WO2020186789A1 - 用户反欺诈实现方法、装置、计算机设备及存储介质 - Google Patents

用户反欺诈实现方法、装置、计算机设备及存储介质 Download PDF

Info

Publication number
WO2020186789A1
WO2020186789A1 PCT/CN2019/118405 CN2019118405W WO2020186789A1 WO 2020186789 A1 WO2020186789 A1 WO 2020186789A1 CN 2019118405 W CN2019118405 W CN 2019118405W WO 2020186789 A1 WO2020186789 A1 WO 2020186789A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
identity
identification
information
fraud
Prior art date
Application number
PCT/CN2019/118405
Other languages
English (en)
French (fr)
Inventor
李晨光
Original Assignee
平安科技(深圳)有限公司
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 平安科技(深圳)有限公司 filed Critical 平安科技(深圳)有限公司
Publication of WO2020186789A1 publication Critical patent/WO2020186789A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR 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
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063Operations research, analysis or management
    • G06Q10/0639Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393Score-carding, benchmarking or key performance indicator [KPI] analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR 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

  • This application relates to the field of intelligent decision-making, and in particular to a method, device, computer equipment and storage medium for implementing user anti-fraud.
  • the embodiments of the application provide a user anti-fraud implementation method, device, computer equipment and storage medium, aiming to solve the problem that the credit analysis for users in the prior art is based on a single dimension, and there is no anti-fraud evaluation based on real-time information. , Leading to inaccurate assessment results.
  • an embodiment of the present application provides a user anti-fraud implementation method, which includes:
  • the identity dimension verification flag is set to the first identification value
  • the user fraud identification initial line vector is input to the pre-trained deep neural network model to obtain the user fraud probability value.
  • an embodiment of the present application provides a user anti-fraud implementation device, which includes:
  • the first identification acquiring unit is configured to receive the identity certification document in the uploaded user file, and if the identity certification document is verified by the identity online verification system, the identity dimension verification flag is set to the first identification value;
  • the second identification acquiring unit is configured to acquire credit investigation information corresponding to the identity certification document, acquire overdue information in the credit investigation information, and set the credit investigation credit dimension identification bit as the first according to the total number of overdue information Two identification value;
  • the third identification acquiring unit is configured to receive the user communication number in the uploaded user file, and if the user communication number is the same as the stored user registration communication number, set the social behavior dimension identification bit to the third identification value;
  • the fourth identification acquisition unit is configured to acquire information about the untrustworthiness of the high court according to the identity certificate, and set the high law untrustworthiness identifier to the fourth identification value according to the information about the untrustworthiness of the high court;
  • An initial row vector obtaining unit configured to concatenate the first identification value, the second identification value, the third identification value, and the fourth identification value to obtain the initial row vector for user fraud identification;
  • the probability value calculation unit is used to input the initial line vector of user fraud identification to the pre-trained deep neural network model to obtain the user fraud probability value.
  • an embodiment of the present application provides a computer device, which includes a memory, a processor, and a computer program stored on the memory and running on the processor, and the processor executes the computer
  • the program implements the user anti-fraud implementation method described in the first aspect.
  • the embodiments of the present application also provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor executes the above-mentioned On the one hand, the user anti-fraud implementation method.
  • FIG. 1 is a schematic diagram of an application scenario of a user anti-fraud implementation method provided by an embodiment of the application
  • FIG. 2 is a schematic flowchart of a user anti-fraud implementation method provided by an embodiment of the application
  • Fig. 3 is a schematic diagram of a sub-flow of a user anti-fraud implementation method provided by an embodiment of the application;
  • FIG. 4 is a schematic diagram of another sub-flow of the method for implementing user anti-fraud provided by an embodiment of the application;
  • FIG. 5 is a schematic diagram of another sub-flow of the method for implementing user anti-fraud provided by an embodiment of the application;
  • FIG. 6 is a schematic diagram of another sub-flow of a user anti-fraud implementation method provided by an embodiment of the application.
  • FIG. 7 is a schematic block diagram of a user anti-fraud implementation device provided by an embodiment of the application.
  • FIG. 8 is a schematic block diagram of subunits of a user anti-fraud implementation device provided by an embodiment of the application.
  • FIG. 9 is a schematic block diagram of another subunit of the user anti-fraud implementation device provided by an embodiment of the application.
  • FIG. 10 is a schematic block diagram of another sub-unit of the user anti-fraud implementation device provided by an embodiment of the application.
  • FIG. 11 is a schematic block diagram of another subunit of the user anti-fraud implementation device provided by an embodiment of the application.
  • FIG. 12 is a schematic block diagram of a computer device provided by an embodiment of the application.
  • Figure 1 is a schematic diagram of an application scenario of a user anti-fraud implementation method provided by an embodiment of the application
  • Figure 2 is a schematic flow diagram of a user anti-fraud implementation method provided by an embodiment of the application. The method is applied to a server, and the method is executed by application software installed in the server.
  • the method includes steps S110 to S160.
  • the server when the server needs to crawl the multi-dimensional user information of a specified user in real time, it needs to upload a user file through the uploader.
  • the user file at least includes the user's name, identity certification documents (such as a scanned photo of an ID card), and the user Communication number, etc.
  • the server receives the user file uploaded by the uploader, it obtains at least four dimensions of information, one is the identity dimension, the other is the credit information dimension, the third is the social behavior dimension, and the fourth is the high law untrustworthy dimension.
  • crawling the user's information in the above four dimensions through the crawler tool the latest information with high real-time nature is obtained, and the user can conduct anti-fraud analysis based on the latest information.
  • the server When obtaining the identity dimension information corresponding to the first dimension, the server needs to upload the user's identity certificate to the online identity verification system for verification, and the identity verification system is used to assist in verifying the authenticity of the user's identity dimension.
  • step S110 includes:
  • S111 Receive the identity certification file in the uploaded user file, and obtain the unique identification code and user name in the identity certification file through image recognition; S112. Upload the unique identification code and user name to the online identity verification system Perform verification; S113, if the verification pass notification message of the online identity verification system is received, set the identity dimension verification flag to the first identity value.
  • the server receives the identity certification file in the user file uploaded by the uploader, and obtains the unique identification code and the user name in the identity certification file through the deep neural network model identification stored in the server. After the unique identification code and the user name in the identification document are identified, the unique identification code and the user name are uploaded to the online identity verification system for verification, if the verification by the online identity verification system is received Through the notification information, the identity dimension verification identification position is set to the corresponding first identification value.
  • the online identity verification system is the identity verification system of the public security system.
  • identity document such as the front and back photos of the ID card
  • the identity verification system can verify the authenticity of the identity document.
  • the identity dimension verification identification position is set to the corresponding first identification value (for example, the first identification value is 1).
  • Mass user identity information data stored in the identity network verification system can assist in verifying the authenticity of user information in the identity dimension.
  • the server when the verification of the authenticity of the user information in the identity dimension is completed, the server then uploads the identity certification document to the credit inquiry system to obtain the credit information corresponding to the identity certification document. After analyzing the credit information, the credit information dimension flag is set as the second identification value according to the total number of overdue information.
  • step S120 includes:
  • S121 If the identity verification request of the credit inquiry system is detected, upload the identity certificate to the credit inquiry system; S122. Receive overdue information sent by the credit inquiry system according to the identity certificate corresponding to the document; S123, Acquire the total number of overdue pieces of the overdue information within the preset query time period; S124. Set the credit information dimension flag to the second identification value according to the total number of overdue pieces of the overdue information.
  • the credit inquiry system is the credit inquiry system of the People’s Bank of China.
  • the user By uploading the user’s identification document (such as the front and back of the ID photo), the user can obtain the specified inquiry time period (in the past 5 years). Overdue information.
  • the server automatically connects to the credit inquiry system and initiates a credit inquiry request, if the server detects the identity verification request of the credit inquiry system, it uploads the identity certification file to the credit inquiry system. After the identity certificate passes the identity verification of the credit inquiry system, the credit information corresponding to the identity certificate is obtained. Since the credit information includes the total number of overdue items of the user in the preset query time period, the credit information dimension flag can be set to the second according to the total number of overdue items of the overdue information. Identification value.
  • the massive user credit information data stored in the credit inquiry system can assist in verifying the creditworthiness of user information in the credit information dimension.
  • the server when the total number of overdue items of user information in the credit investigation dimension is completed, the server then uploads the identification document in the user file to the communication number query system (such as the server of a communication operator) to obtain The stored user registration communication number corresponding to the identification document. If the user communication number is the same as the stored user registration communication number, the social behavior dimension identification bit is set to the third identification value.
  • the communication number query system such as the server of a communication operator
  • step S130 includes:
  • the server uploads the identity certificate to the communication number query system according to the identity certificate in the uploaded user file; the communication number query system obtains the stored user registration communication number corresponding to the identity certificate; if the user registers If the number of communication numbers is greater than 1, it is determined whether there are multiple user registered communication numbers that are the same as the user communication number.
  • the social behavior dimension is identified The bit is set to the corresponding third identification value; specifically, when the social behavior dimension identification bit is set to the corresponding third identification value, the third identification value is 1; if it is determined that the user communication number is registered with the user If the communication numbers are not the same, set the social behavior dimension flag to 0.
