Data processing method and device, equipment, storage medium and program product

Abstract

The invention provides a data processing method which can be applied to the technical field of big data and the technical field of financial science and technology. The data processing method comprises the following steps: based on N types of historical transaction behavior time sequence data sets, extracting a convolutional layer through N first features, and outputting N first time sequence feature vectors; and outputting a transaction behavior prediction result through a first time sequence prediction layer based on the N first time sequence feature vectors, based on the M types of historical fund change time sequence data sets, extracting a convolutional layer through M second features, and outputting M second time sequence feature vectors; and outputting a fund change prediction result through a second time sequence prediction layer based on the M second time sequence feature vectors, and generating a data processing result for representing the transaction risk degree based on the transaction behavior prediction result and the fund change prediction result. The invention further provides a data processing device, equipment, a storage medium and a program product.

Description

Technical Field

[0001] The present disclosure relates to the fields of big data and financial technology, and more specifically to a data processing method, device, equipment, storage medium and program product. Background Art

[0002] Users usually withdraw money or conduct other transactions through preset accounts. However, withdrawals take a long time and the transaction needs to be safe. Although related technologies use intelligent recognition methods such as voiceprints and liveness, or increase online verification operations to ensure the safety of transactions, if users lack financial security awareness, it will also cause losses to users. Related technologies also use prediction models to ensure the safety of transactions. For example, a pre-trained prediction model is used to determine the predicted value of fund deviation. The predicted value is used to indicate the degree of deviation of the number of large-amount fund transactions in a preset time relative to the number of large-amount fund transactions in the same period in history, thereby determining whether the fund limit and the number of large-amount fund transactions are abnormal.

[0003] In the process of realizing the concept of the present disclosure, the inventors found that there are at least the following problems in the related technology: the related methods for ensuring the security of transaction behaviors are not convenient enough, the security is low, and the level of big data is insufficient. Summary of the invention

[0004] In view of the above problems, the present disclosure provides a data processing method, apparatus, device, storage medium and program product.

[0005] According to a first aspect of the present disclosure, a data processing method is provided, comprising: based on N types of pre-acquired historical transaction behavior time series data sets, outputting N first time series feature vectors through N first feature extraction convolution layers corresponding one-to-one to the N types of historical transaction behavior time series data sets; based on the N first time series feature vectors, outputting transaction behavior prediction results through a first time series prediction layer; based on M types of pre-acquired historical funds change time series data sets, outputting M second time series feature vectors through M second feature extraction convolution layers corresponding one-to-one to the M types of historical funds change time series data sets; based on the M second time series feature vectors, outputting funds change prediction results through a second time series prediction layer; based on the transaction behavior prediction results and the funds change prediction results, generating a data processing result for characterizing the degree of transaction risk.

[0006] According to an embodiment of the present disclosure, N types of historical transaction behavior time series data sets include at least one of the following: historical income data sets, historical consumption data sets, and historical repayment data sets; M types of historical funds change time series data sets include at least one of the following: historical single balance change data sets, historical cumulative balance change data sets, historical transfer-in transaction data sets, and historical transfer-out transaction data sets.

[0007] According to an embodiment of the present disclosure, based on the transaction behavior prediction results and the capital change prediction results, generating data processing results for characterizing the degree of transaction risk includes: calculating the deviation of the transaction behavior prediction results relative to the capital change prediction results; generating a first type of data processing result when the deviation meets a preset numerical condition; and generating a second type of data processing result when the deviation does not meet the preset numerical condition.

[0008] According to an embodiment of the present disclosure, the data processing method further includes generating alarm information based on the first type of data processing results.

[0009] According to an embodiment of the present disclosure, outputting N first time series feature vectors includes: performing a first mapping conversion process on N types of historical transaction behavior time series data sets, respectively, to obtain N historical transaction behavior feature data sets corresponding one-to-one to the N types of historical transaction behavior time series data sets; inputting the N historical transaction behavior feature data sets into N first feature extraction convolution layers one-to-one, and outputting N first time series feature vectors.

[0010] According to an embodiment of the present disclosure, outputting M second time series feature vectors includes: performing a second mapping conversion process on M types of historical funds change time series data sets, respectively, to obtain M historical funds change feature data sets corresponding one-to-one to the M types of historical funds change time series data sets; inputting the M historical funds change feature data sets into M second feature extraction convolution layers one-to-one, and outputting M second time series feature vectors.

