WO2020062641A1 - 识别用户角色的方法、用户设备、存储介质及装置 - Google Patents

识别用户角色的方法、用户设备、存储介质及装置 Download PDF

Info

Publication number
WO2020062641A1
WO2020062641A1 PCT/CN2018/122726 CN2018122726W WO2020062641A1 WO 2020062641 A1 WO2020062641 A1 WO 2020062641A1 CN 2018122726 W CN2018122726 W CN 2018122726W WO 2020062641 A1 WO2020062641 A1 WO 2020062641A1
Authority
WO
WIPO (PCT)
Prior art keywords
behavior
preset
type
target
user
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2018/122726
Other languages
English (en)
French (fr)
Inventor
刘中原
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
OneConnect Smart Technology Co Ltd
Original Assignee
OneConnect Smart Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by OneConnect Smart Technology Co Ltd filed Critical OneConnect Smart Technology Co Ltd
Publication of WO2020062641A1 publication Critical patent/WO2020062641A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/03Credit; Loans; Processing thereof

Definitions

  • the present application relates to the field of information processing technologies, and in particular, to a method, a user equipment, a storage medium, and a device for identifying a user role.
  • the current identification methods are: financial institutions may obtain credit information of loan customers, and evaluate the possibility of the loan customer as a fraudulent element of the loan according to the credit status of the loan customer displayed by the credit information. Obviously, the current identification methods have technical problems that cannot accurately identify fraudsters.
  • the above content is only used to assist in understanding the technical solution of the present application, and does not mean that the above content is prior art.
  • the main purpose of this application is to provide a method, a user equipment, a storage medium, and a device for identifying a role of a user, in order to solve the technical problems of current identification methods that cannot accurately identify loan fraudsters.
  • the present application provides a method for identifying a user role.
  • the method for identifying a user role includes the following steps:
  • the target fraud score is greater than or equal to a preset fraud threshold, the user role of the target user is identified as a potential fraudster.
  • the present application further provides a user equipment, where the user equipment includes a memory, a processor, and computer-readable instructions stored on the memory and executable on the processor, the computer
  • the readable instructions are configured as steps of a method of identifying a user role as described above.
  • the present application also proposes a storage medium storing computer-readable instructions stored on the storage medium.
  • the computer-readable instructions are executed by a processor, the method for identifying a user role as described above is implemented. A step of.
  • the present application also proposes a device for identifying a user role.
  • the device for identifying a user role includes:
  • An information acquisition module for acquiring user operation behavior and personal information of the target user
  • a behavior matching module configured to match the user operation behavior with a preset operation behavior in a preset behavior model to obtain a matching result
  • a first score generating module configured to generate a corresponding first fraud score according to the matching result under a preset score evaluation model
  • a network construction module configured to construct a target relationship network based on the personal information under a preset complex network model
  • a user determining module configured to determine an associated user corresponding to the target user through the target relationship network
  • a second score generation module configured to query a corresponding second fraud score in a preset fraud score mapping relationship based on the user role of the associated user, the preset fraud score mapping relationship including a user role and a fraud score Corresponding relationship
  • a target score generating module configured to generate a target fraud score by using the first fraud score and the second fraud score
  • An identity recognition module is configured to identify the user role of the target user as a potential fraud when the target fraud score is greater than or equal to a preset fraud threshold.
  • this application matching is performed with a user operation behavior according to a preset operation behavior in a preset behavior model, a first fraud score is generated according to the matching result, and a second is generated according to a user role of an associated user in the target relationship network.
  • Fraud scores ultimately identifying potential fraudsters through target fraud scores.
  • this application will combine the two identification methods to identify the user's identity. It not only comprehensively evaluates the risk of the target user through the preset behavior model and the preset score evaluation model, but also refers to other users who are associated with the target user. It can more accurately identify fraudsters in each user, so it solves the technical problem of current identification methods that cannot accurately identify loan fraudsters.
  • FIG. 1 is a schematic structural diagram of a user equipment in a hardware operating environment involved in a solution according to an embodiment of the present application
  • FIG. 2 is a schematic flowchart of a first embodiment of a method for identifying a user role in this application
  • FIG. 3 is a schematic diagram of a target relationship network
  • FIG. 4 is a schematic flowchart of a second embodiment of a method for identifying a user role in this application
  • FIG. 5 is a schematic flowchart of a third embodiment of a method for identifying a user role in this application.
  • FIG. 6 is a structural block diagram of a first embodiment of an apparatus for identifying a user role in this application.
  • FIG. 1 is a schematic structural diagram of a user equipment in a hardware operating environment according to an embodiment of the present application.
  • the user equipment may include a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005.
  • the communication bus 1002 is used to implement connection and communication between these components.
  • the user interface 1003 may include a display screen.
  • the optional user interface 1003 may further include a standard wired interface and a wireless interface.
  • the wired interface of the user interface 1003 may be a USB interface in this application.
  • the network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface).
  • the memory 1005 may be a high-speed RAM memory or a non-volatile memory. memory), such as disk storage.
  • the memory 1005 may optionally be a storage device independent of the foregoing processor 1001.
  • FIG. 1 does not constitute a limitation on the user equipment, and may include more or fewer components than shown in the figure, or some components may be combined, or different component arrangements.
  • the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and computer-readable instructions.
  • the network interface 1004 is mainly used to connect to a background server and perform data communication with the background server;
  • the user interface 1003 is mainly used to connect peripherals;
  • the user equipment calls the memory 1005 through the processor 1001
  • FIG. 2 is a schematic flowchart of a first embodiment of a method for identifying a user role in this application.
  • the method for identifying a user role includes the following steps:
  • Step S10 Acquire the user operation behavior and personal information of the target user.
  • a loaned person may be a fraudulent person is also determined based on various user behaviors and relationship networks of the loaned person after handling the loan business.
  • the identity recognition through user behavior refers to the actual operation behavior of the user
  • the identity recognition through relationship network refers to the risk of other users who are associated with the user. Combining these two identification methods can refer to the user and other The risk of users, more accurately identify fraudsters.
  • the execution subject of this embodiment is a user equipment
  • the user equipment may be an electronic device such as a personal computer and an automatic teller machine.
  • user A handles the loan business and obtains the money, he will automatically collect user operation behaviors of user A after receiving the money, such as query behaviors such as querying the transfer information and repayment period.
  • Step S20 Match the user operation behavior with a preset operation behavior in a preset behavior model to obtain a matching result.
  • the user operation behavior can be matched with a predetermined normal operation behavior to determine whether the target user's operation behavior is in line with expectations, and Determine whether the user role of the target user is a normal user or a potential fraudster.
  • normal users are safe users who infer that they will repay normally on the repayment date
  • potential fraudsters are risk users who infer that they may conduct loan fraud.
  • the preset behavior model is a behavior model established based on the behavior logic of normal users, and the preset behavior model will include multiple operation behaviors that are identified as safe users to perform, for example, the preset behavior model
  • the preset operation behaviors include "query behavior of querying repayment period" and "query behavior of querying minimum repayment amount”. If the user operation behavior is "query behavior for querying transfer information", the preset operation behavior in the preset behavior model is different from the user operation behavior, the matching result between the two is considered to be a matching failure; if the user operation behavior is "query "Query behavior of repayment period", the preset operation behavior in the preset behavior model is the same as the user operation behavior, the matching result of the two can be considered as a successful match.
  • the reason why the preset behavior model includes “the query behavior for querying the repayment period” and “the query behavior for querying the minimum repayment amount” and does not include the “query behavior for querying the transfer information” is because of the preset behavior
