ETC (Electronic Toll Collection) declaration risk management method, device, equipment, medium and program product

By constructing a knowledge graph and using abnormal image identification technology, the accuracy and efficiency of assessing the authenticity of user information during the ETC application process have been solved, enabling efficient risk management of ETC application orders and preventing identity theft by criminals.

CN119027125APending Publication Date: 2024-11-26CHINA CONSTRUCTION BANK +1
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202410872907.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-01
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

Existing technologies make it difficult to accurately assess the authenticity of user application information during the ETC application process, leading to misjudgments and low assessment efficiency, and failing to effectively prevent criminals from stealing other people's identities to commit fraud.

Method used

By collecting user data related to ETC application, a knowledge graph is constructed after preprocessing. Combining knowledge graph reasoning and community discovery algorithms, abnormal nodes are screened out, and images in the user data are identified, marked with risk tags, and stored in the risk order database for review.

Benefits of technology

It enables efficient and accurate risk assessment of ETC application orders, identifies document forgery by criminals, improves the accuracy and efficiency of risk management, and reduces the risk of misjudgment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119027125A_ABST
    Figure CN119027125A_ABST
Patent Text Reader

Abstract

The invention relates to the field of artificial intelligence, and particularly discloses an ETC (Electronic Toll Collection) declaration risk management method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: collecting user data related to applying ETC, and preprocessing the user data to obtain preprocessed user data; carrying out interception processing on the ETC application order generated in real time according to a preset interception rule, or marking a risk label; risk calculation is carried out on the preprocessed user data, and whether a risk label is marked on the ETC application order is determined according to a calculation result; storing the orders marked with the risk labels into a risk order database, and auditing each order in the risk order database to obtain an auditing result; wherein the auditing result is used for indicating whether the ETC application order of the corresponding user is passed or not. Therefore, the risk of the order can be evaluated more comprehensively, risk misjudgment is avoided, and the risk management accuracy is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an ETC application risk management method, apparatus, computer equipment, computer-readable storage medium and computer program product. Background Art

[0002] With the development of artificial intelligence technology, various management technologies for information security have emerged. With the popularization of the Electronic Toll Collection (ETC) system, there has been a phenomenon of criminals using other people's identities and driving license information to apply for it. This behavior of using other people's identities to defraud subsidies has had a bad impact on both society and individuals.

[0003] At present, the evaluation of ETC application materials is generally to score users based on information such as user household registration, mobile phone online time, and mobile phone call frequency, and then judge the authenticity of the user's application materials based on the score, and evaluate the user's advance toll payment amount based on the formed user credit rating. However, the relationship between these auxiliary judgment bases and whether the user has the real intention to apply for ETC (whether he or she will use it himself) is weak, which can easily lead to misjudgment of the authenticity of the user's application information. In addition, all the above judgment bases are based on reasoning based on artificial rules, with low accuracy and low evaluation efficiency. Summary of the invention

[0004] Based on this, it is necessary to provide an ETC application risk management method, device, computer equipment, computer-readable storage medium and computer program product that can accurately assess the risk of user application information in response to the above-mentioned technical problems.

[0005] In the first aspect, this application provides an ETC application risk management method, including:

[0006] Collecting user data related to the ETC application, and preprocessing the user data to obtain preprocessed user data; wherein the preprocessed user data includes: structured data and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured;

[0007] Intercept and process the ETC application orders generated in real time according to the preset interception rules, or mark them with risk tags;

[0008] Performing risk calculation on the preprocessed user data, and determining whether to mark a risk label for the ETC application order based on the calculation result; wherein the risk calculation method includes: any one or a combination of two methods of knowledge graph reasoning and community discovery, and abnormal image identification; the reasoning and community discovery of the knowledge graph refer to: constructing a knowledge graph based on the preprocessed user data, and filtering out abnormal nodes based on the knowledge reasoning and community discovery algorithm of the knowledge graph, wherein the abnormal nodes indicate an ETC application order with risks; the abnormal image identification refers to identifying images in the user data to find abnormal images, wherein the abnormal images indicate an ETC application order with risks;

[0009] The orders marked with risk tags are stored in a risk order database, and each order in the risk order database is audited to obtain an audit result; wherein the audit result is used to indicate whether the ETC application order of the corresponding user is approved.

[0010] In one embodiment, the collecting of user data related to the ETC application and preprocessing of the user data to obtain the preprocessed user data includes:

[0011] Collect the application information submitted by the user when applying for ETC, and extract the user data from the application information; wherein the application information includes: acquiring side user ID, application mobile device information, user IP address, WiFi information, ID card information, ID card photo, contact number, driving license information, driving license picture, vehicle front photo, delivery address, debit bank card number, bank reserved mobile phone number, any one or more of the following;

[0012] The extracted user data is cleaned and structured to obtain preprocessed user data.

[0013] In one embodiment, the interception of the real-time generated ETC application order according to the preset interception rules, or marking of risk tags, includes:

[0014] According to the preset interception rules, the real-time generated ETC application orders are compared one by one to see whether they meet the interception conditions. If they meet the conditions, any of the following processing methods is performed according to the pre-configured requirements:

[0015] Method 1: Directly intercept the ETC application order and prompt the user terminal that the application cannot be made;

[0016] Method 2: Mark the ETC application order with a risk tag;

[0017] The preset interception rules include any one or a combination of the following rules:

[0018] Whether the user's real-name verification is passed;

[0019] Whether the number of times the user data has been used has reached the upper limit;

[0020] Whether the user's identity information is consistent with the identity information on the various certificates submitted;

[0021] Whether the user's identity information matches the preset risk list database;

[0022] Whether the license plate complies with the preset naming rules.

[0023] In one embodiment, the performing risk calculation on the pre-processed user data and determining whether to mark a risk tag for the ETC application order according to the calculation result includes:

[0024] dividing the preprocessed user data into structured data and semi-structured data;

[0025] Mapping the structured data directly to nodes of the knowledge graph, and using the association information in the structured data to construct edges of the knowledge graph;

[0026] The semi-structured data is parsed and extracted and then mapped to nodes of the knowledge graph, and the association information in the semi-structured data is used to construct edges of the knowledge graph;

[0027] Performing fusion processing on the knowledge in the constructed knowledge graph, and importing the fused knowledge graph into the database, wherein the fusion processing includes: any one or more of entity alignment, knowledge processing, and knowledge updating;

[0028] Calling the community discovery algorithm in the database to divide each node in the fused knowledge graph into different communities to obtain community aggregation conditions;

[0029] Mark the nodes in the community according to the community aggregation situation and obtain abnormal nodes;

[0030] A risk label is marked on the ETC application order associated with the abnormal node.