  • the massive user communication number data stored in the communication number query system can assist in verifying the credibility of user information in the social behavior dimension.
  • the server After obtaining the credit of the user information in the social behavior dimension, the server then uploads the identity certification document to the high court's untrustworthy inquiry system to obtain the high court's untrustworthy information corresponding to the identity certification document. After analyzing the untrustworthiness information of the high court, the high law untrustworthiness dimension flag is set as the fourth identification value according to the total number of untrustworthiness.
  • step S140 includes:
  • S141 If the identity verification request of the untrustworthy inquiry system of the high court is detected, upload the identity certificate to the credit inquiry system; S142. Receive the untrustworthy high court sent by the untrustworthy inquiry system corresponding to the identity document Information; S143. Obtain the total number of untrustworthy items in the high court's untrustworthy information within a preset query time period; S144. Set the high law untrustworthy dimension flag to be based on the total number of untrustworthy items of the high court untrustworthy information The fourth identification value.
  • the high court's untrustworthy inquiry system is the people's court's untrustworthy inquiry system (commonly known as Laolai inquiry).
  • the user's identification document such as the front and back of the ID photo
  • the server By uploading the user's identification document (such as the front and back of the ID photo), the user can obtain the specified inquiry time Section (in the past 1 year) of untrustworthy information.
  • the server automatically connects to the untrustworthy inquiry system of the high court and initiates a credit inquiry request, if the server detects the identity verification request of the untrustworthy inquiry system of the high court, it uploads the identity certification document to the untrustworthy inquiry system of the high court. After the identity certificate passes the identity verification of the high court's untrustworthy inquiry system, the high court's untrustworthy information corresponding to the identity certificate is obtained.
  • the high-law untrustworthy dimension flag can be set to the fourth identification value according to the total number of untrustworthy items .
  • the massive user high court untrustworthy information data stored in the high court untrustworthy inquiry system can assist in verifying the credibility of user information in the high law untrustworthy dimension.
  • the first identification value, the second identification value, the third identification value, and the fourth identification value are concatenated to obtain an initial row vector for user fraud identification.
  • the first identification value, the second identification value, the third identification value, and the fourth identification value are acquired and constitute the initial row vector for user fraud identification, such as [1 0.5 1 1], Input [1 0.5 1 1] into the pre-selected training deep neural network model.
  • user fraud identification such as [1 0.5 1 1]
  • Input [1 0.5 1 1] into the pre-selected training deep neural network model.
  • the anti-fraud credit risk evaluation index is automatically obtained.
  • the method before step S110, the method further includes:
  • the to-be-trained deep neural network model is trained to obtain a deep neural network model for predicting the probability of user fraud.
  • the input value in each training data in the training set is the initial row vector of historical user fraud recognition, such as [1 0.5 1], [0 0.3 1 0], [1 1 1], [1 0.5 1], [1 0 0 1], etc., and the corresponding output values are 0.9, 0.4, 06, 0.5, etc. respectively.
  • the training can be carried out.
  • DNN Deep Neural Networks
  • the neural network layers inside Deep Neural Networks can be divided into three categories, input layer, hidden layer and output layer, as shown in the example below. Generally speaking, the first layer is the input layer and the last layer is the output. Layers, and the number of layers in the middle are all hidden layers.
  • any neuron in the i-th layer must be connected to any neuron in the i+1th layer.
  • the forward propagation algorithm of DNN is to use a number of weight coefficient matrices W and bias vector b to perform a series of linear operations and activation operations with the input value vector x. Starting from the input layer, the backward calculation layer by layer, until the calculation Go to the output layer and get the output result as a value.
  • the final result is the output aL.
  • Training the deep neural network model means that the input value vector x and the output a L are known , and the matrix W corresponding to all hidden layers and output layers is trained correspondingly, and the bias vector b is enough.
  • step S160 the method further includes:
  • the user file corresponding to the user fraud probability value is automatically added to set the identity of the suspicious user, and the user file corresponding to the user fraud probability value is stored in the suspicious The user's grouping area.
  • the user fraud probability value of each user can be compared with the preset first probability threshold. If one of the users has a user fraud probability value greater than the first probability threshold, The probability threshold indicates that the user may be a suspicious user with fraudulent behavior. At this time, the user files corresponding to the user's fraud probability value greater than the first probability threshold can be stored in the grouping area of the suspicious user, thereby screening multiple fraudulent behaviors Suspicious users. Among them, the grouping area of the suspicious user is set in advance in the storage space in the server.
  • the user file corresponding to the user fraud probability value is automatically added to set the identity of the suspicious user, and the user fraud probability is set After the user file corresponding to the value is stored in the grouping area of the suspicious user, it also includes:
  • the user information corresponding to each user file in the grouping area of the suspicious user wherein the user information includes the unique identification code and the user name; if there is a kinship relationship between user information in the grouping area of the suspicious user, there will be The user information of the relative relationship is stored in the grouping area of the suspicious associated user.
  • the obtaining process is to receive the user information in the uploaded user file.
  • Identity certification documents through image recognition to obtain the unique identification code and user name in the identity certification documents to obtain user information
  • the user's relative relationship graph stored in the server can be obtained (the user's relative relationship graph represents the user Whether there is a kinship between them), so as to analyze the suspicious users in the grouping area of suspicious users, there are also family group frauds.
  • extract the user information with kinship relationships and then store it in the group of suspicious associated users Area, that is, the suspicious users who have been scammed by family gangs have been deeply dug out.
  • This method realizes real-time acquisition of multiple dimensions of user information for anti-fraud analysis, and improves the real-time and accuracy of anti-fraud analysis.
  • the embodiment of the present application also provides a user anti-fraud realization device, which is used to implement any embodiment of the aforementioned user anti-fraud realization method.
  • FIG. 7 is a schematic block diagram of a user anti-fraud implementation device provided by an embodiment of the present application.
  • the user anti-fraud implementation device 100 may be configured in a server.
  • the user anti-fraud implementation device 100 includes a first identification acquisition unit 110, a second identification acquisition unit 120, a third identification acquisition unit 130, a fourth identification acquisition unit 140, an initial row vector acquisition unit 150, and a probability value Calculating unit 160.
  • the first identification acquiring unit 110 is configured to receive the identity certification document in the uploaded user file, and if the identity certification document is verified by the identity online verification system, the identity dimension verification flag is set to the first identification value.
  • the server when the server needs to crawl the multi-dimensional user information of a specified user in real time, it needs to upload a user file through the uploader.
  • the user file at least includes the user's name, identity certification documents (such as a scanned photo of an ID card), and the user Communication number, etc.
  • the server receives the user file uploaded by the uploader, it obtains at least four dimensions of information, one is the identity dimension, the other is the credit information dimension, the third is the social behavior dimension, and the fourth is the high law untrustworthy dimension.
  • crawling the user's information in the above four dimensions through the crawler tool the latest information with high real-time nature is obtained, and the user can conduct anti-fraud analysis based on the latest information.
  • the server When obtaining the identity dimension information corresponding to the first dimension, the server needs to upload the user's identity certificate to the online identity verification system for verification, and the identity verification system is used to assist in verifying the authenticity of the user's identity dimension.
  • the first identification obtaining unit 110 includes:
  • the identity image recognition unit 111 is configured to receive the identity certification document in the uploaded user file, and obtain the unique identification code and the user name in the identity certification document through image recognition; the identity verification unit 112 is configured to uniquely identify the identity The identification code and user name are uploaded to the online identity verification system for verification; the first setting unit 113 is configured to set the identity dimension verification flag to the first identity if the verification pass notification message of the identity online verification system is received value.
  • the server receives the identity certification file in the user file uploaded by the uploader, and obtains the unique identification code and the user name in the identity certification file through the deep neural network model identification stored in the server. After the unique identification code and the user name in the identification document are identified, the unique identification code and the user name are uploaded to the online identity verification system for verification, if the verification by the online identity verification system is received Through the notification information, the identity dimension verification identification position is set to the corresponding first identification value.
  • the online identity verification system is the identity verification system of the public security system.
  • identity document such as the front and back photos of the ID card
  • the identity verification system can verify the authenticity of the identity document.
  • the identity dimension verification identification position is set to the corresponding first identification value (for example, the first identification value is 1).
  • Mass user identity information data stored in the identity network verification system can assist in verifying the authenticity of user information in the identity dimension.
  • the second identification acquiring unit 120 is configured to acquire credit information corresponding to the identity certification document, acquire overdue information in the credit information, and set the credit dimension identification bit according to the total number of overdue information to The second identification value.
  • the server when the verification of the authenticity of the user information in the identity dimension is completed, the server then uploads the identity certification document to the credit inquiry system to obtain the credit information corresponding to the identity certification document. After analyzing the credit information, the credit information dimension flag is set as the second identification value according to the total number of overdue information.
  • the second identification acquiring unit 120 includes:
  • the first uploading unit 121 is configured to upload the identity certification document to the credit inquiry system if the identity verification request of the credit inquiry system is detected; the overdue information acquisition unit 122 is configured to receive the credit inquiry system according to the The overdue information corresponding to the identity certification document is sent; the overdue information analysis unit 123 is used to obtain the total number of overdue items in the preset query time period in the overdue information; the second setting unit 124 is used to The total number of overdue pieces of information is set to the second identification value.
