Object risk evaluation method and training method and device of object risk evaluation model

Abstract

The invention provides an object risk evaluation method and a training method and device of an object risk evaluation model, which can be applied to the fields of Internet technology, artificial intelligence technology, big data technology, information security technology and finance. The object risk evaluation method comprises the following steps: in response to a received object risk evaluation request, obtaining static object risk information and dynamic object risk information corresponding to a to-be-evaluated object identifier according to the to-be-evaluated object identifier in the object risk evaluation request; in response to matching of the dynamic object risk information and a preset rule, determining a preset weight corresponding to the preset rule; respectively processing the static object risk information and the dynamic object risk information to obtain object risk characteristics; determining an object risk evaluation value according to a preset weight and the object risk feature; and determining an object risk evaluation result corresponding to the to-be-evaluated object identifier according to the object risk evaluation value.

Description

Technical Field

[0001] The present disclosure relates to Internet technology, artificial intelligence technology, big data technology, information security technology and financial fields, and more specifically, to an object risk assessment method, a training method and device for an object risk assessment model, an electronic device, a computer-readable storage medium and a computer program product. Background Art

[0002] With the development of Internet technology, financial institutions can use Internet technology to establish financial systems and use financial systems to open a variety of financial businesses.

[0003] Financial services may include corporate online banking, corporate mobile phones, merchant services, etc. As more and more information is generated when users access different financial services, how to effectively process information and manage user risks is an urgent problem to be solved.

[0004] In the process of realizing the concept of the present disclosure, the inventors found that there are at least the following problems in the related art: since the accuracy of risk management for users is low, the information security of users and financial institutions cannot be guaranteed. Summary of the invention

[0005] In view of this, the present disclosure provides an object risk assessment method, a training method and device for an object risk assessment model, an electronic device, a computer-readable storage medium, and a computer program product.

[0006] According to one aspect of the present disclosure, there is provided an object risk assessment method, comprising: in response to receiving an object risk assessment request, obtaining static object risk information and dynamic object risk information corresponding to the object identifier to be evaluated according to the object identifier to be evaluated in the object risk assessment request; in response to the dynamic object risk information matching a preset rule, determining a preset weight corresponding to the preset rule; processing the static object risk information and the dynamic object risk information respectively to obtain an object risk feature; determining an object risk assessment value according to the preset weight and the object risk feature; and determining an object risk assessment result corresponding to the object identifier to be evaluated according to the object risk assessment value.

[0007] According to an embodiment of the present disclosure, the static object risk information includes the object information to be evaluated and the associated object information corresponding to the object identifier to be evaluated, and the dynamic object risk information includes the transaction behavior information corresponding to the object identifier to be evaluated.

[0008] According to an embodiment of the present disclosure, the above-mentioned static object risk information and the above-mentioned dynamic object risk information are processed separately to obtain object risk characteristics, including: performing feature extraction processing on the above-mentioned object information to be evaluated to obtain the characteristics of the object to be evaluated; performing feature extraction processing on the above-mentioned associated object information to obtain associated object characteristics; performing feature extraction processing on the above-mentioned transaction behavior information to obtain transaction behavior characteristics.

[0009] According to an embodiment of the present disclosure, the preset weights include at least a first weight corresponding to the characteristics of the object to be evaluated, a second weight corresponding to the characteristics of the associated object, and a third weight corresponding to the characteristics of the transaction behavior.

[0010] According to an embodiment of the present disclosure, determining the risk assessment value of the object according to the preset weights and the risk characteristics of the object includes: determining a first product value according to the first weight and the characteristics of the object to be evaluated; determining a second product value according to the second weight and the characteristics of the associated object; determining a third product value according to the third weight and the transaction behavior characteristics; determining the object risk assessment value according to the first product value, the second product value and the third product value.

[0011] According to an embodiment of the present disclosure, the above method also includes, before determining the preset weight corresponding to the preset rule in response to the above dynamic object risk information matching the preset rule: determining a first candidate rule and a first candidate weight based on a first number of preset transaction area information; determining a second candidate rule and a second candidate weight based on a second number of preset transaction mode information and preset transaction amount information corresponding to each of the above second number of preset transaction mode information; determining a third candidate rule and a third candidate weight based on a third number of preset transaction frequency information and preset transaction frequency information corresponding to each of the above third number of preset transaction frequency information; and determining a fourth candidate rule and a fourth candidate weight based on a fourth number of preset transaction-related object information.

[0012] According to an embodiment of the present disclosure, the above-mentioned dynamic object risk information includes at least one of the following: transaction area information, transaction mode information, transaction frequency information, transaction amount information and transaction-related object information.

[0013] According to an embodiment of the present disclosure, in response to the dynamic object risk information matching the preset rule, determining the preset weight corresponding to the preset rule includes: in a case where there is no preset transaction area information matching the above-mentioned transaction area information in the above-mentioned first number of preset transaction area information, matching the above-mentioned second number of preset transaction pattern information according to the above-mentioned transaction pattern information; in a case where there is no transaction pattern information matching the above-mentioned transaction pattern information in the above-mentioned second number of preset transaction pattern information, matching the above-mentioned third number of preset transaction frequency information according to the above-mentioned transaction pattern information; in a case where there is no transaction frequency information matching the above-mentioned transaction pattern information in the above-mentioned third number of preset transaction frequency information, matching the above-mentioned fourth number of preset transaction-associated object information according to the above-mentioned transaction-associated object information; in a case where there is preset transaction-associated object information matching the above-mentioned transaction-associated object information in the above-mentioned fourth number of preset transaction-associated object information, determining that the above-mentioned dynamic object risk information matches the above-mentioned fourth candidate preset rule; and determining the above-mentioned fourth candidate weight corresponding to the above-mentioned fourth candidate preset rule as the above-mentioned preset weight.

[0014] According to an embodiment of the present disclosure, the above-mentioned object risk assessment result corresponding to the above-mentioned object identifier to be evaluated based on the above-mentioned object risk assessment value includes: determining the object risk level corresponding to the above-mentioned object identifier to be evaluated based on a predetermined object risk assessment threshold and the above-mentioned object risk assessment value; in response to the above-mentioned object risk level meeting a predetermined condition, obtaining the baseline risk information corresponding to the above-mentioned object identifier to be evaluated; and determining the above-mentioned object risk assessment result based on the above-mentioned baseline risk information and the above-mentioned object risk level.

[0015] According to an embodiment of the present disclosure, the above-mentioned predetermined object risk assessment threshold includes a first predetermined threshold and a second predetermined threshold.

[0016] According to an embodiment of the present disclosure, the above-mentioned determination of the object risk level corresponding to the above-mentioned object identifier to be evaluated based on the predetermined object risk assessment threshold and the above-mentioned object risk assessment value includes: when the above-mentioned object risk assessment value is greater than the above-mentioned first predetermined threshold, determining that the above-mentioned object risk level belongs to the first risk level; when the above-mentioned object risk assessment value is between the above-mentioned first predetermined threshold and the above-mentioned second predetermined threshold, determining that the above-mentioned object risk level belongs to the second risk level; when the above-mentioned object risk assessment value is less than the above-mentioned second predetermined threshold, determining that the above-mentioned object risk level belongs to the third risk level.

[0017] According to an embodiment of the present disclosure, the method further includes, before the above-mentioned response to receiving an object risk assessment request, obtaining the static object risk information and dynamic object risk information corresponding to the above-mentioned object to be evaluated identifier according to the above-mentioned object risk assessment request: in response to detecting an object registration operation initiated by M first candidate objects, obtaining the candidate static object risk information corresponding to each of the M first candidate object identifiers; associating the above-mentioned M first candidate object identifiers and the candidate static object risk information corresponding to each of the M first candidate object identifiers and storing them in a data source; in response to detecting an object transaction operation initiated by N second candidate objects, obtaining the candidate dynamic object risk information corresponding to each of the N second candidate object identifiers, wherein the above-mentioned N second candidate objects belong to the above-mentioned M first candidate objects; associating the above-mentioned N second candidate object identifiers and the candidate dynamic object risk information corresponding to each of the N second candidate object identifiers and storing them in the above-mentioned data source; M and N are both positive integers, and N is less than or equal to M.

[0018] According to another aspect of the present disclosure, a training method for an object risk assessment model is provided, including: obtaining sample static object risk information, sample dynamic object risk information and object risk label information corresponding to the sample object identifier to be evaluated; in response to the sample dynamic object risk information matching a predetermined rule, determining a preset weight corresponding to the predetermined rule; processing the sample static object risk information and the sample dynamic object risk information respectively to obtain sample object risk characteristics; determining a sample object risk assessment value based on the preset weight and the sample object risk characteristics; and training a deep learning model using the sample object risk assessment value and the object risk label information to obtain an object risk assessment model.

[0019] According to an embodiment of the present disclosure, the above-mentioned deep learning model includes a first feature extraction module and a second feature extraction module, the above-mentioned sample static object risk information includes sample object information to be evaluated and sample associated object information corresponding to the above-mentioned sample object identification, and the above-mentioned sample dynamic object risk information includes sample transaction behavior information corresponding to the above-mentioned sample object identification.

[0020] According to an embodiment of the present disclosure, the above-mentioned sample static object risk information and sample dynamic object risk information are processed separately to obtain sample object risk characteristics, including: inputting the above-mentioned sample object information to be evaluated and the above-mentioned sample associated object information into the above-mentioned first feature extraction module to obtain sample object characteristics to be evaluated and sample associated object characteristics; inputting the above-mentioned sample transaction behavior information into the above-mentioned second feature extraction module to obtain sample transaction behavior characteristics.