  • the model is a normal user's behavior model. Therefore, the implicit intention of the behavior contained in the preset behavior model is early repayment. However, the implicit intention of "query behavior of querying transfer information" is to transfer money as soon as possible. The intentions are similar.
  • Step S30 Generate a corresponding first fraud score according to the matching result under a preset score evaluation model.
  • the preset score evaluation model is used to calculate fraud scores based on the matching results of various types of user operation behaviors and preset operation behaviors, and the fraud scores are used to evaluate the target user as a potential fraudster. possibility. Therefore, the greater the number of successful matching results, the closer the target user ’s user behavior is to the preset operation behavior contained in the preset behavior model, the more likely the target user is a normal user, the lower the fraud score. ;
  • the matching result is that the greater the number of matching failure results, the closer it is to the preset operation behavior contained in the preset behavior model, the more likely the target user is a potential fraudster, the higher the fraud score.
  • the value range of the first fraud score may be 0 ⁇ x1 ⁇ 50, where x1 represents the first fraud score, and the first fraud score of the user A is 40.
  • Step S40 Construct a target relationship network according to the personal information under a preset complex network model.
  • the identity can also be identified through the relationship network, and the second fraud score can be obtained.
  • a relationship network can be constructed based on his personal information.
  • personal information includes mobile phone number, contact person, address, email address, company name, and fingerprint.
  • the preset complex network model is used to complete the construction of the relational network, and will be based on graph theory. Considering the characteristics of the complex network model, the network it constructs will consist of multiple nodes and the connections connecting each node. Edges consist of nodes that represent different entities and edges that represent relationships between entities.
  • the nodes in this embodiment represent users, and the edges represent association relationships that exist between users.
  • a relationship network is constructed based on the personal information of user A, and the relationship network is constructed. It will include a pre-existing preset node and node A representing user A. Among them, the preset node represents other users who have already loaned.
  • FIG. 3 is a schematic diagram of a target relationship network. Node B, node C, and node D in the figure are preset nodes.
  • Step S50 Determine an associated user corresponding to the target user through the target relationship network.
  • the target relationship network shown in FIG. 3 is constructed, if there is a connection edge between node A and other nodes, it indicates that there is a cross between personal information of node A and personal information of other nodes. Overlapping situations, for example, the company name recorded in the personal information of node A may be the same as the company name recorded in the personal information of node D, or there may be a social relationship between the user represented by node A and the user represented by node D Wait. Considering that loan fraudsters exist in the form of gangs, the identity of user A can be indirectly inferred based on the identity of the associated user.
  • Step S60 Query a corresponding second fraud score in a preset fraud score mapping relationship based on the user role of the associated user, and the preset fraud score mapping relationship includes a correspondence between the user role and the fraud score.
  • the score can be used to evaluate the possibility that a user is a fraudulent.
  • the fraud score of a normal user is 0, and the fraud score of a potential fraudster is 40. Wait.
  • the value range of the second fraud score may be 0 ⁇ x2 ⁇ 50, where x2 represents the second fraud score. If the user role of the user D indicated by the node D is a potential fraudster, the second fraud score is 40. .
  • Step S70 Generate a target fraud score by using the first fraud score and the second fraud score.
  • the first fraud score of the user A is 40 and the second fraud score is 40
  • the first fraud score and the second fraud score may be accumulated to obtain a target fraud score of 80.
  • the value range of the target fraud score can be 0 ⁇ x ⁇ 100, where x represents the target fraud score.
  • Step S80 When the target fraud score is greater than or equal to a preset fraud threshold, identify the user role of the target user as a potential fraudster.
  • the user A can be classified as a potential fraudster.
  • the fraud score is less than 75, user A can be classified as a normal user.
  • a first fraud score is generated according to a matching result of a preset operation behavior and a user operation behavior, and then a second fraud score is generated according to a user role of an associated user. Finally, a target fraud score is used to complete a potential fraud score.
  • Identification of fraudsters Obviously, this embodiment will combine the two identification methods at the same time, which can more accurately identify the fraudulent elements, and solves the technical problem that the current identification means cannot accurately identify the loan fraudulent elements.
  • FIG. 4 is a schematic flowchart of a second embodiment of a method for identifying a user role in the present application. Based on the first embodiment shown in FIG. 2 described above, a second embodiment of the method for identifying a user role in the present application is proposed.
  • the method before step S20, the method further includes:
  • Step S101 Identify the behavior type of the user operation behavior.
  • the behavior types of the user operation behaviors may be determined first.
  • the behavior types include two categories, which are query behavior types and characterization information that characterize information query behaviors.
  • the input behavior type of the input behavior include “query behavior for querying transfer information”, “query behavior for querying repayment period”, and “query behavior for querying minimum repayment amount”.
  • Step S102 Query a corresponding preset operation behavior in a preset behavior model according to the behavior type.
  • the identified behavior type of user A is a query behavior type
  • a preset operation behavior corresponding to the query behavior type is extracted from a preset behavior model, and obtained
  • the preset operation behavior is used to characterize the similar behavior of normal users. For example, if the user operation behavior is "query behavior for querying transfer information", the preset operation behaviors obtained are “query behavior for querying repayment period” and "query behavior for querying minimum repayment amount”.
  • the step S20 may include:
  • Step S201 Determine a behavior feature type corresponding to the behavior type.
  • the type of behavioral characteristics will also be determined, so as to eventually convert the matching between behaviors into the matching between behavioral characteristics. If the behavior type is query behavior type, the corresponding behavior feature type is query item type; if the behavior type is input behavior type, the corresponding behavior feature type is input duration type, input content type, and input order type.
  • Step S202 extracting a current behavior feature corresponding to the behavior feature type from the user operation behavior, and extracting a target behavior feature corresponding to the behavior feature type from the preset operation behavior.
  • the behavior type is a query behavior type
  • the corresponding behavior characteristic type is a query item type
  • the query item type of "query behavior of querying transfer information" is transfer information
  • query behavior of querying repayment period The type of query item is repayment period
  • the type of query item of "Query behavior of querying minimum repayment amount” is minimum repayment amount.
  • Step S203 Match the current behavior feature with the target behavior feature to obtain a matching result.
  • the matching result is a matching failure.
  • the behavior type includes a query behavior type
  • the behavior characteristic type includes a query item type corresponding to the query behavior type, and the querying a corresponding preset operation in a preset behavior model according to the behavior type
  • the behavior includes: when the behavior type is a query behavior type, querying a corresponding preset operation behavior in a preset behavior model according to the query behavior type;
  • the extracting a current behavior feature corresponding to the behavior feature type from the user operation behavior, and extracting a target behavior feature corresponding to the behavior feature type from the preset operation behavior include: including the behavior feature When the type is a query item type, a current behavior feature corresponding to the query item type is extracted from the user operation behavior, and a target behavior feature corresponding to the query item type is extracted from the preset operation behavior.
  • the preset of behavior type A will be obtained from the preset behavior model first. Operational behavior.
  • a current behavior characteristic under the query item type is extracted from the user operation behavior
  • a target behavior characteristic under the query item type is extracted from the preset operation behavior
  • the current behavior characteristic may be transfer information
  • the target behavior characteristic may be repayment period.
  • the behavior type further includes an input behavior type
  • the behavior characteristic type includes at least one of an input duration type, an input content type, and an input order type corresponding to the input behavior type, and the according to the
  • the behavior type querying the corresponding preset operation behavior in the preset behavior model includes: when the behavior type is an input behavior type, querying the corresponding preset operation behavior in the preset behavior model according to the input behavior type;