  • the credit inquiry system is the credit inquiry system of the People’s Bank of China.
  • the user By uploading the user’s identification document (such as the front and back of the ID photo), the user can obtain the specified inquiry time period (in the past 5 years). Overdue information.
  • the server automatically connects to the credit inquiry system and initiates a credit inquiry request, if the server detects the identity verification request of the credit inquiry system, it uploads the identity certification file to the credit inquiry system. After the identity certificate passes the identity verification of the credit inquiry system, the credit information corresponding to the identity certificate is obtained. Since the credit information includes the total number of overdue items of the user in the preset query time period, the credit information dimension flag can be set to the second according to the total number of overdue items of the overdue information. Identification value.
  • the massive user credit information data stored in the credit inquiry system can assist in verifying the creditworthiness of user information in the credit information dimension.
  • the third identification acquisition unit 130 is configured to receive the user communication number in the uploaded user file, and if the user communication number is the same as the stored user registration communication number, set the social behavior dimension identification bit to the third identification value.
  • the server when the total number of overdue items of user information in the credit investigation dimension is completed, the server then uploads the identification document in the user file to the communication number query system (such as the server of a communication operator) to obtain The stored user registration communication number corresponding to the identification document. If the user communication number is the same as the stored user registration communication number, the social behavior dimension identification bit is set to the third identification value.
  • the communication number query system such as the server of a communication operator
  • the third identification acquiring unit 130 includes:
  • the second uploading unit 131 is configured to upload the identity certification document to the communication number query system according to the identity certification document in the uploaded user file;
  • the registration number obtaining unit 132 is configured to receive information obtained in the communication number query system
  • the third setting unit 133 is used for if the number of stored user registration communication numbers is greater than 1 or equal to 1, and the stored user registration communication numbers There is a user communication number in the uploaded user file, and the social behavior dimension identification bit is set to the third identification value;
  • the fourth setting unit 134 is used to set the number of user registration communication numbers stored in the file is equal to 0.
  • the social behavior dimension flag is set to 0.
  • the server uploads the identity certificate to the communication number query system according to the identity certificate in the uploaded user file; the communication number query system obtains the stored user registration communication number corresponding to the identity certificate; if the user registers If the number of communication numbers is greater than 1, it is determined whether there are multiple user registered communication numbers that are the same as the user communication number.
  • the social behavior dimension is identified The bit is set to the corresponding third identification value; specifically, when the social behavior dimension identification bit is set to the corresponding third identification value, the third identification value is 1; if it is determined that the user communication number is registered with the user If the communication numbers are not the same, set the social behavior dimension flag to 0.
  • the massive user communication number data stored in the communication number query system can assist in verifying the credibility of user information in the social behavior dimension.
  • the fourth identification obtaining unit 140 is configured to obtain information about the untrustworthiness of the high court according to the identity certificate, and set the high law untrustworthiness dimension flag to the fourth identification value according to the information about the untrustworthiness of the high court.
  • the server After obtaining the credit of the user information in the social behavior dimension, the server then uploads the identity certification document to the high court's untrustworthy inquiry system to obtain the high court's untrustworthy information corresponding to the identity certification document. After analyzing the untrustworthiness information of the high court, the high law untrustworthiness dimension flag is set as the fourth identification value according to the total number of untrustworthiness.
  • the fourth identification acquiring unit 140 includes:
  • the third uploading unit 141 is configured to upload the identity certification file to the credit inquiry system if the identity verification request of the high court untrustworthy inquiry system is detected; the untrustworthy information acquisition unit 142 is configured to receive the untrustworthy inquiry system of the high court The high court untrustworthy information correspondingly sent according to the identity certificate; the total untrustworthy information acquisition unit 143, configured to obtain the total untrustworthy information in the high court untrustworthy information within a preset query time period; fifth setting unit 144. It is used to set the high law untrustworthiness dimension flag to a fourth identification value according to the total number of untrustworthy information of the high court.
  • the high court's untrustworthy inquiry system is the people's court's untrustworthy inquiry system (commonly known as Laolai inquiry).
  • the user's identification document such as the front and back of the ID photo
  • the server By uploading the user's identification document (such as the front and back of the ID photo), the user can obtain the specified inquiry time Section (in the past 1 year) of untrustworthy information.
  • the server automatically connects to the untrustworthy inquiry system of the high court and initiates a credit inquiry request, if the server detects the identity verification request of the untrustworthy inquiry system of the high court, it uploads the identity certification document to the untrustworthy inquiry system of the high court. After the identity certificate passes the identity verification of the high court's untrustworthy inquiry system, the high court's untrustworthy information corresponding to the identity certificate is obtained.
  • the high-law untrustworthy dimension flag can be set to the fourth identification value according to the total number of untrustworthy items .
  • the massive user high court untrustworthy information data stored in the high court untrustworthy inquiry system can assist in verifying the credibility of user information in the high law untrustworthy dimension.
  • the initial row vector obtaining unit 150 is configured to concatenate the first identification value, the second identification value, the third identification value, and the fourth identification value to obtain an initial row vector for user fraud identification.
  • the first identification value, the second identification value, the third identification value, and the fourth identification value are acquired and constitute the initial row vector for user fraud identification, such as [1 0.5 1 1], Input [1 0.5 1 1] into the pre-selected training deep neural network model.
  • user fraud identification such as [1 0.5 1 1]
  • Input [1 0.5 1 1] into the pre-selected training deep neural network model.
  • the probability value calculation unit 160 is configured to input the initial line vector for user fraud identification into a pre-trained deep neural network model to obtain a user fraud probability value.
  • the anti-fraud credit risk evaluation index is automatically obtained.
  • the user anti-fraud implementation device 100 further includes:
  • the model training unit is used to obtain the marked historical data as a training set, input the initial row vector of historical user fraud recognition of each piece of training data in the training set to the deep neural network model to be trained, and train each piece of the training set
  • the user fraud probability value of the data is used as the output of the to-be-trained deep neural network model, and the to-be-trained deep neural network model is trained to obtain a deep neural network model for predicting the user’s fraud probability.
  • the input value in each training data in the training set is the initial row vector of historical user fraud recognition, such as [1 0.5 1], [0 0.3 1 0], [1 1 1], [1 0.5 1 1], [1 0 0 1], etc., and the corresponding output values are 0.9, 0.4, 06, 0.5, etc. respectively.
  • the training can be carried out.
  • the device achieves real-time acquisition of multiple dimensions of user information for anti-fraud analysis, and improves the real-time and accuracy of anti-fraud analysis.
  • the foregoing user anti-fraud implementation device may be implemented in the form of a computer program, and the computer program may run on a computer device as shown in FIG. 12.
  • FIG. 12 is a schematic block diagram of a computer device according to an embodiment of the present application.
  • the computer device 500 is a server, and the server may be an independent server or a server cluster composed of multiple servers.
  • the computer device 500 includes a processor 502, a memory, and a network interface 505 connected through a system bus 501, where the memory may include a non-volatile storage medium 503 and an internal memory 504.
  • the non-volatile storage medium 503 can store an operating system 5031 and a computer program 5032.
  • the processor 502 can execute the user anti-fraud implementation method.
  • the processor 502 is used to provide computing and control capabilities, and support the operation of the entire computer device 500.
  • the internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503.
  • the processor 502 can execute the user anti-fraud implementation method.
  • the network interface 505 is used for network communication, such as providing data information transmission.
  • the structure shown in FIG. 12 is only a block diagram of part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device 500 to which the solution of the present application is applied.
  • the specific computer device 500 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.
  • the processor 502 is configured to run a computer program 5032 stored in a memory to implement the user anti-fraud implementation method in the embodiment of the present application.
  • the embodiment of the computer device shown in FIG. 12 does not constitute a limitation on the specific configuration of the computer device.
  • the computer device may include more or fewer components than shown in the figure. Or combine certain components, or different component arrangements.
  • the computer device may only include a memory and a processor. In such an embodiment, the structures and functions of the memory and the processor are consistent with the embodiment shown in FIG. 12, and will not be repeated here.
  • the processor 502 may be a central processing unit (Central Processing Unit, CPU), and the processor 502 may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSPs), Application Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor.
  • a computer-readable storage medium may be a non-volatile computer-readable storage medium.
  • the computer-readable storage medium stores a computer program, where the computer program is executed by a processor to implement the user anti-fraud implementation method of the embodiment of the present application.
  • the storage medium is a physical, non-transitory storage medium, such as a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a magnetic disk, or an optical disk that can store program codes. medium.
  • a physical, non-transitory storage medium such as a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a magnetic disk, or an optical disk that can store program codes. medium.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Finance (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Educational Administration (AREA)
  • Quality & Reliability (AREA)
  • Data Mining & Analysis (AREA)
  • Operations Research (AREA)
  • Technology Law (AREA)
  • Tourism & Hospitality (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Game Theory and Decision Science (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Collating Specific Patterns (AREA)