[0021] According to another aspect of the present disclosure, an object risk assessment device is provided, including: a first acquisition module, used to respond to receiving an object risk assessment request and, according to the object identifier to be evaluated in the above object risk assessment request, obtain static object risk information and dynamic object risk information corresponding to the above object identifier to be evaluated; a first determination module, used to respond to the above dynamic object risk information matching the preset rule, determine the preset weight corresponding to the preset rule; a first processing module, used to process the above static object risk information and the above dynamic object risk information respectively to obtain object risk characteristics; a second determination module, used to determine the object risk assessment value according to the above preset weight and the above object risk characteristics; a third determination module, used to determine the object risk assessment result corresponding to the above object identifier to be evaluated according to the above object risk assessment value.

[0022] According to another aspect of the present disclosure, a training device for an object risk assessment model is provided, including: a second acquisition module, used to acquire sample static object risk information, sample dynamic object risk information and object risk label information corresponding to the sample object identifier to be evaluated; a fourth determination module, used to determine a preset weight corresponding to the predetermined rule in response to the above-mentioned sample dynamic object risk information matching the predetermined rule; a second processing module, used to process the above-mentioned sample static object risk information and the sample dynamic object risk information respectively to obtain the sample object risk characteristics; a fifth determination module, used to determine the sample object risk assessment value according to the above-mentioned preset weight and the above-mentioned sample object risk characteristics; a training module, used to train a deep learning model using the above-mentioned sample object risk assessment value and the above-mentioned object risk label information to obtain an object risk assessment model.

[0023] According to another aspect of the present disclosure, an electronic device is provided, comprising: one or more processors; and a memory for storing one or more instructions, wherein when the one or more instructions are executed by the one or more processors, the one or more processors implement the method described in the present disclosure.

[0024] According to another aspect of the present disclosure, a computer-readable storage medium is provided, on which executable instructions are stored. When the executable instructions are executed by a processor, the processor implements the method described in the present disclosure.

[0025] According to another aspect of the present disclosure, a computer program product is provided. The computer program product includes computer executable instructions. When the computer executable instructions are executed, they are used to implement the method described in the present disclosure.

[0026] According to the embodiments of the present disclosure, since the static object risk information and the dynamic object risk information are automatically obtained according to the identification of the object to be evaluated in the object risk evaluation request, the efficiency of information acquisition is improved. On this basis, since the preset weight is automatically determined according to the predetermined rules when the dynamic object risk information matches the preset rules, the object risk characteristics are obtained by processing the static object risk information and the dynamic object risk information, which is conducive to the subsequent object risk evaluation and risk management of the object to be evaluated according to the preset weight and the object risk characteristics, thereby at least partially overcoming the technical problem of low accuracy in risk management for users in the related art, improving the efficiency and accuracy of determining the object risk evaluation results, and thus facilitating the protection of the information security of users and financial institutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The above and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0028] Figure 1 The system architecture to which the object risk assessment method and the training method of the object risk assessment model can be applied according to the embodiments of the present disclosure is schematically shown;

[0029] Figure 2 The flowchart of the object risk assessment method according to the embodiment of the present disclosure is schematically shown. ;

[0030] Figure 3 An exemplary schematic diagram schematically illustrates a process of determining a preset weight corresponding to a predetermined rule in response to dynamic object risk information matching a preset rule according to an embodiment of the present disclosure;

[0031] Figure 4 An exemplary schematic diagram schematically illustrates a process of determining an object risk evaluation value according to preset weights and object risk characteristics according to an embodiment of the present disclosure;

[0032] Figure 5 An example schematic diagram of a process of determining an object risk assessment result corresponding to an object identifier to be assessed according to an object risk assessment value according to an embodiment of the present disclosure is schematically shown;

[0033] Figure 6 A flowchart schematically illustrates a method for training an object risk assessment model according to an embodiment of the present disclosure;

[0034] Fig. 7A An example schematic diagram schematically illustrates a training process of an object risk assessment model according to an embodiment of the present disclosure;

[0035] Figure 7BAn example schematic diagram schematically shows a training process of an object risk assessment model according to another embodiment of the present disclosure;

[0036] Figure 8 A block diagram of an object risk assessment device according to an embodiment of the present disclosure is schematically shown;

[0037] Fig. 9 A block diagram schematically shows a training device for an object risk assessment model according to an embodiment of the present disclosure; and

[0038] Fig.10 A block diagram of an electronic device suitable for implementing an object risk assessment method and a training method for an object risk assessment model according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0039] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0040] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0041] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0042] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0043] In the technical solution disclosed in the present invention, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations, necessary confidentiality measures are taken, and do not violate public order and good morals.

[0044] In the technical solution of the present disclosure, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

[0045] In the related art, the following methods can be used to implement risk management for users: a risk management method based on before the event, a risk management method based on during the event, and a risk management method based on after the event.

[0046] The ex ante risk management approach refers to the selection and feasibility study of different solutions before the business begins. The in-process risk management approach refers to the real-time management of different solutions during the business execution process. The ex post risk management approach refers to the handling of situations after the business occurs.

[0047] However, since the accuracy of the above-mentioned risk management methods in user risk management is low, the information security of users and financial institutions cannot be guaranteed.

[0048] In order to at least partially solve the technical problems existing in the related art, the present disclosure provides an object risk assessment method, a training method and a device for an object risk assessment model, which can be applied to Internet technology, artificial intelligence technology, big data technology, information security technology and financial fields. The object risk assessment method includes: in response to receiving an object risk assessment request, according to the object identifier to be evaluated in the object risk assessment request, obtaining static object risk information and dynamic object risk information corresponding to the object identifier to be evaluated; in response to the dynamic object risk information matching the preset rule, determining the preset weight corresponding to the preset rule; processing the static object risk information and the dynamic object risk information respectively to obtain the object risk characteristics; determining the object risk assessment value according to the preset weight and the object risk characteristics; and determining the object risk assessment result corresponding to the object identifier to be evaluated according to the object risk assessment value.

[0049] It should be noted that the object risk assessment method, object risk assessment model training method and device provided in the embodiments of the present disclosure can be used in the fields of Internet technology, artificial intelligence technology, big data technology, information security technology and finance, for example, in the field of computer technology. The object risk assessment method, object risk assessment model training method and device provided in the embodiments of the present disclosure can also be used in any field other than the Internet technology, artificial intelligence technology, big data technology, information security technology and finance, for example, in the field of risk assessment technology. The application field of the object risk assessment method, object risk assessment model training method and device provided in the embodiments of the present disclosure is not limited.

[0050] Figure 1The system architecture of the object risk assessment method and the object risk assessment model training method according to the embodiment of the present disclosure is schematically shown. It should be noted that: Figure 1 What is shown is merely an example of a system architecture to which the embodiments of the present disclosure can be applied, in order to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0051] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104 and a server 105. The network 104 is used to provide a medium for a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0052] The user may use at least one of the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only as examples).

[0053] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0054] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0055] It should be noted that the object risk assessment method provided in the embodiment of the present disclosure can generally be executed by the first terminal device 101, the second terminal device 102 or the third terminal device 103, or can also be executed by other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103. Correspondingly, the object risk assessment device provided in the embodiment of the present disclosure can also be set in the first terminal device 101, the second terminal device 102 or the third terminal device 103, or can be set in other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103.

[0056] Alternatively, the object risk assessment method provided in the embodiment of the present disclosure may also be executed by the server 105. Accordingly, the object risk assessment device provided in the embodiment of the present disclosure may generally be arranged in the server 105. The object risk assessment method provided in the embodiment of the present disclosure may also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Accordingly, the object risk assessment device provided in the embodiment of the present disclosure may also be arranged in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105.

[0057] It should be noted that the training method of the object risk assessment model provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the training device of the object risk assessment model provided in the embodiment of the present disclosure can generally be set in the server 105. The training method of the object risk assessment model provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Accordingly, the training device of the object risk assessment model provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105.

[0058] Alternatively, the training method of the object risk assessment model provided in the embodiment of the present disclosure may also be executed by the first terminal device 101, the second terminal device 102 or the third terminal device 103, or may also be executed by other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103. Accordingly, the training device of the object risk assessment model provided in the embodiment of the present disclosure may also be set in the first terminal device 101, the second terminal device 102 or the third terminal device 103, or may be set in other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103.

[0059] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0060] It should be noted that the sequence numbers of the operations in the following method are only used as representations of the operations for the purpose of description, and should not be regarded as representing the execution order of the operations. Unless explicitly stated, the method does not need to be executed completely in the order shown.

[0061] Figure 2 The flowchart of the object risk assessment method according to the embodiment of the present disclosure is schematically shown.

[0062] like Figure 2 As shown, the object risk assessment method 200 includes operations S210 to S250.

[0063] In operation S210, in response to receiving an object risk assessment request, static object risk information and dynamic object risk information corresponding to the object identifier to be assessed are acquired according to the object identifier to be assessed in the object risk assessment request.

[0064] In operation S220, in response to the dynamic object risk information matching the preset rule, a preset weight corresponding to the preset rule is determined.

[0065] In operation S230, the static object risk information and the dynamic object risk information are processed respectively to obtain object risk features.

[0066] In operation S240, a risk assessment value of the object is determined according to a preset weight and the risk characteristics of the object.

[0067] In operation S250, an object risk assessment result corresponding to the identifier of the object to be assessed is determined according to the object risk assessment value.

