Large data turn financial services from extensive management to scientific management

August 19-20th, approved by the Ministry of Industry and Information technology, sponsored by China Communications Society, co-organised by Chinese Telecom, China Mobile and Unicom, 2014 China International Data Conference held successfully in Beijing. 36 Large data (36dsj.com) as the chief guest of the media to participate in the full coverage of the Conference, the following is the executive vice president of UnionPay Chaihong in the "Big Data application and development" venue speech, the main speech titled "China UnionPay Financial data exploration and practice."

Moderator: The next speaker is from China UnionPay Executive Vice President Chaihong professor, his speech theme is China UnionPay financial data exploration and practice, we applaud.

Chaihong: As for us, mainly from the angle of finance or payment, for everyone to serve. I have three aspects to report to you, one is UnionPay on the big data, some understanding of the morning if all the experts and friends to listen to the morning some leaders, some experts speech, this aspect of repeated content is more, I will be faster. The second I will talk about UnionPay in the big data some practice, report to everybody. The third is that we are in the process of large data discussion, there are two points of sentiment, but also share with you.

With the Internet, the International social Network, the original network of electronic business data in all of a sudden drive up. All walks of life at the same time in the process of information fusion, and constantly produce new data, this data we generally in accordance with the academic field, that is, structural, semi-structured and unstructured, just now Professor Stone is also talking about this aspect of things. The actual contact from our own side is the voice, image, video, the amount of data produced by human society has reached the ZB level, especially in these two years with the development of mobile Internet, our experts statistics and explanation, that is, we now two years of data is the human society produces 90% of the amount of data. The pace of further development may be faster, reaching 40ZB or 35ZB by 2020. One is the structure of the Internet and unstructured data, now all the mobile Internet data have a feature, are through the collection of our mobile phones, our PC after the internet generated. A lot of the data that comes out of the back will be generated from the machine to the machine, like the motion ring on my hand, which comes from the machine and then directly into the library.

However, as a large data research, we feel that although we are in the introduction of many different large data products, the formation of a blossoming trend. But on the other hand, the relevant technology of large data is not completely stereotyped, there is still a lot of room for development. This development space, large data at the national strategic level, all over the world are launching big data strategy. We can also see that in terms of the development of large data in the business engine, large data is constantly creating innovation in service patterns. Because of the emergence and development of large data, we are sure to have a lot of hope that entrepreneurs or new companies are also emerging in the wave of large data technology development. Traditional industries also have more reasonable and efficient data collection and processing technology, such as pharmaceuticals, automobiles, finance can be better through scientific data decision-making, to enhance the operational efficiency of enterprises.

We see the data scientists, the corresponding universities, research institutes including enterprises have also set the research direction, like our enterprise also set up the position of data scientists, to promote the study of large data. At the same time, technological innovation drives the improvement of data processing capability, and the methods of data processing are emerging, which can get better results through full amount of data, and make business model innovation be realized. Originally introduced the time, this large data technology, the law of large data is relevance, not causal, just Xu director said not only related, but also pay attention to cause and effect, we can from different angles to consider this problem.

We think that this large data can help to improve the efficiency of all aspects, several aspects we talk about some of our understanding. The first is public management, large data makes the traditional way of public management into the data based scientific work management. We can also see the application of large data in industrialization, and we see Ford Motor Company using car sensor network data and user social network data to analyze user's driving behavior. Large data are also widely used in economic and financial fields, the new economic analysis system based on large data can realize the forecast to the future, take our UnionPay as an example, can effectively predict the situation of the bank card in the next few days, can also through the correlation of weather, can analyze the amount of trading volume today, Indeed, we have done so in this respect.

The role of large data is also reflected in the service of life, in the presence of large data, we live in the services are product-centric, to provide customers with services. And by the large data technology to bring precision marketing, referral systems and other methods, so that today's life services are people-centered, can provide personalized service for everyone. We all know what is the famous Internet thinking? is the user-centric, user first, experience for the king. A simple word difference between customers and users has changed a lot. Because customers are the result of buying and selling relationships, and users are the services you enjoy.