Abstract

本申请公开了用户反欺诈实现方法、装置、计算机设备及存储介质。该方法包括:接收所上传用户文件中的身份证明文件,若身份证明文件通过身份验证,将身份维度验证标识位设置为第一标识值;根据身份证明文件对应获取征信信息,以将征信信用维度标识位设置为第二标识值;若用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值;根据身份证明文件获取高级法院失信信息,以将高法失信维度标识位设置为第四标识值;将第一标识值至第四标识值进行串接得到用户欺诈识别初始行向量,将其输入至深度神经网络模型得到用户欺诈概率值。

Description

用户反欺诈实现方法、装置、计算机设备及存储介质
本申请要求于2019年3月15日提交中国专利局、申请号为201910197847.6、申请名称为“用户反欺诈实现方法、装置、计算机设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及智能决策领域,尤其涉及一种用户反欺诈实现方法、装置、计算机设备及存储介质。
背景技术
目前,针对用户进行信用分析时,往往是停留在单一方面,例如获取用户的金融逾期信息、个人犯罪信息等,并不能从多个角度全方位的反映客户失信风险。而且,现有的反欺诈技术往往都是固定算法,没有基于实时的信息去进行反欺诈评估,也即无法实时的从互联网中获取与用户相关的多维信息以进行反欺诈评估,从而导致评估结果不准确。
发明内容
本申请实施例提供了一种用户反欺诈实现方法、装置、计算机设备及存储介质,旨在解决现有技术中针对用户进行信用分析时是基于单维度,而且没有基于实时信息去进行反欺诈评估,导致评估结果不准确的问题。
第一方面,本申请实施例提供了一种用户反欺诈实现方法,其包括:
接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值;
根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的总条数设置为第二标识值;
接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值;
根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值;
将所述第一标识值、第二标识值、第三标识值、第四标识值进行串接,得到用户欺诈识别初始行向量;以及
将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值。
第二方面,本申请实施例提供了一种用户反欺诈实现装置,其包括:
第一标识获取单元,用于接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值;
第二标识获取单元,用于根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的总条数设置为第二标识值;
第三标识获取单元,用于接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值;
第四标识获取单元,用于根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值;
初始行向量获取单元,用于将所述第一标识值、第二标识值、第三标识值、第四标识值进行串接,得到用户欺诈识别初始行向量;以及
概率值计算单元,用于将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值。
第三方面,本申请实施例又提供了一种计算机设备,其包括存储器、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程序,所述处理器执行所述计算机程序时实现上述第一方面所述的用户反欺诈实现方法。
第四方面,本申请实施例还提供了一种计算机可读存储介质,其中所述计算机可读存储介质存储有计算机程序,所述计算机程序当被处理器执行时使所述处理器执行上述第一方面所述的用户反欺诈实现方法。
附图说明
为了更清楚地说明本申请实施例技术方案,下面将对实施例描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图是本申请的一些实 施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的用户反欺诈实现方法的应用场景示意图;
图2为本申请实施例提供的用户反欺诈实现方法的流程示意图;
图3为本申请实施例提供的用户反欺诈实现方法的子流程示意图;
图4为本申请实施例提供的用户反欺诈实现方法的另一子流程示意图;
图5为本申请实施例提供的用户反欺诈实现方法的另一子流程示意图;
图6为本申请实施例提供的用户反欺诈实现方法的另一子流程示意图;
图7为本申请实施例提供的用户反欺诈实现装置的示意性框图;
图8为本申请实施例提供的用户反欺诈实现装置的子单元示意性框图;
图9为本申请实施例提供的用户反欺诈实现装置的另一子单元示意性框图;
图10为本申请实施例提供的用户反欺诈实现装置的另一子单元示意性框图;
图11为本申请实施例提供的用户反欺诈实现装置的另一子单元示意性框图;
图12为本申请实施例提供的计算机设备的示意性框图。
具体实施方式
下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例是本申请一部分实施例,而不是全部的实施例。基于本申请中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
应当理解,当在本说明书和所附权利要求书中使用时,术语“包括”和“包含”指示所描述特征、整体、步骤、操作、元素和/或组件的存在,但并不排除一个或多个其它特征、整体、步骤、操作、元素、组件和/或其集合的存在或添加。
还应当理解,在此本申请说明书中所使用的术语仅仅是出于描述特定实施例的目的而并不意在限制本申请。如在本申请说明书和所附权利要求书中所使用的那样,除非上下文清楚地指明其它情况,否则单数形式的“一”、“一个”及“该”意在包括复数形式。
还应当进一步理解,在本申请说明书和所附权利要求书中使用的术语“和/或”是指相关联列出的项中的一个或多个的任何组合以及所有可能组合,并且包括这些组合。
请参阅图1和图2,图1为本申请实施例提供的用户反欺诈实现方法的应用场景示意图,图2为本申请实施例提供的用户反欺诈实现方法的流程示意图,该用户反欺诈实现方法应用于服务器中,该方法通过安装于服务器中的应用软件进行执行。
如图2所示,该方法包括步骤S110~S160。
S110、接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值。
在本实施例中,当服务器需实时爬取指定用户多维度的用户信息时,需通过上传端上传用户文件,该用户文件中至少包括用户姓名、身份证明文件(如身份证扫描照片)、用户通讯号码等。当服务器接收了上传端所上传的用户文件后,至少获取四个维度的信息,一是身份维度、二是征信信用维度、三是社会行为维度、四是高法失信维度。通过爬虫工具爬取用户在上述四个维度的信息时,获取的是实时性较高的最新信息,能基于最新信息对用户进行反欺诈分析。
在获取第一个维度对应的身份维度信息时,需服务器将用户的身份证明文件上传至身份联网核查系统进行验证,通过身份联网核查系统来辅助核实用户的身份维度的真实性。
在一实施例中,如图3所示,步骤S110包括:
S111、接收所上传用户文件中的身份证明文件,通过图像识别获取所述身份证明文件中的身份唯一识别码和用户姓名;S112、将所述身份唯一识别码和用户姓名上传至身份联网核查系统进行验证;S113、若接收到所述身份联网核查系统的验证通过通知信息,将所述身份维度验证标识位设置为第一标识值。
即服务器接收上传端所上传的用户文件中身份证明文件,通过服务器中存储的深度神经网络模型识别获取所述身份证明文件中的身份唯一识别码和用户姓名。在识别得到了所述身份证明文件中的身份唯一识别码和用户姓名后,将所述身份唯一识别码和用户姓名上传至身份联网核查系统进行验证,若接收到所述身份联网核查系统的验证通过通知信息,将身份维度验证标识位置为对应的第一标识值。
身份联网核查系统即是公安系统的身份验证系统,当上传了用户的身份证明文件(如身份证的正反面照片)时,身份联网核查系统能对该身份证明文件 的真实性进行验证,一旦通过验证,则将身份维度验证标识位置为对应的第一标识值(例如第一标识值为1)。通过身份联网核查系统中存储的海量用户身份信息数据,能辅助验证用户信息在身份维度的真实性。
S120、根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的总条数设置为第二标识值。
即当完成了用户信息在身份维度的真实性的验证时,此时服务器再将所述身份证明文件上传至征信查询系统,以获取与所述身份证明文件对应的征信信息。通过对征信信息进行解析后,将征信信用维度标识位根据所述逾期信息的总条数设置为第二标识值。
在一实施例中,如图4所示,步骤S120包括:
S121、若检测到征信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统;S122、接收所述征信查询系统根据所述身份证明文件对应发送的逾期信息;S123、获取所述逾期信息中在预设的查询时间段内的逾期总条数;S124、根据所述逾期信息的逾期总条数,将征信信用维度标识位设置为所述第二标识值。
在本实施例中,征信查询系统是人民银行的征信查询系统,通过上传用户的身份证明文件(如身份证照片的正反面),即可获取用户在指定查询时间段(近5年内)的逾期信息。当服务器自动连接征信查询系统并发起征信查询请求时,若服务器检测到征信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统。所述身份证明文件通过征信查询系统的身份验证后,获取与所述身份证明文件对应的征信信息。由于征信信息中包括了该用户在在预设的查询时间段内的逾期总条数,故可根据所述逾期信息的逾期总条数,将征信信用维度标识位设置为所述第二标识值。通过征信查询系统中存储的海量用户征信信息数据,能辅助验证用户信息在征信信用维度的信用度。
S130、接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值。
即当完成了用户信息在征信信用维度的逾期总条数查询时,此时服务器再将所述用户文件中的身份证明文件上传至通讯号码查询系统(如通讯运营商的服务器),以获取与所述身份证明文件对应的已存储的用户登记通讯号码。若 所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值。