[0068] According to an embodiment of the present disclosure, a script for generating an object risk assessment request may be pre-stored in a client. In response to detecting a business transaction operation initiated by an object, the client may obtain an object identifier to be evaluated corresponding to the object and run the script, so as to generate an object risk assessment request based on the object identifier to be evaluated. The object identifier to be evaluated may be used to characterize different objects, and the object may refer to a user to be transacted with. After obtaining the object risk assessment request, the client may send the object risk assessment request to the server, so that the server may perform a risk assessment on the object corresponding to the object identifier to be evaluated according to the object identifier to be evaluated indicated by the object risk assessment request, and obtain an object risk assessment result.

[0069] According to an embodiment of the present disclosure, before executing the object risk assessment method provided by the present disclosure, the object registration operation of any candidate object can be monitored in real time. In response to detecting the object registration operation initiated by the first candidate object, the server can obtain candidate static object risk information used to characterize the first candidate object. The candidate static object risk information may include information obtained through prior monitoring. After obtaining the candidate static object risk information, the first candidate object identifier and the candidate static object risk information corresponding to the first candidate object identifier can be associated and stored in a data source.

[0070] According to an embodiment of the present disclosure, object transaction operations of any candidate object can be monitored in real time. In response to detecting an object transaction operation initiated by a second candidate object, the server can obtain candidate dynamic risk object information used to characterize the second candidate object. The second candidate object may belong to the first candidate object. The candidate dynamic object information may include information obtained through in-process monitoring and post-event monitoring. After obtaining the candidate dynamic object information, the second candidate object identifier and the candidate dynamic object information corresponding to the second candidate object identifier may be associated and stored in a data source.

[0071] According to an embodiment of the present disclosure, after receiving an object risk assessment request, at least one candidate object identifier stored in a data source may be matched according to the identifier of the object to be assessed in the object risk assessment request, so as to determine a target object identifier that matches the identifier of the object to be assessed in the at least one candidate object identifier. On this basis, static object risk information and dynamic object risk information corresponding to the target object identifier may be obtained from the data source.

[0072] According to an embodiment of the present disclosure, after obtaining static object risk information and dynamic object risk information, the dynamic object risk information may be matched according to at least one pre-configured candidate rule, so as to determine whether there is a preset rule matching the dynamic object risk information in at least one candidate rule. For example, in the case where there is a preset rule matching the dynamic object risk information in at least one candidate rule, a preset weight corresponding to the preset rule may be determined according to the matched preset rule.

[0073] According to an embodiment of the present disclosure, the object risk feature may include a static object risk feature and a dynamic object risk feature. After obtaining the static object risk information and the dynamic object risk information, the static object risk information may be subjected to feature extraction processing to obtain the static object risk feature. The dynamic object risk information may be subjected to feature extraction processing to obtain the dynamic object risk feature. The preset weight may include a preset static weight corresponding to the static object risk feature and a preset dynamic weight corresponding to the dynamic object risk feature.

[0074] According to an embodiment of the present disclosure, after obtaining the preset weights and object risk characteristics, a static risk evaluation value corresponding to the static object risk information can be determined according to the static object risk characteristics and the preset static weights. A dynamic risk evaluation value corresponding to the dynamic object risk information can be determined according to the dynamic object risk characteristics and the preset dynamic weights.

[0075] According to an embodiment of the present disclosure, after obtaining a static risk assessment value and a dynamic risk assessment value, the static risk assessment value and the dynamic risk assessment value can be evaluated according to a predetermined object risk assessment threshold value, so as to determine the object risk level corresponding to the object to be evaluated. The object risk level can be set according to actual business needs and is not limited here. For example, the object risk level may include a first risk level for characterizing that the object to be evaluated is a high-risk object, a second risk level for characterizing that the object to be evaluated is a medium-risk object, and a third risk level for characterizing that the object to be evaluated is a low-risk object. After obtaining the object risk level corresponding to the identifier of the object to be evaluated, the object risk assessment result can be determined in combination with the external risk information and the object risk level.

[0076] According to the embodiments of the present disclosure, since the static object risk information and the dynamic object risk information are automatically obtained according to the identification of the object to be evaluated in the object risk evaluation request, the efficiency of information acquisition is improved. On this basis, since the preset weight is automatically determined according to the predetermined rules when the dynamic object risk information matches the preset rules, the object risk characteristics are obtained by processing the static object risk information and the dynamic object risk information, which is conducive to the subsequent object risk evaluation and risk management of the object to be evaluated according to the preset weight and the object risk characteristics, thereby at least partially overcoming the technical problem of low accuracy in risk management for users in the related art, improving the efficiency and accuracy of determining the object risk evaluation results, and thus facilitating the protection of the information security of users and financial institutions.

[0077] Reference below Figure 3 to Figure 5 , the object risk assessment method 200 according to an embodiment of the present invention is further described.

[0078] According to an embodiment of the present disclosure, the object risk assessment method 200 may further include the following operations.

[0079] In response to detecting object registration operations initiated by M first candidate objects, obtain candidate static object risk information corresponding to each of the M first candidate object identifiers. The M first candidate object identifiers and the candidate static object risk information corresponding to each of the M first candidate object identifiers are associated and stored in a data source. In response to detecting object transaction operations initiated by N second candidate objects, obtain candidate dynamic object risk information corresponding to each of the N second candidate object identifiers, wherein the N second candidate objects belong to the M first candidate objects. The N second candidate object identifiers and the candidate dynamic object risk information corresponding to each of the N second candidate object identifiers are associated and stored in a data source.

[0080] According to an embodiment of the present disclosure, M and N are both positive integers, and N is less than or equal to M.

[0081] According to an embodiment of the present disclosure, before executing the object risk assessment method provided by the present disclosure, the object registration operation can be monitored in real time. In response to monitoring the object registration operation initiated by the first candidate object, the candidate static object risk information corresponding to the first candidate object can be obtained. The candidate static object risk information can be used to characterize the basic risk factors of the first candidate object.

[0082] For example, when the first candidate initiates due diligence (know-your-customer, KYC), the static object risk information may include object subject transaction information and associated object information. Associated object information may include shareholders and signatory users of the enterprise object, such as the country, occupation, industry, product, channel and relationship type of the associated object. Associated object information can be used as a risk factor indicator for the first candidate to evaluate the risk level of the first candidate by setting an initial weight.

[0083] According to an embodiment of the present disclosure, before executing the object risk assessment method provided by the present disclosure, object transaction operations can be monitored in real time. In response to monitoring the object transaction operation initiated by the second candidate object, candidate dynamic object risk information corresponding to the second candidate object can be obtained. The candidate dynamic object risk information can be used to characterize the transaction risk factor of the second candidate object.

[0084] For example, when the second candidate object initiates a dynamic account transaction operation, the problem transaction model can be hit according to the screening rules, and on this basis, the initial business parameters and initial transaction indicator weights of the problem transaction model can be set. The candidate dynamic risk object information may include at least one of transaction area information, transaction mode information, transaction frequency information, transaction amount information, and transaction related object information.

[0085] According to an embodiment of the present disclosure, the object risk assessment method 200 may further include the following operations.

[0086] According to the first number of preset transaction area information, a first candidate rule and a first candidate weight are determined. According to the second number of preset transaction mode information and the preset transaction amount information corresponding to the second number of preset transaction mode information, a second candidate rule and a second candidate weight are determined. According to the third number of preset transaction frequency information and the preset transaction frequency information corresponding to the third number of preset transaction frequency information, a third candidate rule and a third candidate weight are determined. According to the fourth number of preset transaction-related object information, a fourth candidate rule and a fourth candidate weight are determined.

[0087] According to an embodiment of the present disclosure, the preset transaction area information may include preset source area identifiers and preset target area identifiers corresponding to the first number of object transaction operations. The preset source area identifier may be used to characterize the preset area from which the transaction is transferred out. The preset target area identifier may be used to characterize the preset area from which the transaction is transferred in. The first candidate rule and the first candidate weight may be determined based on the first number of preset transaction area information. The first candidate rule may refer to a transaction high-risk area model. The first candidate weight may include a source weight corresponding to the preset source area identifier and a target weight corresponding to the preset target area identifier.

[0088] According to an embodiment of the present disclosure, the preset transaction mode information may include a transaction channel identifier corresponding to each of the second number of object transaction operations. The transaction channel identifier may be used to characterize the channel in which the object transaction operation occurs. The preset transaction amount information may be used to characterize the preset transaction limit corresponding to the channel. The second candidate rule and the second candidate weight may be determined based on the second number of preset transaction mode information and the preset transaction amount information corresponding to each of the second number of preset transaction mode information. The second candidate rule may refer to a transaction splitting model. The second candidate weight may include a channel weight corresponding to the transaction channel identifier.

[0089] According to an embodiment of the present disclosure, the preset transaction frequency information may include preset transaction frequencies corresponding to the third number of object transaction operations. The third candidate rule and the third candidate weight may be determined according to the third number of preset transaction frequency information and the preset transaction frequency information corresponding to the third number of preset transaction frequency information. The third candidate rule may refer to a fast-in-fast-out model. The third candidate weight may include a frequency weight corresponding to the transaction frequency.

[0090] According to an embodiment of the present disclosure, the preset transaction-related object information may include preset closeness, preset transaction amount, preset number of transactions, preset transaction frequency, etc. of the transaction-related objects corresponding to each of the fourth number of object transaction operations. The fourth candidate rule and the fourth candidate weight may be determined based on the fourth number of preset transaction-related object information. The fourth candidate rule may refer to a transaction-related object closeness model. The fourth candidate weight may include transaction weights corresponding to each of the above information.

