Data credit: your circle of friends to decide how much money you can borrow from the bank-credit, micro-Alipay-IT information

Last August, Facebook in the United States had successfully applied for a patent, the patent primarily through the analysis of data to a user's friends to do something. One of them reads as follows:

When a user when applying for loans, the lender will review the credit rating of the users social networking friends. The average credit rating of only those friends meet the minimum credit score requirements, loan just will continue to process the loan application. Otherwise, the application shall be rejected.

This patent caused a huge controversy abroad, some call it "loan discrimination". Aside these disputes aside, This is actually in line with an interesting social-economic theory: that is, your income is the average of the 10 people you often.

People category, according to the group, which in today's society, is indeed a more precise structural outline.

The same theory applies also to Tencent micro-data evaluation, and QQ user's credit rating, this is the age of big data, construct the user the basic principles of the credit system. Unlike the Tencent's underlying data from the user app has a strong social connections, and Ali, Jingdong data from user's transactions.

Data credit will enhance risk control capability

Baidu and East of Beijing and saw the United States ZestFinance, founded in 2009, is a machine learning and data technology to personal credit scoring technology Finance Corporation, its rise, and to some extent with the United States and triggered the global financial crisis of 2008 related to the subprime mortgage crisis.

United States traditional financial institutions according to the user's credit score, calculated by models from FICO Corporation. But the FICO credit scoring model used only less than 50 variables, this leads to master the scoring form that consumers can brush. In addition, the information dimension in which FICO single, in the age of big data, the traditional credit risk evaluation system still relies mainly on consumer credit history to credit assessment. Especially during and after the 2008 financial crisis, FICO scores in the United States basically no major changes in the distribution of the population, which after the outbreak of the financial crisis a large number of bad debts, the reality of great change in the financial situation of many people risked. Therefore, FICO's ability to predict risk questioned by industry.

And ZestFinance to big data technology for based, its collection data of source more diverse, which has about 30% of traditional credit records, addition adopted has more may effect user credit of information, as social network information, and user application information even user of writing habits, and reading habits, non-traditional data information, according to user borrowing behavior behind of clues and the clues between of associated sex, eventually to out relative accurate of consumers credit score.

This technology and credit evaluation model, is the place where ZestFinance most valuable investment. The case of Baidu, Taobao, it has no app or the hundreds of millions of people in social or transaction of the underlying data, but through a number of effective data collection and analysis system, users in Baidu's search behavior, even stick registered users speak, information can generate a lot of clues. Of course, this kind of information compared to the effectiveness of micro-credit and PayPal certainly is quite different, so a "smart" identification and analysis of systems is more urgent.

On the Internet, where, will leave a mark. This sentence also applies to areas of credit, especially in the era of the Internet has been all of our lives, most people who understand your credit situation in the future, maybe not your partner or your parents, but your data.

Is the new world or Pandora's box?

Traditional financial institutions credit coverage, but relatively cautious approach to avoiding lending money the "wrong" people, and credit data in the Internet age, under the traditional credit reporting system so that more might have to borrow money is to get more financing.

Open the door at the same time, risks also go hand in hand.

With the rapid rise of Internet financial two years ago, many in the traditional financial institutions to borrow money to individuals and businesses through Internet banking platform can easily borrow money, especially as represented by the P2P platform, rapid expansion of scale. But in the absence of a sound credit system's background, the rapid development of business is tantamount to gather sand from the Tower, a slight touch will be defeated.

Internet convenience and risk control of financial difficulties, which are two sides of a coin. Therefore, when Ali, Tencent, Baidu, Beijing East Giants eye when aimed at the financial sector, the establishment of credit system, becomes an inevitable choice of the financial scene after maturing.

Seen from the means, more corporate data over the Internet in China is similar to the Zestfinance model, from users of consumer behavior, social relationships or other information left to judge on the Internet user's credit rating. In this respect, have a super platform for trading of Sesame imported credit and credit with Super social Portal Tencent holds the great advantage, compete in the future, who have more third-party commercial scene.

It is also consistent with the Chinese Central Bank "data from a third party, use third-party" requirement. For businesses, crediting the source the more risk the more perfect.

Represented by Ali, Tencent's Internet giants has been quietly collecting kindled a fire in this area, now need to wait, or consumer credit licence issued by the Central Bank of China "East wind".