在一实施例中,如图5所示,步骤S130包括:
S131、根据所上传用户文件中的身份证明文件,将所述身份证明文件上传至通讯号码查询系统;S132、接收在所述通讯号码查询系统中获取的与所述身份证明文件对应的已存储的用户登记通讯号码;S133、若已存储的用户登记通讯号码的个数大于1或等于1、且已存储的用户登记通讯号码存在所上传用户文件中的用户通讯号码,将所述社会行为维度标识位设置为第三标识值;S134、若已存储的用户登记通讯号码的个数等于0,将所述社会行为维度标识位设置为0。
即服务器根据所上传用户文件中的身份证明文件,将所述身份证明文件上传至通讯号码查询系统;在通讯号码查询系统根据所述身份证明文件对应获取已存储的用户登记通讯号码;若用户登记通讯号码的个数大于1,判断多个用户登记通讯号码中是否存在与所述用户通讯号码相同的号码,若多个用户登记通讯号码中存在与所述用户通讯号码相同的号码,判定所述用户通讯号码与所述用户登记通讯号码相同;若用户登记通讯号码的个数等于1且所述用户通讯号码与所述用户登记通讯号码相同,判定所述用户通讯号码与所述用户登记通讯号码相同;若用户登记通讯号码的个数小于1,判定所述用户通讯号码与所述用户登记通讯号码不相同;若判定所述用户通讯号码与所述用户登记通讯号码相同,将社会行为维度标识位设置为对应的第三标识值;具体的,将社会行为维度标识位设置为对应的第三标识值时,所述第三标识值为1;若判定所述用户通讯号码与所述用户登记通讯号码不相同,将社会行为维度标识位设置为0。
通过通讯号码查询系统中存储的海量用户通讯号码数据,能辅助验证用户信息在社会行为维度的信用度。
S140、根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值。
即当完成了获取用户信息在社会行为维度的信用度后,此时服务器再将所述身份证明文件上传至高级法院失信查询系统,以获取与所述身份证明文件对应的高级法院失信信息。通过对高级法院失信信息进行解析后,将高法失信维度标识位根据失信总条数设置为第四标识值。
在一实施例中,如图6所示,步骤S140包括:
S141、若检测到高级法院失信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统;S142、接收所述高级法院失信查询系统根据所述身份证明文件对应发送的高级法院失信信息;S143、获取所述高级法院失信信息中在预设的查询时间段内的失信总条数;S144、根据所述高级法院失信信息的失信总条数,将高法失信维度标识位设置为第四标识值。
在本实施例中,高级法院失信查询系统是人民法院的失信查询系统(俗称老赖查询),通过上传用户的身份证明文件(如身份证照片的正反面),即可获取用户在指定查询时间段(近1年内)的失信信息。当服务器自动连接高级法院失信查询系统并发起征信查询请求时,若服务器检测到高级法院失信查询系统的身份验证请求,将所述身份证明文件上传至高级法院失信查询系统。所述身份证明文件通过高级法院失信查询系统的身份验证后,获取与所述身份证明文件对应的高级法院失信信息。由于高级法院失信信息中包括了该用户在在预设的查询时间段内的失信总条数,故可根据所述失信总条数,将高法失信维度标识位设置为所述第四标识值。通过高级法院失信查询系统中存储的海量用户高级法院失信信息数据,能辅助验证用户信息在高法失信维度的信用度。
S150、将所述第一标识值、第二标识值、第三标识值、第四标识值进行串接,得到用户欺诈识别初始行向量。
在本实施例中,将当识别获取了第一标识值、第二标识值、第三标识值、第四标识值并组成用户欺诈识别初始行向量后,如[1 0.5 1 1],此时将[1 0.5 1 1]输入至预选训练的深度神经网络模型。通过上述方式,将用户在四个维度的信息进行了量化,便于对用户进行反欺诈分析。
S160、将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值。
在本实施例中,当将通过深度神经网络模型有效的判断了用户欺诈概率值,自动得出反欺诈信用风险评估指标。
在一实施例中,步骤S110之前还包括:
获取已标注的历史数据作为训练集,将所述训练集中每一条训练数据的历史用户欺诈识别初始行向量输入至待训练深度神经网络模型,将所述训练集中每一条训练数据的用户欺诈概率值作为待训练深度神经网络模型的输出,对所 述待训练深度神经网络模型进行训练,得到用于预测用户欺诈概率的深度神经网络模型。
在预先训练该用于识别用户欺诈概率值的深度神经网络,需要大量已标注的历史数据作为训练集,该训练集中每一条训练数据中输入值是历史用户欺诈识别初始行向量,如[1 0.5 1 1]、[0 0.3 1 0]、[1 1 1 1]、[1 0.5 1 1]、[1 0 0 1]等,与此分别对应的输出值为0.9、0.4、06、0.5等。当已知了待训练的深度神经网络的输入和输出之后,即可进行训练。
深度神经网络(Deep Neural Networks,简称DNN)内部的神经网络层可以分为三类,输入层,隐藏层和输出层,如下图示例,一般来说第一层是输入层,最后一层是输出层,而中间的层数都是隐藏层。
层与层之间是全连接的,也就是说,第i层的任意一个神经元一定与第i+1层的任意一个神经元相连。虽然DNN看起来很复杂,但是从小的局部模型来说,还是和感知机一样,即一个线性关系z=∑wixi+b加上一个激活函数σ(z)。
DNN的前向传播算法也就是利用若干个权重系数矩阵W,偏倚向量b来和输入值向量x进行一系列线性运算和激活运算,从输入层开始,一层层的向后计算,一直到运算到输出层,得到输出结果为值。
即输入:总层数L,所有隐藏层和输出层对应的矩阵W,偏倚向量b,输入值向量x;输出:输出层的输出a L。具体如下:
1)初始化a 1=x
2)for l=2 to L,计算:
a l=σ(z l)=σ(W la l-1+b l)
最后的结果即为输出aL。
训练深度神经网络模型即已知输入值向量x和输出a L,对应训练出所有隐藏层和输出层对应的矩阵W,偏倚向量b即可。
在一实施例中,步骤S160之后还包括:
若所述用户欺诈概率值大于预先设置的第一概率阈值,将所述用户欺诈概率值对应的用户文件自动增加设置可疑用户的标识,并将所述用户欺诈概率值对应的用户文件存储至可疑用户的分组区域。
在本实施例中,当获取了各用户的用户欺诈概率值后,可以各用户的用户欺诈概率值均与预先设置的第一概率阈值进行比较,若其中有用户的用户欺诈 概率值大于第一概率阈值,则表示该用户可能是存在欺诈行为的可疑用户,此时可以将有用户欺诈概率值大于第一概率阈值对应的用户文件存储至可疑用户的分组区域,从而筛选出多个存在欺诈行为的可疑用户。其中,可疑用户的分组区域是在服务器中的存储空间中预先完成设置的。
在一实施例中,所述若所述用户欺诈概率值大于预先设置的第一概率阈值,将所述用户欺诈概率值对应的用户文件自动增加设置可疑用户的标识,并将所述用户欺诈概率值对应的用户文件存储至可疑用户的分组区域之后,还包括:
获取可疑用户的分组区域中各用户文件对应的用户信息;其中,所述用户信息包括身份唯一识别码和用户姓名;若可疑用户的分组区域中存在有用户信息之间存在亲属关联关系,将存在亲属关联关系的用户信息存储至可疑关联用户的分组区域。
在本实施例中,当获取了可疑用户的分组区域中多个可疑用户的用户信息后(其中,所述用户信息包括身份唯一识别码和用户姓名,其获取过程是接收所上传用户文件中的身份证明文件,通过图像识别获取所述身份证明文件中的身份唯一识别码和用户姓名,以得到用户信息),可以获取服务器中已存储的用户亲属关联关系图谱(该用户亲属关联关系图谱表示用户之间是否存在亲属关系),从而分析出在可疑用户的分组区域多个可疑用户中还存在家庭团伙欺诈的行为,此时提取存在亲属关联关系的用户信息,之后再存储至可疑关联用户的分组区域,也即深度挖掘出了家庭团伙诈骗的可疑用户。
该方法实现了实时获取多个维度的用户信息进行反欺诈分析,提高了反欺诈分析的实时性和准确性。
本申请实施例还提供一种用户反欺诈实现装置,该用户反欺诈实现装置用于执行前述用户反欺诈实现方法的任一实施例。具体地,请参阅图7,图7是本申请实施例提供的用户反欺诈实现装置的示意性框图。该用户反欺诈实现装置100可以配置于服务器中。
如图7所示,用户反欺诈实现装置100包括第一标识获取单元110、第二标识获取单元120、第三标识获取单元130、第四标识获取单元140、初始行向量获取单元150、概率值计算单元160。
第一标识获取单元110,用于接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为 第一标识值。
在本实施例中,当服务器需实时爬取指定用户多维度的用户信息时,需通过上传端上传用户文件,该用户文件中至少包括用户姓名、身份证明文件(如身份证扫描照片)、用户通讯号码等。当服务器接收了上传端所上传的用户文件后,至少获取四个维度的信息,一是身份维度、二是征信信用维度、三是社会行为维度、四是高法失信维度。通过爬虫工具爬取用户在上述四个维度的信息时,获取的是实时性较高的最新信息,能基于最新信息对用户进行反欺诈分析。
在获取第一个维度对应的身份维度信息时,需服务器将用户的身份证明文件上传至身份联网核查系统进行验证,通过身份联网核查系统来辅助核实用户的身份维度的真实性。
在一实施例中,如图8所示,第一标识获取单元110包括:
身份图像识别单元111,用于接收所上传用户文件中的身份证明文件,通过图像识别获取所述身份证明文件中的身份唯一识别码和用户姓名;身份验证单元112,用于将所述身份唯一识别码和用户姓名上传至身份联网核查系统进行验证;第一设置单元113,用于若接收到所述身份联网核查系统的验证通过通知信息,将所述身份维度验证标识位设置为第一标识值。
即服务器接收上传端所上传的用户文件中身份证明文件,通过服务器中存储的深度神经网络模型识别获取所述身份证明文件中的身份唯一识别码和用户姓名。在识别得到了所述身份证明文件中的身份唯一识别码和用户姓名后,将所述身份唯一识别码和用户姓名上传至身份联网核查系统进行验证,若接收到所述身份联网核查系统的验证通过通知信息,将身份维度验证标识位置为对应的第一标识值。
身份联网核查系统即是公安系统的身份验证系统,当上传了用户的身份证明文件(如身份证的正反面照片)时,身份联网核查系统能对该身份证明文件的真实性进行验证,一旦通过验证,则将身份维度验证标识位置为对应的第一标识值(例如第一标识值为1)。通过身份联网核查系统中存储的海量用户身份信息数据,能辅助验证用户信息在身份维度的真实性。
第二标识获取单元120,用于根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的总 条数设置为第二标识值。
