Knowledge graph construction method and system

By building a knowledge graph of the financial trading platform, combining history and real-time data, the problem of not being able to quickly correlate and display complex relationships in the existing technology is solved, real-time monitoring and data updates are realized, and more intuitive user relationship network analysis is provided.

CN119026683APending Publication Date: 2024-11-26HAIER CONSUMER FINANCE CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411048247.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The existing risk monitoring system cannot combine real-time big data with complex relationship networks, resulting in the inability to quickly correlate real-time data, the generated graphs are slowly updated, and it is impossible to visually and clearly display complex relationships.

Method used

By obtaining the historical transaction data and real-time transaction data of the financial trading platform, initializing and structuring processing, building node and edge relationships, and using distributed graph database and data flow engine to write knowledge graphs in real time, real-time monitoring and reasoning are realized.

Benefits of technology

Real-time data updates are realized, which can display the user relationship network more intuitively and clearly, quickly understand the relationships and patterns, support fast real-time visual analysis, and realize millisecond data updates.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119026683A_ABST
    Figure CN119026683A_ABST
Patent Text Reader

Abstract

The invention relates to a knowledge graph construction method and system, and belongs to the technical field of financial data processing, and the method comprises the steps: obtaining historical transaction data and real-time transaction data of a financial transaction platform; performing initialization and structured processing on the historical transaction data to obtain structured data, and writing the structured data into a data warehouse for storage; constructing a first node and first edge relationship according to the structured data, writing the relationship into a distributed graph database through a data processing module, and initializing the knowledge graph; a second node and a second edge relation are constructed according to real-time transaction data, the second node and the second edge relation are written into a corresponding first node and a corresponding first edge relation in a distributed graph database in real time through a data flow engine, and a knowledge graph for real-time monitoring is constructed, so that knowledge graph reasoning is carried out, real-time data updating can be realized, and the real-time transaction efficiency is improved. A huge relation network system of the user is displayed more visually and clearly, and potential attributes of the user are analyzed quickly and visually in real time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of financial data processing, and in particular to a knowledge graph construction method and system. Background Art

[0002] At present, fraud is still a difficult problem to eradicate completely in the banking, e-commerce, insurance and even social networking industries. In order to meet this challenge, various industries have adopted a series of technical means to identify potential fraud risks more accurately and in real time, so as to prevent and avoid these risks in advance as much as possible. Through these technical means, various industries strive to ensure the security and reliability of transactions and protect their interests from infringement.

[0003] The risk monitoring systems commonly used in the market still have the following problems: 1. They do not combine real-time big data with complex relationship networks. 2. The real-time data processing steps are complicated, and it is impossible to quickly associate real-time data. The data update is slow, and the generated graphs cannot be quickly monitored. 3. The number of loan users is huge and the relationships are complex. The existing risk monitoring systems often cannot intuitively and clearly display the complex relationships. Summary of the invention

[0004] In view of the shortcomings existing in the relevant technologies, the purpose of the present invention is to provide a knowledge graph construction method and system to solve the technical problems that the existing technology cannot quickly associate real-time data, the generated graph cannot achieve rapid monitoring, and the complex relationships cannot be intuitively and clearly displayed.

[0005] The present invention provides a knowledge graph construction method, comprising the following steps:

[0006] S1. Obtain historical transaction data and real-time transaction data of a financial transaction platform, wherein the historical transaction data and real-time transaction data include transaction operation data, user information data, user relationship data, and device data;

[0007] S2. Initialize and structure the historical transaction data to obtain structured data, and write the structured data into a data warehouse for storage;

[0008] S3. Construct a first node and a first edge relationship according to the structured data stored in the data warehouse, write the first node and the first edge relationship into a distributed graph database through a data processing module, and initialize the knowledge graph;

[0009] S4. Construct a second node and a second edge relationship based on the real-time transaction data, and write the second node and the second edge relationship into the corresponding first node and first edge relationship in a distributed graph database in real time through a data flow engine, so as to construct a knowledge graph for real-time monitoring and perform knowledge graph reasoning.

[0010] The embodiment of the present invention constructs a knowledge graph through historical data and real-time data, which can realize real-time data updating and present the user's huge relationship network system in a more intuitive and clearer way, so that the business can quickly understand the associations and patterns therein, and can quickly and visually analyze the user's potential attributes in real time.

