Internet finance involves a wide range of fields, and the risk control strategies in different fields are not the same, so it cannot be generalized. Here is a general classification and analysis of the application of big data risk control in the field of Internet finance.
First of all, how to understand big data risk control?
The effectiveness of big data risk control not only emphasizes the mass of data, but also lies in the breadth and depth of data used for risk control. These include:
The breadth of data: refers to the diversification of data sources used for risk control. Any Internet finance enterprise cannot expect to solve the risk control problem based on a single massive data. Just like the principle of "cross-validation" emphasized in traditional financial risk control, the risk model should be cross-validated through diversified data. The same is true for the risk control strategy of Internet finance, which may adopt multiple strategies for the same risk event.
Data depth: refers to that the data used for risk control should be based on a complete record of real business scenarios and processes in a vertical field, so as to ensure that the data can restore the real business process logic. For example, many third-party payment platforms have rich real transaction records. However, due to the inability to obtain the detailed information and user identity of the traded goods in most scenarios, the value is greatly reduced when used for risk control, so the integrity and vertical depth of data are very important.
How Internet financial products use big data for risk control generally has the following categories and directions:
1. Conduct risk control according to certain target groups, specific industries, business districts, etc. Due to the deep cultivation of vertical targets such as specific personnel, industries and business districts, it is easier to construct corresponding risk points and risk control strategies.
For example, college students' consumer loans are mainly aimed at the characteristics of college students.
Financing guarantee for agricultural machinery industry.
Wholesale market business circle credit.
2. Risk control is carried out based on self-owned platform identity data, historical transaction data, payment data, credit data, behavior data, blacklist/whitelist and other data.
& gt& gt& gt& gt Identity data: real-name authentication information (name, ID number, mobile phone number, bank card, company, position), industry, home address, company address, relationship circle, etc.
& gt& gt& gt& gt transaction data: for example, the transaction data of B2C/B2B/C2C e-commerce platform, the transaction data of loans and investments of P2P platform, etc.
& gt& gt& gt& gt credit data: for example, the credit data accumulated by the borrowing and repayment behavior of P2P platform, the credit data and credit scores (JD.COM White Bar, Alipay Flower Bank) formed by the e-commerce platform according to the transaction behavior, and the credit data of SNS platform.
& gt& gt& gt& gt behavior data: for example, e-commerce purchase behavior, interaction behavior, real-name authentication behavior (such as Sina Weibo company authentication and friend authentication), and modification of information (such as modification of home and company address, and confirmation of work stability through frequency change).
& gt& gt& gt& gt blacklist/whitelist: credit card blacklist, account whitelist, etc.
3. Based on the services and data of third-party platforms, do Maxmind services for risk control Internet credit platforms (non-PBOC credit reporting), FICO services, retail decision-making (ReD) and industry alliances to share data (such as small loan alliances and P2P alliances).
& gt& gt& gt& gtIP address library, proxy server, stolen/fake card database, malicious address library, etc.
& gt& gt& gt& gt public opinion monitoring and trends, word-of-mouth service. Such as macro-policies, industry trends and case studies.
4. Based on the traditional industry data, do the traditional industry data such as the People's Bank of China, industry and commerce, taxation, housing management, courts, public security, financial institutions, vehicle management offices, telecommunications, public utilities (hydropower and coal) credit information.
5. Off-line field due diligence data
Including self-built risk control team to do offline due diligence mode and cooperation with traditional offline enterprises such as small loan companies, pawn companies and third-party credit management companies to do risk control mode. Offline wind control data is also an important data source and means of big data risk control.
I hope I can help you. If you want to know more, you can pay attention to the micro-signal "big data risk control circle". Many internet industries share information.