  • the extracting a current behavior feature corresponding to the behavior feature type from the user operation behavior, and extracting a target behavior feature corresponding to the behavior feature type from the preset operation behavior include: including the behavior feature When the type is an input duration type, a current behavior feature corresponding to the input duration type is extracted from the user operation behavior, and a target behavior feature corresponding to the input duration type is extracted from the preset operation behavior.
  • the operation behavior refers to the behavior of user A entering information when operating the user device, for example, the input behavior of inputting information such as home address, name, and ID number. .
  • the input duration type refers to the statistical duration of the input ID number
  • the input content type refers to the input ID number
  • the input sequence type is Refers to the input order of inputting each sub-information when there are multiple sub-informations to be inputted; correspondingly, the preset behavior model will store the input behaviors of inputting ID numbers that are regarded as normal user behaviors, for example, normal input time
  • the feature is 20 seconds
  • the normal input content type is 12345
  • the normal input order type is to input sub-information in order from left to right from top to bottom.
  • the matching operation may be that if the duration indicated by the current behavior feature is greater than or equal to the target behavior feature, the match is deemed to have failed; if the duration indicated by the current behavior feature is less than the target behavior feature, the match is deemed successful.
  • matching the current behavior feature with the target behavior feature to obtain a matching result includes: querying a corresponding matching rule according to the behavior feature type in a preset behavior model; and based on the matching rule, the current behavior feature and the target are matched. Match behavioral features to get matching results.
  • the differences between behavior characteristics lead to differences in the matching methods between behavior characteristics, so different matching rules can be set in a targeted manner.
  • the behavior feature type is the input content type
  • the corresponding matching rule is "calculate similarity and match based on similarity". Specifically, based on the matching rule, the similarity between the current behavior feature and the target behavior feature is calculated. The similarity is compared with a preset similarity threshold to obtain a comparison result, and the comparison result is determined as a matching result corresponding to a matching rule.
  • the input content of the current behavioral feature characterization is the ID card number 12346 and the input content of the target behavioral feature characterization is the ID card number 12345, and calculate the similarity between the two, the higher the similarity is Those are more similar. If the calculated similarity is 0.8, the preset similarity threshold is 0.7, and the similarity is greater than the preset similarity threshold, the comparison result is "similarity is greater than or equal to the preset similarity threshold", and the matching result is a successful match.
  • the behavior type of the user's operation behavior is determined in advance, and then the behavior characteristics corresponding to the behavior characteristic type are extracted, and the matching between the behaviors is converted into the matching between the characteristics, which can more accurately complete the user's operation behavior and The comparison between the normal user's operation behavior, and then complete the judgment of the user's identity.
  • FIG. 5 is a schematic flowchart of a third embodiment of a method for identifying a user role in the present application. Based on the first embodiment shown in FIG. 2 described above, a third embodiment of the method for identifying a user role in this application is proposed.
  • the step S40 may include:
  • Step S401 Obtain a preset relationship network generated from preset personal information.
  • a preset relationship network in order to build a target relationship network, a preset relationship network may be obtained first.
  • the preset relationship network is composed of each preset node and a connecting edge connecting each preset node. See FIG. 3, the preset relationship network includes Node B representing user B, node C representing user C, node D representing user D, and the edges connecting these three nodes to each other.
  • the preset personal information represents personal information of the loaned user.
  • Step S402 Create a target node corresponding to the target user according to the personal information of the target user under a preset complex network model.
  • Step S403 Add the target node to the preset relationship network, and set the preset relationship network after the node is added as the target relationship network.
  • node A after obtaining the preset relationship network, node A can be created, and node A is added to the preset relationship network, thereby obtaining the target relationship network shown in FIG. 3.
  • the preset relationship network includes each preset node and a connecting edge connecting each preset node, and the target node is added to the preset relationship network, and the preset relationship after the node is added
  • the network is set as a target relationship network, which includes: matching personal information of the target user with preset personal information; and when the matching is successful, determining a preset node corresponding to the successfully matched preset personal information and setting it as an associated node; Adding the target node to the preset relationship network, and adding a connecting edge connecting the associated node to the target node; and setting the modified preset relationship network as a target relationship network.
  • the connecting edges will represent the association relationship between the connected nodes. If the association relationship is determined as the information similarity between the personal information of the user represented by the connected node, the matching operation of the personal information may be performed first.
  • the matching operation of personal information is specifically that the company names of user A and user B can be matched first. If the names are not the same, no connection edge is added between node A and node B; the company names of user A and user D are performed. Match, if the names are the same, it is considered that there is a certain information similarity between user A and user D. If there is an association between the two, a connection edge can be added between node A and node B.
  • the personal information of the target user includes sub-information of each information type
  • matching the personal information of the target user with preset personal information includes: matching sub-information in the personal information of the target user Compare with the sub-information in the preset personal information, and count the number of target items with the same sub-information; when the number of target items is greater than or equal to the number of preset standard items, determine that the personal information of the target user and the It is assumed that the personal information matches successfully.
  • the matching operation for personal information can also comprehensively consider all types of personal information, so that the finally selected associated nodes and targets Nodes have more reliable node associations.
  • the personal information of each type of information is a sub-information.
  • the number of items of the same sub-information in the personal information of user A and other users may be counted, for example, the company name, address, and mobile phone number of user A are the same as those of user D.
  • the target number of items is 3, and the preset standard number of items is set to 2. Because the number of target items is greater than the number of preset standard items, it can be considered that there is an association relationship between user A and user D, and a connection edge between node A and node D is established.
  • the construction of the target relationship network is completed based on the original preset relationship network, and the construction efficiency of constructing the target relationship network is improved.
  • an embodiment of the present application further provides a storage medium.
  • the storage medium may be a non-volatile readable storage medium.
  • the storage medium stores computer-readable instructions, and the computer-readable instructions are executed by a processor.
  • an embodiment of the present application further provides a device for identifying a user role.
  • the device for identifying a user role includes:
  • An information acquisition module 10 configured to acquire user operation behavior and personal information of a target user
  • a behavior matching module 20 configured to match the user operation behavior with a preset operation behavior in a preset behavior model to obtain a matching result
  • a first score generating module 30, configured to generate a corresponding first fraud score according to the matching result under a preset score evaluation model
  • a network construction module 40 configured to construct a target relationship network according to the personal information under a preset complex network model
  • a user determining module 50 configured to determine an associated user corresponding to the target user through the target relationship network
  • a second score generation module 60 is configured to query a corresponding second fraud score in a preset fraud score mapping relationship based on a user role of the associated user, where the preset fraud score mapping relationship includes a user role and a fraud score. Correspondence of values;
  • a target score generating module 70 configured to generate a target fraud score by using the first fraud score and the second fraud score
  • the identity recognition module 80 is configured to identify the user role of the target user as a potential fraud when the target fraud score is greater than or equal to a preset fraud threshold.

