The revelation of the Big data credit system of the open sesame

January 28, Ant Financial Services Group (hereinafter referred to as "Ant") under the first Test of Sesame credit system, this is also January 5, the central bank issued to allow 8 institutions to carry out personal credit service, the first company to launch the system. According to the 21st century economic reporter learned that, on January 5, after the central bank issued a notice, ant gold Clothing quickly registered company, the formation of a team. Its backbone members are mostly data scientists, and have started to recruit people on a large scale. Sesame Credit general manager from China Merchants Bank "dug" to Hu Yu as. It is reported that these 8 institutions are now in the preparatory stage, there is no administrative license, there is no licence. In accordance with the requirements of the notification, the preparation period is 6 months.

Sesame credits take and pay the treasure purse in a cooperative way that users can see this feature in the "Wealth" column of Alipay 8.5. But currently only users who participate in the test can see credit scores. Compared with the traditional central bank credit reporting, the form of sesame credits is simpler-credit score. Sesame Credit score of the lowest 350 points, the highest 950 points, the higher the score represents the credibility of the better, the lower the likelihood of default.

The introduction of Sesame Credit test version means that the market-oriented personal credit is officially started in China, and more and more institutions will build up a set of trust files for each person, which will become a card for each user.

But for now, both users and credit agencies will face a long process of cognition and construction. How to collect personal information, how to form credit scores, how to use personal credit, what personal credit means to users is a new challenge.

Building Credit Dimensions

Sesame Credit backed by Alibaba, and Alibaba currently covers hundreds of millions of of Internet users, only Alipay has more than 300 million users. At the same time, the Sesame credit has Ali's electricity trader transaction data and the ant Gold Clothing Internet financial data, is more gold than the pure network social data, this also becomes the Sesame credit to enter the letter industry a big unique superiority.

Traditional financial institutions use more data in the field of credit and have limited coverage. According to a previous statistic, as of the end of 2013, the bank's personal credit system contains about 320 million per cent of natural persons with credits recorded, and less than 1/4 of the total population, while the coverage of the system has reached 85% in the United States.

"In the field of credit we are more of the central Bank to supplement the letter." "Ant Gold Clothing credit business development director Deng in a small range of communication. This also means that the market-oriented, Internet personal credit will greatly expand the coverage of the population and increase the personal credibility of the evaluation dimension.

But in practice, how to build such a large and important system, at present there is no mature case reference. Although Alibaba has such a large number of data, the formation of products will still be a systematic project.

First of all, the dimensions, ant Golden Service public Relations Zhang Daosheng accepted the 21st century Economic Report reporter interview, said sesame Credit points this seemingly simple score, behind is the sesame credit to the massive information data comprehensive processing and the appraisal, the Sesame Credit establishes the personal credit dimension aspect also to refer to the United States famous letter organization Fico.

Ant Gold Service chief credit data scientist Yu Wujie introduced some foreign experience, summed up can be divided into, one, basic information, including occupation, income, job stability, block, vacation and residential stability; second, loan information, including car loans, mortgages, credit cards and student loans; third, consumer information, shopping malls and network consumption frequency, Iv. Public information. Court decisions, tax arrears, subway fare evasion, traffic violations, relatives and friends arrears. This means that personal credit covers every aspect of everyone's life and work. He also cited, for example, the user's public interest behavior will also affect personal credit.

It is understood that on this basis, Sesame Credit has adopted five dimensions: User credit history, behavioral preferences, performance ability, identity characteristics, personal relationships, respectively, corresponding to the past credit account records and credit account history; preferences and stability in shopping, payment, transfer, financial management, etc. To enjoy all types of credit services and to ensure timely compliance, leaving sufficient and reliable personal information in the use of related services, the identity characteristics of friends, and the level of interaction with friends.

Data sources and calculations

Alibaba has Taobao, cat, a good deal and other electric platform, as well as the Internet financial Platform ant service, at the same time the business began to deepen the electricity and water charges, taxis, medical and other fields, itself has accumulated a large number of data.