[0031] In one embodiment, the performing risk calculation on the pre-processed user data and determining whether to mark a risk tag for the ETC application order according to the calculation result includes:

[0032] Filtering pictures from the preprocessed user data;

[0033] Identify the image to determine whether there is any abnormality in the image.

[0034] If there are any abnormal situations, the ETC application order will be marked with a risk tag;

[0035] The abnormal situations include: overlap with pictures in the photo library, tampering with pictures, and any one or more of the following situations: the vehicle brand predicted based on the picture does not match the vehicle brand filled in by the user.

[0036] In one embodiment, the orders marked with risk tags are stored in a risk order database, and each order in the risk order database is audited to obtain audit results, including:

[0037] Manually review each order in the risk order database; wherein the manual review includes: return visit;

[0038] Delete orders that have passed manual review from the risky order database.

[0039] In the second aspect, the present application also provides an ETC application risk management device, including:

[0040] A data collection module is used to collect user data related to ETC application, and pre-process the user data to obtain pre-processed user data; wherein the pre-processed user data includes: structured data and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured;

[0041] The interception and marking module is used to intercept and process the ETC application orders generated in real time according to the preset interception rules, or mark the risk tags;

[0042] A risk calculation module, used for performing risk calculation on the pre-processed user data, and determining whether to mark a risk label for the ETC application order according to the calculation result; wherein the risk calculation method includes: any one or a combination of two methods of knowledge graph reasoning and community discovery, and abnormal image identification; the reasoning and community discovery of the knowledge graph refer to: constructing a knowledge graph according to the pre-processed user data, and filtering out abnormal nodes based on the knowledge reasoning and community discovery algorithm of the knowledge graph, wherein the abnormal nodes indicate an ETC application order with risks; the abnormal image identification refers to identifying images in the user data to find abnormal images, wherein the abnormal images indicate an ETC application order with risks;

[0043] The order review module is used to store orders marked with risk tags into a risk order database, and review each order in the risk order database to obtain a review result; wherein the review result is used to indicate whether the ETC application order of the corresponding user is approved.

[0044] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0045] Collecting user data related to the ETC application, and preprocessing the user data to obtain preprocessed user data; wherein the preprocessed user data includes: structured data and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured;

[0046] Intercept and process the ETC application orders generated in real time according to the preset interception rules, or mark them with risk tags;

[0047] Performing risk calculation on the preprocessed user data, and determining whether to mark a risk label for the ETC application order based on the calculation result; wherein the risk calculation method includes: any one or a combination of two methods of knowledge graph reasoning and community discovery, and abnormal image identification; the reasoning and community discovery of the knowledge graph refer to: constructing a knowledge graph based on the preprocessed user data, and filtering out abnormal nodes based on the knowledge reasoning and community discovery algorithm of the knowledge graph, wherein the abnormal nodes indicate an ETC application order with risks; the abnormal image identification refers to identifying images in the user data to find abnormal images, wherein the abnormal images indicate an ETC application order with risks;

[0048] The orders marked with risk tags are stored in a risk order database, and each order in the risk order database is audited to obtain an audit result; wherein the audit result is used to indicate whether the ETC application order of the corresponding user is approved.

[0049] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0050] Collecting user data related to the ETC application, and preprocessing the user data to obtain preprocessed user data; wherein the preprocessed user data includes: structured data and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured;

[0051] Intercept and process the ETC application orders generated in real time according to the preset interception rules, or mark them with risk tags;

[0052] Performing risk calculation on the preprocessed user data, and determining whether to mark a risk label for the ETC application order based on the calculation result; wherein the risk calculation method includes: any one or a combination of two methods of knowledge graph reasoning and community discovery, and abnormal image identification; the reasoning and community discovery of the knowledge graph refer to: constructing a knowledge graph based on the preprocessed user data, and filtering out abnormal nodes based on the knowledge reasoning and community discovery algorithm of the knowledge graph, wherein the abnormal nodes indicate an ETC application order with risks; the abnormal image identification refers to identifying images in the user data to find abnormal images, wherein the abnormal images indicate an ETC application order with risks;

[0053] The orders marked with risk tags are stored in a risk order database, and each order in the risk order database is audited to obtain an audit result; wherein the audit result is used to indicate whether the ETC application order of the corresponding user is approved.

[0054] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0055] Collecting user data related to the ETC application, and preprocessing the user data to obtain preprocessed user data; wherein the preprocessed user data includes: structured data and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured;

[0056] Intercept and process the ETC application orders generated in real time according to the preset interception rules, or mark them with risk tags;

[0057] Perform risk calculation on the pre-processed user data, and determine whether to mark a risk tag on the ETC application order based on the calculation result;

[0058] The orders marked with risk tags are stored in a risk order database, and each order in the risk order database is audited to obtain an audit result; wherein the audit result is used to indicate whether the ETC application order of the corresponding user is approved.

[0059] The above-mentioned ETC application risk management method, device, computer equipment, computer-readable storage medium and computer program product collect user data related to ETC application and pre-process the user data to obtain pre-processed user data; wherein, the pre-processed user data includes: structured data, and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured; thereby, data closely related to ETC application can be collected more comprehensively, and by pre-processing the data, it is convenient to perform subsequent risk calculation on the data and improve data processing efficiency. By intercepting and processing the real-time generated ETC application order according to the preset interception rules, or marking the risk label; thereby, in the process of real-time order generation, the orders that do not meet the requirements can be quickly screened out according to the preset interception rules for interception, thereby improving processing efficiency, reducing the processing volume of subsequent audits, and quickly marking the orders with risks and pushing them to the subsequent audit process. By performing risk calculation on the pre-processed user data and determining whether to mark the risk label for the ETC application order according to the calculation result; thereby, the risk of the order can be evaluated more comprehensively, with high recognition efficiency and high accuracy. The orders marked with risk labels are stored in the risk order database, and each order in the risk order database is reviewed to obtain the review result; wherein the review result is used to indicate whether the ETC application order of the corresponding user is approved. In this way, the orders marked with risk labels can be reviewed to avoid misjudgment of risks and improve the accuracy of risk management. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0061] Figure 1 An application environment diagram of an ETC application risk management method in an embodiment;