[0011] In some embodiments of the present invention, step S3 further comprises:

[0012] A first node and a first edge relationship are constructed according to the structured data, and corresponding point attributes and edge attributes are configured based on the first node and the first edge relationship, wherein the first node includes a user information node, a spending information node, a credit information node, an address information node, a contact book information node, a call record information node, a device information node, a mobile phone number information node and a GPS information node, and the first edge relationship is an association relationship between any two of the first nodes.

[0013] In some embodiments of the present invention, the step S4 further comprises: constructing a second node and a second edge relationship according to the real-time transaction data through the real-time generation logic of the data stream engine, and writing the second node and the second edge relationship to the corresponding first node and the first edge relationship in the distributed graph database in real time,

[0014] The second node includes real-time user information node, expenditure information node, credit information node, address information node, address book information node, call record information node, device information node, mobile phone number information node and GPS information node, and the second edge relationship is an association relationship between any two of the second nodes.

[0015] In some embodiments of the present invention, the point attributes include node constraints and / or key attributes thereof, and the edge attributes include the start point and end point of the edge and key attributes thereof.

[0016] In some embodiments of the present invention, the present invention further includes:

[0017] S5. intercept all associated nodes centered on the user information node and generate a subgraph, and analyze the user's relationship network based on the subgraph to screen out bad users.

[0018] Some embodiments of the present invention further provide a knowledge graph construction system, including:

[0019] A data acquisition module, which acquires historical transaction data and real-time transaction data of a financial transaction platform, wherein the historical transaction data and real-time transaction data include transaction operation data, user information data, user relationship data, and device data;

[0020] A data processing module, which performs initialization and structural processing on the historical transaction data to obtain structured data, and writes the structured data into a data warehouse for storage;

[0021] A graph initialization module, which constructs a first node and a first edge relationship according to the structured data stored in the data warehouse, writes the first node and the first edge relationship into a distributed graph database through a data processing module, and initializes the knowledge graph;

[0022] A graph construction module constructs a second node and a second edge relationship according to the real-time transaction data, and writes the second node and the second edge relationship into the corresponding first node and first edge relationship in a distributed graph database in real time through a data flow engine, so as to construct a knowledge graph for real-time monitoring and perform knowledge graph reasoning.

[0023] In some embodiments of the present invention, the graph initialization module further comprises:

[0024] A first node and a first edge relationship are constructed according to the structured data, and corresponding point attributes and edge attributes are configured based on the first node and the first edge relationship, wherein the first node includes a user information node, a spending information node, a credit information node, an address information node, a contact book information node, a call record information node, a device information node, a mobile phone number information node and a GPS information node, and the first edge relationship is an association relationship between any two of the first nodes.

[0025] In some embodiments of the present invention, the graph construction module further comprises: constructing a second node and a second edge relationship according to the real-time transaction data through the real-time generation logic of the data flow engine, and writing them into the corresponding first node and first edge relationship in the distributed graph database in real time,

[0026] The second node includes real-time user information node, expenditure information node, credit information node, address information node, address book information node, call record information node, device information node, mobile phone number information node and GPS information node, and the second edge relationship is an association relationship between any two of the second nodes.

[0027] In some embodiments of the present invention, the point attributes include node constraints and / or key attributes thereof, and the edge attributes include the start point and end point of the edge and key attributes thereof.

[0028] In some embodiments of the present invention, the present invention further includes:

[0029] The graph analysis module intercepts all related nodes centered on the user information node and generates a subgraph, and analyzes the user's relationship network based on the subgraph to screen out bad users.

[0030] The beneficial effects of the present invention are:

[0031] 1. Display the user's huge network system in a more intuitive and clear way, so that the business can quickly understand the associations and patterns;

[0032] 2. It can quickly and visually analyze the potential attributes of loan users in real time, giving great freedom and flexibility, and can conduct in-depth data exploration and decision support according to needs;

[0033] 3. By combining real-time data with historical data, it is possible to achieve rapid monitoring and rapid response, and realize millisecond-level data updates. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the specific embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0035] Figure 1 A flowchart of a knowledge graph construction method provided in an embodiment of the present invention;

[0036] Figure 2 An entity relationship diagram of a knowledge graph construction method provided by an embodiment of the present invention;

[0037] Figure 3 A schematic diagram of the structure of a knowledge graph construction system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0039] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0040] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0041] It should be noted that the financial transaction platform in this application clarifies the content of the entire life cycle of user transaction data and personal information from collection to use, storage, sharing, and outbound transfer, while guaranteeing the exercise of users' rights to review, correct, and delete; when a user uses / opens the financial transaction platform for the first time, the user is prompted to read in a prominent manner (pop-up window embedded link, etc.), and is provided with click-to-agree, check-box settings, etc. to obtain the user's authorization and consent to the user agreement and privacy policy.

