User abnormal transaction probability identification method based on block chain, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a block chain-based user abnormal transaction probability identification method, electronic equipment and a storage medium. The method comprises the following steps: performing iterative training on an initial transaction recognition model through a locally stored user transaction data set to obtain a first transaction recognition model corresponding to the trained initial transaction recognition model; uploading a first training parameter carried by the first transaction identification model to cloud service equipment through encryption transmission; under the condition that an updating instruction fed back by the cloud service equipment is received, downloading a second training parameter from the cloud service equipment; and adjusting the first transaction identification model according to the second training parameter to obtain a target transaction identification model, and identifying locally updated credit card transaction data of the to-be-identified user through the target identification model to obtain an abnormal transaction probability corresponding to the to-be-identified user.

Description

Technical Field

[0001] The present application relates to the field of financial security, and in particular to a method, electronic device, and storage medium for identifying the probability of abnormal user transactions based on blockchain. Background Art

[0002] Abnormal credit card consumption behavior mainly refers to abnormal consumption or transfer behavior related to credit cards. For example, non-cardholders use credit cards in the name of cardholders without the cardholder's consent or authorization, use fake credit cards for consumption, overdraw credit cards, etc. all fall into the category of abnormal credit card consumption behavior. Abnormal credit card consumption behavior not only endangers financial order, but also poses a hidden danger to the property safety of the public.

[0003] In the financial sector, banks, as core financial institutions, bear the crucial responsibility of ensuring fund security and preventing financial risks. With the rapid development of financial technology, transactions are becoming increasingly complex and diverse, making the identification of abnormal transactions a key component of bank risk management. However, currently, banks typically store user information and transaction data independently, lacking cross-institutional data sharing mechanisms. This results in significant differences in the ability of different banks to identify the same abnormal transaction. Banks with limited sample data are unable to identify abnormal transactions promptly, resulting in losses.

[0004] Therefore, there is an urgent need for a method to optimize the performance of the abnormal transaction identification model and improve the accuracy of model recognition. Summary of the Invention

[0005] The embodiments of the present application provide a blockchain-based method for identifying the probability of abnormal user transactions, an electronic device, and a storage medium, so as to optimize the performance of the abnormal transaction identification model and improve the accuracy of model identification.

[0006] In a first aspect, an embodiment of the present application provides a method for identifying the probability of abnormal user transactions based on blockchain, comprising: iteratively training an initial transaction identification model through a locally stored user transaction data set to obtain a first transaction identification model corresponding to the training of the initial transaction identification model, wherein the initial identification model is downloaded through a cloud service device, and the initial identification model is used to identify the transaction probability of abnormal user transactions; uploading the first training parameters carried by the first transaction identification model to the cloud service device through encrypted transmission; upon receiving an update instruction fed back by the cloud service device, downloading the second training parameters from the cloud service device, wherein the second training parameters are obtained by the cloud service device adjusting the initial parameters of the initial model based on the training parameters uploaded by multiple bank terminals; adjusting the first transaction identification model according to the second training parameters to obtain a target transaction identification model, and identifying the locally updated credit card transaction data of the user to be identified through the target identification model to obtain the abnormal transaction probability corresponding to the user to be identified.

[0007] In a second aspect, an embodiment of the present application provides a device for identifying the probability of abnormal user transactions based on blockchain, comprising: a training module for iteratively training an initial transaction identification model through a locally stored user transaction data set to obtain a first transaction identification model corresponding to the training of the initial transaction identification model, wherein the initial identification model is downloaded through a cloud service device, and the initial identification model is used to identify the transaction probability of abnormal user transactions; a transmission module for uploading the first training parameters carried by the first transaction identification model to the cloud service device through encrypted transmission; a download module for downloading second training parameters from the cloud service device upon receiving an update instruction fed back by the cloud service device, wherein the second training parameters are obtained by the cloud service device adjusting the initial parameters of the initial model based on the training parameters uploaded by multiple bank terminals; an identification module for adjusting the first transaction identification model according to the second training parameters to obtain a target transaction identification model, and identifying the locally updated credit card transaction data of the user to be identified through the target identification model to obtain the abnormal transaction probability corresponding to the user to be identified.