[0062] Figure 2 A schematic diagram of a process of an ETC application risk management method in an embodiment;

[0063] Figure 3 is a schematic flow chart of step 203 in one embodiment;

[0064] Figure 4 A schematic diagram of the structure of an ETC application risk management system in one embodiment;

[0065] Figure 5A schematic diagram of the principle of an ETC application risk management system in one embodiment;

[0066] Figure 6 It is a structural block diagram of an ETC application risk management device in one embodiment;

[0067] Figure 7 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail 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 used to limit the present application. It should be noted that in the embodiments of the present application, certain software, groups, models and other existing solutions in the industry may be mentioned, which should be considered as exemplary, and their purpose is only to illustrate the feasibility of the implementation of the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0069] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations. The acquisition, storage, use, and processing of data in the technical solution of this application are in compliance with relevant provisions of national laws and regulations.

[0070] The ETC application risk management method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The user submits an ETC application request to the server 104 through the terminal 102, and the server 104 collects user data related to the ETC application, and pre-processes the user data to obtain the pre-processed user data; wherein the pre-processed user data includes: structured data, and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured; according to the preset interception rules, the real-time generated ETC application order is intercepted and processed, or a risk tag is marked; the pre-processed user data is risk calculated, and it is determined whether to mark the ETC application order with a risk tag according to the calculation result; the order marked with the risk tag is stored in the risk order database, and each order in the risk order database is reviewed to obtain the review result; wherein the review result is used to indicate whether the ETC application order of the corresponding user is passed. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car devices, projection devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server 104 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0071] In an exemplary embodiment, Figure 2 As shown in the figure, a risk management method for ETC application is provided, which is applied to Figure 1 The server in the example is used to illustrate, including the following steps 201 to 206. Among them:

[0072] Step 201, collect user data related to the ETC application, and pre-process the user data to obtain the pre-processed user data.

[0073] In this embodiment, the user's application information can be collected and stored, and then the collected application information can be extracted to extract user data, and then the user data can be pre-processed. For example, for pictures uploaded by users, the picture information can be OCR recognized to obtain text information, and then the unstructured information can be converted into structured information as much as possible. If it cannot be converted into structured data in a standardized manner, it can be converted into semi-structured data. Figure 1 In the application scenario shown, ETC application users can fill in and upload application materials through various terminals (visit the application webpage). Among them, structured data refers to data that can be represented and stored in a database, showing a two-dimensional form, and can obtain corresponding information through inherent key values; semi-structured data refers to data that can obtain corresponding information through flexible key value adjustment, and the data format is not fixed. The information stored under the same key value may be numerical, textual, dictionary or list.

[0074] Exemplarily, the application information submitted by the user when applying for ETC can be collected, and user data can be extracted from the application information; wherein the application information includes: user ID on the acquiring side, application mobile device information, user IP address, WiFi information, ID card information, ID card photo, contact number, driving license information, driving license picture, vehicle front photo, delivery address, debit bank card number, bank reserved mobile phone number, any one or more of the above; then the extracted user data is cleaned and structured to obtain the pre-processed user data.

[0075] In this embodiment, the extraction of application information may include the following processing:

[0076] 1) Split the mobile device information into device brand, device model, device resolution, etc.

[0077] 2) For uploaded ID card images, in order to prevent criminals from using photoshopped images to modify the ID card number to apply for ETC, it is necessary to identify the ID card information through OCR (Optical Character Recognition) to obtain the user's ID card number province, gender, age, ID card address, etc.

[0078] 3) For the uploaded driving license image, in order to prevent criminals from using Photoshop to modify the driving license information, OCR recognition is required to obtain the driving license information, including: vehicle license plate, vehicle model, vehicle owner, etc.

[0079] 4) For the delivery address entered by the user, the user may not enter the address in a standardized manner. In this case, the user-entered address needs to be processed, and the province, city and other information corresponding to the address need to be parsed. The address information is then converted into longitude and latitude information through a third-party longitude and latitude query interface.

[0080] 5) For IP addresses, convert the IP addresses into longitude and latitude information through a third-party longitude and latitude query interface.

[0081] 6) For the license plate image, identify the province corresponding to the license plate.

[0082] Step 202, intercept and process the ETC application order generated in real time according to the preset interception rules, or mark it with a risk tag.

[0083] In this embodiment, the ETC application orders can be screened by using the preset interception rules, and those that meet the conditions recorded in the interception rules are directly intercepted. For example, for problems that are not very serious, risk tags can be marked first, and the orders can be pushed to the review stage.

[0084] Exemplarily, according to the preset interception rules, the real-time generated ETC application orders are compared one by one to see whether they meet the interception conditions. If they meet the conditions, any of the following processing methods is performed according to the pre-configured requirements:

[0085] Method 1: Directly intercept the ETC application order and prompt the user terminal that the application cannot be made;

[0086] Method 2: Mark the ETC application order with a risk tag.

[0087] Exemplarily, the preset interception rules include but are not limited to: whether the user real-name verification is passed; whether the user data used has reached the upper limit of the number of times; whether the user identity information is consistent with the identity information on the various submitted certificates; whether the user identity information matches the preset risk list library; whether the license plate complies with the preset naming rules.

[0088] In this embodiment, when the user's application information meets a certain interception rule, the user's application order is intercepted, and the user is prompted that the application cannot be made, or the user is pushed to the manual review system for secondary review.

[0089] For example, whether the user's real-name verification is passed means that when the user uses the acquiring software to handle it, the acquiring side returns the user's real-name information, such as uid (user account), etc. When the operator configures the real-name, only users who have undergone real-name verification on the acquiring software can apply.

[0090] For example, whether the number of times the user data used reaches the upper limit refers to: limiting the upper limit of the number of times the ID card, acquiring software account, contracted bank card and other information used by the user to apply for ETC. For example, if the upper limit of the number of times the same ID card can be used is set to 3, the same ID card can only be used to apply for ETC for 3 cars, and so on.

[0091] For example, whether the user's identity information is consistent with the identity information on the various certificates submitted refers to whether the acquiring account, ID card, driving license and bank card owner used by the restricted user are the same or multiple conditions are the same. You can select the certificate and the restriction conditions of the certificate. If the ID card and driving license are configured to be the same, the ID card owner and the driving license owner must be the same person, otherwise the application does not meet the application conditions and will not be processed.

