Big Data Credit Rating and Risk Management


Apply big data credit rating and evaluation and risk management technologies to construct online, offline or combined to-C credit loan (citizen loan, farmer loan, credit card, etc.) and to-B credit loan services (operating loan, business zone loan, invoice loan, etc.). The target is to release loans sooner, lower risk, improve return per customer, lower bad debt rate, lower service cost and improve customer experience.


  • Online Risk Management


Risk management capability is the core of banking digital transformation. Our risk management solutions apply technologies including face recognition, device fingerprints, biological detection and GPS positioning, together with financial technologies including big data, cloud computing, artificial intelligence to establish smart risk management decision making models.


In terms of decision-making method, we transform “causal decision making” into “correlation decision making”. We record all behaviour features of “good customers” and “bad customers” into big data smart risk management system and perform correlation analysis to obtain general conclusions. When a new customer arrives, we apply the general rule to the customer to obtain a decision about this new customer. This method focuses on users’ behaviour data.


In terms of decision-making model, we transform “expert experience decision” to “smart-model decision”. We apply machine learning models including GBDT and random forest in smart decisions making, to improve the accuracy and effectiveness of decision making.


  • Build Autonomous Risk Management System


We aim to establish autonomous risk management system including pre-loan credit admission, initial loan amount deciding, risk pricing, on-loan account management (including risk alerting, amount adjustment, re-pricing, freezing) and post-loan collection


  • Application Anti-fraud


Pre-screen customers based on target customer segmentation and risk tolerance, in combination with internal and third-party information. This can help pre-exclude non-compliant and risky applications, lower down default rate, avoid credit risk and reduce approval cost.


  • Pre-loan Management


Pre-loan management includes admission (credit granting approval, credit withdrawal approval), credit amount and pricing strategy. Our solution will assess applications against entry criteria, and provide suggestions over accept/reject decision, and further advise on credit amount and risk / pricing.


  • On-loan Alerting


Our solutions will observe the changing trend of borrowers’ repayment willingness and repayment capability. We also closely follow factors such as the loan itself, lenders and collaterals, to identify any loan issues in order to take necessary actions. In such a way, we can prevent, control and mitigate loan risks and improve quality of credit assets.


  • Post-loan Collection


For those accounts with overdue loans, we will identify which account to collect repayments upon, when and how.


  • Risk Monitoring


We do 7 * 24 end-to-end real-time monitoring on the pre-loan, on-loan and post-loan data performance and the risk control model performance, extract risk alerting tags, analyse results of strategy executions through dashboard monitoring and gather statistics for iterating models and strategies.


Based on existing bank internal data, third-party data, post-loan check data, credit inquiry data and internet behaviour data, we establish application evaluation models, application anti-fraud models, initial credit amount and risk/pricing models as well as behaviour rating models, which vary on different users across different stages. We also perform data analysis over risk control results, continuously manage and optimize risk control models and strategies in order to prevent risks, fill in gaps and reduce bad debt rate.



Case Study

1. Individual customer online credit loan solution

2. Online auto credit granting approval system