Just now we understand all aspects of large data, and ultimately point to the data to guide life. We quantify life through large data technology, record everything, quantify everything, through data improvement, service for the core, through large data technology, the formation of large data thinking, all based on data, in the data to explore and learn, and ultimately bring scientific decision-making, meticulous production, predictable operation and personalized service.

The following report to you on the financial data of the practice of UnionPay, we all know that UnionPay was established in 2002, the bank card organization, now we are 400 member agencies, 400 banks are our partners. UnionPay has been established for 12 years, has become the world's first card issuer of the bank card organization, network size has been spread around the world's 142 countries, the world's second trading scale, UnionPay industry, achievement partners, benefiting society, human society from paper money to electronic currency this process, UnionPay to exercise this responsibility. When UnionPay was founded in 2002, I had an assertion that all of you in this Chamber, you see my assertion after 12 years of reincarnation, actually we have achieved it. At that time, my judgment was that our China, through the rapid development of the bank card industry, we can cross the developed countries paper personal check stage, after 12 years, this assertion including everyone's support, we have achieved. China UnionPay has a wealth of large data resources, involving 4.3 billion bank cards, more than 900 million of cardholders, more than 10 million households, every day 70 million transactions data, the daily core transaction data are more than TB, our UnionPay data resources both macro-level and micro-level have high value. Where does this value come from? I'm going to say it. The data value of UnionPay, the characteristics of payment data is more referential, they can measure the real purchase behavior, thus promote user orientation, personalized pricing, product recommendation, user loyalty and loss modeling strategy, This is my report on the value of payment data from our payment industry from this angle to everyone. Why is the report worth? The last of the cooperation will be discussed.

There are many background factors for UnionPay to carry out large data work, from the payment data of UnionPay, it also meets the challenge in the process of data processing of UnionPay. UnionPay's original processing method has met the needs of large data, from its own business look, many fine business also need large data strong support. From the perspective of partners, partners on the diversity of data services demand, from the e-commerce industry, through the mining of large data to enhance the level of industry development. From the National Demonstration project construction, large data related work is also the focus of China UnionPay to undertake the national project, UnionPay also in this respect increased the intensity. UnionPay has a guiding ideology for big data, first of all, we hope to integrate all kinds of internal and external data, through our cooperation, based on these data, the establishment of corresponding large data infrastructure, you can make these data can be safely and easily accessible, based on large data platform, we will carry out large-span data analysis and in-depth data mining work, These analysis and excavation work is done externally to the internal partners.

I first introduce the data of the General Assembly at four levels, this level is the basic data, basic platform, basic data, model research and application services, building a large data platform is the first step for UnionPay to carry out large data work, there is a cloud platform, full collection of data, we integrate common distributed machine learning algorithm, We have supported more than 10 business application systems in actual production. We and the Xinhua News Agency jointly released the bank card consumer Confidence Index (BCCI), which can read the changes in household consumption and consumption structure. The bank's data are also reflected in the BCCI index. We use large data, but also for the cardholder to provide data services, the first cardholder can through the UnionPay wallet this mobile phone app platform, inquires the history of their own bank card transactions. At the same time, the cardholder can provide a richer card holders of the billing services. Customer segmentation is also the future of E-commerce and electronic payment industry hotspot and Focus, Bank of the General Assembly data practice, based on external data and data quantification indicators and the overall characteristics of data, based on data quantification indicators, we can analyze the individual characteristics of each cardholder, based on the cardholder's individual characteristics and overall characteristics, We can label cardholders on a variety of types. such as business travel labels, medical labels, car tags, forming each of our purchase habits, habits of the image. In addition to the cardholder dimension, we have done a corresponding job in the merchant dimension, we provide business intelligence analysis for the merchant, can enable the merchant to realize its own business situation, we also provide the merchant Merchant's comparison, provides the best cooperation object for the merchant, also is he this chain cooperation object's joint service. At the same time, we also provide customers with customer loyalty, loss of merchant analysis, repeat customers analysis and so on.

At the same time, we also apply large data to the field of risk control. Using the method of machine learning, we excavate the historical data of UnionPay and get the judgment model of the seven major transactions, and can make a real-time judgment on the UnionPay transfer trading system. There is also a practice in the bank's data collection work, which stems from our understanding of the different stages of large data. We think that the large data from 1.0, 2.0, 3.0 different stages of development, the maturity of each stage gradually improve, we are very clear that the bank data resources have a high quality and value, but some deficiencies. So we have a limited ability to reach the end-user, so we want to work through the data collection, integration of internal and external data, so that the value of UnionPay's data is increased.