[0092] Exemplarily, whether the user identity information matches the preset risk list library means: saving the historical application risk list information, and when the user information matches the risk list library, the user is immediately intercepted or pushed to the manual review system according to the configuration. Among them, the risk list information can be provided by the risk calculation system.

[0093] Exemplarily, whether the license plate complies with the preset naming rules refers to: determining whether the license plate complies with my country's license plate naming rules.

[0094] Step 203, perform risk calculation on the pre-processed user data, and determine whether to mark the ETC application order with a risk tag based on the calculation result.

[0095] In this embodiment, risk calculation refers to judging the authenticity of the collected user data, whether it has been used repeatedly (impersonated), tampered with, and other situations. For example, the risk of information can be analyzed through knowledge graph construction and knowledge reasoning, or abnormalities can be identified on pictures uploaded by users to judge the risk of applying for an order.

[0096] In an optional embodiment, if Figure 3 As shown, the above step 203 may include the following steps:

[0097] Step 2031, dividing the pre-processed user data into structured data and semi-structured data.

[0098] In this embodiment, the knowledge graph is constructed by converting the acquired user data into entity-relationship triples, combining entities with related entities through relationship connections, importing the entity and relationship formats into a graph database, converting entity tables into vertices, and converting connections between entity tables into edges.

[0099] For example, since the collected data are basically structured and semi-structured data, a top-down approach is chosen when constructing the knowledge graph.

[0100] Step 2032, directly map the structured data to the nodes of the knowledge graph, and use the associated information in the structured data to construct the edges of the knowledge graph.

[0101] In this embodiment, structured data is directly mapped to nodes of the knowledge graph. For example, license plate numbers, user names, IP (Internet Protocol, a protocol for interconnecting networks) information, etc. can be directly used as nodes in the graph. The associated information in the structured data can be used to construct edges in the graph. For example, the purchase relationship between vehicles and owners, the relationship between users and delivery addresses, etc., these associated information can be used to construct edges in the graph.

[0102] Step 2033, after parsing and extracting the semi-structured data, map it to the nodes of the knowledge graph, and use the associated information in the semi-structured data to construct the edges of the knowledge graph.

[0103] In this embodiment, for semi-structured data, the information in the semi-structured data may not be standardized and needs to be parsed and extracted before it can be mapped to the nodes of the graph. The associated information in the semi-structured data is used to construct the edges in the graph.

[0104] Step 2034, fuse the knowledge in the constructed knowledge graph, and import the fused knowledge graph into the database.

[0105] In this embodiment, the fusion processing includes: any one or more of entity alignment, knowledge processing, and knowledge updating. Since knowledge usually has duplication, conflict, uneven quality, etc., it is necessary to fuse the knowledge in the knowledge graph, mainly including entity alignment, knowledge processing, and knowledge updating. Among them, entity alignment is also called entity matching or entity resolution, which is mainly used to eliminate inconsistencies such as entity conflicts and unclear pointing in heterogeneous data. Knowledge processing means that through entity alignment, a series of basic fact expressions or preliminary ontology prototypes can be obtained. After knowledge processing, a large-scale knowledge system is formed from the level to manage knowledge in a unified manner. Knowledge updating means that the content of the knowledge graph also needs to keep pace with the times. Whether it is a general knowledge graph or an industry knowledge graph, it needs to be continuously iterated and updated to expand existing knowledge and add new knowledge.

[0106] Step 2035, calling the community discovery algorithm in the database to divide each node in the fused knowledge graph into different communities to obtain community aggregation status.

[0107] In this embodiment, Neo4j database can be selected to store knowledge graphs. Neo4j provides rich functions and tools to conveniently store and manage knowledge graph data and supports query, analysis and visualization operations on knowledge graphs. The knowledge extracted from structured data and semi-structured data is represented as nodes and edges of the graph database and imported into the graph database.

[0108] It should be noted that this embodiment does not limit the type of database, and those skilled in the art can select databases with different functions to implement community division as needed.

[0109] Specifically, after completing knowledge fusion, knowledge reasoning is performed on the data, and fraud groups are discovered through community discovery algorithms. In order to ensure the robustness of knowledge reasoning results, multiple algorithms can also be used for knowledge reasoning, and the calculation results are output, and the optimal solution is selected to assist decision-making. The algorithms used include: Louvain algorithm, label propagation algorithm (LPA), etc.

[0110] Exemplarily, the Louvain algorithm is a community discovery algorithm based on modularity. Its basic idea is that nodes in the network try to traverse the community labels of all neighbors and select the community label that maximizes the modularity increment. After maximizing the modularity, each community is regarded as a new node, and the process is repeated until the modularity no longer increases. Initially, each vertex is regarded as a community, and the number of communities is the same as the number of vertices. Each vertex is merged with its adjacent vertices in turn, and their maximum modularity gain is calculated to see if it is greater than 0. If it is greater than 0, the node is placed in the community where the adjacent node with the largest modularity increment is located. This embodiment directly uses the Louvain algorithm in the graph algorithm library of Neo4j to perform community discovery, and uses the Cypher language to call the Louvain algorithm to divide the nodes in the graph into different communities. Specifically, the following steps may be included:

[0111] Initialization: Treat each node as an independent community.

[0112] Iterative optimization: For each node, calculate the modularity gain of moving it to the adjacent community. Move the node to the adjacent community that maximizes the modularity gain until the modularity can no longer be increased.

[0113] Merge communities: Merge nodes with the same community ID into a super node to build a new graph. Iterate and optimize the new graph until the stopping condition is reached.

[0114] Stop condition: no nodes can be moved, or the modularity gain is less than a certain threshold.

[0115] Output: Output the final community division results.

[0116] The core of the Louvain algorithm is the modularity function, which is used to measure the quality of community division. The modularity function is defined as follows:

[0117]

[0118] Among them: Aij represents the weight of the edge between node i and node j, ki represents the degree of node i (i.e., the number of edges connected to node i), m is half of the sum of the total weights of all edges in the graph, δ(ci,cj) represents the indicator function, which is 1 when node i and node j belong to the same community, otherwise it is 0, and ci and cj represent the community identifiers of node i and node j.