即当完成了用户信息在身份维度的真实性的验证时,此时服务器再将所述身份证明文件上传至征信查询系统,以获取与所述身份证明文件对应的征信信息。通过对征信信息进行解析后,将征信信用维度标识位根据所述逾期信息的总条数设置为第二标识值。
在一实施例中,如图9所示,第二标识获取单元120包括:
第一上传单元121,用于若检测到征信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统;逾期信息获取单元122,用于接收所述征信查询系统根据所述身份证明文件对应发送的逾期信息;逾期信息解析单元123,用于获取所述逾期信息中在预设的查询时间段内的逾期总条数;第二设置单元124,用于根据所述逾期信息的逾期总条数,将征信信用维度标识位设置为所述第二标识值。
在本实施例中,征信查询系统是人民银行的征信查询系统,通过上传用户的身份证明文件(如身份证照片的正反面),即可获取用户在指定查询时间段(近5年内)的逾期信息。当服务器自动连接征信查询系统并发起征信查询请求时,若服务器检测到征信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统。所述身份证明文件通过征信查询系统的身份验证后,获取与所述身份证明文件对应的征信信息。由于征信信息中包括了该用户在在预设的查询时间段内的逾期总条数,故可根据所述逾期信息的逾期总条数,将征信信用维度标识位设置为所述第二标识值。通过征信查询系统中存储的海量用户征信信息数据,能辅助验证用户信息在征信信用维度的信用度。
第三标识获取单元130,用于接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值。
即当完成了用户信息在征信信用维度的逾期总条数查询时,此时服务器再将所述用户文件中的身份证明文件上传至通讯号码查询系统(如通讯运营商的服务器),以获取与所述身份证明文件对应的已存储的用户登记通讯号码。若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值。
在一实施例中,如图10所示,第三标识获取单元130包括:
第二上传单元131,用于根据所上传用户文件中的身份证明文件,将所述身份证明文件上传至通讯号码查询系统;登记号码获取单元132,用于接收在所述通讯号码查询系统中获取的与所述身份证明文件对应的已存储的用户登记通讯号码;第三设置单元133,用于若已存储的用户登记通讯号码的个数大于1或等于1、且已存储的用户登记通讯号码存在所上传用户文件中的用户通讯号码,将所述社会行为维度标识位设置为第三标识值;第四设置单元134,用于若已存储的用户登记通讯号码的个数等于0,将所述社会行为维度标识位设置为0。
即服务器根据所上传用户文件中的身份证明文件,将所述身份证明文件上传至通讯号码查询系统;在通讯号码查询系统根据所述身份证明文件对应获取已存储的用户登记通讯号码;若用户登记通讯号码的个数大于1,判断多个用户登记通讯号码中是否存在与所述用户通讯号码相同的号码,若多个用户登记通讯号码中存在与所述用户通讯号码相同的号码,判定所述用户通讯号码与所述用户登记通讯号码相同;若用户登记通讯号码的个数等于1且所述用户通讯号码与所述用户登记通讯号码相同,判定所述用户通讯号码与所述用户登记通讯号码相同;若用户登记通讯号码的个数小于1,判定所述用户通讯号码与所述用户登记通讯号码不相同;若判定所述用户通讯号码与所述用户登记通讯号码相同,将社会行为维度标识位设置为对应的第三标识值;具体的,将社会行为维度标识位设置为对应的第三标识值时,所述第三标识值为1;若判定所述用户通讯号码与所述用户登记通讯号码不相同,将社会行为维度标识位设置为0。
通过通讯号码查询系统中存储的海量用户通讯号码数据,能辅助验证用户信息在社会行为维度的信用度。
第四标识获取单元140,用于根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值。
即当完成了获取用户信息在社会行为维度的信用度后,此时服务器再将所述身份证明文件上传至高级法院失信查询系统,以获取与所述身份证明文件对应的高级法院失信信息。通过对高级法院失信信息进行解析后,将高法失信维度标识位根据失信总条数设置为第四标识值。
在一实施例中,如图11所示,第四标识获取单元140包括:
第三上传单元141,用于若检测到高级法院失信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统;失信信息获取单元142,用于接收所 述高级法院失信查询系统根据所述身份证明文件对应发送的高级法院失信信息;失信总条数获取单元143,用于获取所述高级法院失信信息中在预设的查询时间段内的失信总条数;第五设置单元144,用于根据所述高级法院失信信息的失信总条数,将高法失信维度标识位设置为第四标识值。
在本实施例中,高级法院失信查询系统是人民法院的失信查询系统(俗称老赖查询),通过上传用户的身份证明文件(如身份证照片的正反面),即可获取用户在指定查询时间段(近1年内)的失信信息。当服务器自动连接高级法院失信查询系统并发起征信查询请求时,若服务器检测到高级法院失信查询系统的身份验证请求,将所述身份证明文件上传至高级法院失信查询系统。所述身份证明文件通过高级法院失信查询系统的身份验证后,获取与所述身份证明文件对应的高级法院失信信息。由于高级法院失信信息中包括了该用户在在预设的查询时间段内的失信总条数,故可根据所述失信总条数,将高法失信维度标识位设置为所述第四标识值。通过高级法院失信查询系统中存储的海量用户高级法院失信信息数据,能辅助验证用户信息在高法失信维度的信用度。
初始行向量获取单元150,用于将所述第一标识值、第二标识值、第三标识值、第四标识值进行串接,得到用户欺诈识别初始行向量。
在本实施例中,将当识别获取了第一标识值、第二标识值、第三标识值、第四标识值并组成用户欺诈识别初始行向量后,如[1 0.5 1 1],此时将[1 0.5 1 1]输入至预选训练的深度神经网络模型。通过上述方式,将用户在四个维度的信息进行了量化,便于对用户进行反欺诈分析。
概率值计算单元160,用于将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值。
在本实施例中,当将通过深度神经网络模型有效的判断了用户欺诈概率值,自动得出反欺诈信用风险评估指标。
在一实施例中,用户反欺诈实现装置100还包括:
模型训练单元,用于获取已标注的历史数据作为训练集,将所述训练集中每一条训练数据的历史用户欺诈识别初始行向量输入至待训练深度神经网络模型,将所述训练集中每一条训练数据的用户欺诈概率值作为待训练深度神经网络模型的输出,对所述待训练深度神经网络模型进行训练,得到用于预测用户欺诈概率的深度神经网络模型。
在预先训练该用于识别用户欺诈概率值的深度神经网络,需要大量已标注的历史数据作为训练集,该训练集中每一条训练数据中输入值是历史用户欺诈识别初始行向量,如[1 0.5 1 1]、[0 0.3 1 0]、[1 1 1 1]、[1 0.5 1 1]、[1 0 0 1]等,与此分别对应的输出值为0.9、0.4、06、0.5等。当已知了待训练的深度神经网络的输入和输出之后,即可进行训练。
该装置实现了实时获取多个维度的用户信息进行反欺诈分析,提高了反欺诈分析的实时性和准确性。
上述用户反欺诈实现装置可以实现为计算机程序的形式,该计算机程序可以在如图12所示的计算机设备上运行。
请参阅图12,图12是本申请实施例提供的计算机设备的示意性框图。该计算机设备500是服务器,服务器可以是独立的服务器,也可以是多个服务器组成的服务器集群。
参阅图12,该计算机设备500包括通过系统总线501连接的处理器502、存储器和网络接口505,其中,存储器可以包括非易失性存储介质503和内存储器504。
该非易失性存储介质503可存储操作系统5031和计算机程序5032。该计算机程序5032被执行时,可使得处理器502执行用户反欺诈实现方法。
该处理器502用于提供计算和控制能力,支撑整个计算机设备500的运行。
该内存储器504为非易失性存储介质503中的计算机程序5032的运行提供环境,该计算机程序5032被处理器502执行时,可使得处理器502执行用户反欺诈实现方法。
该网络接口505用于进行网络通信,如提供数据信息的传输等。本领域技术人员可以理解,图12中示出的结构,仅仅是与本申请方案相关的部分结构的框图,并不构成对本申请方案所应用于其上的计算机设备500的限定,具体的计算机设备500可以包括比图中所示更多或更少的部件,或者组合某些部件,或者具有不同的部件布置。
其中,所述处理器502用于运行存储在存储器中的计算机程序5032,以实现本申请实施例的用户反欺诈实现方法。
本领域技术人员可以理解,图12中示出的计算机设备的实施例并不构成对计算机设备具体构成的限定,在其他实施例中,计算机设备可以包括比图示更 多或更少的部件,或者组合某些部件,或者不同的部件布置。例如,在一些实施例中,计算机设备可以仅包括存储器及处理器,在这样的实施例中,存储器及处理器的结构及功能与图12所示实施例一致,在此不再赘述。
应当理解,在本申请实施例中,处理器502可以是中央处理单元(Central Processing Unit,CPU),该处理器502还可以是其他通用处理器、数字信号处理器(Digital Signal Processor,DSP)、专用集成电路(Application Specific Integrated Circuit,ASIC)、现成可编程门阵列(Field-Programmable Gate Array,FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。其中,通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
在本申请的另一实施例中提供计算机可读存储介质。该计算机可读存储介质可以为非易失性的计算机可读存储介质。该计算机可读存储介质存储有计算机程序,其中计算机程序被处理器执行时实现本申请实施例的用户反欺诈实现方法。
所述存储介质为实体的、非瞬时性的存储介质,例如可以是U盘、移动硬盘、只读存储器(Read-Only Memory,ROM)、磁碟或者光盘等各种可以存储程序代码的实体存储介质。
所属领域的技术人员可以清楚地了解到,为了描述的方便和简洁,上述描述的设备、装置和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
以上所述,仅为本申请的具体实施方式,但本申请的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本申请揭露的技术范围内,可轻易想到各种等效的修改或替换,这些修改或替换都应涵盖在本申请的保护范围之内。因此,本申请的保护范围应以权利要求的保护范围为准。