Landscapes

  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本申请公开了识别用户角色的方法、用户设备、存储介质及装置。本申请中获取用户操作行为和个人信息;将用户操作行为与预设操作行为进行匹配,以获得匹配结果;根据匹配结果生成对应的第一欺诈分值;根据个人信息构建目标关系网络;通过目标关系网络确定关联用户;基于关联用户的用户角色查询对应的第二欺诈分值;通过第一欺诈分值与第二欺诈分值生成目标欺诈分值;在目标欺诈分值大于或等于预设欺诈阈值时,将用户角色识别为潜在欺诈分子。明显地,本申请能够全面地评估目标用户的风险性,也参考到了关联用户的风险性,能够准确地识别出欺诈分子,所以,解决了现行的识别手段存在的不能准确地识别出贷款欺诈分子的技术问题。

Description

识别用户角色的方法、用户设备、存储介质及装置
本申请要求于2018年09月26日提交中国专利局,申请号为201811125685.7,发明名称为“识别用户角色的方法、用户设备、存储介质及装置”的中国专利申请的优先权,其全部内容通过引用结合在申请中。
技术领域
本申请涉及信息处理技术领域,尤其涉及识别用户角色的方法、用户设备、存储介质及装置。
背景技术
针对互联网金融机构的贷款业务而言,由于欺诈贷款并不及时归还贷款的欺诈分子不断增多,这对于金融机构的正常运营而言会带来诸多麻烦。可是,常规的识别手段并不能准确地识别出各贷款客户中隐藏着的欺诈分子。
现行的识别手段有,金融机构可调取贷款客户的征信信息,根据征信信息显示的贷款客户的信用状况来评估贷款客户为贷款欺诈分子的可能性。明显地,现行的识别手段存在不能准确地识别出欺诈分子的技术问题。上述内容仅用于辅助理解本申请的技术方案,并不代表承认上述内容是现有技术。
发明内容
本申请的主要目的在于提供识别用户角色的方法、用户设备、存储介质及装置,旨在解决现行的识别手段存在的不能准确地识别出贷款欺诈分子的技术问题。
为实现上述目的,本申请提供一种识别用户角色的方法,所述识别用户角色的方法包括以下步骤:
获取目标用户的用户操作行为和个人信息;
将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果;
在预设分值评定模型下根据所述匹配结果生成对应的第一欺诈分值;
在预设复杂网络模型下根据所述个人信息构建目标关系网络;
通过所述目标关系网络确定与所述目标用户对应的关联用户;
基于所述关联用户的用户角色在预设欺诈分映射关系中查询对应的第二欺诈分值,所述预设欺诈分映射关系中包括用户角色与欺诈分值的对应关系;
通过所述第一欺诈分值与所述第二欺诈分值生成目标欺诈分值;
在所述目标欺诈分值大于或等于预设欺诈阈值时,将所述目标用户的用户角色识别为潜在欺诈分子。
此外,为实现上述目的,本申请还提出一种用户设备,所述用户设备包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机可读指令,所述计算机可读指令配置为实现如上文所述的识别用户角色的方法的步骤。
此外,为实现上述目的,本申请还提出一种存储介质,所述存储介质上存储有计算机可读指令,所述计算机可读指令被处理器执行时实现如上文所述的识别用户角色的方法的步骤。
此外,为实现上述目的,本申请还提出一种识别用户角色的装置,所述识别用户角色的装置包括:
信息获取模块,用于获取目标用户的用户操作行为和个人信息;
行为匹配模块,用于将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果;
第一分值生成模块,用于在预设分值评定模型下根据所述匹配结果生成对应的第一欺诈分值;
网络构建模块,用于在预设复杂网络模型下根据所述个人信息构建目标关系网络;
用户确定模块,用于通过所述目标关系网络确定与所述目标用户对应的关联用户;
第二分值生成模块,用于基于所述关联用户的用户角色在预设欺诈分映射关系中查询对应的第二欺诈分值,所述预设欺诈分映射关系中包括用户角色与欺诈分值的对应关系;
目标分值生成模块,用于通过所述第一欺诈分值与所述第二欺诈分值生成目标欺诈分值;
身份识别模块,用于在所述目标欺诈分值大于或等于预设欺诈阈值时,将所述目标用户的用户角色识别为潜在欺诈分子。
在本申请中根据预设行为模型中的预设操作行为进行与用户操作行为的匹配,根据匹配结果来生成第一欺诈分值,再根据目标关系网络中的关联用户的用户角色来生成第二欺诈分值,最终通过目标欺诈分值来完成对于潜在欺诈分子的认定。明显地,本申请将同时联合两种识别手段来识别用户身份,既通过预设行为模型与预设分值评定模型全面地评估目标用户的风险性,也参考到了与目标用户存在关联的其他用户的风险性,能够更加准确地识别出各用户中的欺诈分子,所以,解决了现行的识别手段存在的不能准确地识别出贷款欺诈分子的技术问题。
附图说明
图1是本申请实施例方案涉及的硬件运行环境的用户设备结构示意图;
图2为本申请识别用户角色的方法第一实施例的流程示意图;
图3为目标关系网络的示意图;
图4为本申请识别用户角色的方法第二实施例的流程示意图;
图5为本申请识别用户角色的方法第三实施例的流程示意图;
图6为本申请识别用户角色的装置第一实施例的结构框图。
本申请目的的实现、功能特点及优点将结合实施例,参照附图做进一步说明。
具体实施方式
应当理解,此处所描述的具体实施例仅仅用以解释本申请,并不用于限定本申请。参照图1,图1为本申请实施例方案涉及的硬件运行环境的用户设备结构示意图。
如图1所示,该用户设备可以包括:处理器1001,例如CPU,通信总线1002、用户接口1003,网络接口1004,存储器1005。其中,通信总线1002用于实现这些组件之间的连接通信。用户接口1003可以包括显示屏(Display),可选用户接口1003还可以包括标准的有线接口、无线接口,对于用户接口1003的有线接口在本申请中可为USB接口。网络接口1004可选的可以包括标准的有线接口、无线接口(如WI-FI接口)。存储器1005可以是高速RAM存储器,也可以是稳定的存储器(non-volatile memory),例如磁盘存储器。存储器1005可选的还可以是独立于前述处理器1001的存储装置。
本领域技术人员可以理解,图1中示出的结构并不构成对用户设备的限定,可以包括比图示更多或更少的部件,或者组合某些部件,或者不同的部件布置。
如图1所示,作为一种计算机存储介质的存储器1005中可以包括操作系统、网络通信模块、用户接口模块以及计算机可读指令。在图1所示的用户设备中,网络接口1004主要用于连接后台服务器,与所述后台服务器进行数据通信;用户接口1003主要用于连接外设;所述用户设备通过处理器1001调用存储器1005中存储的计算机可读指令,并执行本申请实施例提供的识别用户角色的方法。
基于上述硬件结构,提出本申请识别用户角色的方法的实施例。
参照图2,图2为本申请识别用户角色的方法第一实施例的流程示意图。
在第一实施例中,所述识别用户角色的方法包括以下步骤:
步骤S10:获取目标用户的用户操作行为和个人信息。
可以理解的是,本实施例将同时基于已贷款人员在办理贷款业务之后的各种用户行为以及关系网络来判断该人员是否可能为欺诈分子。通过用户行为来进行身份识别参考到用户的实际操作行为,通过关系网络来进行身份识别参考到与用户存在关联性的其他用户的风险性,联合这两种识别手段可以同时参考到本用户与其他用户的风险性,更加准确地识别出欺诈分子。
在具体实现中,本实施例的执行主体为用户设备,而用户设备可为个人电脑以及自动取款机等电子设备。当用户A在办理贷款业务并获得款项后,将自动采集用户A在获得款项之后的用户操作行为,用户操作行为比如为查询转账信息以及还款期限等的查询行为。
步骤S20:将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果。
应当理解的是,在采集到用户A在获得款项之后实际发生的用户操作行为后,可将该用户操作行为与预先确定的正常操作行为进行匹配,以判断目标用户的操作行为是否符合预期,进而判断目标用户的用户角色究竟为正常用户或者潜在欺诈分子。其中,正常用户为推断其将在还款日正常还款的安全用户,潜在欺诈分子为推断其可能进行贷款欺诈的风险性用户。
在具体实现中,预设行为模型为基于正常用户的行为逻辑建立的行为模型,预设行为模型内将包含有多个被认定为安全用户将执行的操作行为,比如,预设行为模型内的预设操作行为包括有“查询还款期限的查询行为”与“查询还款最低额度的查询行为”。若用户操作行为为“查询转账信息的查询行为”,则预设行为模型内的预设操作行为与用户操作行为不同,则可认为二者的匹配结果为匹配失败;若用户操作行为为“查询还款期限的查询行为”,预设行为模型内的预设操作行为与用户操作行为相同,则可认为二者的匹配结果为匹配成功。
可以理解的是,预设行为模型内包含“查询还款期限的查询行为”与“查询还款最低额度的查询行为”而不包含“查询转账信息的查询行为”的原因在于,由于预设行为模型是正常用户的行为模型,所以,预设行为模型内包含的行为的隐含意图为尽早还款,但是,“查询转账信息的查询行为”的隐含意图为尽快转账,这与欺诈份子的意图较为类似。
步骤S30:在预设分值评定模型下根据所述匹配结果生成对应的第一欺诈分值。
应当理解的是,预设分值评定模型用于根据各种类型的用户操作行为与预设操作行为的匹配结果来计算出欺诈分值,而欺诈分值用于评估目标用户为潜在欺诈分子的可能性。所以,匹配结果为匹配成功的结果个数越多,表示目标用户的用户操作行为越接近预设行为模型内包含的预设操作行为,则目标用户越可能为正常用户,则欺诈分值越低;匹配结果为匹配失败的结果个数越多,表示越不接近预设行为模型内包含的预设操作行为,则目标用户越可能为潜在欺诈分子,则欺诈分值越高。比如,第一欺诈分值的取值范围可为0≤x1≤50,x1表示第一欺诈分值,而用户A的第一欺诈分值为40。