[0008] In a third aspect, an embodiment of the present application provides a device for identifying the probability of abnormal user transactions based on blockchain, comprising: a memory, a processor;

[0009] The memory stores computer-executable instructions;

[0010] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.

[0011] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the first aspect above and / or various possible implementation methods of the first aspect.

[0012] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementation methods of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0014] Figure 1 Schematic diagram of the process of identifying the probability of abnormal user transactions based on blockchain provided by this application Figure 1 ;

[0015] Figure 2 Schematic diagram of the process of identifying the probability of abnormal user transactions based on blockchain provided by this application Figure 2 ;

[0016] Figure 3 This is a schematic diagram of the structure of the device for identifying the probability of abnormal user transactions based on blockchain provided by this application;

[0017] Figure 4 Schematic diagram of the structure of the device for identifying the probability of abnormal user transactions based on blockchain provided in this application.

[0018] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0019] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0020] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0021] Figure 1 Schematic diagram of the process of identifying the probability of abnormal user transactions based on blockchain provided by this application Figure 1 ,like Figure 1 As shown, the method includes:

[0022] S101. Iteratively train an initial transaction recognition model using a locally stored user transaction dataset to obtain a first transaction recognition model corresponding to the trained initial transaction recognition model, wherein the initial recognition model is downloaded from a cloud service device and is used to identify the transaction probability of abnormal user transactions;

[0023] The user transaction dataset contains historical user transaction records, including transaction time, amount, and type. Specifically, it includes information such as the credit card holder, the location at the time of the credit card transaction, and the transaction limit. The initial transaction recognition model is a pre-trained model used to identify abnormal transactions. This model is downloaded from the cloud service. The first transaction recognition model is an iteratively trained model of the initial transaction recognition model, and it more accurately identifies abnormal credit card transactions than the initial model.

[0024] By iteratively training using local user-specific data, the model can be better adapted to the user's trading habits and patterns, thereby improving the accuracy of identifying abnormal transactions.

[0025] S102: Upload the first training parameter carried by the first transaction recognition model to the cloud service device through encrypted transmission;

[0026] Optionally, encrypted transmission refers to encrypting data using encryption technology (such as SSL / TLS, AES, etc.) during data transmission to ensure that the data is not stolen or tampered with during transmission. The purpose of encrypted transmission is to protect the confidentiality and integrity of the first training parameters during the process of uploading the first training parameters to the cloud service device.

[0027] Encrypted transmission ensures that the first training parameters cannot be stolen or tampered with by third parties during the upload process, protecting the privacy and security of user data. The first transaction recognition model parameters obtained through local training are securely transmitted to the cloud to enable data synchronization, model optimization, distributed computing, and collaboration. Encrypted transmission ensures data security; uploaded parameters provide a foundation for subsequent model improvements and cloud services.

[0028] For example, encrypted transmission can be achieved through blockchain technology. The immutable nature of blockchain ensures that uploaded model parameters cannot be maliciously modified or tampered with during storage and transmission, ensuring the integrity and authenticity of the data. After encrypting the first training parameters and uploading them to the blockchain network, only authorized users can decrypt and use the data, effectively protecting data privacy and preventing data leakage to unauthorized third parties.

[0029] S103. Upon receiving an update instruction fed back by the cloud service device, downloading second training parameters from the cloud service device, wherein the second training parameters are obtained by the cloud service device adjusting initial parameters of the initial model based on training parameters uploaded by multiple bank terminals;

[0030] The second training parameters are obtained by adjusting the initial model parameters using training parameters uploaded by multiple bank terminals. This adjustment is based on data from multiple bank terminals, incorporating personalized information from different bank terminals and enabling better adaptation to a wider range of transaction scenarios.

[0031] By downloading the second training parameters, a globally optimized model is generated, improving the accuracy and generalization of abnormal transaction identification. The training parameters uploaded by multiple bank terminals reflect the transaction habits and patterns of different users. By integrating this data, the cloud generates a more universal and robust model. Downloading the second training parameters allows users to benefit from the data and experience of other bank terminals, avoiding the limitations of relying solely on local data.