[0119] The goal of the Louvain algorithm is to maximize the modularity function, that is, to find a community partitioning scheme that maximizes the modularity. Through iterative optimization, nodes are constantly moved and communities are merged, and finally a local optimal community partition is achieved.

[0120] Exemplarily, the label propagation algorithm is a local community discovery algorithm based on label propagation. The basic idea is that the label of a node depends on the label information of its neighboring nodes, the degree of influence is determined by the node similarity, and stability is achieved through propagation iterative updates. After initializing each node with a unique label, the algorithm repeatedly socializes a node's label into the label that appears most frequently among the node's neighboring nodes. The algorithm stops when the label of each node appears most frequently among its neighboring nodes. It should be noted that the algorithm is asynchronous because each node is updated without waiting for the rest of the nodes to be updated. The Neo4j graph database provides a plug-in called Graph DataScience Library (GDS), which contains a variety of graph algorithms, including the label propagation algorithm.

[0121] For example, the core idea of ​​the label propagation algorithm is to achieve community discovery through information transmission between nodes. The steps of the algorithm are as follows:

[0122] Initialization: Assign a unique label to each node.

[0123] Information passing: Iteratively pass the label information of a node to its neighboring nodes. Specifically, each node uses the most frequently appearing label in its neighboring nodes as its own label.

[0124] Convergence judgment: When the labels of all nodes no longer change, the algorithm converges.

[0125] Community division: Nodes are divided into different communities based on their labels.

[0126] In this embodiment, the convergence condition of the label propagation algorithm is usually that the algorithm converges when the labels of all nodes no longer change. The label propagation algorithm is essentially an iterative algorithm. It does not need to explicitly build a model structure of the graph, but instead transmits information based on the connection relationship between nodes. The calculation formula of the label propagation algorithm is relatively simple, mainly updating the label of the node based on the label that appears most frequently in the node's neighbor nodes. The specific formula is as follows:

[0127]

[0128] Among them, new_labl(v) represents the new label of node v; argmax l Represents the label l that maximizes the following expression; ∑ u∈N(v) It means to sum all neighbors u of node v; I(l,label(u)) represents the indicator function, which is 1 when l is equal to the label of u, otherwise it is 0.

[0129] For example, when the above two algorithms are used simultaneously to find communities, the accuracy and robustness of the community structure can be increased. The method for confirming the community list finally considers the results of the two algorithms and merges, screens and verifies them.

[0130] Exemplarily, merging results refers to merging the community discovery results of the two algorithms to obtain an overall community list. Screening communities refers to screening out communities containing very few nodes, which may be caused by noise or abnormal situations. Overlapping community processing means that if there is overlap in the communities discovered by the two algorithms, consider merging the overlapping parts into one community, or select the result of one of the algorithms as the final community according to certain rules. Verifying community quality refers to verifying the merged communities to ensure that they have a certain quality and significance, and evaluating the quality of the community through indicators such as the tightness of the connection within the community and the sparseness of the connection between the community and the outside.

[0131] Step 2036: Mark the nodes in the community according to the community aggregation situation to obtain abnormal nodes.

[0132] Step 2037, mark the ETC application order associated with the abnormal node with a risk tag.

[0133] In this embodiment, for the community situation discovered in the knowledge reasoning step, its related nodes will be marked with risk tags, and the relevant information involved will be synchronously pushed to the risk list library.

[0134] In another optional implementation, images are screened out from the pre-processed user data; the images are identified to determine whether there are any abnormalities in the images. If there are any abnormalities, a risk label is marked on the ETC application order; wherein the abnormalities include: overlap with images in the photo library, image tampering, and the vehicle brand predicted based on the image is inconsistent with the vehicle brand filled in by the user, any one or more of the following situations.

[0135] In this embodiment, criminals may forge ID card images by modifying the ID card number and name information, the license plate information and headshot information in the driving license, and use false information to apply for an ID card. In view of the fact that criminals may forge ID card images by modifying images, using a single image for multiple purposes, and uploading them to the system, this subsystem will identify the images uploaded by users, and needs to identify whether the images have been modified, whether the images are consistent with the uploaded information, and whether the images have been used multiple times.

[0136] For example, to determine whether there is overlap with the pictures in the photo library, the similarity between the pictures can be calculated. Each newly uploaded photo needs to be compared with the pictures in the original photo library to find out whether there is overlap and calculate the image similarity. The similarity calculation method combines deep learning features and local feature descriptors to calculate the image similarity by feature extraction and similarity comparison. The specific steps are as follows:

[0137] 1) Feature extraction: Use a deep learning model to extract the global features of the image, and use a local feature descriptor to extract the local features of the image.

[0138] 2) Feature combination: Combine global features and local features to obtain rich feature representation.

[0139] 3) Similarity calculation: Use similarity measurement method to calculate the similarity between images.

[0140] Specifically, the global features of the image are extracted using a deep learning model, which has a strong feature representation capability and can capture the overall information of the image. At the same time, the local features of the image are extracted using local feature descriptors, which can effectively extract the local information of the image. By combining global features with local features, a richer and more robust feature representation is obtained, thereby improving the accuracy and robustness of image similarity calculation. The specific formula for calculating the similarity sim(A,B) is as follows:

[0141]

[0142] Where A and B represent the feature vectors of the two images, Ai and Bi represent the i-th element in the feature vector, and n represents the dimension of the feature vector.

[0143] For example, the judgment is made by setting a preset similarity threshold. When the image similarity exceeds the threshold, the abnormal image is marked. Since the similarity between ID cards and driving licenses is high, the probability of overlap is very high. If the photos are found to be overlapped, the overlapping images are marked as "suspected risk images". If the front photo of the car is found to be overlapped, it is marked as a "risk image".

[0144] For example, to determine whether an image has been tampered with, a third-party image detection interface can be accessed to detect the image uploaded by the customer. The interface returns the result of whether the image has been tampered with and locates the tampered area of ​​the tampered image. The tampered image is marked as a "risk image".

[0145] For example, based on the vehicle brand predicted by the image, the vehicle brand information and license plate number information are extracted in this link, and this information is compared with the driving license information. For the extraction of license plate number information, the third-party OCR recognition interface is directly used to call the interface to obtain the license plate number information. For vehicle brand recognition, it is identified through the self-built vehicle brand library.