We try to set up the data collection from two angles, first using the open external technology, we use the reptile technology to find the cardholder and the Merchant's information on the Internet. In addition, the data of the partner are set up, such as the data of the communication operator, everyone knows about the Snowden incident, saying that if you use one months of your phone information and one months of credit card information, basically analyze what kind of person the person is, what behavior people are, what family background people can analyze, We are now doing the work with the communications operator.

Let's talk about our outlook later. This picture everyone uses a lot of, first big data is in the development of perfect. In many ways embodies the two sides, large data on the one hand in many industries are widely used, on the one hand not mature enough. Large data use a new technical framework, but sometimes not suitable for traditional scenarios, large data concepts have been fully interpreted, but the application of high value is not rich. I wonder if you have seen the play in the card room, as if it were made with big data. But I asked a friend of the United States, asked the director, he said it was exaggerated, this is also real. Including this is on the MBA case, but the actual question he is another story. Large data on the one hand blossom, on the one hand, the application of docking is not smooth phenomenon. We think that after a period of effort, the value of large data can be further highlighted, but it is not a universal technology that remains to be considered.

The second sentiment, regarding the information security problem in the data opening, the data opening brings the data value enhancement, brings the 1+1>2 effect, but also may bring the information security question, like the personal privacy leakage question, the better enhancement big data value, we need the legal policy level guidance. More needs the government and the enterprise's many attempts, but also needs everybody here the friends all to support mutually, can produce the cooperation.

As for the big data on finance, we still think that this will be one of the most promising areas for big data. Financial institutions can be said to be large data-born collaborators, on the one hand, there is a strong use of technology dividends to bring the income impulse, on the other hand, there is a better information base in the country, from the data level, the financial field has high-quality data resources, from the technical level, the financial enterprise's technical team also has a strong strength, from the thinking perspective , the financial industry has the most professional financial capacity, can study and develop the most professional financial data products and services. Financial data for the industry to bring changes will be all-round, for example, in the credit risk assessment, Professor Shi repeated the credit scoring this piece, the actual acquisition system has yet to be further excavation to serve the community. At the same time can also be in customer service based on large data technology, but also to achieve intelligent customer service. In the intelligent operation, large data technology can also analyze the data of financial enterprises, so as to help financial institutions to make operational decisions, reduce costs, in product innovation, large data to participate in the design of new products, combined with the data, so that the enterprise product innovation. In short, financial services will shift from extensive management to scientific management, from profit-centric to customer-centric. The data of the UN General Assembly will revolve around customer service, combine the internal and external resources, form the cardholder's cognition, such as consumption habit, life habit, based on the cardholder's cognition, we can form the cardholder's response forecast. For example, precision marketing, personalized services, at the merchant level, we will also be based on the resources of large data to form a deeper understanding of the various aspects of the merchant. such as the customer group business situation of the merchant, based on these knowledge, we will form the merchant's response forecast, such as the Merchant's development forecast, the credit appraisal.

The data work of the UN General Assembly will focus on providing stable, efficient and rich data services, and we will improve the stability and efficiency of large data systems. We believe that in the context of the rapid development of large data, cooperation and mutual win is the trend of the general situation, cooperation and win the first aspect is the complementarity of data, data sharing, can be integrated to create new value. The second aspect is the complementarity of resources, the cooperation of resources will bring about the improvement of mining ability, and the third aspect of cooperation and win is the complementary business. The complementarity and mutual use of the business will give full play to the maximum effectiveness of large data, through the cooperation of data, resources and so on, it is inevitable that many joint cross-border innovation results can be produced. Can give everybody big data construction to bring a broader imagination development space. Co-win is the consistent attitude of UnionPay, our areas of cooperation can be in many areas, including technical exchange and cooperation, open sharing of data, model model research, industry best practices, and data application cooperation, UnionPay is willing to work together partners to build a large data ecosystem of the bank card industry. Thank you for listening, thank you.