[0032] S104. Adjust the first transaction recognition model according to the second training parameters to obtain a target transaction recognition model. Identify the locally updated credit card transaction data of the user to be identified using the target recognition model to obtain an abnormal transaction probability corresponding to the user to be identified.

[0033] By applying the second training parameters to the first transaction recognition model, a target transaction recognition model is generated. This model retains the personalized characteristics of local data while incorporating global optimization experience, thereby improving recognition accuracy and generalization. By combining local and global optimization parameters, the target transaction recognition model can better adapt to user trading habits while avoiding overfitting and underfitting. This allows for more accurate identification of abnormal transactions and reduces false positives and missed negatives.

[0034] Optionally, the execution entity of the above steps is a terminal device of a bank.

[0035] It should be noted that the data in the above-mentioned user transaction dataset can only be used for model training if the user authorizes its use.

[0036] The blockchain-based method for identifying the probability of abnormal transactions of users provided in the embodiment of the present application is to obtain a first transaction identification model by iteratively training the initial transaction identification model after it is downloaded from the cloud and combined with the transaction data set stored locally by the user. By utilizing the transaction habits and historical data of local users, the model can capture the personalized characteristics of users, thereby more accurately identifying abnormal transactions. The training parameters of the first transaction identification model are uploaded to the cloud through encrypted transmission to ensure data security without leaking local data. The cloud globally adjusts the initial model based on the parameters uploaded by multiple bank terminals to generate the second training parameters. The first transaction identification model is adjusted according to the second training parameters to obtain the target transaction identification model. By integrating local personalized features and global optimization experience, the ability to identify new abnormal transactions is improved, the performance of the abnormal transaction identification model is optimized, and the accuracy of model recognition is improved.

[0037] Figure 2 Schematic diagram of the process of identifying the probability of abnormal user transactions based on blockchain provided by this application Figure 2 ,like Figure 2 As shown, this embodiment Figure 2 Based on the embodiment, a method for identifying the probability of abnormal user transactions based on blockchain is described in detail. The method includes:

[0038] S201. Send the identity information of the first bank terminal to the cloud service device; upon receiving a pass reminder from the cloud service device based on the identity information, determine that the first bank terminal has passed the identity authentication; upload the encrypted parameters obtained after encrypting the first training parameters to the cloud service device; upon receiving a fail reminder from the cloud service device based on the identity information, determine that the first bank terminal has not passed the identity authentication; initiate an identity allocation instruction to the cloud service device, wherein the allocation instruction is used to instruct the cloud service device to allocate new identity information to the first bank terminal.

[0039] Optionally, the identity information of the first bank terminal is sent to the cloud service device for authentication, ensuring that only authorized terminals can upload training parameters or download updates. If the authentication is successful (a "pass" notification is received), the encryption parameters are allowed to be uploaded, ensuring the security and legitimacy of the data. If the authentication fails (a "fail" notification is received), the upload operation is blocked to prevent unauthorized terminals from accessing the cloud service.

[0040] The first training parameters are encrypted and uploaded to the cloud, ensuring they are not stolen or tampered with during transmission, thus protecting the privacy and integrity of user data. It should be noted that the use of encrypted parameters complies with data security regulations to prevent the leakage of sensitive information. If identity authentication fails, an identity allocation instruction is initiated, requesting the cloud service device to assign new identity information to the first bank terminal. This mechanism can address the issue of expired, lost, or stolen identity information, ensuring that each terminal has a unique and valid identity.

[0041] Through authentication and encrypted transmission, secure communication between the terminal and the cloud is achieved, ensuring the reliability of the upload and download process of training parameters.

[0042] S202. Receive a verification result of the first training parameter fed back by the cloud service device; if the verification result indicates that the first training parameter is abnormal, initiate a review reminder of the first training parameter to the target object; and re-encrypt the first training parameter after review by the target object and upload it to the cloud service device.

