At present, the application of knowledge graph in the financial field is the most popular, involving risk control, marketing, forecasting and other key links in finance.
Application of knowledge graph to risk control in financial field
-
1. Anti-fraud application:
In recent years, the forms of financial fraud are diverse, providing false information, gang fraud, internal and external collusion and other methods are more and more “sophisticated”.
In this case, the original method of anti-fraud through single point breakthrough is far from enough, we need to establish a positive and effective knowledge map.
The core of anti-fraud is people. First of all, it is necessary to open up all data sources related to borrowers and build knowledge maps containing multiple data sources, so as to integrate them into structured data that can be understood by machines.
Here, we can not only integrate the basic information of borrowers, but also integrate the consumption records, behavior records, relationship information and online log information of borrowers into the anti-fraud knowledge map for analysis and prediction.
In addition to the application stage of fraud, fraud by building known elements (such as mobile phones, equipment, accounts, regional) the relationship between the map and comprehensive understanding of customer’s massive risk data statistical analysis, according to the thematic elements to collect feedback, as a result of risk operation signature database building customer risk, risk optimization model and rules, also can do trading phase of the fraud.
-
2. Application of internal audit and internal control
Similarly, with the help of relationship mining and knowledge mapping, we can also help financial institutions improve the efficiency and accuracy of internal audit and internal control systems.
Help financial institutions to prevent internal and external collusion, for example, conduct data mining on emails and account transactions of regulated personnel and build relevant networks, so as to timely discover violations of colluding with external personnel or abnormal account transactions.
-
3. Anti-money laundering application
In the field of anti-money laundering, we can also help the regulatory authorities to carry out effective monitoring. Through level by level mining of related accounts, we can find the money laundering accounts hidden behind.
Compared with the identification of individual accounts and relationships, it is more difficult to excavate anti-money laundering gangs. Such organizations are often hidden in a very complex network of relationships and difficult to be discovered.
Only by sorting out the relationship network and analyzing it from time and space, can we identify the potential risks and discover the hidden anti-money laundering gangs.
Marketing application of knowledge graph in financial field
-
1. Explore potential customers:
Mining potential customers has always been an important application concerned by the financial industry. Accurate and rapid identification of potential customers of relevant businesses through existing data and external data will be of great benefit to the improvement of banking business.
We can build a social network knowledge graph based on existing bank customers, and define the relationship model of the graph according to different communication modes and frequency. Conduct social mining for customers’ relatives, friends, colleagues, classmates, strangers, etc., and evaluate the closeness of the relationship.
For example, based on existing VIP customers, dig out relevant contacts and hobbies, or find an organization with common interests among existing customers, so as to develop marketing strategies for a certain part or a group of people.
-
2. Dig into potential customer needs:
In addition to exploring potential customers, we also need to explore the needs of existing customers and push relevant products according to their characteristics.
When we establish a knowledge graph system based on bank customer relationship, it can be flexibly expanded, such as adding vehicle information, personal hobbies, behaviors, etc. Combined with multiple data sources, it can more accurately analyze customer behavior, understand customer potential demand, and make precise push.
The above business can not only for individual customers, but also for enterprise customers. Analyze capital relationship, legal person relationship, upstream and downstream investment relationship, business relationship of similar enterprises, etc., and recommend appropriate products and services for enterprises.
Predictive applications of knowledge graphs in finance
-
1. Industry prediction of potential risks:
Based on multi-dimensional data, we can establish a closely related knowledge graph among customers, enterprises and industries to predict risks from the dimension of industry association.
By subdividing the industry, according to the loan information, industry information building relationships mining model, and through the machine learning model of training, can show every industry and some of the industry, with its highest correlation if one industry industry risk or high risk event happened, we can predict the future in time with the potential risk of related industries, In this way, financial institutions can predict the risks of related industries and find and avoid risks as soon as possible.
-
2. Forecast of potential risk customers:
By establishing knowledge maps of customers, enterprises and industries, data between industries and enterprises can also be connected, and enterprise customers associated with industry risks and systemic risks can be found in a timely manner based on the prediction of potential risks of the industry.
For example, there are several overdue loans in a certain industry in a province recently. By analyzing the knowledge graph of the industry and customers, we can timely find other customers with potential risks that may be located in related industries or upstream and downstream.
References:
- www.infoq.cn/article/App…
- Baijiahao.baidu.com/s?id=157232…
- www.sohu.com/a/207380688…