[0011] According to an embodiment of the present disclosure, based on N first time series feature vectors, through the first time series prediction layer, outputting the transaction behavior prediction result includes: performing a first splicing operation on the N first time series feature vectors to obtain a first target feature vector; inputting the first target feature vector into the first time series prediction layer, and outputting the transaction behavior prediction result.

[0012] According to an embodiment of the present disclosure, based on M second time series feature vectors, through the second time series prediction layer, outputting the prediction result of capital changes includes: performing a second splicing operation on the M second time series feature vectors to obtain a second target feature vector; inputting the second target feature vector into the second time series prediction layer, and outputting the prediction result of capital changes.

[0013] The second aspect of the present disclosure provides a data processing device, including: a first output module, which is used to output N first time series feature vectors based on N types of historical transaction behavior time series data sets acquired in advance, through N first feature extraction convolution layers corresponding one-to-one to the N types of historical transaction behavior time series data sets; a second output module, which is used to output transaction behavior prediction results through a first time series prediction layer based on the N first time series feature vectors; and a third output module, which is used to output M second time series feature vectors based on M types of historical funds change time series data sets acquired in advance, through M second feature extraction convolution layers corresponding one-to-one to the M types of historical funds change time series data sets; a fourth output module, which is used to output funds change prediction results through a second time series prediction layer based on the M second time series feature vectors; and a generation module, which is used to generate a data processing result for characterizing the degree of transaction risk based on the transaction behavior prediction results and the funds change prediction results.

[0014] A third aspect of the present disclosure provides an electronic device, comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the above method.

[0015] The fourth aspect of the present disclosure further provides a computer-readable storage medium having a computer program or instructions stored thereon, which implements the steps of the above method when the above computer program or instructions are executed by a processor.

[0016] The fifth aspect of the present disclosure further provides a computer program product, including a computer program or instructions, which implement the steps of the above method when the above computer program or instructions are executed by a processor.

[0017] According to the embodiments of the present disclosure, by outputting N first time series feature vectors through N first feature extraction convolution layers, and by outputting M second time series feature vectors through M second feature extraction convolution layers, it is possible to focus on the self-correlation within different types of data, and fully explore the distribution laws and characteristics of various types of data themselves; by outputting the transaction behavior prediction results through the first time series prediction layer, and by outputting the capital change prediction results through the second time series prediction layer, it is possible to fully focus on the global information of the historical transaction behavior time series data and the historical capital change time series data, thereby improving the accuracy of the transaction behavior prediction results and the capital change prediction results, and at the same time, improving the convenience and big data degree of the data processing method. At the same time, the data processing results are generated based on the transaction behavior prediction results and the capital change prediction results, which can improve the accuracy of the data processing results. Therefore, at least partially overcome the technical problems of insufficient convenience, low security, and insufficient big data degree of the method for ensuring the safety of transaction behavior, thereby achieving the technical effect of improving the convenience, big data degree and accuracy of the data processing method. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above contents and other purposes, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0019] Figure 1 The application scenario diagram of the data processing method, apparatus, device, medium and program product according to the embodiments of the present disclosure is schematically shown;

[0020] Figure 2 A flowchart schematically shows a data processing method according to an embodiment of the present disclosure;

[0021] Figure 3 A flowchart schematically shows a data processing method according to another embodiment of the present disclosure;

[0022] Figure 4 A schematic diagram schematically shows a schematic diagram of outputting transaction behavior prediction results according to an embodiment of the present disclosure;

[0023] Figure 5 A schematic diagram schematically shows a result of outputting a prediction of fund changes according to an embodiment of the present disclosure;

[0024] Figure 6 A structural block diagram schematically shows a data processing device according to an embodiment of the present disclosure; and

[0025] Figure 7 A block diagram of an electronic device suitable for implementing a data processing method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0026] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0027] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0028] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0029] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0030] It should be noted that the data processing method and device disclosed herein can be used in the fields of big data technology and financial technology technology, and can also be used in any field other than the fields of big data technology and financial technology technology. The application field of the data processing method and device disclosed herein is not limited.