步骤S40:在预设复杂网络模型下根据所述个人信息构建目标关系网络。
可以理解的是,在根据用户行为得出第一欺诈分值后,还可通过关系网络来进行身份识别,并得出第二欺诈分值。
在具体实现中,当用户A办理了贷款业务后,由于办理贷款业务时用户A会填写其个人信息,可基于其个人信息来构建关系网络。其中,个人信息包括手机号、联系人、地址、邮箱、公司名称以及指纹等。
应当理解的是,预设复杂网络模型用于完成对于关系网络的构建,将以图论作为基础,而考虑到复杂网络模型的特性,其构建出的网络将由多个节点以及连接各节点的连接边组成,其中,节点用来代表不同的实体,而边则用来表示实体间的关系。
在具体实现中,本实施例中的节点表示用户,而边表示用户之间存在的关联关系,具体而言,本实施例将基于用户A的个人信息构建出关系网络,构建出的关系网络中将包括预先存在的预设节点以及代表用户A的节点A。其中,预设节点表示已贷款的其他用户。具体可参见图3,图3为目标关系网络的示意图,图中的节点B、节点C以及节点D均为预设节点。
步骤S50:通过所述目标关系网络确定与所述目标用户对应的关联用户。
应当理解的是,在构建出如图3所示的目标关系网络后,若节点A与其他节点之间存在着连接边,则表示节点A的个人信息与其他节点的个人信息之间存在着交叉重叠的情况,比如,可能节点A的个人信息中记录的公司名称可能与节点D的个人信息中记录的公司名称相同,也可能节点A表示的用户与节点D表示的用户之间存在着社交关系等。考虑到贷款欺诈分子多以团伙的形式存在,可基于关联用户的身份来间接推断用户A的身份。
步骤S60:基于所述关联用户的用户角色在预设欺诈分映射关系中查询对应的第二欺诈分值,所述预设欺诈分映射关系中包括用户角色与欺诈分值的对应关系。
应当理解的是,不同的用户角色将对应不同的分值,可通过分值来评估用户为欺诈分子的可能性,比如,正常用户的欺诈分值为0,潜在欺诈分子的欺诈分值为40等。其中,第二欺诈分值的取值范围可为0≤x2≤50,x2表示第二欺诈分值,若节点D表示的用户D的用户角色为潜在欺诈分子,则第二欺诈分值为40。
步骤S70:通过所述第一欺诈分值与所述第二欺诈分值生成目标欺诈分值。
在具体实现中,用户A的第一欺诈分值为40,第二欺诈分值为40,则可将第一欺诈分值与第二欺诈分值进行累加操作,以获得目标欺诈分值80。其中,目标欺诈分值的取值范围可为0≤x≤100,x表示目标欺诈分值。
步骤S80:在所述目标欺诈分值大于或等于预设欺诈阈值时,将所述目标用户的用户角色识别为潜在欺诈分子。
在具体实现中,若预设欺诈阈值为75,由于目标欺诈分值大于75,则可将用户A归类为潜在欺诈分子。当然,若欺诈分值小于75,则可将用户A归类为正常用户。
在本实施例中根据预设操作行为与用户操作行为的匹配结果来生成第一欺诈分值,再根据关联用户的用户角色来生成第二欺诈分值,最终通过目标欺诈分值来完成对于潜在欺诈分子的认定。明显地,本实施例将同时联合两种识别手段,能够更加准确地识别出欺诈分子,解决了现行的识别手段存在的不能准确地识别出贷款欺诈分子的技术问题。
参照图4,图4为本申请识别用户角色的方法第二实施例的流程示意图,基于上述图2所示的第一实施例,提出本申请识别用户角色的方法的第二实施例。
第二实施例中,所述步骤S20之前,还包括:
步骤S101:识别所述用户操作行为的行为类型。
可以理解的是,考虑到用户A的用户操作行为存在多种类型,可先确定用户操作行为的行为类型,比如,行为类型包括两大类,分别为表征信息查询行为的查询行为类型以及表征信息输入行为的输入行为类型。其中,与查询行为类型对应的操作行为有“查询转账信息的查询行为”、“查询还款期限的查询行为”以及“查询还款最低额度的查询行为”等。
步骤S102:根据所述行为类型在预设行为模型中查询对应的预设操作行为。
在具体实现中,若识别出的用户A的行为类型为查询行为类型,为了实现行为之间的匹配,将从预设行为模型中提取出与该查询行为类型对应的预设操作行为,获取到的预设操作行为用于表征正常用户的同类行为。比如,若用户操作行为为“查询转账信息的查询行为”,则获取到的预设操作行为为“查询还款期限的查询行为”与“查询还款最低额度的查询行为”等。
所述步骤S20,可以包括:
步骤S201:确定与所述行为类型对应的行为特征类型。
应当理解的是,在确定行为类型后,还将确定行为特征类型,以最终将行为之间的匹配转换为行为特征之间的匹配。若行为类型为查询行为类型,则对应的行为特征类型为查询事项类型;若行为类型为输入行为类型,则对应的行为特征类型为输入时长类型、输入内容类型与输入次序类型等。
步骤S202:从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征。
应当理解的是,由于行为类型为查询行为类型,则对应的行为特征类型为查询事项类型,而“查询转账信息的查询行为”的查询事项类型为转账信息,“查询还款期限的查询行为”的查询事项类型为还款期限,“查询还款最低额度的查询行为”的查询事项类型为还款最低额度。
步骤S203:将所述当前行为特征与所述目标行为特征进行匹配,以获得匹配结果。
可以理解的是,由于当前行为特征为转账信息,目标行为特征为还款期限与还款最低额度不含有转账信息,则匹配结果为匹配失败。
进一步地,所述行为类型包括查询行为类型,所述行为特征类型中包括与所述查询行为类型对应的查询事项类型,所述根据所述行为类型在预设行为模型中查询对应的预设操作行为,包括:在所述行为类型为查询行为类型时,根据所述查询行为类型在预设行为模型中查询对应的预设操作行为;
所述从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征,包括:在所述行为特征类型为查询事项类型时,从所述用户操作行为中提取与所述查询事项类型对应的当前行为特征,从所述预设操作行为中提取与所述查询事项类型对应的目标行为特征。
在具体实现中,若将查询行为类型简记为行为类型A,且与行为类型A对应的行为特征类型为查询事项类型,所以,将先从预设行为模型中获取出行为类型A的预设操作行为。在进行用户操作行为与预设操作行为的匹配时,将从用户操作行为中提取出查询事项类型下的当前行为特征,从所述预设操作行为中提取出查询事项类型下的目标行为特征,其中,当前行为特征可为转账信息,目标行为特征可为还款期限。
进一步地,所述行为类型还包括输入行为类型,所述行为特征类型中包括与所述输入行为类型对应的输入时长类型、输入内容类型与输入次序类型中的至少一项,所述根据所述行为类型在预设行为模型中查询对应的预设操作行为,包括:在所述行为类型为输入行为类型时,根据所述输入行为类型在预设行为模型中查询对应的预设操作行为;
所述从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征,包括:在所述行为特征类型为输入时长类型时,从所述用户操作行为中提取与所述输入时长类型对应的当前行为特征,从所述预设操作行为中提取与所述输入时长类型对应的目标行为特征。
在具体实现中,若操作行为的行为类型为输入行为类型,则该操作行为是指用户A在操作用户设备时输入信息的行为,比如,输入家庭地址、姓名以及身份证号码等信息的输入行为。
可以理解的是,若用户A的用户操作行为为输入身份证号码的输入行为,则输入时长类型是指输入身份证号码的统计时长,输入内容类型是指输入的身份证号码,输入次序类型是指当存在多个待输入的子信息时输入各子信息的输入顺序;对应地,预设行为模型中将存储有认定为正常用户行为的输入身份证号码的输入行为,比如,正常的输入时间特征为20秒,正常的输入内容类型为12345,正常的输入次序类型为从左至右从上至下依次输入子信息。
应当理解的是,若用户A输入身份证号码的实际统计时长为40秒,则可将实际统计时长40秒与正常的输入时间特征20秒进行匹配,以获得匹配结果。当然,此处的匹配操作可为,若当前行为特征表示的时长大于或等于目标行为特征,则认定为匹配失败;若当前行为特征表示的时长小于目标行为特征,则认定为匹配成功。
进一步地,所述将当前行为特征与目标行为特征进行匹配,以获得匹配结果,包括:在预设行为模型中根据行为特征类型查询对应的匹配规则;基于所述匹配规则对当前行为特征与目标行为特征进行匹配,以获得匹配结果。
可以理解的是,行为特征之间的差异导致行为特征之间的匹配方式也存在着差异,所以,可有针对性地设置不同的匹配规则。若行为特征类型为输入内容类型,则对应的匹配规则为“计算相似度并根据相似度进行匹配”,具体而言,基于该匹配规则计算当前行为特征与目标行为特征之间的相似度,将相似度与预设相似度阈值进行比较,以获得比较结果,并将比较结果认定为与匹配规则对应的匹配结果。
在具体实现中,比如,若当前行为特征表征的输入内容为身份证号12346,而目标行为特征表征的输入内容为身份证号12345,并计算二者的相似度,相似度越高则表明二者较为相近。若计算出的相似度为0.8,预设相似度阈值为0.7,相似度大于预设相似度阈值,则比较结果为“相似度大于或等于预设相似度阈值”,匹配结果为匹配成功。
在本实施例中通过预先确定用户操作行为的行为类型,进而提取出与行为特征类型对应的行为特征,将行为之间的匹配转换为特征之间的匹配,能够更加精准地完成用户操作行为与正常用户的操作行为之间比较,进而完成了对于用户身份的判定。
参照图5,图5为本申请识别用户角色的方法第三实施例的流程示意图,基于上述图2所示的第一实施例,提出本申请识别用户角色的方法的第三实施例。
第三实施例中,所述步骤S40,可以包括:
步骤S401:获取由预设个人信息生成的预设关系网络。