Zhang Daosheng said and the traditional credit data mainly originates from the loan domain different, derives from the Internet the Sesame credits data source to be more broad, the kind is richer, the timeliness is also stronger, covers the credit card repayment, the net buys, the transfer, the financial management, the electricity and water coal payment, the rental information, the address relocation history

The biggest feature of the data generated by the Internet is real-time, with traces and easy tracking, which is more comprehensive and accurate than the traditional offline collection and integration.

As for Alibaba's business, it has accumulated huge data through its electricity quotient, payment, internet finance, Taxi, "Future hospital" plan and other lines of business. But to achieve a more comprehensive coverage, this amount of data is clearly not enough. Zhang Daosheng said that the data sources of sesame credits can be divided into three parts, some of which are Ali's existing data, and some of the data from the public service (including some government departments); future Sesame credits will also launch a user-submitted data platform based on the data.

However, there are two major problems in obtaining external data, first, how to provide external data, free or paid; second, different types of data vary from one organization to another.

Yu Wujie in foreign countries, for example, there is a fee and payment model, in addition to a resource exchange, such as banks, a bank to the credit agencies to collect data, foreign countries have a reciprocal treaty, banks must be some default information to the credit agencies, or credit agencies will

If banks are not allowed to collect data, the cost is very high, and if they do not use the information of the company, it is a kind of exchange. In addition, the credit agencies will collect their own electricity from the coal and water utilities and pay for them relatively cheaply.

For data types, it needs to be resolved in a technical way.

After collecting the data, how to calculate the user's credit through the data model, this is the letter organization is also the user most concerned matter. Zhang Daosheng explained that not a change in the behavior of a user will lead to the changes in Sesame credit, unless there is a major default event, are calculated through a large number of data, so not the more you spend on Taobao, Sesame credit score higher. At the same time, the user's default record will have a retention period, if the period of credit record is positive, credit points will be promoted.

He said that the ant in the Internet financial field has accumulated for many years, including the previous micro-loans, days of cat installment, Flower chant and other products, behind the large data for users to establish a personal credit file, so micro-loan in a minute can be issued loans. At present, some banks have recognized the data of ant, in some large loans, will refer to these data, the cooperation of the Bank has more than 10.

Use Scenarios and security

January 28, the first day of the test, a lot of participating users in the circle of friends to bask in their credit score, but when the personal credit is really established after how to use? Traditional credit may still be the biggest application scenario, but in addition, Sesame Credit has been extended to hotels, car rental and other fields. Ultimately, Deng says, he hopes to penetrate personal credit into the lives of users.

Deng puts forward several kinds of credit insufficiency way, deposit, advance authorization, guarantee. And once personal credit is established, in some scenarios, these methods will be canceled. It is understood that the current Sesame credit and Ali's tourism brands go to AH cooperation, in 3,000 hotels using sesame credit points, full 600 points can not pay a deposit; At the same time with the cooperation of the Shenzhou car rental will also be on line, as long as the sesame credit points to reach a certain level, without the use of

He also takes visas and recruiting abroad as an example visas to go abroad, banks need to freeze a part of the funds, but if the personal credit is established, according to the credit can be handled; recruitment, many companies will do the back, in essence, is a credit intermediary concept, and personal credit can do this intermediary things. Deng that the next most interesting application is to share the economy, including PP car rental (carpool), pp rental, everyone express and so on.

It is noteworthy that once a personal credit file is established, it also means that each user's data will be lumped together, and how to secure data security may be the biggest challenge in the future. Yu Wujie said that with the data, there is a calculation method, but the personal credit is the most important or safe, because personal credit is involved in each individual, so privacy protection, data security has become the most significant.

He said that in Ali the same data is left in different places, one place with any problems, another city machine can be restarted. Physical security is access control, two access control, two road power, to ensure business integrity. Also have their own network services to prevent hacker attacks. And for the information itself, also has the classification, from the information security own angle consideration, has divided many layers, first, the bank card information related, is absolutely the topmost, the most important; second, consumer information, and desensitization of data, not to any one person, is the group information. In the information transmission and preservation are through the complete encryption technology, for credit-related, must have special protection, how to authorize, let who to see, this is very important.

Market-oriented large-scale personal credit has started, but for credit agencies, how to build such a large model of the system, output products, still need time, and for users, acceptance and use of personal credits will face a long-term popularization process.

(Responsible editor: Mengyishan)