[0031] In the technical solution of the present disclosure, the user information (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with relevant laws, regulations and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0032] In the scenario of using personal information for automated decision-making, the methods, devices, and systems provided by the embodiments of the present disclosure provide users with corresponding operation portals for users to choose to agree or reject the automated decision-making results; if the user chooses to reject, the expert decision-making process will be entered. The expression "automated decision-making" here refers to the activity of automatically analyzing and evaluating a person's behavioral habits, interests and hobbies, or economic, health, credit status, etc. through computer programs, and making decisions. The expression "expert decision-making" here refers to the activity of making decisions by people who specialize in a certain field, have specialized experience, knowledge and skills, and have reached a certain level of professionalism.

[0033] The transaction behavior of the user includes withdrawal behavior, deposit behavior, etc. It is usually necessary to predict the transaction behavior of the user to determine the degree of transaction risk corresponding to the transaction behavior. In the process of realizing the concept of the present disclosure, the inventor found that there are at least the following problems in the related technology: on the one hand, the convenience of the related methods for ensuring the safety of transaction behavior is insufficient, for example: the related methods ensure the safety of transactions by limiting the time of transactions such as transfers, but it will bring inconvenience to the user's daily use; on the other hand, the availability of the related methods for ensuring the safety of transaction behavior has certain limitations, for example, for a single transaction, the related methods can only verify whether it is operated by the person himself, if not operated by the person himself, it is a high-risk behavior, or a single large transaction is checked and restricted, but it is impossible to verify whether the above transaction has high-risk behavior assisted by the user, so the security is low; in addition, the big data level of the related methods for ensuring the safety of transaction behavior is insufficient, for example, in the case of modeling supervision of large-scale fund transfers, the technology used is relatively old and cannot cope with abnormal operations such as small-amount cumulative transfers.

[0034] In view of this, an embodiment of the present disclosure provides a data processing method, including: based on N types of pre-acquired historical transaction behavior time series data sets, through N first feature extraction convolution layers corresponding one-to-one to the N types of historical transaction behavior time series data sets, output N first time series feature vectors; based on the N first time series feature vectors, through a first time series prediction layer, output transaction behavior prediction results; based on M types of pre-acquired historical funds change time series data sets, through M second feature extraction convolution layers corresponding one-to-one to the M types of historical funds change time series data sets, output M second time series feature vectors; based on the M second time series feature vectors, through a second time series prediction layer, output funds change prediction results; based on the transaction behavior prediction results and the funds change prediction results, generate a data processing result for characterizing the degree of transaction risk.

[0035] Figure 1 The application scenario diagram of the data processing method, apparatus, device, medium and program product according to the embodiments of the present disclosure is schematically shown.

[0036] like Figure 1 As shown, the application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used to provide a medium for a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or optical fiber cables, etc.

[0037] The user can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only for example).

[0038] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0039] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0040] It should be noted that the data processing method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the data processing device provided in the embodiment of the present disclosure can generally be set in the server 105. The data processing method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Accordingly, the data processing device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105.

[0041] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0042] The following will be based on Figure 1 The scene described by Figure 2~Figure 5 The data processing method of the embodiment of the present disclosure is described in detail.

[0043] Figure 2 The flowchart of the data processing method according to the embodiment of the present disclosure is schematically shown.

[0044] like Figure 2 As shown, the data processing method of this embodiment includes operations S210 to S250.

[0045] In operation S210, based on the pre-acquired N types of historical transaction behavior time series data sets, N first time series feature vectors are output through N first feature extraction convolution layers corresponding one-to-one to the N types of historical transaction behavior time series data sets.

[0046] In operation S220, based on the N first time series feature vectors, a transaction behavior prediction result is output through a first time series prediction layer.

[0047] In operation S230, based on the pre-acquired M types of historical fund change time series data sets, M second time series feature vectors are output through M second feature extraction convolution layers corresponding one-to-one to the M types of historical fund change time series data sets.

[0048] In operation S240, based on the M second time series feature vectors, a fund change prediction result is output through a second time series prediction layer.

[0049] In operation S250, a data processing result for characterizing the degree of transaction risk is generated based on the transaction behavior prediction result and the capital change prediction result.

[0050] According to an embodiment of the present disclosure, in an embodiment of the present disclosure, before obtaining the user's information, the user's consent or authorization may be obtained. For example, before operation S210, a request to obtain the user's information may be issued to the user. When the user agrees or authorizes that the user's information may be obtained, operation S210 is performed.

[0051] According to an embodiment of the present disclosure, in operation S210, multiple historical transaction behavior time series data of the same user within a predetermined historical time interval may be obtained in advance, and the historical transaction behavior time series data may be divided into N historical transaction behavior time series data sets according to K predetermined transaction categories. One historical transaction behavior time series data set corresponds to one predetermined transaction category, and the predetermined transaction category may include income category, consumption category, repayment category, etc.