大数据征信：你的朋友圈决定你能从银行借到多少钱 - 征信,微信,支付宝 - IT资讯

去年8月，Facebook在美国成功申请了一个专利，该专利主要是通过分析某个用户的好友数据来做一些事情。其中一项内容如下：

当一个用户申请贷款的时候，贷款方会审查该用户社交网络好友的信用等级。只有这些好友的平均信用等级达到了最低的信用分要求，贷款方才会继续处理贷款申请。否则的话，该申请即被拒绝。

这个专利在海外引起了巨大争议，有人称之为“贷款歧视”。抛开这些争议不谈，这其实符合一个有趣的社交经济学理论：即你的收入是你经常联系的10个人的平均值。

人以类聚、物以群分，这在当今社会，确实是一个较为精准的结构性概括。

同样的理论，其实也适用于腾讯以微信、QQ的用户数据为基础评价用户的信用等级，这就是大数据时代构建用户征信系统的基本原理。不同的是，腾讯的基础数据来自于对用户有着强关系社交连接的微信，而阿里、京东的数据来自于用户的交易行为。

大数据征信将提升风控能力

被百度和京东同时看中的美国ZestFinance，成立于2009年，是一家通过机器学习和大数据技术进行个人信用评分的科技金融公司，它的崛起，一定程度上与美国2008年爆发并进而引发全球金融危机的次贷危机有关。

此前美国传统金融机构针对用户的信用评分，计算方法模型都来自FICO公司。但FICO信用评估模型仅使用不到50条变量，这导致了掌握评分套路的消费者可以进行刷分。此外，FICO所采用的信息维度较为单一，在大数据时代，这一传统信用风险评估体系仍主要依赖消费者的信贷记录去进行信用评估。特别是在2008年金融危机前后，FICO评分在美国人口中的分布基本上没有大的改变，而这与金融危机爆发之后出现大量坏账、许多人的财务状况发生极大改变的现实严重不符。因此，FICO预测风险的能力受到业界质疑。

而ZestFinance以大数据技术为基础，其采集数据的来源更多样，其中有大约30%的传统信贷记录，另外采纳了更多可能影响用户信用的信息，如社交网络信息、用户申请信息甚至用户的写作习惯、阅读习惯等非传统数据信息，根据用户借款行为背后的线索及线索间的关联性，最终给出相对准确的消费者信用评分。

这套技术和信用评价模式，应该就是ZestFinance最值得投资的地方。以百度为例，它没有微信或淘宝那样基于数亿人的社交或交易行为产生的基础数据，但通过一些有效的数据采集和分析系统，用户在百度的搜索行为，甚至贴吧注册用户的发言，都是可以产生大量线索的信息。当然，这种信息的有效性相比微信和支付宝无疑有较大差别，因此一套“精明”的识别和分析系统就显得更加迫切。

在互联网上，凡走过的，必留下痕迹。这句话同样适用于征信领域，尤其是在互联网已经深入我们生活方方面面的时代，未来最了解你的信用情况的人，也许不是你的伴侣或你的父母，而是你的大数据。

是新大陆还是潘多拉盒子？

传统金融机构的征信方式虽然覆盖面窄，但相对谨慎的方式可以避免将钱借给“不对”的人，而互联网时代的大数据征信，可以让更多以前可能在传统征信体系下借不到钱的人也能得到更多融资。

在打开大门的同时，风险也如影随形。

前两年随着互联网金融迅速兴起，许多在传统金融机构借不到钱的个人或企业，通过互联网金融平台可以很方便地借到钱，特别是以P2P为代表的平台，规模急速扩张。但在缺乏健全征信体系的背景下，这种飞速发展的业态无异于聚沙而成的塔，稍一碰触便会溃散。

互联网金融的便利性与风险控制的难度，这是一枚硬币的两面。因此，当阿里、腾讯、百度、京东等巨头纷纷将目光瞄向了金融领域时，征信体系的建立，也就成了金融场景日益成熟后的必然选择。

从手段上看，中国互联网企业的大数据征信更多类似于Zestfinance的模式，从用户的消费行为、社交关系或其他互联网上留下的信息去评价用户的信贷评级。在这方面，拥有超级交易平台入口的芝麻信用和拥有超级社交入口的腾讯征信占有较大优势，未来要比拼的，是谁拥有更多的第三方商业化场景。

这也符合中国中央银行“数据来源于第三方，使用于第三方”的要求。对企业来说，授信源越多，风控越完善。

以阿里、腾讯为代表的互联网巨头们，已经悄然在征信这个领域燃起了战火，如今需要等待的，也就是中国中央银行发放个人征信牌照的“东风”了。