[0091] According to the embodiments of the present disclosure, taking the authorized account opening object as an example, the above fields corresponding to each of the 100 account opening objects in a certain region can be selected, and analysis and calculation can be performed to analyze the proportion of the above fields that match the candidate rules. For example, if the proportion of the above fields that match the candidate rules is 10%, then 10% can be used as a boundary, and it is considered that the situation where the hit ratio of the account opening object field in the region is greater than 10% is high risk, and the account opening object needs to be monitored in detail. Alternatively, it is considered that the situation where the hit ratio of the account opening object field in the region is less than or equal to 10% is low risk, and the fields that must be monitored and repeatedly judged to be risk-free can be manually processed, and other fields can be automatically released for processing. Model analysis can be used to mine various characteristics of various businesses of related objects. After continuous learning of the model, when the fields that must be landed are also risk-free, they can be automatically processed instead of manual processing. Thereby improving processing efficiency.

[0092] According to an embodiment of the present disclosure, the dynamic object risk information includes at least one of the following: transaction area information, transaction mode information, transaction frequency information, transaction amount information and transaction-related object information.

[0093] According to an embodiment of the present disclosure, operation S220 may include the following operations.

[0094] In the case where there is no preset transaction area information matching the transaction area information in the first number of preset transaction area information, the second number of preset transaction pattern information is matched according to the transaction pattern information. In the case where there is no transaction pattern information matching the transaction pattern information in the second number of preset transaction pattern information, the third number of preset transaction frequency information is matched according to the transaction pattern information. In the case where there is no transaction frequency information matching the transaction pattern information in the third number of preset transaction frequency information, the fourth number of preset transaction associated object information is matched according to the transaction associated object information. In the case where there is preset transaction associated object information matching the transaction associated object information in the fourth number of preset transaction associated object information, it is determined that the dynamic object risk information matches the fourth candidate preset rule. The fourth candidate weight corresponding to the fourth candidate preset rule is determined as the preset weight.

[0095] According to an embodiment of the present disclosure, the transaction area information may include a source area identifier and a target area identifier corresponding to the object transaction operation. The source area identifier may be used to characterize the area from which the transaction is transferred out. The target area identifier may be used to characterize the area from which the transaction is transferred in. In the case where there is preset transaction area information matching the transaction area information in the first number of preset transaction area information, it may be determined that the dynamic object risk information matches the first candidate preset rule. On this basis, the first candidate weight corresponding to the first candidate preset rule may be determined as the preset weight.

[0096] According to an embodiment of the present disclosure, the transaction mode information may include a transaction channel identifier corresponding to the object transaction operation. The transaction channel identifier may be used to characterize the channel in which the object transaction operation occurs. The transaction amount information may be used to characterize the transaction limit corresponding to the channel. In the case where there is transaction mode information matching the transaction mode information in the second number of preset transaction mode information, it may be determined that the dynamic object risk information matches the second candidate preset rule. On this basis, the second candidate weight corresponding to the second candidate preset rule may be determined as the preset weight.

[0097] According to an embodiment of the present disclosure, the transaction frequency information may include the transaction frequency corresponding to the object transaction operation. In the case where there is transaction frequency information matching the transaction frequency information in the third number of preset transaction frequency information, it can be determined that the dynamic object risk information matches the third candidate preset rule. On this basis, the third candidate weight corresponding to the third candidate preset rule can be determined as the preset weight.

[0098] According to an embodiment of the present disclosure, the transaction-related object information may include the closeness, transaction amount, number of transactions, and transaction frequency of the transaction-related objects corresponding to the object transaction operations. In the case where there is no preset transaction-related object information matching the transaction-related object information in the fourth number of preset transaction-related object information, it can be determined that the dynamic object risk information does not match any candidate rule. In this case, it can be determined that the object to be evaluated corresponding to the object identifier to be evaluated belongs to the object risk evaluation result of the security object.

[0099] Figure 3 An example schematic diagram of a process of determining a preset weight corresponding to a predetermined rule in response to dynamic object risk information matching a preset rule according to an embodiment of the present disclosure is schematically shown.

[0100] like Figure 3 As shown in 300 , the dynamic object risk information 301 may include transaction area information 301_1 , transaction mode information 301_2 , transaction frequency information 301_3 , and transaction associated object information 301_4 .

[0101] After the transaction area information 301_1 and the preset transaction area information 302 are obtained, operation S310 may be performed according to the transaction area information 301_1 and the preset transaction area information 302.

[0102] In operation S310, it is determined whether there is preset transaction area information matching the transaction area information 301_1 in the first number of preset transaction area information 302?

[0103] If yes, it can be determined that the dynamic object risk information 301 matches the first candidate preset rule, and on this basis, the first candidate weight 303 corresponding to the first candidate preset rule can be determined as the preset weight. If no, operation S320 can be performed according to the transaction mode information 301_2 and the preset transaction mode information 304.

[0104] In operation S320 , it is determined whether there is transaction mode information matching the transaction mode information 301_2 in the second number of preset transaction mode information 304 .

[0105] If yes, it can be determined that the dynamic object risk information 301 matches the second candidate preset rule, on which basis, the second candidate weight 305 corresponding to the second candidate preset rule can be determined as the preset weight. If no, operation S330 can be performed according to the transaction frequency information 301_3 and the preset transaction frequency information 306.

[0106] In operation S330 , it is determined whether there is transaction frequency information matching the transaction frequency information 301_3 in the third number of preset transaction frequency information 306 ?

[0107] If yes, it can be determined that the dynamic object risk information 301 matches the third candidate preset rule, on which basis, the third candidate weight 307 corresponding to the third candidate preset rule can be determined as the preset weight. If no, operation S340 can be performed according to the transaction-related object information 301_4 and the preset transaction-related object information 308.

[0108] In operation S340 , is it determined whether there is any preset transaction-associated object information matching the transaction-associated object information 301_4 in the fourth number of preset transaction-associated object information 308 ?

[0109] If yes, it can be determined that the dynamic object risk information 301 matches the fourth candidate preset rule, and on this basis, the fourth candidate weight 309 corresponding to the fourth candidate preset rule can be determined as the preset weight. If no, the process can be terminated.

[0110] According to an embodiment of the present disclosure, static object risk information includes object information to be evaluated and associated object information corresponding to the object identifier to be evaluated, and dynamic object risk information includes transaction behavior information corresponding to the object identifier to be evaluated.

[0111] According to an embodiment of the present disclosure, operation S230 may include the following operations.

[0112] Perform feature extraction on the information of the object to be evaluated to obtain the features of the object to be evaluated. Perform feature extraction on the information of the associated object to obtain the features of the associated object. Perform feature extraction on the information of the transaction behavior to obtain the features of the transaction behavior.

[0113] According to the embodiments of the present disclosure, the model for performing feature extraction processing can be set according to actual business needs and is not limited here. For example, the predetermined model may include a Wide & Deep Learning model, in which case the predetermined model may include a Wide sub-model and a Deep sub-model.

[0114] According to the embodiments of the present disclosure, the Wide sub-model is used to analyze the information of the object to be evaluated and the information of the associated object, which may include using the Wide sub-model to perform feature extraction processing on the information of the object to be evaluated to obtain the features of the object to be evaluated, and using the Wide sub-model to perform feature extraction processing on the information of the associated object to obtain the features of the associated object. On this basis, the business indicators such as the features of the object to be evaluated and the features of the associated object can be analyzed and calculated to obtain the initial object risk assessment plan. On this basis, the Deep sub-model transaction behavior information can be used to perform feature extraction processing to obtain transaction behavior features.

[0115] According to an embodiment of the present disclosure, in response to detecting an object registration operation initiated by object A, static object risk information corresponding to object A can be obtained. The static object risk information corresponding to object A can include object information to be evaluated corresponding to object A and associated object information corresponding to object A. For example, the associated object information can include authorized account opening object A1, corporate beneficiary object A2, corporate director A3, corporate executive A4 and loan guarantee object A5. It can be assumed that the closeness of transaction-related objects is authorized account opening object A1>corporate beneficiary object A2>corporate director A3>corporate executive A4>loan guarantee object A5. In this case, the hit rule judgment corresponding to object A can include 5 business indicators.

[0116] According to an embodiment of the present disclosure, the preset weights include at least a first weight corresponding to the characteristics of the object to be evaluated, a second weight corresponding to the characteristics of the associated object, and a third weight corresponding to the characteristics of the transaction behavior.

[0117] According to an embodiment of the present disclosure, operation S240 may include the following operations.

[0118] The first product value is determined based on the first weight and the characteristics of the object to be evaluated. The second product value is determined based on the second weight and the characteristics of the associated object. The third product value is determined based on the third weight and the transaction behavior characteristics. The object risk evaluation value is determined based on the first product value, the second product value and the third product value.

[0119] According to an embodiment of the present disclosure, after obtaining the preset weights and object risk characteristics output by the Wide sub-model and the Deep sub-model, the preset weights and object risk characteristics can be weighted and summed through a common model to obtain the final weights of each specified business rule. On this basis, the object risk assessment value can be generated according to the final weights of each specified business rule.

[0120] According to an embodiment of the present disclosure, the object feature to be evaluated may include a fifth number of object sub-features, and the first weight may include first sub-weights corresponding to the fifth number of object sub-features. For each object sub-feature in the fifth number of object sub-features, a first sub-weight corresponding to the object sub-feature may be determined, and a first sub-product value may be determined based on the first sub-weight and the object sub-feature. On this basis, the first sub-product values ​​corresponding to the fifth number of object sub-features may be accumulated to obtain the first product value.