[0042] The technical solution of the present invention is described in detail below in conjunction with specific embodiments and the accompanying drawings.

[0043] Embodiment 1

[0044] As attached Figure 1 As shown, in an exemplary embodiment of a knowledge graph construction method of the present invention, the following steps are included:

[0045] S1. Obtain historical transaction data and real-time transaction data of the financial transaction platform. The historical transaction data and real-time transaction data include transaction operation data, user information data, user relationship data, and device data;

[0046] S2. Perform data cleaning, initialization operations after deduplication and structural processing on historical transaction data to obtain structured data, and write the structured data into the data warehouse for storage;

[0047] S3. Logically process the structured data stored in the data warehouse through SQL to construct the first node and the first edge relationship, write the first node and the first edge relationship into the distributed graph database through a data processing module, and initialize the knowledge graph; optionally, the data processing module is Spark SQL. After logically processing the structured data, the generated data information is written into the knowledge graph in batches at one time through Spark SQL, thereby completing the data initialization work. Spark SQL provides users with a powerful tool to process and analyze structured data in Spark, query data using the standard SQL language, and execute these queries using Spark's distributed computing capabilities, thereby supporting data analysis, report generation, data mining, and integration with other Spark components.

[0048] S4. Construct the second node and the second edge relationship according to the real-time transaction data through the real-time generation logic of the data flow engine, and write the second node and the second edge relationship into the corresponding first node and the first edge relationship in the distributed graph database in real time through the data flow engine to construct a knowledge graph for real-time monitoring, so as to perform knowledge graph reasoning; Optionally, the data flow engine is Flink, which can process large-scale data streams in real time and generate results in real time, and has the characteristics of low latency, high throughput and accuracy, and is suitable for scenarios requiring fast response and real-time decision-making, such as real-time alarm, real-time monitoring, and real-time recommendation.

[0049] Based on the above method, the knowledge graph constructed through historical transaction data and real-time transaction data can realize real-time data update, and present the user's huge relationship network system in a more intuitive and clear way, so that the business can quickly gain insight into the associations and patterns therein. The initialization process is imported through spark sql, and the real-time data is obtained through fl ink real-time logic processing, which greatly improves work efficiency.

[0050] Combined with reference Figure 2 As shown, the first node includes a user information node, a spending information node, a credit information node, an address information node, a contact information node, a call record information node, a device information node, a mobile phone number information node and a GPS information node, and the first edge relationship is an association relationship between any two of the first nodes;

[0051] The second node includes real-time user information nodes, expenditure information nodes, credit information nodes, address information nodes, address book information nodes, call record information nodes, device information nodes, mobile phone number information nodes and GPS information nodes. The second edge relationship is the association relationship between any two of the second nodes. The edge between the user information node and the credit information node is expressed as: user information->credit information; similarly, the association relationship between any two of the second nodes is listed in sequence as: user information->expenditure information, user information->mobile phone number information, user information->device information, device information->address book information, device information->call record information, credit information->address information, credit information->mobile phone number information, credit information->device information, expenditure information->address information, expenditure information->mobile phone number information, credit information->device information, address book information->mobile phone number information, call record information->mobile phone number information, credit information->GPS information and expenditure information->GPS information.

[0052] Further, based on the first node configuration corresponding to the point attribute, the point attribute includes the node constraint and / or its key attribute, wherein:

[0053] User information should be uniquely constrained by the user code and contain at least the following key attributes: user name, user gender, user ID number, user registration time, user risk level, and whether the user is on the blacklist;

[0054] The credit information shall be based on the credit code as the only constraint and shall contain at least the following key attributes: user code, credit result, credit time, and credit amount;

[0055] The expenditure information shall be uniquely constrained by the expenditure code and shall contain at least the following key attributes: user code, expenditure status, expenditure time, repayment status, and expenditure amount;

[0056] Address information should use the detailed address as the only constraint and contain at least the following key attributes: province, city, district, county, and city level;

[0057] Mobile phone number information should use the mobile phone number as the only constraint and include at least the following key attributes: user code, mobile phone type (landline, mobile, other), province, city, and city level;

[0058] Device information should be uniquely constrained by the device serial number and contain at least the following key attributes: device brand, whether it is blacklisted;

[0059] GPS information should use longitude and latitude as joint constraints and contain at least the following key attributes: whether it is blacklisted;

[0060] The address book information should be uniquely constrained by the device serial number plus the fixed suffix of the address book information;

[0061] The call record information should be uniquely constrained by the device serial number plus the fixed suffix of the call record information.