Claims (20)

  1. 一种用户反欺诈实现方法,包括:
    接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值;
    根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的总条数设置为第二标识值;
    接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值;
    根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值;
    将所述第一标识值、第二标识值、第三标识值、第四标识值进行串接,得到用户欺诈识别初始行向量;以及
    将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值。
  2. 根据权利要求1所述的用户反欺诈实现方法,其中,所述接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值,包括:
    接收所上传用户文件中的身份证明文件,通过图像识别获取所述身份证明文件中的身份唯一识别码和用户姓名;
    将所述身份唯一识别码和用户姓名上传至身份联网核查系统进行验证;
    若接收到所述身份联网核查系统的验证通过通知信息,将所述身份维度验证标识位设置为第一标识值。
  3. 根据权利要求1所述的用户反欺诈实现方法,其中,所述根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的逾期总条数设置为第二标识值,包括:
    若检测到征信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统;
    接收所述征信查询系统根据所述身份证明文件对应发送的逾期信息;
    获取所述逾期信息中在预设的查询时间段内的逾期总条数;
    根据所述逾期信息的逾期总条数,将征信信用维度标识位设置为所述第二标识值。
  4. 根据权利要求1所述的用户反欺诈实现方法,其中,所述接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值,包括:
    根据所上传用户文件中的身份证明文件,将所述身份证明文件上传至通讯号码查询系统;
    接收在所述通讯号码查询系统中获取的与所述身份证明文件对应的已存储的用户登记通讯号码;
    若已存储的用户登记通讯号码的个数大于1或等于1、且已存储的用户登记通讯号码存在所上传用户文件中的用户通讯号码,将所述社会行为维度标识位设置为第三标识值;
    若已存储的用户登记通讯号码的个数等于0,将所述社会行为维度标识位设置为0。
  5. 根据权利要求1所述的用户反欺诈实现方法,其中,所述根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值,包括:
    若检测到高级法院失信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统;
    接收所述高级法院失信查询系统根据所述身份证明文件对应发送的高级法院失信信息;
    获取所述高级法院失信信息中在预设的查询时间段内的失信总条数;
    根据所述高级法院失信信息的失信总条数,将高法失信维度标识位设置为第四标识值。
  6. 根据权利要求1所述的用户反欺诈实现方法,其中,所述接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值之前,还包括:
    获取已标注的历史数据作为训练集,将所述训练集中每一条训练数据的历史用户欺诈识别初始行向量输入至待训练深度神经网络模型,将所述训练集中每一条训练数据的用户欺诈概率值作为待训练深度神经网络模型的输出,对所 述待训练深度神经网络模型进行训练,得到用于预测用户欺诈概率的深度神经网络模型。
  7. 根据权利要求1所述的用户反欺诈实现方法,其中,所述将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值之后,还包括:
    若所述用户欺诈概率值大于预先设置的第一概率阈值,将所述用户欺诈概率值对应的用户文件自动增加设置可疑用户的标识,并将所述用户欺诈概率值对应的用户文件存储至可疑用户的分组区域。
  8. 根据权利要求7所述的用户反欺诈实现方法,其中,所述若所述用户欺诈概率值大于预先设置的第一概率阈值,将所述用户欺诈概率值对应的用户文件自动增加设置可疑用户的标识,并将所述用户欺诈概率值对应的用户文件存储至可疑用户的分组区域之后,还包括:
    获取可疑用户的分组区域中各用户文件对应的用户信息;其中,所述用户信息包括身份唯一识别码和用户姓名;
    若可疑用户的分组区域中存在有用户信息之间存在亲属关联关系,将存在亲属关联关系的用户信息存储至可疑关联用户的分组区域。
  9. 一种用户反欺诈实现装置,包括:
    第一标识获取单元,用于接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值;
    第二标识获取单元,用于根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的总条数设置为第二标识值;
    第三标识获取单元,用于接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值;
    第四标识获取单元,用于根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值;
    初始行向量获取单元,用于将所述第一标识值、第二标识值、第三标识值、第四标识值进行串接,得到用户欺诈识别初始行向量;以及
    概率值计算单元,用于将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值。
  10. 根据权利要求9所述的用户反欺诈实现装置,其中,还包括:
    模型训练单元,用于获取已标注的历史数据作为训练集,将所述训练集中每一条训练数据的历史用户欺诈识别初始行向量输入至待训练深度神经网络模型,将所述训练集中每一条训练数据的用户欺诈概率值作为待训练深度神经网络模型的输出,对所述待训练深度神经网络模型进行训练,得到用于预测用户欺诈概率的深度神经网络模型。
  11. 一种计算机设备,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程序,所述处理器执行所述计算机程序时实现以下步骤:
    接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值;
    根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的总条数设置为第二标识值;
    接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值;
    根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值;
    将所述第一标识值、第二标识值、第三标识值、第四标识值进行串接,得到用户欺诈识别初始行向量;以及
    将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值。
  12. 根据权利要求11所述的计算机设备,其中,所述接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值,包括:
    接收所上传用户文件中的身份证明文件,通过图像识别获取所述身份证明文件中的身份唯一识别码和用户姓名;
    将所述身份唯一识别码和用户姓名上传至身份联网核查系统进行验证;
    若接收到所述身份联网核查系统的验证通过通知信息,将所述身份维度验 证标识位设置为第一标识值。
  13. 根据权利要求11所述的计算机设备,其中,所述根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的逾期总条数设置为第二标识值,包括:
    若检测到征信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统;
    接收所述征信查询系统根据所述身份证明文件对应发送的逾期信息;
    获取所述逾期信息中在预设的查询时间段内的逾期总条数;
    根据所述逾期信息的逾期总条数,将征信信用维度标识位设置为所述第二标识值。
  14. 根据权利要求11所述的计算机设备,其中,所述接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值,包括:
    根据所上传用户文件中的身份证明文件,将所述身份证明文件上传至通讯号码查询系统;
    接收在所述通讯号码查询系统中获取的与所述身份证明文件对应的已存储的用户登记通讯号码;
    若已存储的用户登记通讯号码的个数大于1或等于1、且已存储的用户登记通讯号码存在所上传用户文件中的用户通讯号码,将所述社会行为维度标识位设置为第三标识值;
    若已存储的用户登记通讯号码的个数等于0,将所述社会行为维度标识位设置为0。
  15. 根据权利要求11所述的计算机设备,其中,所述根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值,包括:
    若检测到高级法院失信查询系统的身份验证请求,将所述身份证明文件上传至征信查询系统;
    接收所述高级法院失信查询系统根据所述身份证明文件对应发送的高级法院失信信息;
    获取所述高级法院失信信息中在预设的查询时间段内的失信总条数;
    根据所述高级法院失信信息的失信总条数,将高法失信维度标识位设置为第四标识值。
  16. 根据权利要求11所述的计算机设备,其中,所述接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值之前,还包括:
    获取已标注的历史数据作为训练集,将所述训练集中每一条训练数据的历史用户欺诈识别初始行向量输入至待训练深度神经网络模型,将所述训练集中每一条训练数据的用户欺诈概率值作为待训练深度神经网络模型的输出,对所述待训练深度神经网络模型进行训练,得到用于预测用户欺诈概率的深度神经网络模型。
  17. 根据权利要求11所述的计算机设备,其中,所述将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值之后,还包括:
    若所述用户欺诈概率值大于预先设置的第一概率阈值,将所述用户欺诈概率值对应的用户文件自动增加设置可疑用户的标识,并将所述用户欺诈概率值对应的用户文件存储至可疑用户的分组区域。
  18. 根据权利要求17所述的计算机设备,其中,所述若所述用户欺诈概率值大于预先设置的第一概率阈值,将所述用户欺诈概率值对应的用户文件自动增加设置可疑用户的标识,并将所述用户欺诈概率值对应的用户文件存储至可疑用户的分组区域之后,还包括:
    获取可疑用户的分组区域中各用户文件对应的用户信息;其中,所述用户信息包括身份唯一识别码和用户姓名;
    若可疑用户的分组区域中存在有用户信息之间存在亲属关联关系,将存在亲属关联关系的用户信息存储至可疑关联用户的分组区域。
  19. 一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,所述计算机程序当被处理器执行以下操作:
    接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值;
    根据所述身份证明文件对应获取征信信息,获取所述征信信息中的逾期信息,将征信信用维度标识位根据所述逾期信息的总条数设置为第二标识值;
    接收所上传用户文件中的用户通讯号码,若所述用户通讯号码与已存储的用户登记通讯号码相同,将社会行为维度标识位设置为第三标识值;
    根据所述身份证明文件对应获取高级法院失信信息,根据所述高级法院失信信息,将高法失信维度标识位设置为第四标识值;
    将所述第一标识值、第二标识值、第三标识值、第四标识值进行串接,得到用户欺诈识别初始行向量;以及
    将所述用户欺诈识别初始行向量输入至预先训练的深度神经网络模型,得到用户欺诈概率值。
  20. 根据权利要求19所述的计算机可读存储介质,其中,所述接收所上传用户文件中的身份证明文件,若所述身份证明文件通过身份联网核查系统的验证,将身份维度验证标识位设置为第一标识值,包括:
    接收所上传用户文件中的身份证明文件,通过图像识别获取所述身份证明文件中的身份唯一识别码和用户姓名;
    将所述身份唯一识别码和用户姓名上传至身份联网核查系统进行验证;
    若接收到所述身份联网核查系统的验证通过通知信息,将所述身份维度验证标识位设置为第一标识值。
PCT/CN2019/118405 2019-03-15 2019-11-14 用户反欺诈实现方法、装置、计算机设备及存储介质 WO2020186789A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910197847.6A CN110070431A (zh) 2019-03-15 2019-03-15 用户反欺诈实现方法、装置、计算机设备及存储介质
CN201910197847.6 2019-03-15