[0121] According to an embodiment of the present disclosure, the associated object feature may include a sixth number of associated object sub-features, and the second weight may include second sub-weights corresponding to each of the sixth number of associated object sub-features. For each of the sixth number of associated object sub-features, a second sub-weight corresponding to the associated object sub-feature may be determined, and a second sub-product value may be determined based on the second sub-weight and the associated object sub-feature. On this basis, the second sub-product values ​​corresponding to each of the sixth number of associated object sub-features may be accumulated to obtain a second product value.

[0122] According to an embodiment of the present disclosure, the transaction behavior feature may include a seventh number of transaction behavior sub-features, and the third weight may include a third sub-weight corresponding to each of the seventh number of transaction behavior sub-features. For each of the seventh number of transaction behavior sub-features, a second sub-weight corresponding to the associated object sub-feature may be determined, and a second sub-product value may be determined based on the second sub-weight and the associated object sub-feature. On this basis, the second sub-product values ​​corresponding to each of the sixth number of associated object sub-features may be accumulated to obtain a second product value.

[0123] According to an embodiment of the present disclosure, after obtaining the first product value, the second product value, and the third product value, the object risk assessment value can be determined according to the first product value, the second product value, and the third product value. For example, the first product value, the second product value, and the third product value can be accumulated and the object risk assessment value can be stored. Alternatively, weights can be set for the first product value, the second product value, and the third product value, respectively, and on this basis, the object risk assessment value can be determined by multiplying each product value and the corresponding weight.

[0124] According to the embodiments of the present disclosure, the specific calculation method of the object risk assessment value can be set according to the actual business needs and is not limited here. For example, object risk assessment value = (authorized account opening object + enterprise beneficiary object + enterprise director + enterprise executive + loan guarantee object) * (first candidate rule or second candidate rule or third candidate rule or fourth candidate rule). Alternatively, object risk assessment value = (authorized account opening object + enterprise beneficiary object + enterprise director + enterprise executive + loan guarantee object) * (first candidate rule + second candidate rule + third candidate rule + fourth candidate rule).

[0125] Figure 4 The following is a schematic diagram showing an example of a process of determining a risk assessment value of an object according to preset weights and risk characteristics of the object according to an embodiment of the present disclosure.

[0126] like Figure 4As shown in 400 , the preset weights may include a first weight 401 corresponding to a feature 402 of the object to be evaluated, a second weight 403 corresponding to a feature 404 of the associated object, and a third weight 405 corresponding to a feature 406 of the transaction behavior.

[0127] A first product value 407 can be determined according to the first weight 401 and the feature 402 of the object to be evaluated. A second product value 408 can be determined according to the second weight 403 and the feature 404 of the associated object. A third product value 409 can be determined according to the third weight 405 and the feature 406 of the transaction behavior. On this basis, an object risk evaluation value 410 can be determined according to the first product value 407, the second product value 408 and the third product value 409.

[0128] According to an embodiment of the present disclosure, operation S250 may include the following operations.

[0129] According to the predetermined object risk assessment threshold and the object risk assessment value, the object risk level corresponding to the object identifier to be assessed is determined. In response to the object risk level meeting the predetermined condition, the baseline risk information corresponding to the object identifier to be assessed is obtained. According to the baseline risk information and the object risk level, the object risk assessment result is determined.

[0130] According to an embodiment of the present disclosure, after obtaining the object risk assessment value, the object risk level can be determined according to the object risk assessment value. After obtaining the object risk level, the risk level can be evaluated according to predetermined conditions. When the object risk level meets the predetermined conditions, it can be considered that it needs to be reported to the local supervision, and the benchmark risk information corresponding to the identification of the object to be evaluated can be obtained. The benchmark risk information can refer to the local supervision result record and the system initial identification record.

[0131] According to the embodiments of the present disclosure, the corresponding risk business parameters and weights can be optimized by means of machine learning, and additional regulatory elements can be introduced in response to changes in current regulatory requirements, so that a standard hit processing repository can be continuously learned and generated to reduce the number of system false hits.

[0132] According to an embodiment of the present disclosure, the predetermined object risk assessment threshold includes a first predetermined threshold and a second predetermined threshold.

[0133] According to an embodiment of the present disclosure, determining the object risk level corresponding to the to-be-evaluated object identifier according to a predetermined object risk assessment threshold and an object risk assessment value may include the following operations.

[0134] When the object risk assessment value is greater than the first predetermined threshold, the object risk level is determined to belong to the first risk level. When the object risk assessment value is between the first predetermined threshold and the second predetermined threshold, the object risk level is determined to belong to the second risk level. When the object risk assessment value is less than the second predetermined threshold, the object risk level is determined to belong to the third risk level.

[0135] According to an embodiment of the present disclosure, the object risk level is used to characterize whether the object to be evaluated corresponding to the object identifier to be evaluated has a risk and the degree of risk. The object risk level can construct different risk levels based on different object states. For example, the first risk level can be set to high risk, the second risk level can be set to medium risk, and the third risk level can be set to low risk.

[0136] According to an embodiment of the present disclosure, the first risk level can be used to characterize that the object to be evaluated corresponding to the object to be evaluated identifier is a high-risk object. The second risk level can be used to characterize that the object to be evaluated corresponding to the object to be evaluated identifier is a medium-risk object. The third risk level can be used to characterize that the object to be evaluated corresponding to the object to be evaluated identifier is a low-risk object.

[0137] For example, the first predetermined threshold can be set to 100 and the second predetermined threshold can be set to 50. When the object risk assessment value corresponding to the object identifier to be evaluated is 120, it can be determined that the object risk assessment value is greater than the first predetermined threshold, and then it can be determined that the object risk level corresponding to the object identifier to be evaluated belongs to the first risk level, that is, the object to be evaluated corresponding to the object identifier to be evaluated is a high-risk object.

[0138] Alternatively, when the object risk assessment value corresponding to the object identifier to be evaluated is 80, it can be determined that the object risk assessment value is between the first predetermined threshold and the second predetermined threshold, and then it can be determined that the object risk level corresponding to the object identifier to be evaluated belongs to the second risk level, that is, the object to be evaluated corresponding to the object identifier to be evaluated is a medium-risk object.

[0139] Alternatively, when the object risk assessment value corresponding to the object identifier to be evaluated is 30, it can be determined that the object risk assessment value is less than the second predetermined threshold, and then it can be determined that the object risk level corresponding to the object identifier to be evaluated belongs to the third risk level, that is, the object to be evaluated corresponding to the object identifier to be evaluated is a low-risk object.

[0140] According to the embodiments of the present disclosure, by combining pre-monitoring, in-process monitoring and post-monitoring, the object risk level of the object to be evaluated is obtained, and the rule parameters of the model are intelligently optimized to reduce the misidentification of suspicious objects, so as to help compliance personnel more effectively identify suspicious objects and report them to local supervision.

[0141] Figure 5 The exemplary schematic diagram schematically shows a process of determining an object risk assessment result corresponding to an object identifier to be assessed according to an object risk assessment value according to an embodiment of the present disclosure.

[0142] like Figure 5 As shown, in 500, after obtaining the object risk evaluation value 501, operation S310 may be performed. In operation S310, it is determined whether the object risk evaluation value 501 is greater than a first predetermined threshold. If so, it is determined that the object risk level belongs to the first risk level 502. If not, operation S320 may be performed.

[0143] In operation S320, is it determined whether the object risk evaluation value 501 is greater than a second predetermined threshold? If yes, it can be determined that the object risk level belongs to the second risk level 503. If no, it can be determined that the object risk level belongs to the third risk level 504. Based on this, operation S340 can be performed.

[0144] In operation S340, it is determined whether the object risk level meets the predetermined condition. If yes, the reference risk information 505 corresponding to the object identifier to be evaluated can be obtained, and the object risk evaluation result 506 can be determined according to the reference risk information 505 and the object risk level. If no, the process can be terminated.

[0145] The above are only exemplary embodiments, but are not limited thereto and may also include other object risk assessment methods 200 known in the art, as long as they can improve the efficiency and accuracy of determining the object risk assessment results and help protect the information security of users and financial institutions.

[0146] Figure 6 The flowchart of the training method of the object risk assessment model according to the embodiment of the present disclosure is schematically shown.

[0147] like Figure 6 As shown, the object risk assessment model training method 600 includes operations S610 to S650.

[0148] In operation S610, sample static object risk information, sample dynamic object risk information, and object risk label information corresponding to the sample to-be-evaluated object identifier are obtained.

[0149] In operation S620, in response to the sample dynamic object risk information matching the predetermined rule, a preset weight corresponding to the predetermined rule is determined.

[0150] In operation S630, the sample static object risk information and the sample dynamic object risk information are processed respectively to obtain the sample object risk characteristics.

[0151] In operation S640, a risk evaluation value of the sample object is determined according to a preset weight and a risk feature of the sample object.

[0152] In operation S650, a deep learning model is trained using the sample object risk assessment value and the object risk label information to obtain an object risk assessment model.

[0153] According to the embodiments of the present disclosure, for the description of the sample object identification to be evaluated, the sample static object risk information, the sample dynamic object risk information, the sample object risk characteristics and the sample object risk evaluation value, please refer to the relevant content above regarding the object identification to be evaluated, the static object risk information, the dynamic object risk information, the object risk characteristics and the object risk evaluation value, which will not be repeated here.