在具体实现中,为了构建出目标关系网络,可先获取预设关系网络,预设关系网络由各预设节点与连接各预设节点的连接边构成,可参见图3,预设关系网络包括表示用户B的节点B、表示用户C的节点C、表示用户D的节点D以及这三个节点彼此之间的连接边。预设个人信息表示已贷款用户的个人信息。
步骤S402:在预设复杂网络模型下根据所述目标用户的个人信息创建与所述目标用户对应的目标节点。
步骤S403:将所述目标节点添加至所述预设关系网络中,并将添加节点后的预设关系网络设为目标关系网络。
可以理解的是,在获取到预设关系网络后,可创建出节点A,并将节点A添加至预设关系网络中,从而将获得如图3所示的目标关系网络。
进一步地,所述预设关系网络包括各预设节点与连接各预设节点的连接边,所述将所述目标节点添加至所述预设关系网络中,并将添加节点后的预设关系网络设为目标关系网络,包括:将所述目标用户的个人信息与预设个人信息进行匹配;在匹配成功时,确定与匹配成功的预设个人信息对应的预设节点并设为关联节点;将所述目标节点添加至所述预设关系网络中,并添加连接所述关联节点与所述目标节点的连接边;将改动后的预设关系网络设为目标关系网络。
可以理解的是,考虑到各节点之间存在着连接边,而连接边将表示已连接节点之间的关联关系。若将该关联关系认定为已连接节点表示的用户的个人信息之间的信息相似性,可先执行个人信息的匹配操作。个人信息的匹配操作具体为,可先将用户A与用户B的公司名称进行匹配,若名称不相同,则不在节点A与节点B之间添加连接边;将用户A与用户D的公司名称进行匹配,若名称相同,则认为用户A与用户D之间存在一定的信息相似性。若二者之间存在关联关系,则可在节点A与节点B之间添加连接边。
进一步地,所述目标用户的个人信息包括各信息类型的子信息,所述将所述目标用户的个人信息与预设个人信息进行匹配,包括:将所述目标用户的个人信息中的子信息与预设个人信息中的子信息进行比较,并统计子信息相同的目标项数;当所述目标项数大于或等于预设标准项数时,认定所述目标用户的个人信息与所述预设个人信息匹配成功。
可以理解的是,针对个人信息的匹配操作,除了判断用户A的公司名称与用户D的公司名称是否相同的方式外,还可综合考虑全部类型的个人信息,使得最终挑选出的关联节点与目标节点之间具有较为可信的节点关联性。
在具体实现中,由于个人信息存在多种信息类型,每一种信息类型的个人信息都为一项子信息。在评判用户A与其他用户之间的关联关系时,可统计用户A与其他用户的个人信息中相同的子信息的项数,比如,用户A的公司名称、地址以及手机号均与用户D的相同,则目标项数为3,并设置预设标准项数为2。由于目标项数大于预设标准项数,则可认为用户A与用户D之间存在着关联关系,并建立节点A与节点D之间的连接边。
在本实施例中将基于原有的预设关系网络来完成目标关系网络的构建,提高了构建出目标关系网络的构建效率。
此外,本申请实施例还提出一种存储介质,所述存储介质可以为非易失性可读存储介质,所述存储介质上存储有计算机可读指令,所述计算机可读指令被处理器执行时实现如上文所述的识别用户角色的方法的步骤。
此外,参照图6,本申请实施例还提出一种识别用户角色的装置,所述识别用户角色的装置包括:
信息获取模块10,用于获取目标用户的用户操作行为和个人信息;
行为匹配模块20,用于将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果;
第一分值生成模块30,用于在预设分值评定模型下根据所述匹配结果生成对应的第一欺诈分值;
网络构建模块40,用于在预设复杂网络模型下根据所述个人信息构建目标关系网络;
用户确定模块50,用于通过所述目标关系网络确定与所述目标用户对应的关联用户;
第二分值生成模块60,用于基于所述关联用户的用户角色在预设欺诈分映射关系中查询对应的第二欺诈分值,所述预设欺诈分映射关系中包括用户角色与欺诈分值的对应关系;
目标分值生成模块70,用于通过所述第一欺诈分值与所述第二欺诈分值生成目标欺诈分值;
身份识别模块80,用于在所述目标欺诈分值大于或等于预设欺诈阈值时,将所述目标用户的用户角色识别为潜在欺诈分子。
上述本申请实施例序号仅仅为了描述,不代表实施例的优劣。在列举了若干装置的单元权利要求中,这些装置中的若干个可以是通过同一个硬件项来具体体现。词语第一、第二、以及第三等的使用不表示任何顺序,可将这些词语解释为名称。以上仅为本申请的优选实施例,并非因此限制本申请的专利范围,凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换,或直接或间接运用在其他相关的技术领域,均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种识别用户角色的方法,其特征在于,所述识别用户角色的方法包括以下步骤:
    获取目标用户的用户操作行为和个人信息;
    将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果;
    在预设分值评定模型下根据所述匹配结果生成对应的第一欺诈分值;
    在预设复杂网络模型下根据所述个人信息构建目标关系网络;
    通过所述目标关系网络确定与所述目标用户对应的关联用户;
    基于所述关联用户的用户角色在预设欺诈分映射关系中查询对应的第二欺诈分值,所述预设欺诈分映射关系中包括用户角色与欺诈分值的对应关系;
    通过所述第一欺诈分值与所述第二欺诈分值生成目标欺诈分值;
    在所述目标欺诈分值大于或等于预设欺诈阈值时,将所述目标用户的用户角色识别为潜在欺诈分子。
  2. 如权利要求1所述的识别用户角色的方法,其特征在于,所述将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果的步骤之前,还包括:
    识别所述用户操作行为的行为类型;
    根据所述行为类型在预设行为模型中查询对应的预设操作行为;
    所述将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果,包括:
    确定与所述行为类型对应的行为特征类型;
    从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征;
    将所述当前行为特征与所述目标行为特征进行匹配,以获得匹配结果。
  3. 如权利要求2所述的识别用户角色的方法,其特征在于,所述行为类型包括查询行为类型,所述行为特征类型中包括与所述查询行为类型对应的查询事项类型;
    所述根据所述行为类型在预设行为模型中查询对应的预设操作行为,包括:
    在所述行为类型为查询行为类型时,根据所述查询行为类型在预设行为模型中查询对应的预设操作行为;
    所述从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征,包括:
    在所述行为特征类型为查询事项类型时,从所述用户操作行为中提取与所述查询事项类型对应的当前行为特征,从所述预设操作行为中提取与所述查询事项类型对应的目标行为特征。
  4. 如权利要求3所述的识别用户角色的方法,其特征在于,所述行为类型还包括输入行为类型,所述行为特征类型中包括与所述输入行为类型对应的输入时长类型、输入内容类型与输入次序类型中的至少一项;
    所述根据所述行为类型在预设行为模型中查询对应的预设操作行为,包括:
    在所述行为类型为输入行为类型时,根据所述输入行为类型在预设行为模型中查询对应的预设操作行为;
    所述从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征,包括:
    在所述行为特征类型为输入时长类型时,从所述用户操作行为中提取与所述输入时长类型对应的当前行为特征,从所述预设操作行为中提取与所述输入时长类型对应的目标行为特征。
  5. 如权利要求1所述的识别用户角色的方法,其特征在于,所述在预设复杂网络模型下根据所述个人信息构建目标关系网络,包括:
    获取由预设个人信息生成的预设关系网络;
    在预设复杂网络模型下根据所述目标用户的个人信息创建与所述目标用户对应的目标节点;
    将所述目标节点添加至所述预设关系网络中,并将添加节点后的预设关系网络设为目标关系网络。
  6. 如权利要求2所述的识别用户角色的方法,其特征在于,所述在预设复杂网络模型下根据所述个人信息构建目标关系网络,包括:
    获取由预设个人信息生成的预设关系网络;
    在预设复杂网络模型下根据所述目标用户的个人信息创建与所述目标用户对应的目标节点;
    将所述目标节点添加至所述预设关系网络中,并将添加节点后的预设关系网络设为目标关系网络。
  7. 如权利要求5所述的识别用户角色的方法,其特征在于,所述预设关系网络包括各预设节点与连接各预设节点的连接边;
    所述将所述目标节点添加至所述预设关系网络中,并将添加节点后的预设关系网络设为目标关系网络,包括:
    将所述目标用户的个人信息与预设个人信息进行匹配;
    在匹配成功时,确定与匹配成功的预设个人信息对应的预设节点并设为关联节点;
    将所述目标节点添加至所述预设关系网络中,并添加连接所述关联节点与所述目标节点的连接边;
    将改动后的预设关系网络设为目标关系网络。
  8. 如权利要求6所述的识别用户角色的方法,其特征在于,所述预设关系网络包括各预设节点与连接各预设节点的连接边;
    所述将所述目标节点添加至所述预设关系网络中,并将添加节点后的预设关系网络设为目标关系网络,包括:
    将所述目标用户的个人信息与预设个人信息进行匹配;