Publications (1)

Publication Number Publication Date
WO2020186789A1 true WO2020186789A1 (zh) 2020-09-24

Family

ID=67366273

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/118405 WO2020186789A1 (zh) 2019-03-15 2019-11-14 用户反欺诈实现方法、装置、计算机设备及存储介质

Country Status (2)

Country Link
CN (1) CN110070431A (zh)
WO (1) WO2020186789A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112200583A (zh) * 2020-10-28 2021-01-08 交通银行股份有限公司 一种基于知识图谱的欺诈客户识别方法
CN112598508A (zh) * 2020-12-28 2021-04-02 中国农业银行股份有限公司 征信数据使用方法及系统
CN113657902A (zh) * 2021-08-03 2021-11-16 浙江创邻科技有限公司 基于图数据库的金融安全管理方法、系统及存储介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110070431A (zh) * 2019-03-15 2019-07-30 平安科技(深圳)有限公司 用户反欺诈实现方法、装置、计算机设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN109033139A (zh) * 2018-06-06 2018-12-18 中国平安人寿保险股份有限公司 客户信息查询方法、装置、计算机设备和存储介质
CN109191129A (zh) * 2018-07-18 2019-01-11 阿里巴巴集团控股有限公司 一种风控方法、系统及计算机设备
CN110070431A (zh) * 2019-03-15 2019-07-30 平安科技(深圳)有限公司 用户反欺诈实现方法、装置、计算机设备及存储介质

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107103548A (zh) * 2011-11-17 2017-08-29 阿里巴巴集团控股有限公司 网络行为数据的监控方法和系统以及风险监控方法和系统
CN109308615B (zh) * 2018-08-02 2020-12-29 同济大学 基于统计序列特征的实时欺诈交易检测方法、系统、存储介质及电子终端
CN109389494B (zh) * 2018-10-25 2021-11-05 北京芯盾时代科技有限公司 借贷欺诈检测模型训练方法、借贷欺诈检测方法及装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN109033139A (zh) * 2018-06-06 2018-12-18 中国平安人寿保险股份有限公司 客户信息查询方法、装置、计算机设备和存储介质
CN109191129A (zh) * 2018-07-18 2019-01-11 阿里巴巴集团控股有限公司 一种风控方法、系统及计算机设备
CN110070431A (zh) * 2019-03-15 2019-07-30 平安科技(深圳)有限公司 用户反欺诈实现方法、装置、计算机设备及存储介质

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112200583A (zh) * 2020-10-28 2021-01-08 交通银行股份有限公司 一种基于知识图谱的欺诈客户识别方法
CN112200583B (zh) * 2020-10-28 2023-12-19 交通银行股份有限公司 一种基于知识图谱的欺诈客户识别方法
CN112598508A (zh) * 2020-12-28 2021-04-02 中国农业银行股份有限公司 征信数据使用方法及系统
CN112598508B (zh) * 2020-12-28 2024-01-19 中国农业银行股份有限公司 征信数据使用方法及系统
CN113657902A (zh) * 2021-08-03 2021-11-16 浙江创邻科技有限公司 基于图数据库的金融安全管理方法、系统及存储介质
CN113657902B (zh) * 2021-08-03 2024-03-22 浙江创邻科技有限公司 基于图数据库的金融安全管理方法、系统及存储介质

Also Published As

Publication number Publication date
CN110070431A (zh) 2019-07-30

Similar Documents

Publication Publication Date Title
WO2020186789A1 (zh) 用户反欺诈实现方法、装置、计算机设备及存储介质
US11695755B2 (en) Identity proofing and portability on blockchain
US11847199B2 (en) Remote usage of locally stored biometric authentication data
US20210406354A1 (en) System and method for user enrollment in a secure biometric verification system
US10924514B1 (en) Machine learning detection of fraudulent validation of financial institution credentials
US10037533B2 (en) Systems and methods for detecting relations between unknown merchants and merchants with a known connection to fraud
TWI706333B (zh) 欺詐交易識別方法、裝置、伺服器及儲存媒體
US20170140386A1 (en) Transaction assessment and/or authentication
JP2021508889A (ja) 本人確認方法および装置
US20190347425A1 (en) Method and apparatus for identity authentication
US20220131844A1 (en) Identity access management using access attempts and profile updates
JP2017524998A (ja) 本人照合を行うための方法およびシステム
US20160283603A1 (en) Methods and systems for testing performance of biometric authentication systems
CN113034145B (zh) 用户异常加密数字资产交易类别判断方法、装置
WO2021098274A1 (zh) 针对隐私数据泄漏的风险评估方法及装置
US11823197B2 (en) Authenticating based on user behavioral transaction patterns
US11423177B2 (en) Systems and methods for establishing trust online
WO2022126975A1 (zh) 客户信息校验方法、装置、计算机设备及存储介质
US20230134651A1 (en) Synchronized Identity, Document, and Transaction Management
US20140150097A1 (en) System and method for email fraud risk assessment
US20220164423A1 (en) Method and apparatus for user recognition
US20240135379A1 (en) Authenticating Based on Behavioral Transactional Patterns
CN116777294A (zh) 区块链协助下基于联邦学习的众包质量安全评估方法
WO2019143360A1 (en) Data security using graph communities
JP7176158B1 (ja) 学習モデル評価システム、学習モデル評価方法、及びプログラム

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: 19920150

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19920150

Country of ref document: EP

Kind code of ref document: A1