[0043] Optionally, the verification result indicates the result returned by the cloud service device after checking the first training parameters, which is used to determine whether the parameters meet the requirements (such as format correctness, integrity, security, etc.). The verification result indicates a status where there is a problem with the first training parameters, such as parameter damage, format error, or non-compliance with safety standards. The target object may be the entity responsible for reviewing the first training parameters, which may be an administrator or a dedicated auditor. When there is an abnormality in the first training parameters, a notification is sent to the target object, prompting it to manually check and confirm the parameters. The first training parameters that have been reviewed and confirmed by the target object are encrypted again and then uploaded to the cloud service device.

[0044] By reviewing and re-uploading, we ensure that the training parameters received by the cloud service device are valid data that has been manually confirmed, avoiding cloud model optimization errors caused by abnormal parameters.

[0045] S203: Compare the second training parameter with the first training parameter to determine a parameter difference between the second training parameter and the first training parameter; replace the first training parameter with the parameter difference to obtain a target transaction recognition model corresponding to the first transaction recognition model.

[0046] By calculating the difference between the second and first training parameters, the difference between the global optimization parameters and the local parameters is extracted. Replacing the first training parameters with the parameter difference is equivalent to applying the global optimization adjustments to the local model, resulting in a target model that retains local personalized characteristics while incorporating global optimization experience. The parameter difference represents the direction of improvement to the initial model based on data from multiple bank terminals in the cloud. Applying this to the local model can rapidly improve the model's accuracy and generalization capabilities. This replacement avoids training the model from scratch, saving computing resources, and ensuring that the model is adaptable to a wider range of transaction scenarios.

[0047] S204. Determine the numerical interval corresponding to the probability value of the abnormal transaction probability; when the probability value corresponds to the first numerical interval, determine that the user to be identified is a normal transaction user; when the probability value corresponds to the second numerical interval, determine that the user to be identified is a potential abnormal transaction user; when the probability value corresponds to the third numerical interval, determine that the user to be identified is an abnormal transaction user.

[0048] Optionally, the abnormal transaction probability is a numerical value generated by the target transaction recognition model for each transaction, indicating the possibility of the transaction being determined to be abnormal, for example, a value between 0 and 1 (such as 0.3, 0.7, etc.) is set.

[0049] The probability of abnormal transactions is divided into different ranges to distinguish normal transactions, potential abnormal transactions and abnormal transactions.

[0050] For example:

[0051] First numerical range: The probability value is low, indicating normal trading.

[0052] Second numerical range: medium probability value, indicating potential abnormal transactions.

[0053] The third numerical range: The probability value is higher, indicating abnormal transactions.

[0054] Normal transaction users: users whose transaction behaviors conform to normal patterns and are not judged as abnormal by the model.

[0055] Potentially abnormal trading users: users whose trading behaviors have certain abnormal characteristics but have not yet reached the high-risk threshold.

[0056] Abnormal transaction users: users whose transaction behaviors are judged by the model to be high-risk and abnormal.

[0057] By dividing the probability of abnormal transactions into different numerical ranges, the abstract probability value is converted into a specific risk level (normal, potential abnormal, abnormal), which facilitates subsequent decision-making and processing. This division makes risk assessment more intuitive and operational.

[0058] By categorizing the probability of abnormal transactions into different numerical ranges, the model's output of abstract probability values ​​is converted into specific risk levels (normal, potentially abnormal, abnormal), enabling tiered processing of trading behavior. This mechanism improves risk control efficiency, reduces the risk of false positives and false negatives, and provides differentiated processing for users of different risk levels. By properly setting numerical ranges, the system can more accurately identify abnormal transactions and ensure transaction security.

[0059] S205. Determine a first accuracy rate of the first recognition model in recognizing a locally stored user transaction data set, wherein the training parameter corresponding to the first recognition model is a first training parameter; determine a second accuracy rate of the target recognition model in recognizing a locally stored user transaction data set; if the first accuracy rate is greater than the second accuracy rate, initiate a request to the cloud service device to refuse to use the target recognition model.