[0154] According to an embodiment of the present disclosure, object risk label information can be used to characterize the actual risk situation of the sample object to be evaluated corresponding to the sample object to be evaluated identifier. After obtaining the preset weights and sample object risk characteristics, the preset weights and sample object risk characteristics can be input into the deep learning model to obtain the sample object evaluation value. The sample object evaluation value and the object risk label information can be input into the loss function to obtain the loss function value. On this basis, the model parameters of the deep learning model can be adjusted according to the loss function value until the predetermined conditions are met. The deep learning model can be configured according to actual business needs as long as it can realize the object risk evaluation function, which is not limited here.

[0155] For example, the model parameters of the deep learning model can be adjusted according to the back propagation algorithm or the stochastic gradient descent algorithm until the predetermined conditions are met. The deep learning model obtained when the predetermined conditions are met is determined as the object risk assessment model. The predetermined conditions may include at least one of the convergence of the loss function value and the number of training rounds reaching the maximum number of training rounds.

[0156] According to the embodiments of the present disclosure, since the sample static object risk information, the sample dynamic object risk information and the object risk label information are automatically obtained according to the sample to-be-evaluated object identification, the efficiency of information acquisition is improved. On this basis, the deep learning model is trained by training the sample object risk evaluation value obtained by processing the preset weights and the sample object risk characteristics according to the object risk label information, so that the obtained object risk evaluation model can automatically realize the object risk evaluation of the object to be evaluated, thereby improving the efficiency and accuracy of the object risk evaluation, which is beneficial to protecting the information security of users and financial institutions.

[0157] Reference below Fig. 7A and Figure 7B , the training method 600 of the object risk assessment model according to an embodiment of the present invention is further described.

[0158] According to an embodiment of the present disclosure, the deep learning model includes a first feature extraction module and a second feature extraction module, the sample static object risk information includes sample object information to be evaluated and sample associated object information corresponding to the sample object identification to be evaluated, and the sample dynamic object risk information includes sample transaction behavior information corresponding to the sample object identification to be evaluated.

[0159] According to an embodiment of the present disclosure, operation S630 may include the following operations.

[0160] The sample object to be evaluated information and the sample associated object information are input into the first feature extraction module to obtain the sample object to be evaluated features and the sample associated object features. The sample transaction behavior information is input into the second feature extraction module to obtain the sample transaction behavior features.

[0161] According to an embodiment of the present disclosure, the first feature extraction module may include a Wide sub-model and a Deep sub-model. After obtaining the sample object information to be evaluated, the sample associated object information and the sample transaction behavior information, the sample object information to be evaluated may be input into the Wide sub-model, and the sample object features to be evaluated may be output. The sample associated object information may be input into the Wide sub-model, and the sample associated object features may be output. The sample transaction behavior information may be input into the Deep sub-model, and the sample transaction behavior features may be output.

[0162] According to the embodiments of the present disclosure, after obtaining the characteristics of the sample object to be evaluated, the characteristics of the sample associated object, and the characteristics of the sample transaction behavior, the comprehensive score of the customer can be calculated, and then the risk of the object can be further identified based on manual offline investigation and reported to the local supervision. On this basis, the list of suspicious customers can be confirmed based on the results of the supervision feedback, and the deep learning model can be trained based on this as a sample to facilitate the continuous optimization of the object risk parameters.

[0163] According to the embodiments of the present disclosure, the deep learning model can be continuously learned and corrected through pre-monitoring, in-process monitoring and post-monitoring. After locating suspicious data, the relevant database is registered and combined with manual offline identification to determine the object risk level of the object to be evaluated, and determine whether it is a suspicious transaction and whether it should be reported to local supervision. The result records reported to the local supervision are compared with the records of the initial identification of the system. If the result of the local supervision returns a risk-free result that can be released, the deep learning model can be continuously corrected according to the result to reduce the false hit rate.

[0164] Fig. 7A An example schematic diagram of a training process of an object risk assessment model according to an embodiment of the present disclosure is schematically shown.

[0165] like Fig. 7AAs shown, in 700A, sample static object risk information 701, sample dynamic object risk information 702 and object risk label information 703 corresponding to the sample to-be-evaluated object identifier are obtained.

[0166] The sample static object risk information 701 and the sample dynamic object risk information 702 can be input into the deep learning model 704, and the sample object risk evaluation value 705 can be output. Based on the first loss function 706, according to the sample object risk evaluation value 705 and the object risk label information 703, the first loss function value 707 is obtained.

[0167] After obtaining the first loss function value 707, the model parameters of the deep learning model 704 can be adjusted according to the first loss function value 707 until a predetermined end condition is met to obtain an object risk assessment model.

[0168] Figure 7B An example schematic diagram of a training process of an object risk assessment model according to another embodiment of the present disclosure is schematically shown.

[0169] like Figure 7B As shown, in 700B, sample static object risk information, sample dynamic object risk information and object risk label information 716 corresponding to the sample to-be-evaluated object identifier are obtained. The sample static object risk information may include sample to-be-evaluated object information 708 and sample associated object information 709. The sample dynamic object risk information may include sample transaction behavior information 710.

[0170] The deep learning model 711 may include a first feature extraction module 711_1, a second feature extraction module 711_2, and a fusion module 711_3. The sample object information 708 and the sample associated object information 709 may be input into the first feature extraction module 711_1, and the sample object features 712 corresponding to the sample object information 708 and the sample associated object features 713 corresponding to the sample associated object information 709 may be output. The sample transaction behavior information 710 may be input into the second feature extraction module 711_2, and the sample transaction behavior features 714 may be output.

[0171] On this basis, the sample object to be evaluated feature 712, the sample associated object feature 713 and the sample transaction behavior feature 714 can be input into the fusion module 711_3 to output the sample object risk evaluation value 715. Based on the second loss function 717, according to the sample object risk evaluation value 715 and the object risk label information 716, the second loss function value 718 is obtained.

[0172] After obtaining the second loss function value 718, the model parameters of the deep learning model 711 can be adjusted according to the second loss function value 718 until a predetermined end condition is met to obtain an object risk assessment model.

[0173] The above is only an exemplary embodiment, but is not limited thereto and may also include other object risk assessment model training methods 600 known in the art, as long as the efficiency and accuracy of object risk assessment can be improved and the information security of users and financial institutions can be protected.

[0174] Figure 8 A block diagram of an object risk assessment device according to an embodiment of the present disclosure is schematically shown.

[0175] like Figure 8 As shown, the object risk assessment device 800 may include a first acquisition module 810, a first determination module 820, a first processing module 830, a second determination module 840 and a third determination module 850.

[0176] The first acquisition module 810 is used to, in response to receiving an object risk assessment request, acquire static object risk information and dynamic object risk information corresponding to the object identifier to be assessed according to the object identifier to be assessed in the object risk assessment request.

[0177] The first determination module 820 is configured to determine a preset weight corresponding to a preset rule in response to the dynamic object risk information matching the preset rule.

[0178] The first processing module 830 is used to process the static object risk information and the dynamic object risk information respectively to obtain object risk features.

[0179] The second determination module 840 is used to determine the object risk assessment value according to the preset weight and the object risk characteristics.

[0180] The third determination module 850 is used to determine the object risk assessment result corresponding to the identifier of the object to be assessed according to the object risk assessment value.

[0181] According to an embodiment of the present disclosure, static object risk information includes object information to be evaluated and associated object information corresponding to the object identifier to be evaluated, and dynamic object risk information includes transaction behavior information corresponding to the object identifier to be evaluated.

[0182] According to an embodiment of the present disclosure, the first processing module 830 may include a first feature extraction processing unit, a second feature extraction processing unit, and a third feature extraction processing unit.

[0183] The first feature extraction processing unit is used to perform feature extraction processing on the information of the object to be evaluated to obtain features of the object to be evaluated.

[0184] The second feature extraction processing unit is used to perform feature extraction processing on the associated object information to obtain associated object features.

[0185] The third feature extraction processing unit is used to perform feature extraction processing on the transaction behavior information to obtain transaction behavior features.

[0186] According to an embodiment of the present disclosure, the preset weights include at least a first weight corresponding to the characteristics of the object to be evaluated, a second weight corresponding to the characteristics of the associated object, and a third weight corresponding to the characteristics of the transaction behavior.

[0187] According to an embodiment of the present disclosure, the second determining module 840 may include a first determining unit, a second determining unit, a third determining unit, and a fourth determining unit.

[0188] The first determining unit is used to determine a first product value according to the first weight and the characteristics of the object to be evaluated.

[0189] The second determining unit is used to determine a second product value according to the second weight and the associated object feature.

[0190] The third determining unit is used to determine a third product value according to the third weight and the transaction behavior characteristics.

[0191] The fourth determining unit is used to determine the object risk assessment value according to the first product value, the second product value and the third product value.

[0192] According to an embodiment of the present disclosure, the object risk assessment device 800 may further include a sixth determination module, a seventh determination module, an eighth determination module and a ninth determination module.

[0193] The sixth determination module is used to determine a first candidate rule and a first candidate weight according to information of a first number of preset transaction areas.

[0194] The seventh determination module is used to determine the second candidate rule and the second candidate weight according to the second number of preset transaction mode information and the preset transaction amount information respectively corresponding to the second number of preset transaction mode information.

[0195] The eighth determination module is used to determine a third candidate rule and a third candidate weight according to a third number of preset transaction frequency information and preset transaction frequency information corresponding to each of the third number of preset transaction frequency information.

[0196] The ninth determination module is used to determine a fourth candidate rule and a fourth candidate weight according to a fourth number of preset transaction-related object information.

