From credit to investment big data change finance

"You can imagine that all the information that can be collected can be used for finance, and the key is how to have innovative tools and ideas to make the data in your hands worthwhile." "The 2013 Internet Financial Innovation Summit, the Silicon Valley dialogue Beijing Forum, co-sponsored by Idg-accel, the new Financial Research Center for Finance and Economics, Long, said yesterday.

Long was once the chief architect of the Shanghai Stock Exchange trading system, it sounds cool that two-thirds of all Chinese stock transactions are executed by his program every day, but instead of helping people achieve deals with cold computer code, Long is doing what it is doing-the nuggets are more imaginative in big numbers.

Every minute of the internet has hot data fresh out, as long as there is electricity on the earth, the data on the Internet will continue to produce, then what can these data for the traditional financial industry to do?

Credit Score

All data are credit data. Many entrepreneurs from Peer-to-peer industries have invariably expressed the same view on the forums: Big data can be evaluated for credit.

"The United States to apply for a credit card, the name may be all lowercase, it may be all uppercase, which in our view credit is completely different, if a person can know when the size of his name, in a way, the name index is better, with the education background formed a positive correlation." For example, the Honda Accord and Nissan 350Z are, to a certain extent, a different degree of risk appetite: Nissan tends to be more aggressive and the repayment level is faster. "Turbo Financial group chief risk officer Gu Lingyun, for example.

Also using large data for risk assessment and credit evaluation are the US peer-to-peer lending industry's leading lending Club.

"With big data, the most important thing is how to find the most basic information you need to predict early warning fraud risk and credit risk." Lending Club chief Wind control officer Chen Sumei said, "We will also ask borrowers to provide a lot of other information, including why they want to borrow, the amount of money they want, education, careers and so on, in addition to the data on credit statistics." Third-party ratings include his mail, phone number and address, and computer IP addresses, which are operated online. Not only credit but also fraud risk, verification is also a very important part of this part of our manual operation accumulated experience, the future step-by-step automation. ”

And the Internet's evolving data can do more than just provide a static score for credit ratings. Using large data to do credit evaluation is mainly to observe two aspects: first there is no repayment will, the second has the ability to repay, but the two can not be perfect coordination. "The reason is very simple, because there is a lag, and the solution is to the discrete score into a continuous, hope that the resulting version is based on different data sources, every minute of change, not wait two or three months of information to change once." "Gu Lingyun said.

It is noteworthy that in the "financial" before the "internet" or "large data" modifier is not to ensure the success of financial innovation panacea, in fact, the Internet-driven model innovation also need to carry traditional financial technology to better development.

"We need to use the idea and technology of the Internet to upgrade and change traditional finance." "Long said," domestic peer-to-peer credit has thousands of companies, because this model is very good copy, really do good only a handful of few, but in addition to the model itself, there is a core business ability to grasp the credit, and the real master of this business ability is very few enterprises, This is the essence of traditional financial business. ”

Investment Guide

In addition to doing credit evaluation, big data is also starting to make a mark in directing investment. Thasos Group is a hedge fund that uses big data technology to invest, and for Wall Street investors who like to diversify, this new idea has received a lot of favor.

"As far as I know, Thasos Group is the only hedge fund that uses big data to invest, and yields exceed average hedge-fund trading levels outside of high-frequency trading, and they are able to perform well because of the scientific and full use of big data." Thasos Group chief scientist Pan said.

"Traditional investment institutions basically use financial and financial data to invest, and we use large data mining, through these data mining to accurately judge the behavior of American consumers, and then understand the trend of the U.S. macro-economic operation, so as to make the right investment decisions." Pan said in an interview with the first financial daily.

The use of financial data and large data is the pan distinguishing features of Thasos group and other hedge funds.

While some financial institutions have previously invested through the use of social networks for customer tracking, Pan believes that the way in which investment is based on social networks such as Twitter and Facebook has largely failed, "because these social networks provide data based on unstructured language, It's hard to pinpoint and ultimately bring benefits. "Most of the data on the Internet is semi-structured and unstructured.

Pan's technology is not the case, however, in view of market competition, Pan refused to disclose to our correspondent the large data sources or mining methods, but its language revealed that Thasos group's data mining is concerned with macroeconomic fundamentals and industry assessment, to accurately locate the company or industry to invest.

"The continued excavation of big data will change the investment industry as a whole, and the most direct investment approach, including the use of large data to create new credit rating agencies, will bring huge opportunities and markets." "Pan said.

In addition, from the market aspect, Pan told our correspondent, in the United States, in view of the financial innovation behind technological innovation, "once you have a different approach, many institutional investors will pounce on you, as long as we can prove that we differ, which helps these institutional investors to reduce their portfolio relevance, to achieve diversification of investment, We can find investors. ”

Yesterday's forum, including ZestFinance and other institutions have also explained the big Data mining on credit ratings, credit matching and other aspects of the potential for the great opportunities.

However, Pan also said that the current large data technology mining and investment also faces many from the regulatory, technical and market environment challenges.

"The tracking and use of data can have a certain amount of harm to individuals or societies, including the infiltration of privacy, and the regulatory body's avoidance of these injuries is, to some extent, a barrier to large data," he said. Pan told the newspaper reporter.

In addition, the nature of the large data itself determines that some of the conclusions are scientifically worthy of being screened, "as long as the data is large enough, you may find out what conclusions are possible, but these conclusions are likely to be wrong, and how to discriminate is a challenge, including the relationship between relativity and causality, which is more valuable. ”

Although Pan is very confident of big data changes in the investment industry, he believes that China's market still needs to be nurtured, "Big data investment does not apply to China's equity investment market, including China's debt level two market, money market, derivatives market, etc. are immature." ”

(Responsible editor: The good of the Legacy)