[0060] By comparing the accuracy of the first recognition model (local initial model) and the target recognition model (adjusted model) on the local dataset, we can evaluate whether the second training parameters provided by the cloud have truly improved model performance. If the first accuracy is higher than the second, it means that the adjusted model performs worse on the local data, possibly because the global optimization parameters do not match the local data characteristics.

[0061] Optionally, the first training parameters include at least one of the following: a first loss function and a first model gradient corresponding to the initial transaction recognition model after training; the second training parameters include at least one of the following: a second loss function and a second gradient model obtained after the cloud service device adjusts the initial parameters of the initial model based on the training parameters uploaded by multiple bank terminals.

[0062] The local dataset has personalized characteristics for each user, while the second training parameters in the cloud are generated based on data from multiple bank terminals and may not fully adapt to the transaction patterns of specific users. By comparing the accuracy rates, we ensure that the object recognition model does not sacrifice the adaptability of local data due to over-reliance on global parameters.

[0063] By comparing the accuracy of the initial recognition model and the target recognition model on the local dataset, we verify whether the second training parameters provided by the cloud have truly improved model performance. If the local initial model performs better, the target model is rejected to ensure transaction recognition accuracy and adaptability to local data. This mechanism achieves a dynamic balance between local and cloud performance, leveraging global optimization experience while avoiding performance degradation caused by parameter mismatch.

[0064] The blockchain-based method for identifying the probability of abnormal transactions of users provided in the embodiment of the present application is to obtain a first transaction identification model by iteratively training the initial transaction identification model after it is downloaded from the cloud and combined with the transaction data set stored locally by the user. By utilizing the transaction habits and historical data of local users, the model can capture the personalized characteristics of users, thereby more accurately identifying abnormal transactions. The training parameters of the first transaction identification model are uploaded to the cloud through encrypted transmission to ensure data security without leaking local data. The cloud globally adjusts the initial model based on the parameters uploaded by multiple bank terminals to generate the second training parameters. The first transaction identification model is adjusted according to the second training parameters to obtain the target transaction identification model. By integrating local personalized features and global optimization experience, the ability to identify new abnormal transactions is improved, the performance of the abnormal transaction identification model is optimized, and the accuracy of model recognition is improved.

[0065] Figure 3 This is a schematic diagram of the structure of the device for identifying the probability of abnormal user transactions based on blockchain provided by this application, as shown in Figure 3 As shown, the blockchain-based user abnormal transaction probability identification device 30 provided in this embodiment includes:

[0066] A training module 301 is configured to iteratively train an initial transaction recognition model using a locally stored user transaction dataset to obtain a first transaction recognition model corresponding to the trained initial transaction recognition model, wherein the initial recognition model is downloaded from a cloud service device and is used to identify the transaction probability of abnormal user transactions;

[0067] A transmission module 302 is configured to upload the first training parameters carried by the first transaction recognition model to the cloud service device via encrypted transmission;

[0068] a downloading module 303 configured to download second training parameters from the cloud service device upon receiving an update instruction fed back by the cloud service device, wherein the second training parameters are obtained by the cloud service device adjusting initial parameters of the initial model based on training parameters uploaded by multiple bank terminals;

[0069] The identification module 304 is used to adjust the first transaction identification model according to the second training parameters to obtain a target transaction identification model, identify the locally updated credit card transaction data of the user to be identified through the target identification model, and obtain the abnormal transaction probability corresponding to the user to be identified.

[0070] In one possible implementation, the identification module 304 is used to compare the second training parameter with the first training parameter to determine a parameter difference between the second training parameter and the first training parameter; and replace the first training parameter with the parameter difference to obtain a target transaction identification model corresponding to the first transaction identification model.

[0071] In one possible implementation, the transmission module 302 is used to receive the verification result of the first training parameter fed back by the cloud service device; when the verification result indicates that there is an abnormality in the first training parameter, initiate a review reminder of the first training parameter to the target object; and re-encrypt the first training parameter after review by the target object and upload it to the cloud service device.