[0146] For example, in order to extract vehicle brand information, a convolutional neural network (CNN) in the deep learning method can be used to learn the features in the image and classify them. CNN is composed of convolutional layers, pooling layers, and fully connected layers. It achieves image classification by extracting features from the image and learning the relationship between the features. During the training process, CNN adjusts the network parameters through the back propagation algorithm so that the network can output the correct classification results. The specific implementation process includes the following steps:

[0147] 1) Data preparation: Prepare an image dataset containing vehicle brands. The images should include photos of the front of the vehicle and annotate the brand category corresponding to each image.

[0148] 2) Model construction: Construct a CNN model, including convolutional layers, pooling layers, and fully connected layers. The convolutional layers are used to extract image features, the pooling layers are used to reduce feature dimensions, and the fully connected layers are used for classification.

[0149] 3) Model training: Use the prepared data set to train the CNN model. During the training process, the model optimizes the network parameters through the back propagation algorithm so that the model can accurately identify the vehicle brand.

[0150] 4) Model evaluation: Use an independent test data set to evaluate the trained model and calculate the model's accuracy, precision, recall and other indicators on the test set.

[0151] 5) Model application: Use the trained model to predict and identify the vehicle in front of the photo. Input the vehicle photo into the model, and the model outputs the prediction result of the vehicle brand.

[0152] In this embodiment, when the license plate brand prediction result does not match the license plate information filled in by the user, it is marked as a "risky image". Through risk calculation, the relevant content marked as "risky" in the knowledge graph construction, knowledge reasoning, and image anomaly recognition is pushed to the manual review background for further judgment by humans.

[0153] It should be noted that in order to improve the accuracy of risk calculation, the methods in the above two optional embodiments can be combined, that is, knowledge graph reasoning and discovery, as well as abnormal image identification can be used to realize risk calculation of ETC application orders.

[0154] Step 204: Store the orders marked with risk tags into a risk order database, and review each order in the risk order database to obtain review results.

[0155] Exemplarily, each order in the risk order database may be manually reviewed; and then the orders that have passed the manual review may be deleted from the risk order database; wherein the manual review includes: a return visit.

[0156] In this embodiment, the operator manually reviews the newly entered orders in the risk order database. For example, when the operator reviews the risk order, the user can be revisited and verified according to the user's actual situation. For users who fail the revisit, their application is prohibited, and the user's relevant information is added to the risk list library. For users who have normal revisits, the manual review is passed, the risk tag marked on the order is deleted, and the order is deleted from the risk order database at the same time.

[0157] In the above-mentioned ETC application risk management method, by collecting user data related to the ETC application and preprocessing the user data, the preprocessed user data is obtained; wherein, the preprocessed user data includes: structured data, and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured; thereby, data closely related to the ETC application can be collected more comprehensively, and by preprocessing the data, it is convenient to perform subsequent risk calculations on the data, thereby improving data processing efficiency. By intercepting and processing the real-time generated ETC application orders according to the preset interception rules, or marking risk tags; thereby, in the process of real-time order generation, orders that do not meet the requirements can be quickly screened out according to the preset interception rules for interception, thereby improving processing efficiency, reducing the processing volume of subsequent audits, and quickly marking orders with risks, and pushing them to subsequent audit processes. By performing risk calculation on the preprocessed user data, and determining whether to mark the risk tag for the ETC application order according to the calculation results; thereby, the risk of the order can be evaluated more comprehensively, with high recognition efficiency and high accuracy. The orders marked with risk labels are stored in the risk order database, and each order in the risk order database is reviewed to obtain the review result; wherein the review result is used to indicate whether the ETC application order of the corresponding user is approved. In this way, the orders marked with risk labels can be reviewed to avoid misjudgment of risks and improve the accuracy of risk management.

[0158] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0159] In an exemplary embodiment, Figure 4 As shown, an ETC application risk management system is provided, which may include multiple subsystems: information collection system 401, prior risk limitation system 402, risk calculation system 403, and manual review system 404. Generally speaking, the main steps for users to apply for ETC are filling in information, reviewing the information, and passing the review. Among them:

[0160] The information collection system 401 is used to collect user data and establish structured and semi-structured databases.

[0161] The pre-risk limitation system 402 mainly acts on the step of users filling in information, and is used for real-time order interception or risk marking during the user application process. The information input by the user is compared with the preset rules through the preset interception rules. When the interception rules are met, the order will be intercepted.

[0162] The risk calculation system 403 is mainly used for risk calculation after users collect information, that is, the information review stage. If user aggregation is found, it is considered that a suspicious community has been identified and image tampering has been discovered.

[0163] The manual review system 404 is used in the data review stage to conduct manual review or customer service return visits for risky orders identified in the risk calculation system. If the order is not intercepted by the prior risk limit system during the data filling stage, and is not identified as abnormal by the risk calculation system, the order is a normal order, which will be directly reviewed and approved, and the user will complete the ETC application.

[0164] It should be noted that the data transmission between the above systems uses an encrypted transmission method, and the data is decrypted when needed.

[0165] For example, Figure 5 The cooperation principle between the information collection system 401, the prior risk limitation system 402, the risk calculation system 403, and the manual review system 404 is shown in the figure, which can be used to achieve Figure 2 , Figure 3 The individual steps in the method shown.

[0166] In this embodiment, machine learning technology is used to convert all the user's application information into structured and unstructured data, and a knowledge graph is constructed to realize intelligent risk identification. This intelligent method can greatly improve the recognition efficiency and accuracy, and is more efficient than traditional methods. By analyzing all the information uploaded by the user, including text information and picture information, the user's risk can be fully identified, not just limited to a single source of information, which improves the comprehensiveness and accuracy of identification. In addition, the method in this embodiment does not need to obtain user information for auxiliary identification from the outside, but analyzes the information uploaded by the user for autonomous identification. This autonomous analysis method reduces the dependence on external data, making the entire data processing process more independent and reliable. For the pictures uploaded by the user, not only the picture similarity calculation is performed, but also the picture is monitored whether it has been modified, and the front photo of the car is identified. Through this multi-dimensional analysis, the authenticity of the picture can be more comprehensively understood, and the lawbreakers can be effectively prevented from forging application materials by modifying the picture. The method in this embodiment also uses a community discovery algorithm to identify risky orders, which also means that potential risks can be identified by analyzing the relationship and behavior patterns between users, improving the ability to discover abnormal situations.