[0052] For example, multiple historical transaction behavior time series data of a certain user within a predetermined historical time interval are obtained in advance, and the multiple historical transaction behavior time series data include, for example:

[0053] Historical transaction behavior time series data 1, including: consumption of 500 yuan on the first day of month X;

[0054] Historical transaction behavior time series data 2, including: consumption of 400 yuan on month 2;

[0055] Historical transaction behavior time series data 3, including: income of 500 yuan on the first day of month X;

[0056] Historical transaction behavior time series data 4, including: income of 400 yuan on month 3.

[0057] Three predetermined transaction categories can be pre-set: income category, consumption category, and repayment category. Based on the above three predetermined transaction categories, the above historical transaction behavior time series data 1 to 4 are divided to obtain two historical transaction behavior time series data sets, wherein the historical transaction behavior time series data set A includes the historical transaction behavior time series data 1 to 2, and the corresponding transaction category is consumption category; the historical transaction behavior time series data set B includes the historical transaction behavior time series data 3 to 4, and the corresponding transaction category is income category. Since the multiple historical transaction behavior time series data do not include repayment data, the repayment category historical transaction behavior time series data set is not generated.

[0058] According to an embodiment of the present disclosure, the historical transaction behavior time series data set can be locally preprocessed. The localized preprocessing can be performed by outputting N first time series feature vectors through N first feature extraction convolution layers corresponding one-to-one to N types of historical transaction behavior time series data sets.

[0059] Specifically, the first feature extraction convolutional layer can be a temporal prediction convolutional layer (Temporal Convolutional Attention Network, abbreviated as TCAN). Since different types of historical trading behavior time series data sets have different characteristics, if N types of historical trading behavior time series data sets are directly input into the first time series prediction layer at the same time, the self-correlation within the different types of historical trading behavior time series data sets will be ignored, and the characteristics and laws of the various types of historical trading behavior time series data sets themselves cannot be fully mined. However, by first inputting N types of historical trading behavior time series data sets into one-to-one corresponding N first feature extraction convolutional layers, the entire time series of various types of historical trading behavior data can be independently embedded into the features, expanding the local receptive field and paying attention to the self-correlation within different types of data. As a result, the local features of different types of historical trading behavior time series data sets can be extracted respectively, so as to fully mine the characteristics and laws of various types of historical trading behavior time series data sets themselves.

[0060] According to an embodiment of the present disclosure, in operation S220, based on the N first time series feature vectors, through the first time series prediction layer, outputting the transaction behavior prediction result can be: splicing the N first time series feature vectors, and inputting the spliced ​​time series feature vectors into the first time series prediction layer, thereby outputting the transaction behavior prediction result. Specifically, the first time series prediction layer can be a transormer model. The transormer model can model the global information of the historical transaction behavior time series data, thereby capturing the relationship across variables and better performing feature processing.

[0061] When using the transformer model for time series prediction, local vision is usually ignored due to excessive globalization. Therefore, when using the transformer directly to process the fused data, for example, when N types of historical trading behavior time series data sets are directly input into the transformer model, the self-correlation within different types of historical trading behavior time series data sets will be ignored, resulting in inaccurate output trading behavior prediction results. However, based on the N types of historical trading behavior time series data sets, N first feature extraction convolution layers corresponding to the N types of historical trading behavior time series data sets are used to output N first time series feature vectors, so that the entire time series of each variable can be independently embedded into the feature to expand the local receptive field; based on the N first time series feature vectors, the trading behavior prediction results are output through the first time series prediction layer, which can overcome the previous transformer's poor performance in time series prediction due to excessive globalization and neglect of local vision, thereby better extracting information and avoiding the problem of missing associated information due to excessive globalization and mismatched timestamps caused by embedding data early.

[0062] According to the embodiments of the present disclosure, predicting transaction risk based only on transaction behavior time series data may result in inaccurate transaction risk prediction results. To improve the accuracy of transaction risk prediction results, transaction risk can be predicted based on both transaction behavior time series data and capital change data.

[0063] Specifically, in operation S230, multiple historical fund change time series data of the same user within a predetermined historical time interval may be obtained in advance, and the multiple historical fund change time series data may be divided into M types of historical fund change time series data sets according to L predetermined fund change categories. Among them, one historical fund change time series data set corresponds to one predetermined fund change category. The predetermined fund change category may include, for example, a transfer-in transaction category, a transfer-out transaction category, etc.