[0072] In one possible implementation, the transmission module 302 is used to send the identity information of the first bank terminal to the cloud service device; upon receiving a pass reminder feedback from the cloud service device based on the identity information, determine that the first bank terminal has passed the identity authentication; upload the encrypted parameters obtained after encrypting the first training parameters to the cloud service device; upon receiving a fail reminder feedback from the cloud service device based on the identity information, determine that the first bank terminal has failed the identity authentication; and initiate an identity allocation instruction to the cloud service device, wherein the allocation instruction is used to instruct the cloud service device to allocate new identity information to the first bank terminal.

[0073] In one possible implementation, the first training parameters include at least one of the following: a first loss function and a first model gradient corresponding to the initial transaction recognition model after training; the second training parameters include at least one of the following: a second loss function and a second gradient model obtained by the cloud service device adjusting the initial parameters of the initial model based on the training parameters uploaded by multiple bank terminals.

[0074] In one possible implementation, the device also includes: a first determination module (not shown in the figure), which is used to determine the numerical interval corresponding to the probability value of the abnormal transaction probability; when the probability value corresponds to the first numerical interval, the user to be identified is determined to be a normal transaction user; when the probability value corresponds to the second numerical interval, the user to be identified is determined to be a potential abnormal transaction user; when the probability value corresponds to the third numerical interval, the user to be identified is determined to be an abnormal transaction user.

[0075] In one possible implementation, the device also includes: a second determination module (not shown in the figure), used to determine a first accuracy rate of the first recognition model in identifying a locally stored user transaction data set, wherein the training parameter corresponding to the first recognition model is a first training parameter; determine a second accuracy rate of the target recognition model in identifying a locally stored user transaction data set; and when the first accuracy rate is greater than the second accuracy rate, initiate a request to the cloud service device to refuse to use the target recognition model.

[0076] The device for identifying the probability of abnormal user transactions based on blockchain provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effects are similar, and are not described in detail in this embodiment.

[0077] Figure 4 This is a schematic diagram of the structure of the device for identifying the probability of abnormal user transactions based on blockchain provided by this application. Figure 4 As shown, the electronic device 40 provided in this embodiment includes: at least one processor 401 and a memory 402. Optionally, the device 40 further includes a communication component 403. The processor 401, the memory 402 and the communication component 403 are connected via a bus 404.

[0078] In a specific implementation process, at least one processor 401 executes the computer-executable instructions stored in the memory 402, so that the at least one processor 401 performs the above method.

[0079] The specific implementation process of the processor 401 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0080] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASICs), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly executed by a hardware processor or by a combination of hardware and software modules in the processor.

[0081] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (NVM), such as at least one disk memory.

[0082] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be classified as address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0083] The present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0084] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0085] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile memory device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0086] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium can also exist in a device as discrete components.

[0087] The division of units is merely a logical functional division; actual implementations may employ alternative divisions, such as combining or integrating multiple units or components into another system, or omitting or disabling certain features. Furthermore, any direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units, either through an interface, electrical, mechanical, or other means.

[0088] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0089] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0090] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0091] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0092] Finally, it should be noted that those skilled in the art will readily identify other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art not disclosed herein. The present invention is not limited to the precise structure described above and illustrated in the accompanying drawings, and various modifications and variations may be made without departing from the scope thereof. The scope of the present invention is limited solely by the appended claims.

Claims

1. A method for identifying the probability of abnormal user transactions based on blockchain, characterized in that: The method is applied to a bank terminal in a blockchain, wherein the blockchain includes a cloud service device and multiple bank terminals. The method includes: Iteratively training an initial transaction recognition model using a locally stored user transaction dataset to obtain a first transaction recognition model corresponding to the trained initial transaction recognition model, wherein the initial recognition model is downloaded from a cloud service device and is used to identify the transaction probability of abnormal user transactions; Uploading the first training parameter carried by the first transaction recognition model to the cloud service device through encrypted transmission; Upon receiving an update instruction fed back by the cloud service device, downloading second training parameters from the cloud service device, wherein the second training parameters are obtained by the cloud service device adjusting initial parameters of the initial model based on training parameters uploaded by multiple bank terminals; The first transaction recognition model is adjusted according to the second training parameter to obtain a target transaction recognition model. The locally updated credit card transaction data of the user to be identified is identified through the target recognition model to obtain the abnormal transaction probability corresponding to the user to be identified.