[0062] Based on the above configuration, this embodiment can extract entity nodes and important attributes of the nodes that are worthy of attention based on historical transaction data and real-time transaction data. For the configuration of important attributes, priority is given to attributes and related attributes that may pose risks during the transaction process. For example but not limited to, let t_cust be user information, and its point attributes are mainly user number, user name, user age, user gender, user identity information, and whether it is on the blacklist; t_device is device information, and its point attributes are mainly device serial number, device brand, and whether it is on the blacklist.

[0063] Furthermore, the corresponding edge attributes are configured based on the first edge relationship, and the edge attributes include the start point, the end point and key attributes of the edge, wherein:

[0064] User information->credit information starts and ends with user code and credit code, and contains at least the following key attributes: user code, credit code, application time, and application channel;

[0065] User information->Expenditure information starts and ends with user code and expenditure code, and contains at least the following key attributes: user code, expenditure code, application time, and application channel;

[0066] User information->mobile phone number information starts and ends with user code and mobile phone number, and contains at least the following key attributes: user code, mobile phone number, binding time;

[0067] User information->device information starts and ends with the user code and device serial number, and contains at least the following key attributes: user code, device serial number, binding time;

[0068] Device information->Address book information starts and ends with the device serial number and the device serial number plus the address book information suffix, and contains at least the following key attributes: device serial number, binding time;

[0069] Device Information->Call Record Information starts and ends with the device serial number and the device serial number plus the call record information suffix, and contains at least the following key attributes: device serial number, binding time;

[0070] Credit information->Address information starts and ends with the credit code and detailed address, and contains at least the following key attributes: credit code, detailed address, address type (home address, company address, other), binding time;

[0071] Credit information->mobile phone number information starts and ends with the credit code and mobile phone number, and contains at least the following key attributes: credit code, mobile phone number, binding time;

[0072] Credit information->Device information starts and ends with the credit code and device serial number, and contains at least the following key attributes: credit code, device serial number, binding time;

[0073] Spending information->address information starts and ends with the spending code and detailed address, and contains at least the following key attributes: spending code, detailed address, address type (home address, company address, other), binding time;

[0074] Spending information->mobile phone number information starts and ends with the spending code and mobile phone number, and contains at least the following key attributes: spending code, mobile phone number, binding time;

[0075] Spending information->device information starts and ends with the spending code and device serial number, and contains at least the following key attributes: spending code, device serial number, binding time;

[0076] Contact information->Mobile phone number information starts and ends with the device serial number plus the contact information suffix and mobile phone number, and contains at least the following key attributes: device serial number, mobile phone number, note name, creation time;

[0077] Call record information->mobile phone number information starts and ends with the device serial number plus the call record information suffix and the mobile phone number, and contains at least the following key attributes: device serial number, mobile phone number, remark name, call type (caller, called), call start time, and call duration;

[0078] Credit information->GPS information uses the credit code and longitude and latitude as the starting and ending points, and contains at least the following key attributes: credit code, longitude, latitude, and binding time;

[0079] Spending information->GPS information uses the spending code and longitude and latitude as the starting and ending points, and contains at least the following key attributes: spending code, longitude, latitude, and binding time.

[0080] Based on the above, by combing through various point relationships, the association relationship of each entity bearing point can be defined, and the connection attributes between points, that is, the edge relationship attributes, can be established. For example but not limited to, let e_device_id be the relationship established by user information->device information. By generating an association relationship, that is, a certain user has used a certain device to operate, the user's potential attributes can be analyzed point-to-point quickly and in real time. Visualization.

[0081] Based on the knowledge graph constructed by the first node, the second node and their edge relationships and attributes, that is, the obtained real-time complex relationship network graph about the visualized user, business personnel and business systems can create customized application rules and create real-time risk control decisions and perform visual output. For example, but not limited to, through the association relationship between user information->credit information->GPS information, it can be inferred that the risk control decision rules about the user and the risk control decision rules about the user's GPS are the number of days since the user's last credit application, the number of days since the user's earliest credit application, whether the GPS is on the blacklist when the credit is granted, etc.; through user information->device information->address book information->mobile phone number information, it can be inferred that the risk control decision rules about the user's device and the risk control decision rules about the user's contacts are whether the device is on the blacklist, the distance the device has moved in the last hour, the number of blacklisted user contacts, the number of credits granted to user contacts in the last three months, etc., thereby enabling point-to-point rapid real-time visual analysis of the potential attributes of loan users and realizing monitoring of the user's relationship network.