[0064] According to an embodiment of the present disclosure, in an embodiment of the present disclosure, before obtaining the user's information, the user's consent or authorization may be obtained. For example, before operation S230, a request to obtain the user's information may be issued to the user. When the user agrees or authorizes that the user's information may be obtained, operation S230 is performed.

[0065] According to an embodiment of the present disclosure, in operation S230 to operation S240, the second feature extraction convolution layer may be a second time series prediction convolution layer. In order to extract the type features of each type of historical fund change data set, the M types of historical fund change data sets may be input into the M second feature extraction convolution layers one by one, and M second time series feature vectors may be output, so that the characteristics of each type of historical fund change data set may be fully mined. After outputting the M second time series feature vectors, in order to extract the global information corresponding to the M types of historical fund change time series data sets, the M types of historical fund change time series data sets may be input into the second time series prediction layer. Among them, the second time series prediction layer may be a transformer model.

[0066] According to an embodiment of the present disclosure, in operation S250, the transaction behavior prediction result can be used to predict the type and transaction amount of the user's transaction behavior; the fund change prediction result can be used to predict the user's transfer amount and total funds. The transfer amount can include the transfer amount and the transfer amount, and the total funds can include the total account balance.

[0067] According to an embodiment of the present disclosure, based on the transaction behavior prediction results and the capital change prediction results, the data processing results generated to characterize the degree of transaction risk may include, for example: when the transaction behavior prediction results obviously deviate from the capital change prediction results, the data processing results are used to characterize that the transaction risk is higher; when the transaction behavior prediction results are relatively close to the capital change prediction results, the data processing results are used to characterize that the transaction risk is lower.

[0068] Among them, the transaction behavior prediction result obviously deviates from the capital change prediction result, for example, the transaction amount obviously deviates from the total account balance. Specifically, the deviation between the transaction amount and the total account balance can be calculated. When the deviation between the transaction amount and the total account balance is greater than a preset deviation threshold, a data processing result is generated to characterize the transaction risk as high risk; correspondingly, when the deviation between the transaction amount and the total account balance is less than or equal to the preset deviation threshold, a data processing result is generated to characterize the transaction risk as low risk.

[0069] For example: based on the N-type historical transaction behavior time series data set of user A, a transaction behavior prediction result is output. Based on the above transaction behavior prediction result, it can be predicted that user A's next transaction behavior will be a deduction of 5,000 yuan; based on the M-type historical fund change time series data set of user A, a fund change prediction result is output. Based on the above fund change prediction result, it can be predicted that user A's total fund balance is 1,000 yuan. Since the transaction behavior prediction result is significantly higher than the fund change prediction result, the transaction risk of user A's next transaction is higher.

[0070] According to the embodiments of the present disclosure, N first time series feature vectors are outputted through N first feature extraction convolution layers corresponding to N types of historical transaction behavior time series data sets, and M second time series feature vectors are outputted through M second feature extraction convolution layers corresponding to M types of historical capital change data sets, so that the self-correlation within different types of data can be paid attention to, and the distribution law and characteristics of each type of data can be fully explored; then, the transaction behavior prediction results are outputted through the first time series prediction layer, and the capital change prediction results are outputted through the second time series prediction layer, so that the global information of the historical transaction behavior time series data and the historical capital change time series data can be fully paid attention to, and the cross-variable relationship can be captured, and the feature processing can be better performed, thereby improving the accuracy of the transaction behavior prediction results and the capital change prediction results, and at the same time, the convenience of the data processing method can be improved, and the big data degree of the data processing method can be improved. In addition, the data processing results are generated based on the transaction behavior prediction results and the capital change prediction results at the same time, which can improve the accuracy of the data processing results.

[0071] According to an embodiment of the present disclosure, N types of historical transaction behavior time series data sets include at least one of the following: historical income data sets, historical consumption data sets, and historical repayment data sets; M types of historical funds change time series data sets include at least one of the following: historical single balance change data sets, historical cumulative balance change data sets, historical transfer-in transaction data sets, and historical transfer-out transaction data sets.

[0072] According to the embodiments of the present disclosure, the transaction behavior data of the user's income, consumption, deduction, etc. within a predetermined time interval can be obtained, and the above transaction data can be converted into a time series behavior sequence. The above transaction behavior data can be organized into a unified JSON format to obtain a historical transaction behavior time series data set.