[0197] According to an embodiment of the present disclosure, the dynamic object risk information includes at least one of the following: transaction area information, transaction mode information, transaction frequency information, transaction amount information and transaction-related object information.

[0198] According to an embodiment of the present disclosure, the first determination module 820 may include a first matching unit, a second matching unit, a third matching unit, a fourth matching unit, and a fifth determination unit.

[0199] The first matching unit is used to match the second number of preset transaction mode information according to the transaction mode information when there is no preset transaction area information matching the transaction area information in the first number of preset transaction area information.

[0200] The second matching unit is used to match the third number of preset transaction mode information according to the transaction mode information when there is no transaction mode information matching the transaction mode information in the second number of preset transaction mode information.

[0201] The third matching unit is configured to match a fourth number of preset transaction-associated object information according to the transaction-associated object information when there is no transaction frequency information matching the transaction frequency information in the third number of preset transaction frequency information.

[0202] The fourth matching unit is configured to determine that the dynamic object risk information matches a fourth candidate preset rule when there is preset transaction-associated object information matching the transaction-associated object information in a fourth number of preset transaction-associated object information.

[0203] The fifth determining unit is configured to determine the fourth candidate weight corresponding to the fourth candidate preset rule as the preset weight.

[0204] According to an embodiment of the present disclosure, the third determining module 850 may include a sixth determining unit, an acquiring unit, and a seventh determining unit.

[0205] The sixth determining unit is used to determine the object risk level corresponding to the identification of the object to be evaluated according to a predetermined object risk evaluation threshold and an object risk evaluation value.

[0206] The acquisition unit is used to acquire the benchmark risk information corresponding to the identifier of the object to be evaluated in response to the object risk level meeting the predetermined condition.

[0207] The seventh determining unit is used to determine the object risk assessment result according to the baseline risk information and the object risk level.

[0208] According to an embodiment of the present disclosure, the predetermined object risk assessment threshold includes a first predetermined threshold and a second predetermined threshold.

[0209] According to an embodiment of the present disclosure, the sixth determining unit may include a first determining subunit, a second determining subunit, and a third determining subunit.

[0210] The first determining subunit is used to determine that the risk level of the object belongs to the first risk level when the risk evaluation value of the object is greater than a first predetermined threshold.

[0211] The second determining subunit is configured to determine that the risk level of the object belongs to the second risk level when the risk evaluation value of the object is between the first predetermined threshold and the second predetermined threshold.

[0212] The third determining subunit is used to determine that the risk level of the object belongs to the third risk level when the risk evaluation value of the object is less than the second predetermined threshold.

[0213] According to an embodiment of the present disclosure, the object risk assessment device 800 may further include a third acquisition module, a first association storage module, a fourth acquisition module and a second association storage module.

[0214] The third acquisition module is configured to acquire, in response to detecting object registration operations initiated by the M first candidate objects, candidate static object risk information corresponding to each of the M first candidate object identifiers.

[0215] The first association storage module is used to associate and store the M first candidate object identifiers and the candidate static object risk information corresponding to each of the M first candidate object identifiers in a data source.

[0216] The fourth acquisition module is used to acquire the candidate dynamic object risk information corresponding to each of the N second candidate object identifiers in response to detecting the object transaction operation initiated by N second candidate objects, wherein the N second candidate objects belong to the M first candidate objects.

[0217] The second association storage module is used to associate and store the N second candidate object identifiers and the candidate dynamic object risk information corresponding to each of the N second candidate object identifiers in a data source.

[0218] According to an embodiment of the present disclosure, M and N are both positive integers, and N is less than or equal to M.

[0219] Fig. 9 A block diagram of a training device for an object risk assessment model according to an embodiment of the present disclosure is schematically shown.

[0220] like Fig. 9 As shown, the training device 900 for the object risk assessment model may include a second acquisition module 910 , a fourth determination module 920 , a second processing module 930 , a fifth determination module 940 and a training module 950 .

[0221] The second acquisition module 910 is used to acquire the sample static object risk information, the sample dynamic object risk information and the object risk label information corresponding to the sample to-be-evaluated object identifier.

[0222] The fourth determination module 920 is configured to determine a preset weight corresponding to a predetermined rule in response to the sample dynamic object risk information matching the predetermined rule.

[0223] The second processing module 930 is used to process the sample static object risk information and the sample dynamic object risk information respectively to obtain the sample object risk characteristics.

[0224] The fifth determination module 940 is used to determine the risk evaluation value of the sample object according to the preset weight and the risk characteristics of the sample object.

[0225] The training module 950 is used to train the deep learning model using the sample object risk assessment value and the object risk label information to obtain the object risk assessment model.

[0226] According to an embodiment of the present disclosure, the deep learning model includes a first feature extraction module and a second feature extraction module, the sample static object risk information includes the sample to-be-evaluated object information and the sample associated object information corresponding to the sample to-be-evaluated object identifier, and the sample dynamic object risk information includes the sample transaction behavior information corresponding to the sample to-be-evaluated object identifier.

[0227] According to an embodiment of the present disclosure, the second processing module 930 may include a first input unit and a second input unit.

[0228] The first input unit is used to input the sample to-be-evaluated object information and the sample associated object information into the first feature extraction module to obtain the sample to-be-evaluated object feature and the sample associated object feature.

[0229] The second input unit is used to input the sample transaction behavior information into the second feature extraction module to obtain the sample transaction behavior features.

[0230] According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits, or at least part of the functions of any one of them can be implemented in one module. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be split into multiple modules for implementation. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be at least partially implemented as hardware circuits, such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems on chips, systems on substrates, systems on packages, application specific integrated circuits (ASICs), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, according to the embodiments of the present invention, one or more of the modules, submodules, units, and subunits can be at least partially implemented as computer program modules, and when the computer program modules are run, the corresponding functions can be performed.

[0231] For example, any multiple of the first acquisition module 810, the first determination module 820, the first processing module 830, the second determination module 840, and the third determination module 850 can be combined in one module / unit / subunit for implementation, or any one of the modules / units / subunits can be split into multiple modules / units / subunits. Alternatively, at least part of the functions of one or more of these modules / units / subunits can be combined with at least part of the functions of other modules / units / subunits and implemented in one module / unit / subunit. According to an embodiment of the present disclosure, at least one of the first acquisition module 810, the first determination module 820, the first processing module 830, the second determination module 840, and the third determination module 850 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or by any one of the three implementation methods of software, hardware, and firmware, or by a suitable combination of any of them. Alternatively, at least one of the first acquisition module 810, the first determination module 820, the first processing module 830, the second determination module 840 and the third determination module 850 may be at least partially implemented as a computer program module, which may perform corresponding functions when executed.

[0232] It should be noted that the object risk assessment device part in the embodiment of the present disclosure corresponds to the object risk assessment method part in the embodiment of the present disclosure. The description of the object risk assessment device part specifically refers to the object risk assessment method part, which will not be repeated here.

[0233] For example, any of the second acquisition module 910, the fourth determination module 920, the second processing module 930, the fifth determination module 940, and the training module 950 can be combined in one module / unit / subunit for implementation, or any one of the modules / units / subunits can be split into multiple modules / units / subunits. Alternatively, at least part of the functions of one or more of these modules / units / subunits can be combined with at least part of the functions of other modules / units / subunits and implemented in one module / unit / subunit. According to an embodiment of the present disclosure, at least one of the second acquisition module 910, the fourth determination module 920, the second processing module 930, the fifth determination module 940, and the training module 950 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or by any one of the three implementation methods of software, hardware, and firmware, or by a suitable combination of any of them. Alternatively, at least one of the second acquisition module 910, the fourth determination module 920, the second processing module 930, the fifth determination module 940 and the training module 950 may be at least partially implemented as a computer program module, which may perform a corresponding function when executed.

[0234] It should be noted that the training device part of the object risk assessment model in the embodiment of the present disclosure corresponds to the training method part of the object risk assessment model in the embodiment of the present disclosure. The description of the training device part of the object risk assessment model specifically refers to the training method part of the object risk assessment model, which will not be repeated here.

[0235] Fig.10 A block diagram of an electronic device suitable for implementing an object risk assessment method and a training method for an object risk assessment model according to an embodiment of the present disclosure is schematically shown. Fig.10 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0236] like Fig.10As shown, the computer electronic device 1000 according to the embodiment of the present disclosure includes a processor 1001, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1002 or the program loaded from the storage part 1009 to the random access memory (RAM) 1003. The processor 1001 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 1001 may also include an onboard memory for caching purposes. The processor 1001 may include a single processing unit or multiple processing units for performing different actions of the method flow according to the embodiment of the present disclosure.

[0237] In RAM1003, various programs and data required for the operation of electronic device 1000 are stored. Processor 1001, ROM1002 and RAM1003 are connected to each other via bus 1004. Processor 1001 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM1002 and / or RAM1003. It should be noted that the program may also be stored in one or more memories other than ROM1002 and RAM1003. Processor 1001 may also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in the one or more memories.

[0238] According to an embodiment of the present disclosure, the electronic device 1000 may further include an input / output (I / O) interface 1005, which is also connected to the bus 1004. The electronic device 1000 may further include one or more of the following components connected to the input / output (I / O) interface 1005: an input portion 1006 including a keyboard, a mouse, etc.; an output portion 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 1008 including a hard disk, etc.; and a communication portion 1009 including a network interface card such as a LAN card, a modem, etc. The communication portion 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the input / output (I / O) interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed, so that a computer program read therefrom is installed into the storage portion 1008 as needed.

