growth of the whole nation in coming years.3Accordingly, our proposed GRF project is a direct response to these calls for strategic FinTechdevelopment. More specifically, the proposed research aims to design a novel online creditscoring framework which leverages big social media data to assess the potential credit risks ofindividuals. Online credit scoring is a supplement rather than a replacement of the traditional creditassessment method, and it lays the solid foundation to support an array of Internet Finance servicessuch as online peer-to-peer (P2P) lending, online micro loans, personal wealth management, etc.Although some prior studies about online credit scoring have been reported in literature, none ofthem examines the challenging problem of extracting and analyzing credit intelligence from bigdata. Moreover, very few studies have explored the problem of analyzing individuals’ behavioralfootprints left on online social media for the prediction of their credit risks. From a technologicalperspective, traditional computational methods cannot effectively resolve the volume, velocity,variety, and veracity issues of big social media data. For instance, applying traditional NaturalLanguage Processing (NLP) techniques to parse the semantics of individuals’ online postingsinvolves polynomial time complexity. Given billions of user-contributed online messages each day,traditional computational methods simply cannot scale up. On the other hand, none of the workreported in literature has examined the problem of mining and analyzing the characteristics of anindividual’s social circle for the assessment of their credit risk.Guided by the Design Science research methodology and the theoretical underpinning of SocialCapital, the goal of the proposed research project is to fill the aforementioned research gaps bydesigning a novel framework which can leverage big social media data to enhance online creditscoring. In particular, the aims of our research are as follows.1. To design a novel framework which can effectively and efficiently analyze individuals’behavioral footprints and social circles as captured on online social media for enhancing onlinecredit scoring;2. To empirically evaluate the effectiveness, efficiency, and usability of the proposed onlinecredit scoring framework through a series of controlled experiments and a field test involvingreal users.3. To reflect and synthesize sound design theories related to both the design processes and thedesign artifacts for building a class of big social media analytics services for financialapplications.The online credit scoring framework developed through this project could be adopted by Internet Finance service providers such as WeBank, Dianrong, Jimubox, etc. and traditional financial institutions. We plan for technology transfer via our applied research collaborator,—one of the largest Internet Finance companies in the Greater China area. Theproposed research project will contribute to develop novel FinTech services in the region. Thereby,it will facilitate Hong Kong’s technological innovation and economic growth, and develop HongKong as an international financial hub in the long run.1


Project number9042420
Grant typeGRF
Effective start/end date1/01/173/06/21

    Research areas

  • Online Credit Scoring , Financial Technology , Internet Finance , Big Data Stream Analytics , Design Science