[0082] In order to intercept loan application requests from bad users, all related nodes centered on the user information node are intercepted and a subgraph is generated. The user's relationship network is analyzed based on the subgraph. For example but not limited to, if you want to get the number of contacts on the blacklist in the user's address book, you can get a user subgraph through the knowledge graph: user information->device information->address book information->mobile phone number information, and check whether the blacklist attribute is true in all the matched mobile phone number information. You can get the number of blacklisted contacts by summing up. If you want to know whether the GPS has been marked as blacklisted when the user makes a payment, you can get a user subgraph through the knowledge graph: user information->spending information->GPS information, and check whether the blacklist attribute is true in the matched GPS information. If the blacklist attribute is true, the user is on the GPS blacklist when making a payment, thereby screening out bad users and being able to quickly reject their loan application requests in a specific aspect.

[0083] Embodiment 2

[0084] Reference Figure 3 As shown, in an exemplary embodiment of a knowledge graph construction system of the present invention, it includes:

[0085] Data acquisition module, which acquires historical transaction data and real-time transaction data of the financial transaction platform. Historical transaction data and real-time transaction data include transaction operation data, user information data, user relationship data, and device data;

[0086] The data processing module initializes and structures the historical transaction data to obtain structured data, and writes the structured data into the data warehouse for storage;

[0087] A graph initialization module, which constructs a first node and a first edge relationship according to the structured data stored in the data warehouse, writes the first node and the first edge relationship into a distributed graph database through a data processing module, and initializes the knowledge graph;

[0088] A graph construction module, which constructs a second node and a second edge relationship according to real-time transaction data, writes the second node and the second edge relationship to the corresponding first node and the first edge relationship in a distributed graph database in real time through a data stream engine, and constructs a knowledge graph for real-time monitoring, thereby performing knowledge graph reasoning;

[0089] The graph analysis module intercepts all related nodes centered on the user information node and generates a subgraph, and analyzes the user's relationship network based on the subgraph to screen out bad users.

[0090] Among them, the graph initialization module further includes: constructing a first node and a first edge relationship according to structured data, and configuring corresponding point attributes and edge attributes based on the first node and the first edge relationship, wherein the first node includes a user information node, a spending information node, a credit information node, an address information node, a contact book information node, a call record information node, a device information node, a mobile phone number information node and a GPS information node, and the first edge relationship is an association relationship between any two of the first nodes.

[0091] Among them, the graph construction module further includes: constructing a second node and a second edge relationship according to the real-time transaction data through the real-time generation logic of the data flow engine, and writing it in real time to the corresponding first node and first edge relationship in the distributed graph database, the second node includes real-time user information node, expenditure information node, credit information node, address information node, address book information node, call record information node, device information node, mobile phone number information node and GPS information node, and the second edge relationship is the association relationship between any two of the second nodes.

[0092] Furthermore, the point attributes include node constraints and / or key attributes thereof, and the edge attributes include the start point, end point, and key attributes of the edge.

[0093] It should be noted that the above is a reference method for constructing a knowledge graph and a system, and the present invention is not limited thereto.

[0094] The embodiments of the present invention abstract historical transaction data and real-time transaction data into entity bearing points and certain important attributes of the points that deserve attention, and then define the association relationship between each entity bearing point to achieve a more intuitive and clearer way to display the user's huge relationship network system, so that the business can quickly understand the association relationship and pattern therein; and by combining real-time transaction data with historical transaction data, it can achieve the purpose of rapid monitoring and rapid response, and realize millisecond-level data update, which solves the problem that the real-time data processing steps in the prior art are complicated, the real-time data cannot be quickly associated, the data update is slow, and the generated map cannot achieve rapid monitoring; regarding the large number of loan users and complex relationships, the existing risk monitoring system often cannot intuitively and clearly display the complex relationships.

[0095] Finally, it should be noted that: the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0096] The above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or some technical features can be replaced by equivalents without departing from the spirit of the technical solution of the present invention, which should be included in the scope of the technical solution for protection of the present invention.