[0239] According to an embodiment of the present disclosure, the method flow according to an embodiment of the present disclosure can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1009, and / or installed from the removable medium 1011. When the computer program is executed by the processor 1001, the above-mentioned functions defined in the system of the embodiment of the present disclosure are executed. According to an embodiment of the present disclosure, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0240] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0241] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include, but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0242] For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 1002 and / or the RAM 1003 described above and / or one or more memories other than the ROM 1002 and the RAM 1003 .

[0243] An embodiment of the present disclosure also includes a computer program product, which includes a computer program, which contains program code for executing the method provided by the embodiment of the present disclosure. When the computer program product is run on an electronic device, the program code is used to enable the electronic device to implement the object risk assessment method and the object risk assessment model training method provided by the embodiment of the present disclosure.

[0244] When the computer program is executed by the processor 1001, the above functions defined in the system / device of the embodiment of the present disclosure are executed. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0245] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 1009, and / or installed from the removable medium 1011. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0246] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0247] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments and / or claims of the present disclosure can be combined and / or combined in a variety of ways, even if such a combination or combination is not explicitly recorded in the present disclosure. In particular, without departing from the spirit and teaching of the present disclosure, the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0248] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. The scope of the present disclosure is defined by the attached claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A method for object risk assessment, comprising: In response to receiving an object risk assessment request, acquiring static object risk information and dynamic object risk information corresponding to the object identifier to be assessed according to the object identifier to be assessed in the object risk assessment request; In response to the dynamic object risk information matching a preset rule, determining a preset weight corresponding to the preset rule; Processing the static object risk information and the dynamic object risk information respectively to obtain object risk features; Determining a risk assessment value of the object according to the preset weight and the risk characteristics of the object; as well as According to the object risk assessment value, an object risk assessment result corresponding to the identification of the object to be assessed is determined.

2. The method according to claim 1, wherein: The static object risk information includes the object information to be evaluated and the associated object information corresponding to the object identifier to be evaluated, and the dynamic object risk information includes the transaction behavior information corresponding to the object identifier to be evaluated; The processing of the static object risk information and the dynamic object risk information to obtain the object risk features includes: Performing feature extraction processing on the information of the object to be evaluated to obtain features of the object to be evaluated; Performing feature extraction processing on the associated object information to obtain associated object features; and Perform feature extraction processing on the transaction behavior information to obtain transaction behavior features.

3. The method according to claim 1, wherein: The preset weights at least include a first weight corresponding to the characteristics of the object to be evaluated, a second weight corresponding to the characteristics of the associated object, and a third weight corresponding to the characteristics of the transaction behavior; Determining the object risk assessment value according to the preset weight and the object risk feature includes: Determining a first product value according to the first weight and the feature of the object to be evaluated; Determining a second product value according to the second weight and the associated object feature; determining a third product value according to the third weight and the transaction behavior feature; and The object risk assessment value is determined according to the first product value, the second product value and the third product value.

4. The method according to any one of claims 1 to 3, further comprising, before determining a preset weight corresponding to a predetermined rule in response to the dynamic object risk information matching the predetermined rule: Determining a first candidate rule and a first candidate weight according to information of a first number of preset transaction areas; Determining a second candidate rule and a second candidate weight according to a second number of preset transaction mode information and preset transaction amount information corresponding to each of the second number of preset transaction mode information; Determining a third candidate rule and a third candidate weight according to a third number of preset transaction pattern information and preset transaction frequency information corresponding to each of the third number of preset transaction pattern information; as well as A fourth candidate rule and a fourth candidate weight are determined according to a fourth number of preset transaction-related object information.

5. The method according to claim 4, wherein: The dynamic object risk information includes at least one of the following: transaction area information, transaction mode information, transaction frequency information, transaction amount information and transaction-related object information; In response to the dynamic object risk information matching the preset rule, determining the preset weight corresponding to the preset rule includes: If there is no preset transaction area information matching the transaction area information in the first number of preset transaction area information, matching the second number of preset transaction mode information according to the transaction mode information; When there is no transaction mode information matching the transaction mode information in the second number of preset transaction mode information, matching the third number of preset transaction mode information according to the transaction mode information; If there is no transaction frequency information matching the transaction frequency information in the third number of preset transaction frequency information, matching the fourth number of preset transaction association object information according to the transaction association object information; In the case where there is preset transaction-associated object information matching the transaction-associated object information in the fourth number of preset transaction-associated object information, determining that the dynamic object risk information matches the fourth candidate preset rule; and The fourth candidate weight corresponding to the fourth candidate preset rule is determined as the preset weight.

6. The method according to any one of claims 1 to 3, wherein: Determining the object risk assessment result corresponding to the identification of the object to be assessed according to the object risk assessment value includes: Determining the object risk level corresponding to the object identifier to be evaluated according to a predetermined object risk evaluation threshold and the object risk evaluation value; In response to the object risk level meeting a predetermined condition, obtaining baseline risk information corresponding to the identifier of the object to be evaluated; and The object risk assessment result is determined according to the baseline risk information and the object risk level.

7. The method according to claim 6, wherein: The predetermined object risk assessment threshold includes a first predetermined threshold and a second predetermined threshold; The determining, according to a predetermined object risk assessment threshold and the object risk assessment value, the object risk level corresponding to the identification of the object to be assessed comprises: When the object risk evaluation value is greater than the first predetermined threshold, determining that the object risk level belongs to the first risk level; When the object risk evaluation value is between the first predetermined threshold and the second predetermined threshold, determining that the object risk level belongs to the second risk level; and When the object risk evaluation value is less than the second predetermined threshold, it is determined that the object risk level belongs to the third risk level.

8. The method according to any one of claims 1 to 3, further comprising, before obtaining, in response to receiving the object risk assessment request, the static object risk information and the dynamic object risk information corresponding to the object identifier to be assessed according to the object identifier to be assessed in the object risk assessment request: In response to detecting object registration operations initiated by M first candidate objects, acquiring candidate static object risk information corresponding to each of the M first candidate object identifiers; Associating and storing the M first candidate object identifiers and the candidate static object risk information corresponding to each of the M first candidate object identifiers in a data source; In response to detecting object transaction operations initiated by N second candidate objects, acquiring candidate dynamic object risk information corresponding to each of the N second candidate object identifiers, wherein the N second candidate objects belong to the M first candidate objects; and Associating and storing the N second candidate object identifiers and the candidate dynamic object risk information corresponding to each of the N second candidate object identifiers in the data source; Wherein, M and N are both positive integers, and N is less than or equal to M.

9. A method for training an object risk assessment model, comprising: Obtaining sample static object risk information, sample dynamic object risk information and object risk label information corresponding to the sample to-be-evaluated object identifier; In response to the sample dynamic object risk information matching a predetermined rule, determining a preset weight corresponding to the predetermined rule; Processing the sample static object risk information and the sample dynamic object risk information respectively to obtain sample object risk features; Determine a risk assessment value of the sample object according to the preset weight and the risk characteristics of the sample object; as well as The sample object risk assessment value and the object risk label information are used to train a deep learning model to obtain an object risk assessment model.

10. The method according to claim 9, wherein: The deep learning model includes a first feature extraction module and a second feature extraction module, the sample static object risk information includes sample to-be-evaluated object information and sample associated object information corresponding to the sample to-be-evaluated object identifier, and the sample dynamic object risk information includes sample transaction behavior information corresponding to the sample to-be-evaluated object identifier; The sample static object risk information and the sample dynamic object risk information are processed separately to obtain the sample object risk features, including: Inputting the sample to-be-evaluated object information and the sample associated object information into the first feature extraction module to obtain the sample to-be-evaluated object feature and the sample associated object feature; as well as The sample transaction behavior information is input into the second feature extraction module to obtain sample transaction behavior features.

11. An object risk assessment device, comprising: A first acquisition module, configured to, in response to receiving an object risk assessment request, acquire static object risk information and dynamic object risk information corresponding to the object identifier to be assessed according to the object identifier to be assessed in the object risk assessment request; A first determination module, configured to determine a preset weight corresponding to a predetermined rule in response to the dynamic object risk information matching the predetermined rule; A first processing module, configured to process the static object risk information and the dynamic object risk information respectively to obtain object risk features; A second determination module, used to determine the object risk assessment value according to the preset weight and the object risk feature; as well as The third determination module is used to determine the object risk assessment result corresponding to the identification of the object to be assessed according to the object risk assessment value.

12. A training device for an object risk assessment model, comprising: The second acquisition module is used to acquire the sample static object risk information, the sample dynamic object risk information and the object risk label information corresponding to the sample to-be-evaluated object identifier; A fourth determination module, configured to determine a preset weight corresponding to the predetermined rule in response to the sample dynamic object risk information matching the predetermined rule; A second processing module is used to process the sample static object risk information and the sample dynamic object risk information respectively to obtain the sample object risk characteristics; A fifth determination module, used to determine a risk evaluation value of the sample object according to the preset weight and the risk characteristics of the sample object; as well as The training module is used to train a deep learning model using the sample object risk assessment value and the object risk label information to obtain an object risk assessment model.

13. An electronic device comprising: one or more processors; a memory for storing one or more instructions, Wherein, when the one or more instructions are executed by the one or more processors, the one or more processors implement the method of any one of claims 1 to 8 or any one of claims 9 to 10.

14. A computer-readable storage medium having executable instructions stored thereon, wherein when the executable instructions are executed by a processor, the processor is enabled to implement the method of any one of claims 1 to 8 or any one of claims 9 to 10.

Images