2. The method according to claim 1, characterized in that Adjusting the first transaction recognition model according to the second training parameter to obtain a target transaction recognition model includes: Comparing the second training parameter with the first training parameter to determine a parameter difference between the second training parameter and the first training parameter; The first training parameter is replaced by the parameter difference to obtain a target transaction recognition model corresponding to the first transaction recognition model.

3. The method according to claim 1, characterized in that After uploading the first training parameter carried by the first transaction recognition model to the cloud service device through encrypted transmission, the method further includes: receiving a verification result of the first training parameter fed back by the cloud service device; If the verification result indicates that the first training parameter is abnormal, initiating a review reminder of the first training parameter to the target subject; The first training parameters that have passed the review of the target object are re-encrypted and uploaded to the cloud service device.

4. The method according to claim 1, wherein Uploading the first training parameter carried by the first transaction recognition model to the cloud service device through encrypted transmission includes: Sending the identity information of the first bank terminal to the cloud service device; Upon receiving a passing reminder fed back by the cloud service device based on the identity information, determining that the first bank terminal has passed the identity authentication; Uploading the encrypted parameters obtained by encrypting the first training parameters to the cloud service device; Upon receiving a failure reminder fed back by the cloud service device based on the identity information, determining that the first bank terminal has failed identity authentication; An identity allocation instruction is initiated to the cloud service device, wherein the allocation instruction is used to instruct the cloud service device to allocate new identity information to the first bank terminal.

5. The method according to claim 1, wherein include: The first training parameter includes at least one of the following: a first loss function and a first model gradient corresponding to the initial transaction recognition model after training; The second training parameters include at least one of the following: a second loss function and a second gradient model obtained by the cloud service device adjusting the initial parameters of the initial model based on the training parameters uploaded by multiple bank terminals.

6. The method according to any one of claims 1 to 5, characterized in that After obtaining the abnormal transaction probability corresponding to the user to be identified, the method further includes: Determine a numerical range corresponding to the probability value of the abnormal transaction probability; When the probability value corresponds to the first numerical range, determining that the user to be identified is a normal transaction user; When the probability value corresponds to the second numerical range, determining that the user to be identified is a potential abnormal transaction user; When the probability value corresponds to the third numerical interval, it is determined that the user to be identified is an abnormal transaction user.

7. The method according to any one of claims 1 to 5, characterized in that After obtaining the abnormal transaction probability corresponding to the user to be identified, the method further includes: Determining a first accuracy rate of a first recognition model for recognizing a locally stored user transaction dataset, wherein a training parameter corresponding to the first recognition model is a first training parameter; determining a second accuracy rate of the target recognition model in recognizing the locally stored user transaction dataset; In a case where the first accuracy rate is greater than the second accuracy rate, a request to reject use of the target recognition model is initiated to the cloud service device.

8. A device for identifying the probability of abnormal user transactions based on blockchain, characterized in that: include: a training module, configured to iteratively train an initial transaction recognition model using a locally stored user transaction dataset to obtain a first transaction recognition model corresponding to the trained initial transaction recognition model, wherein the initial recognition model is downloaded from a cloud service device and is used to identify the transaction probability of abnormal user transactions; a transmission module, configured to upload the first training parameters carried by the first transaction recognition model to the cloud service device through encrypted transmission; a downloading module, configured to download second training parameters from the cloud service device upon receiving an update instruction fed back by the cloud service device, wherein the second training parameters are obtained by the cloud service device adjusting initial parameters of the initial model based on training parameters uploaded by multiple bank terminals; The identification module is used to adjust the first transaction identification model according to the second training parameters to obtain a target transaction identification model, identify the locally updated credit card transaction data of the user to be identified through the target identification model, and obtain the abnormal transaction probability corresponding to the user to be identified.

9. A device for identifying the probability of abnormal user transactions based on blockchain, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

Images