    在匹配成功时,确定与匹配成功的预设个人信息对应的预设节点并设为关联节点;
    将所述目标节点添加至所述预设关系网络中,并添加连接所述关联节点与所述目标节点的连接边;
    将改动后的预设关系网络设为目标关系网络。
  9. 如权利要求7所述的识别用户角色的方法,其特征在于,所述目标用户的个人信息包括各信息类型的子信息;
    所述将所述目标用户的个人信息与预设个人信息进行匹配,包括:
    将所述目标用户的个人信息中的子信息与预设个人信息中的子信息进行比较,并统计子信息相同的目标项数;
    当所述目标项数大于或等于预设标准项数时,认定所述目标用户的个人信息与所述预设个人信息匹配成功。
  10. 一种用户设备,其特征在于,所述用户设备包括:存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机可读指令,其中所述计算机可读指令被所述处理器执行时,实现如下步骤:
    获取目标用户的用户操作行为和个人信息;
    将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果;
    在预设分值评定模型下根据所述匹配结果生成对应的第一欺诈分值;
    在预设复杂网络模型下根据所述个人信息构建目标关系网络;
    通过所述目标关系网络确定与所述目标用户对应的关联用户;
    基于所述关联用户的用户角色在预设欺诈分映射关系中查询对应的第二欺诈分值,所述预设欺诈分映射关系中包括用户角色与欺诈分值的对应关系;
    通过所述第一欺诈分值与所述第二欺诈分值生成目标欺诈分值;
    在所述目标欺诈分值大于或等于预设欺诈阈值时,将所述目标用户的用户角色识别为潜在欺诈分子。
  11. 如权利要求10所述的用户设备,其特征在于,所述将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果的步骤之前,还包括:
    识别所述用户操作行为的行为类型;
    根据所述行为类型在预设行为模型中查询对应的预设操作行为;
    所述将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果,包括:
    确定与所述行为类型对应的行为特征类型;
    从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征;
    将所述当前行为特征与所述目标行为特征进行匹配,以获得匹配结果。
  12. 如权利要求11所述的用户设备,其特征在于,所述行为类型包括查询行为类型,所述行为特征类型中包括与所述查询行为类型对应的查询事项类型;
    所述根据所述行为类型在预设行为模型中查询对应的预设操作行为,包括:
    在所述行为类型为查询行为类型时,根据所述查询行为类型在预设行为模型中查询对应的预设操作行为;
    所述从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征,包括:
    在所述行为特征类型为查询事项类型时,从所述用户操作行为中提取与所述查询事项类型对应的当前行为特征,从所述预设操作行为中提取与所述查询事项类型对应的目标行为特征。
  13. 如权利要求12所述的用户设备,其特征在于,所述行为类型还包括输入行为类型,所述行为特征类型中包括与所述输入行为类型对应的输入时长类型、输入内容类型与输入次序类型中的至少一项;
    所述根据所述行为类型在预设行为模型中查询对应的预设操作行为,包括:
    在所述行为类型为输入行为类型时,根据所述输入行为类型在预设行为模型中查询对应的预设操作行为;
    所述从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征,包括:
    在所述行为特征类型为输入时长类型时,从所述用户操作行为中提取与所述输入时长类型对应的当前行为特征,从所述预设操作行为中提取与所述输入时长类型对应的目标行为特征。
  14. 如权利要求10所述的用户设备,其特征在于,所述在预设复杂网络模型下根据所述个人信息构建目标关系网络,包括:
    获取由预设个人信息生成的预设关系网络;
    在预设复杂网络模型下根据所述目标用户的个人信息创建与所述目标用户对应的目标节点;
    将所述目标节点添加至所述预设关系网络中,并将添加节点后的预设关系网络设为目标关系网络。
  15. 一种存储介质,其特征在于,所述存储介质上存储有计算机可读指令,所述计算机可读指令被处理器执行时,实现如下步骤:
    获取目标用户的用户操作行为和个人信息;
    将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果;
    在预设分值评定模型下根据所述匹配结果生成对应的第一欺诈分值;
    在预设复杂网络模型下根据所述个人信息构建目标关系网络;
    通过所述目标关系网络确定与所述目标用户对应的关联用户;
    基于所述关联用户的用户角色在预设欺诈分映射关系中查询对应的第二欺诈分值,所述预设欺诈分映射关系中包括用户角色与欺诈分值的对应关系;
    通过所述第一欺诈分值与所述第二欺诈分值生成目标欺诈分值;
    在所述目标欺诈分值大于或等于预设欺诈阈值时,将所述目标用户的用户角色识别为潜在欺诈分子。
  16. 如权利要求15所述的存储介质,其特征在于,所述将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果的步骤之前,还包括:
    识别所述用户操作行为的行为类型;
    根据所述行为类型在预设行为模型中查询对应的预设操作行为;
    所述将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果,包括:
    确定与所述行为类型对应的行为特征类型;
    从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征;
    将所述当前行为特征与所述目标行为特征进行匹配,以获得匹配结果。
  17. 如权利要求16所述的存储介质,其特征在于,所述行为类型包括查询行为类型,所述行为特征类型中包括与所述查询行为类型对应的查询事项类型;
    所述根据所述行为类型在预设行为模型中查询对应的预设操作行为,包括:
    在所述行为类型为查询行为类型时,根据所述查询行为类型在预设行为模型中查询对应的预设操作行为;
    所述从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征,包括:
    在所述行为特征类型为查询事项类型时,从所述用户操作行为中提取与所述查询事项类型对应的当前行为特征,从所述预设操作行为中提取与所述查询事项类型对应的目标行为特征。
  18. 如权利要求17所述的存储介质,其特征在于,所述行为类型还包括输入行为类型,所述行为特征类型中包括与所述输入行为类型对应的输入时长类型、输入内容类型与输入次序类型中的至少一项;
    所述根据所述行为类型在预设行为模型中查询对应的预设操作行为,包括:
    在所述行为类型为输入行为类型时,根据所述输入行为类型在预设行为模型中查询对应的预设操作行为;
    所述从所述用户操作行为中提取与所述行为特征类型对应的当前行为特征,从所述预设操作行为中提取与所述行为特征类型对应的目标行为特征,包括:
    在所述行为特征类型为输入时长类型时,从所述用户操作行为中提取与所述输入时长类型对应的当前行为特征,从所述预设操作行为中提取与所述输入时长类型对应的目标行为特征。
  19. 如权利要求15所述的存储介质,其特征在于,所述在预设复杂网络模型下根据所述个人信息构建目标关系网络,包括:
    获取由预设个人信息生成的预设关系网络;
    在预设复杂网络模型下根据所述目标用户的个人信息创建与所述目标用户对应的目标节点;
    将所述目标节点添加至所述预设关系网络中,并将添加节点后的预设关系网络设为目标关系网络。
  20. 一种识别用户角色的装置,其特征在于,所述识别用户角色的装置包括:
    信息获取模块,用于获取目标用户的用户操作行为和个人信息;
    行为匹配模块,用于将所述用户操作行为与预设行为模型中的预设操作行为进行匹配,以获得匹配结果;
    第一分值生成模块,用于在预设分值评定模型下根据所述匹配结果生成对应的第一欺诈分值;
    网络构建模块,用于在预设复杂网络模型下根据所述个人信息构建目标关系网络;
    用户确定模块,用于通过所述目标关系网络确定与所述目标用户对应的关联用户;
    第二分值生成模块,用于基于所述关联用户的用户角色在预设欺诈分映射关系中查询对应的第二欺诈分值,所述预设欺诈分映射关系中包括用户角色与欺诈分值的对应关系;
    目标分值生成模块,用于通过所述第一欺诈分值与所述第二欺诈分值生成目标欺诈分值;
    身份识别模块,用于在所述目标欺诈分值大于或等于预设欺诈阈值时,将所述目标用户的用户角色识别为潜在欺诈分子。
PCT/CN2018/122726 2018-09-26 2018-12-21 识别用户角色的方法、用户设备、存储介质及装置 Ceased WO2020062641A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811125685.7A CN109670933A (zh) 2018-09-26 2018-09-26 识别用户角色的方法、用户设备、存储介质及装置
CN201811125685.7 2018-09-26