[0167] This embodiment effectively addresses the problem of criminals stealing other people's identities and driving license information to apply for ETC through intelligent identification, comprehensive identification, autonomous analysis, multi-dimensional analysis and community discovery, and provides a more reliable risk management method for society and individuals.

[0168] Based on the same inventive concept, the embodiment of the present application also provides an ETC application risk management device for implementing the ETC application risk management method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more ETC application risk management device embodiments provided below can refer to the limitations of the ETC application risk management method above, and will not be repeated here.

[0169] In an exemplary embodiment, Figure 6 As shown, an ETC application risk management device is provided, including: a data collection module 601, an interception and marking module 602, a risk calculation module 603 and an order review module 604, wherein:

[0170] The data collection module 601 is used to collect user data related to the ETC application, and pre-process the user data to obtain pre-processed user data; wherein the pre-processed user data includes: structured data and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured;

[0171] The interception and marking module 602 is used to intercept and process the ETC application order generated in real time according to the preset interception rules, or mark the risk label;

[0172] The risk calculation module 603 is used to perform risk calculation on the pre-processed user data and determine whether to mark a risk tag on the ETC application order according to the calculation result;

[0173] The order review module 604 is used to store orders marked with risk tags into a risk order database, and review each order in the risk order database to obtain a review result; wherein the review result is used to indicate whether the ETC application order of the corresponding user is approved.

[0174] Exemplarily, the data collection module 601 is specifically used to: collect application information submitted by users when applying for ETC, and extract user data from the application information; wherein the application information includes: user ID on the acquiring side, application mobile device information, user IP address, WiFi information, ID card information, ID card photo, contact number, driving license information, driving license picture, vehicle front photo, delivery address, debit bank card number, bank reserved mobile phone number, any one or more of the above; clean and structure the extracted user data to obtain pre-processed user data.

[0175] Exemplarily, the interception and marking module 602 is specifically used to: compare the real-time generated ETC application orders one by one to see whether they meet the interception conditions according to the preset interception rules, and if they meet the conditions, perform any of the following processing methods according to the pre-configured requirements:

[0176] Method 1: Directly intercept the ETC application order and prompt the user terminal that the application cannot be made;

[0177] Method 2: Mark the ETC application order with a risk tag;

[0178] The preset interception rules include any one or a combination of the following rules:

[0179] Whether the user's real-name verification is passed;

[0180] Whether the number of times the user data has been used has reached the upper limit;

[0181] Whether the user's identity information is consistent with the identity information on the various certificates submitted;

[0182] Whether the user's identity information matches the preset risk list database;

[0183] Whether the license plate complies with the preset naming rules.

[0184] Exemplarily, the risk calculation module 603 is specifically used to: divide the pre-processed user data into structured data and semi-structured data; directly map the structured data to the nodes of the knowledge graph, and use the associated information in the structured data to construct the edges of the knowledge graph; parse and extract the semi-structured data and map them to the nodes of the knowledge graph, and use the associated information in the semi-structured data to construct the edges of the knowledge graph; fuse the knowledge in the constructed knowledge graph, and import the fused knowledge graph into the database, wherein the fusion processing includes: any one or more of entity alignment, knowledge processing, and knowledge updating; call the community discovery algorithm in the database to divide each node in the fused knowledge graph into different communities to obtain community aggregation; mark the nodes in the community according to the community aggregation to obtain abnormal nodes; mark the ETC application orders associated with the abnormal nodes with risk labels.

[0185] Exemplarily, the risk calculation module 603 is also used to: filter out pictures from the pre-processed user data; identify the pictures to determine whether there are any abnormalities in the pictures, and if there are any abnormalities, mark the ETC application order with a risk label; wherein the abnormalities include: overlap with pictures in the photo library, picture tampering, and the vehicle brand predicted based on the picture is inconsistent with the vehicle brand filled in by the user, any one or more of the following situations.

[0186] Exemplarily, the order review module 604 is specifically used to: manually review each order in the risk order database; wherein the manual review includes: returning a visit; and deleting orders that have passed the manual review from the risk order database.

[0187] Each module in the above-mentioned ETC application risk management device can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0188] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 7As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface, the display unit and the input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a method for managing the risk of applying for ETC is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0189] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0190] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned various method embodiments when executing the computer program.

[0191] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0192] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0193] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0194] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0195] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A risk management method for ETC application. It is characterized in that The method comprises: Collecting user data related to the ETC application, and preprocessing the user data to obtain preprocessed user data; wherein the preprocessed user data includes: structured data and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured; Intercept and process the ETC application orders generated in real time according to the preset interception rules, or mark them with risk tags; Performing risk calculation on the preprocessed user data, and determining whether to mark a risk label for the ETC application order based on the calculation result; wherein the risk calculation method includes: any one or a combination of two methods of knowledge graph reasoning and community discovery, and abnormal image identification; the reasoning and community discovery of the knowledge graph refer to: constructing a knowledge graph based on the preprocessed user data, and filtering out abnormal nodes based on the knowledge reasoning and community discovery algorithm of the knowledge graph, wherein the abnormal nodes indicate an ETC application order with risks; the abnormal image identification refers to identifying images in the user data to find abnormal images, wherein the abnormal images indicate an ETC application order with risks; The orders marked with risk tags are stored in a risk order database, and each order in the risk order database is audited to obtain an audit result; wherein the audit result is used to indicate whether the ETC application order of the corresponding user is approved.

2. The method according to claim 1, It is characterized in that The collecting of user data related to the ETC application and preprocessing of the user data to obtain the preprocessed user data includes: Collect the application information submitted by the user when applying for ETC, and extract the user data from the application information; wherein the application information includes: acquiring side user ID, application mobile device information, user IP address, WiFi information, ID card information, ID card photo, contact number, driving license information, driving license picture, vehicle front photo, delivery address, debit bank card number, bank reserved mobile phone number, any one or more of the following; The extracted user data is cleaned and structured to obtain preprocessed user data.

3. The method according to claim 1, It is characterized in that The method of intercepting and processing the ETC application order generated in real time according to the preset interception rules, or marking the risk label, includes: According to the preset interception rules, the real-time generated ETC application orders are compared one by one to see whether they meet the interception conditions. If they meet the conditions, any of the following processing methods is performed according to the pre-configured requirements: Method 1: Directly intercept the ETC application order and prompt the user terminal that the application cannot be made; Method 2: Mark the ETC application order with a risk tag; The preset interception rules include any one or a combination of the following rules: Whether the user's real-name verification is passed; Whether the number of times the user data has been used has reached the upper limit; Whether the user's identity information is consistent with the identity information on the various certificates submitted; Whether the user's identity information matches the preset risk list database; Whether the license plate complies with the preset naming rules.