[0073] According to an embodiment of the present disclosure, a historical income data set includes at least one of the following: income amount, income transaction generation time, income transaction category, and time interval between each income transaction data. Among them, the income transaction category may include, for example, salary category, branch manuscript fee category, etc. A historical consumption data set may include at least one of the following: consumption amount, consumption time, merchant category code, and time interval between each consumption transaction data; a historical repayment data set may include at least one of the following: repayment amount, repayment time, repayment transaction category, and time interval between each repayment transaction data. Among them, the repayment transaction category may include, for example, credit card repayment, loan repayment, utility deduction, and other categories.

[0074] According to the embodiments of the present disclosure, it is also possible to obtain funds change data such as the user's account balance within a predetermined time interval, the cumulative account balance change data, the historical maximum number of funds transferred in, the historical maximum number of funds transferred out, the maximum amount of funds transferred into the account, the historical maximum amount of funds transferred out, etc., and convert the above-mentioned funds change data into a time series behavior sequence, and the above-mentioned funds information can be organized into a unified JSON format to obtain a historical funds change time series data set.

[0075] According to an embodiment of the present disclosure, the historical single balance change data set includes at least one of the following: total balance, balance change time, and time interval between each single balance change data.

[0076] The historical cumulative balance change data set may include at least one of the following: the cumulative change total balance, the cumulative balance change time, and the time interval between each cumulative balance change data.

[0077] The historical transfer-in transaction data set may include at least one of the following: transfer-in amount, transfer-in time, transfer-in account number, and time interval between each transfer-in data. Specifically, the historical transfer-in transaction data set may include: the highest transfer-in amount within a predetermined period, and the transfer-in time, transfer-in account number, and time interval between each transfer-in data corresponding to the highest transfer-in amount within the predetermined period; and / or the transfer-in amount with the highest frequency of occurrence within the predetermined period, and the transfer-in time, transfer-in account number, and time interval between each transfer-in data corresponding to the transfer-in amount with the highest frequency of occurrence within the predetermined period.

[0078] The historical outgoing transaction data set includes at least one of the following: an outgoing amount, an outgoing time, an outgoing account number, and a time interval between each outgoing data. Specifically, the historical outgoing transaction data set may include: the highest outgoing amount within a predetermined period, and the outgoing time, the outgoing account number, and the time interval between each outgoing data corresponding to the highest outgoing amount within the predetermined period; and / or the outgoing amount with the highest frequency within the predetermined period, and the outgoing time, the outgoing account number, and the time interval between each outgoing data corresponding to the outgoing amount with the highest frequency within the predetermined period.

[0079] According to an embodiment of the present disclosure, based on the transaction behavior prediction results and the capital change prediction results, generating data processing results for characterizing the degree of transaction risk includes: calculating the deviation of the transaction behavior prediction results relative to the capital change prediction results; generating a first type of data processing result when the deviation meets a preset numerical condition; and generating a second type of data processing result when the deviation does not meet the preset numerical condition.

[0080] According to an embodiment of the present disclosure, calculating the deviation of the transaction behavior prediction results relative to the funds change prediction results may include: calculating the similarity between the transaction behavior prediction results and the funds change prediction results, and determining the deviation of the transaction behavior prediction results relative to the funds change prediction results based on the similarity.

[0081] According to an embodiment of the present disclosure, the deviation degree satisfies the preset numerical condition that the deviation degree is greater than or equal to a preset threshold value. When the deviation degree is greater than or equal to the preset threshold value, it indicates that the transaction behavior prediction result and the capital change prediction result are greatly different, indicating that the transaction risk may be greater, and therefore a first type of data processing result can be generated to characterize a higher transaction risk.

[0082] Correspondingly, the deviation not meeting the preset numerical condition may be that the deviation is less than a preset threshold. When the deviation is less than the preset threshold, it indicates that the difference between the transaction behavior prediction result and the capital change prediction result is small, indicating that the transaction risk may be small, so the second type of data processing result can be generated, and the second type of data processing result can represent that the transaction risk is low.

[0083] For example, the preset threshold is 10%. When the transaction behavior prediction result is higher or lower than the fund change prediction result, and the deviation of the transaction behavior prediction result from the fund change prediction result is greater than 10%, it means that the transaction behavior prediction result differs greatly from the fund change prediction result, so the first type of data processing result can be generated. The first type of data processing result is used to characterize that the transaction risk is high, indicating that the transaction may be an abnormal transaction.