Publications (1)

Publication Number Publication Date
WO2020062641A1 true WO2020062641A1 (zh) 2020-04-02

Family

ID=66142104

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/122726 Ceased WO2020062641A1 (zh) 2018-09-26 2018-12-21 识别用户角色的方法、用户设备、存储介质及装置

Country Status (2)

Country Link
CN (1) CN109670933A (zh)
WO (1) WO2020062641A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113807940A (zh) * 2020-06-17 2021-12-17 马上消费金融股份有限公司 信息处理和欺诈行为识别方法、装置、设备及存储介质
CN117216332A (zh) * 2023-09-14 2023-12-12 北京国信华源科技有限公司 一种用于监测设备的信息扩展呈现方法及装置

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110135804B (zh) * 2019-04-29 2024-03-29 深圳市元征科技股份有限公司 数据处理方法及装置
CN111369264A (zh) * 2020-02-13 2020-07-03 平安科技(深圳)有限公司 实体关联方法、装置、设备及计算机可读存储介质
CN111415241A (zh) * 2020-02-29 2020-07-14 深圳壹账通智能科技有限公司 欺诈人员识别方法、装置、设备和存储介质
CN113553369B (zh) * 2020-04-26 2024-06-25 清华大学 可视化用户分类方法、服务方法、系统、装置及存储介质
CN115564576A (zh) * 2022-11-08 2023-01-03 吉林亿联银行股份有限公司 业务请求的风险识别方法及装置、存储介质和处理器

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105956823A (zh) * 2016-04-28 2016-09-21 中国建设银行股份有限公司 一种业务数据办理系统
CN106875271A (zh) * 2017-02-17 2017-06-20 齐鲁工业大学 基于参考人关系网络上随机漫步的信用评价方法
CN108364224A (zh) * 2018-01-12 2018-08-03 深圳壹账通智能科技有限公司 信贷风险联控方法、装置、设备及可读存储介质

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8145560B2 (en) * 2006-11-14 2012-03-27 Fmr Llc Detecting fraudulent activity on a network
CN107644098A (zh) * 2017-09-29 2018-01-30 马上消费金融股份有限公司 一种欺诈行为识别方法、装置、设备及存储介质

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105956823A (zh) * 2016-04-28 2016-09-21 中国建设银行股份有限公司 一种业务数据办理系统
CN106875271A (zh) * 2017-02-17 2017-06-20 齐鲁工业大学 基于参考人关系网络上随机漫步的信用评价方法
CN108364224A (zh) * 2018-01-12 2018-08-03 深圳壹账通智能科技有限公司 信贷风险联控方法、装置、设备及可读存储介质

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113807940A (zh) * 2020-06-17 2021-12-17 马上消费金融股份有限公司 信息处理和欺诈行为识别方法、装置、设备及存储介质
CN113807940B (zh) * 2020-06-17 2024-04-12 马上消费金融股份有限公司 信息处理和欺诈行为识别方法、装置、设备及存储介质
CN117216332A (zh) * 2023-09-14 2023-12-12 北京国信华源科技有限公司 一种用于监测设备的信息扩展呈现方法及装置

Also Published As

Publication number Publication date
CN109670933A (zh) 2019-04-23

Similar Documents

Publication Publication Date Title
WO2020062641A1 (zh) 识别用户角色的方法、用户设备、存储介质及装置
WO2020078053A1 (zh) 医疗数据异常检测方法、装置、设备及存储介质
WO2021027143A1 (zh) 信息推送方法、装置、设备及计算机可读存储介质
WO2020119116A1 (zh) 基于数据分析的医保审核方法、装置、设备和存储介质
WO2020015067A1 (zh) 数据采集方法、装置、设备及存储介质
WO2020006852A1 (zh) 差旅费自助核销处理方法、装置、设备和计算机存储介质
WO2020015064A1 (zh) 系统故障处理方法、装置、设备及存储介质
WO2020258657A1 (zh) 异常检测方法、装置、计算机设备及存储介质
WO2020073495A1 (zh) 基于人工智能的复审方法、装置、设备及存储介质
WO2020078058A1 (zh) 医疗数据异常识别方法、装置、终端及存储介质
WO2020224250A1 (zh) 智能合约的触发方法、装置、设备及存储介质
WO2020119115A1 (zh) 数据审核方法、装置、设备及存储介质
WO2019144738A1 (zh) 金融业务的验证方法、装置、设备和计算机存储介质
WO2020147385A1 (zh) 数据录入方法、装置、终端及计算机可读存储介质
WO2020233060A1 (zh) 事件通知方法、事件通知服务器、存储介质及装置
WO2020119384A1 (zh) 基于大数据分析的医保异常检测方法、装置、设备和介质
WO2020087981A1 (zh) 风控审核模型生成方法、装置、设备及可读存储介质
WO2020119176A1 (zh) 报销数据的排查方法、识别服务端及存储介质
WO2020062658A1 (zh) 合同生成方法、装置、设备及存储介质
WO2018205373A1 (zh) 人伤理赔定损费用测算方法、装置、服务器和介质
WO2020233089A1 (zh) 测试用例生成方法、装置、终端及计算机可读存储介质
WO2020087704A1 (zh) 信贷信息管理方法、装置、设备和存储介质
WO2020108111A1 (zh) 医保欺诈行为的识别方法、装置、设备及可读存储介质
WO2020114221A1 (zh) 基于生物特征识别的定点医药机构的监管方法及相关设备
WO2020015060A1 (zh) 用电量异常评估方法、装置、设备和计算机存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18935612

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 08/07/2021)

122 Ep: pct application non-entry in european phase

Ref document number: 18935612

Country of ref document: EP

Kind code of ref document: A1