4. The method according to claim 1, It is characterized in that The performing risk calculation on the pre-processed user data and determining whether to mark a risk tag for the ETC application order according to the calculation result includes: dividing the preprocessed user data into structured data and semi-structured data; Mapping the structured data directly to nodes of the knowledge graph, and using the association information in the structured data to construct edges of the knowledge graph; The semi-structured data is parsed and extracted and then mapped to nodes of the knowledge graph, and the association information in the semi-structured data is used to construct edges of the knowledge graph; Performing fusion processing on the knowledge in the constructed knowledge graph, and importing the fused knowledge graph into the database, wherein the fusion processing includes: any one or more of entity alignment, knowledge processing, and knowledge updating; Calling the community discovery algorithm in the database to divide each node in the fused knowledge graph into different communities to obtain community aggregation conditions; Mark the nodes in the community according to the community aggregation situation and obtain abnormal nodes; A risk label is marked on the ETC application order associated with the abnormal node.

5. The method according to claim 1, It is characterized in that The performing risk calculation on the pre-processed user data and determining whether to mark a risk tag for the ETC application order according to the calculation result includes: Filtering pictures from the preprocessed user data; Identify the image to determine whether there is any abnormality in the image. If there are any abnormal situations, the ETC application order will be marked with a risk tag; The abnormal situations include: overlap with pictures in the photo library, tampering with pictures, and any one or more of the following situations: the vehicle brand predicted based on the picture does not match the vehicle brand filled in by the user.

6. The method according to any one of claims 1 to 5, It is characterized in that The orders marked with risk tags are stored in a risk order database, and each order in the risk order database is reviewed to obtain review results, including: Manually review each order in the risk order database; wherein the manual review includes: return visit; Delete orders that have passed manual review from the risky order database.

7. An ETC application risk management device, It is characterized in that The device comprises: A data collection module is used to collect user data related to ETC application, and pre-process the user data to obtain pre-processed user data; wherein the pre-processed user data includes: structured data and / or semi-structured data; the semi-structured data refers to a data form between structured and unstructured; The interception and marking module is used to intercept and process the ETC application orders generated in real time according to the preset interception rules, or mark the risk tags; A risk calculation module, used for performing risk calculation on the pre-processed user data, and determining whether to mark a risk label for the ETC application order according to the calculation result; wherein the risk calculation method includes: any one or a combination of two methods of knowledge graph reasoning and community discovery, and abnormal image identification; the reasoning and community discovery of the knowledge graph refer to: constructing a knowledge graph according to the pre-processed user data, and filtering out abnormal nodes based on the knowledge reasoning and community discovery algorithm of the knowledge graph, wherein the abnormal nodes indicate an ETC application order with risks; the abnormal image identification refers to identifying images in the user data to find abnormal images, wherein the abnormal images indicate an ETC application order with risks; The order review module is used to store orders marked with risk tags into a risk order database, and review each order in the risk order database to obtain a review result; wherein the review result is used to indicate whether the ETC application order of the corresponding user is approved.

8. The device according to claim 7, It is characterized in that The data collection module is specifically used for: Collect the application information submitted by the user when applying for ETC, and extract the user data from the application information; wherein the application information includes: acquiring side user ID, application mobile device information, user IP address, WiFi information, ID card information, ID card photo, contact number, driving license information, driving license picture, vehicle front photo, delivery address, debit bank card number, bank reserved mobile phone number, any one or more of the following; The extracted user data is cleaned and structured to obtain preprocessed user data.

9. The device according to claim 7, It is characterized in that The interception and marking module is specifically used to: compare the real-time generated ETC application orders one by one to see whether they meet the interception conditions according to the preset interception rules, and if they do, perform any of the following processing methods according to the pre-configured requirements: Method 1: Directly intercept the ETC application order and prompt the user terminal that the application cannot be made; Method 2: Mark the ETC application order with a risk tag; The preset interception rules include any one or a combination of the following rules: Whether the user's real-name verification is passed; Whether the number of times the user data has been used has reached the upper limit; Whether the user's identity information is consistent with the identity information on the various certificates submitted; Whether the user's identity information matches the preset risk list database; Whether the license plate complies with the preset naming rules.

10. The device according to claim 7, It is characterized in that The risk calculation module is specifically used for: dividing the preprocessed user data into structured data and semi-structured data; Mapping the structured data directly to nodes of the knowledge graph, and using the association information in the structured data to construct edges of the knowledge graph; The semi-structured data is parsed and extracted and then mapped to nodes of the knowledge graph, and the association information in the semi-structured data is used to construct edges of the knowledge graph; Performing fusion processing on the knowledge in the constructed knowledge graph, and importing the fused knowledge graph into the database, wherein the fusion processing includes: any one or more of entity alignment, knowledge processing, and knowledge updating; Calling the community discovery algorithm in the database to divide each node in the fused knowledge graph into different communities to obtain community aggregation conditions; Mark the nodes in the community according to the community aggregation situation and obtain abnormal nodes; A risk label is marked on the ETC application order associated with the abnormal node.

11. The device according to claim 7, It is characterized in that The risk calculation module is further used for: Filtering pictures from the preprocessed user data; Identify the image to determine whether there is any abnormality in the image. If there are any abnormal situations, the ETC application order will be marked with a risk tag; The abnormal situations include: overlap with pictures in the photo library, tampering with pictures, and any one or more of the following situations: the vehicle brand predicted based on the picture does not match the vehicle brand filled in by the user.

12. The device according to any one of claims 1 to 5, It is characterized in that The order review module is specifically used for: Manually review each order in the risk order database; wherein the manual review includes: return visit; Delete orders that have passed manual review from the risky order database.

13. A computer device comprising a memory and a processor, wherein the memory stores a computer program. It is characterized in that When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

14. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

15. A computer program product comprising a computer program, It is characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Citation Information

Cited By

  • Truck ETC (Electronic Toll Collection) application system and method based on multi-modal data fusion dynamic detection

    CN121526280A

  • A truck ETC application system and method based on multi-modal data fusion dynamic detection

    CN121526280B