Claims

1. A knowledge graph construction method, characterized in that: The steps include: S1. Obtain historical transaction data and real-time transaction data of a financial transaction platform, wherein the historical transaction data and real-time transaction data include transaction operation data, user information data, user relationship data, and device data; S2. Initialize and structure the historical transaction data to obtain structured data, and write the structured data into a data warehouse for storage; S3. Construct a first node and a first edge relationship according to the structured data stored in the data warehouse, write the first node and the first edge relationship into a distributed graph database through a data processing module, and initialize the knowledge graph; S4. Construct a second node and a second edge relationship based on the real-time transaction data, and write the second node and the second edge relationship into the corresponding first node and first edge relationship in a distributed graph database in real time through a data flow engine, so as to construct a knowledge graph for real-time monitoring and perform knowledge graph reasoning.

2. The knowledge graph construction method according to claim 1, characterized in that: The step S3 further comprises: A first node and a first edge relationship are constructed according to the structured data, and corresponding point attributes and edge attributes are configured based on the first node and the first edge relationship, wherein the first node includes a user information node, a spending information node, a credit information node, an address information node, a contact book information node, a call record information node, a device information node, a mobile phone number information node and a GPS information node, and the first edge relationship is an association relationship between any two of the first nodes.

3. The knowledge graph construction method according to claim 2, characterized in that: The step S4 further comprises: constructing a second node and a second edge relationship according to the real-time transaction data through the real-time generation logic of the data stream engine, and writing the second node and the second edge relationship to the corresponding first node and the first edge relationship in the distributed graph database in real time, The second node includes real-time user information node, expenditure information node, credit information node, address information node, address book information node, call record information node, device information node, mobile phone number information node and GPS information node, and the second edge relationship is an association relationship between any two of the second nodes.

4. The method for constructing a knowledge graph according to claim 2, characterized in that: The point attributes include node constraints and / or key attributes thereof, and the edge attributes include the start point and end point of the edge and key attributes thereof.

5. The knowledge graph construction method according to claim 2, characterized in that: Also includes: S5. intercept all associated nodes centered on the user information node and generate a subgraph, and analyze the user's relationship network based on the subgraph to screen out bad users.

6. A knowledge graph construction system, characterized in that: include: A data acquisition module, which acquires historical transaction data and real-time transaction data of a financial transaction platform, wherein the historical transaction data and real-time transaction data include transaction operation data, user information data, user relationship data, and device data; A data processing module, which performs initialization and structural processing on the historical transaction data to obtain structured data, and writes the structured data into a data warehouse for storage; A graph initialization module, which constructs a first node and a first edge relationship according to the structured data stored in the data warehouse, writes the first node and the first edge relationship into a distributed graph database through a data processing module, and initializes the knowledge graph; A graph construction module constructs a second node and a second edge relationship according to the real-time transaction data, and writes the second node and the second edge relationship into the corresponding first node and first edge relationship in a distributed graph database in real time through a data flow engine, so as to construct a knowledge graph for real-time monitoring and perform knowledge graph reasoning.

7. The knowledge graph construction method according to claim 6, characterized in that: The graph initialization module further includes: A first node and a first edge relationship are constructed according to the structured data, and corresponding point attributes and edge attributes are configured based on the first node and the first edge relationship, wherein the first node includes a user information node, a spending information node, a credit information node, an address information node, a contact book information node, a call record information node, a device information node, a mobile phone number information node and a GPS information node, and the first edge relationship is an association relationship between any two of the first nodes.

8. The knowledge graph construction method according to claim 7, characterized in that: The graph construction module further includes: constructing a second node and a second edge relationship according to the real-time transaction data through the real-time generation logic of the data flow engine, and writing them into the corresponding first node and first edge relationship in the distributed graph database in real time, The second node includes real-time user information node, expenditure information node, credit information node, address information node, address book information node, call record information node, device information node, mobile phone number information node and GPS information node, and the second edge relationship is an association relationship between any two of the second nodes.

9. The method for constructing a knowledge graph according to claim 7, characterized in that: The point attributes include node constraints and / or key attributes thereof, and the edge attributes include the start point and end point of the edge and key attributes thereof.

10. The knowledge graph construction method according to claim 7, characterized in that: Also includes: The graph analysis module intercepts all related nodes centered on the user information node and generates a subgraph, and analyzes the user's relationship network based on the subgraph to screen out bad users.

Citation Information

Patent Citations

  • System and method for realizing intelligent early warning of fraudulent transactions

    CN110148001A

  • Method and system for identifying fraud gang in real time based on search engine

    CN110956547A

  • Knowledge graph display method and device

    CN114020852A

  • Construction method for multi-source heterogeneous data knowledge graph of power distribution network

    CN117273133A

  • Abnormal user identification system and method, electronic equipment and storage medium

    CN117670500A