The core of financial technology is to effectively integrate big data technology and use the power of big data to promote financial enterprises to continuously improve efficiency and service capabilities throughout the life cycle of the financial industry.However, the combination of fintech and big data cannot be seen as a panacea.At present, big data still has its limitations and can only be used as a supplement to financial risk control.

Why use big data risk monitoring?

不管是银行还是消费金融公司，其他金融机构如互联网小额贷款公司，金融机构一般都有风险监控的需求。基础业务逻辑几乎相同，但金融产品和风险偏好存在差异。

Traditional institutions such as banks are inherently risky. First of all, the regulatory authorities on the financial institutions of the risk control capacity put forward very high requirements. Second, risk monitoring directly affects the profit level of financial institutions.

Therefore, big data risk monitoring directly solves the core needs of financial institutions, has the greatest value. Big data risk monitoring can greatly improve the efficiency and risk control capabilities of financial institutions in areas such as user profiles, anti-fraud and credit ratings, this is the development of financial enterprises must be combined with a technology.

Wind-controlled coverage process for big data

Big data covers all processes in the credit area, focusing on customer acquisition, authentication and credit in and after credit.

In the customer acquisition stage, create user profiles to track the complete life cycle of users;

In the identity verification process, techniques such as identity verification and live identification are used to resolve the question of whether the applicant is himself.Correlation analysis is the use of graph correlation technology to find fraud gangs;

In the credit granting stage, multiple data sources are collected, and risks are priced through modeling.Fintech service providers output credit scores to institutions for use

After the loan, the main is to check abnormal customers, timely alarm and overdue customers lost contact repair.

大数据行业简介

There are three main types of players in the big data industry at this stage:

Data institutions such as PBOC, Pengyuan Credit, Qianhai Credit and UnionPay Smart Strategy are characterized by cooperation with traditional banks, the Ministry of Public Security, the Administration for Industry and Commerce, airlines and social security bureaus and other state organs.It is characterized by providing external data query, basic ID card information, bank card information, air travel information and corporate business information.Data is rich and valuable, but the disadvantage is that risk control products are weak.

Internet companies such as Ant Financial, Tencent Credit and Baidu Finance are characterized by their large amounts of data for e-commerce, social networking and search, as well as some external data to form their own risk monitoring products and data output capabilities.At the beginning, these Internet companies only cooperated with their strategic partners to export risk control, and now they are slowly providing 2B risk control products.

When the internet giant does not provide external wind control technology and the wind control technology of traditional data institutions is not strong, such start-up technology companies as tongdun technology, Baifinancial services, Bangsheng technology, juxinli technology, and Shumei technology, etc. , they can make up for the huge demand for risk-monitoring products from peer-to-peer Finance and cash lending. Their data is the integration of multiple data sources, 2B Enterprises continue to provide wind control model and data, and accumulated some online lending data.

Analyze the value of big data risk monitoring

1.数据

Data is at the heart of big data risk monitoring.Only data that directly tells financial institutions that the target customers are blacklisted customers.It is simpler and more effective for serious customers who have expired.

It is best to have massive data that can reach enough users; User data value has high density, low noise, easy data cleaning; User data has multiple dimensions and can form rich user portraits; Its own business scenarios can obtain valuable data.

2. Technology

For some financial institutions, if the risk control standards are very strict, it is not difficult to identify customers who cannot be accepted.However, for most financial institutions, risk monitoring and business are mutually exclusive.In order to increase business volume, it is necessary to lower access standards, but also to prevent risks.This requires technical means to evaluate white households through anti-fraud modeling and credit modeling, and assess customer credit levels to determine acceptance.

The technical requirements have strong underlying technical architecture functions, good enterprise-level product output capabilities, and big data cleaning and modeling capabilities.In the future, it is necessary to combine technologies such as Al to form an intelligent risk monitoring and anti-fraud platform.

3. Scenes

Financial services such as wealth management, insurance, auto finance, and cash lending have different requirements for modeling, requiring a good understanding of the customer's business scenario so that the model can be adapted to industry characteristics. Need experienced modeling team and industry expert team; have customer experience in service industry, understand customer business situation; and deeply understand business requirements.

Application of big data risk control in credit

The current credit approval process is mainly divided into manual review and automatic review.For customers with good qualifications and good standing, the system will automatically approve as long as they can pass negative information, fraudulent information, and credit assessments.For customers who fail negative information and risk of fraud, the system can automatically reject or request manual review.For customers with low credit scores, a manual intervention is required.

Big data industry data

Central Bank Credit Report: in general, licensed financial institutions have access to central bank credit, including records of individual professional qualifications, records of administrative awards and penalties, records of court proceedings and enforcement actions, records of tax arrears, etc. .

Judicial information: The latest published list of courts at all levels of the SPC and provinces and cities, including information such as the enforcement court, the time for filing the case, the enforcement case number, the subject of enforcement, the status of the case, the basis for enforcement, the enforcement agency, the obligations specified in the effective legal documents, the performance of the person subject to enforcement, and the conduct of the untrustworthy person subject to enforcement.

Public Security Information: information about persons involved, at large or with a criminal record in the public security system, including the time of the incident, details of the case, such as cases of fraud/production, sales of fake drugs, etc. .

Credit card information: information such as bank deposit/credit card purchases, income, and overdue.

Travel information: Contains data such as flight cities, number of flights, and seat levels for each quarter of the past year.

Social information: including social account match type, social account gender, number of social account followers, etc.

Carrier information: Check Carrier account duration, network status and consumption level.

Online loan blacklist: Verify whether there are overdue online loans and blacklist information based on personal names and ID numbers.

There are also driver's license status, rental car blacklist, and e-commerce consumption records.

Problems in the big data industry

The major issue facing the big data industry today is customer privacy. Because the information such as public security and court is very sensitive, they are actually blank of legal supervision.

Before the establishment of Baihang Credit, the data of various data agencies has not been opened, and the validity of the data will be reduced.It is expected that after the hundred lines of credit investigation data comes out, the data will be more coherent due to the combination of the length of the data of each company.

Different big data companies have different data collection and cleaning methods, which can cause data pollution, so the output data will be somewhat inaccurate.

Today, citizen data mainly comes from offline collection and online behavior records.There is a certain lag in the data, and there is a certain delay in the data collected offline.

Big data is still in the early stages of development.Today's bigger problems are that the amount of data is not large and complete, and how to reconcile the contradiction between data disclosure and citizens' privacy.In the future, it will be necessary to combine other technologies such as artificial intelligence, blockchain, and the Internet of Things.Realize the ability of data not to be tampered with, timely data collection, etc., so as to better serve the financial industry.