[0084] According to the embodiments of the present disclosure, data processing results are generated based on the deviation of trading behavior prediction results from capital change prediction results and preset numerical conditions, rather than directly generating data processing results based only on trading behavior prediction results or capital change prediction results. This can fully consider the historical trading behavior patterns of users and improve the accuracy of data processing results. By judging whether the deviation meets the preset numerical conditions to determine whether to generate the first type of data processing results or the second type of data processing results, trading risks can be flexibly determined.

[0085] Figure 3 The flowchart of the data processing method according to another embodiment of the present disclosure is schematically shown. Figure 3 As shown, the data processing method of this embodiment includes operations S310 to S370.

[0086] In operation S310, transaction behavior data within a predetermined time interval is obtained.

[0087] In operation S320, the transaction behavior data within the predetermined time interval is classified and processed based on the predetermined transaction category to generate a historical transaction behavior time series data set.

[0088] In operation S330, based on the historical transaction behavior time series data set, a transaction behavior prediction result is output.

[0089] In operation S340, capital change data within a predetermined time interval is obtained.

[0090] In operation S350, the capital change data within the predetermined time interval is classified and processed based on the predetermined capital change category to generate a historical capital change time series data set.

[0091] In operation S360, based on the historical fund change time series data set, a fund change prediction result is output.

[0092] In operation S370, a data processing result is generated based on the transaction behavior prediction result and the capital change prediction result.

[0093] According to an embodiment of the present disclosure, the data processing method further includes: generating alarm information based on the first type of data processing results.

[0094] According to the embodiment of the present disclosure, when the first type of data processing result is generated, it indicates that the user may conduct an abnormal transaction, so an alarm message can be generated. Based on the alarm message, the staff can solve the abnormal transaction according to the predetermined business operation rules.

[0095] Specifically, after the alarm information is generated, the user's current transaction can be frozen urgently, and the staff can be reminded to handle the transaction offline. For example, the transaction can be quickly handled by adding verification code verification, human body detection authentication, etc., or by freezing the account in time. Relevant staff can also be reminded to intervene manually. For example, relevant staff can judge the specific situation of the transaction through telephone communication. If the risk of the transaction is manually determined to be within the normal range, the transaction can be processed normally.

[0096] According to an embodiment of the present disclosure, when generating the first type of data processing results, the quality of the historical transaction behavior time series data set and the quality of the historical funds change time series data set can be controlled by staff with data processing authority. For example, the historical transaction behavior time series data set and the historical funds change time series data set can be reviewed and screened to determine whether there is abnormal data in the historical transaction behavior time series data set and the historical funds change time series data set. Abnormal data may, for example, include historical transaction behavior time series data and historical funds change time series data corresponding to abnormal exchanges, and may also include historical transaction behavior time series data and historical funds change time series data corresponding to sudden transaction situations. Among them, abnormal transactions may be transactions with higher transaction risks. Sudden transaction situations may be transactions with lower transaction frequencies and lower risks, such as buying cars and houses.

[0097] In the case of abnormal data, the historical transaction behavior time series data set and the historical capital change time series data set can be optimized. For example, the abnormal data can be removed from the historical transaction behavior time series data set and the historical capital change time series data set to obtain an optimized historical transaction behavior time series data set and an optimized historical capital change time series data set. The first feature extraction convolution layer, the second feature extraction convolution layer, the first time series prediction layer and the second time series prediction layer can be retrained based on the optimized historical transaction behavior time series data set and the optimized historical capital change time series data set.

[0098] In the absence of abnormal data, that is, when the transaction is normal, the historical transaction behavior time series data set and the historical fund change time series data set can be directly used as training sample sets to train the first feature extraction convolution layer, the second feature extraction convolution layer, the first time series prediction layer and the second time series prediction layer, so as to calibrate the first feature extraction convolution layer, the second feature extraction convolution layer, the first time series prediction layer and the second time series prediction layer. Among them, normal transactions can include regular transfers, etc.

[0099] According to the embodiments of the present disclosure, alarm information is generated based on the first type of data processing results, so that abnormal transactions can be discovered in time, thereby reducing transaction risks and ensuring account security.

[0100] According to an embodiment of the present disclosure, training the first feature extraction convolution layer, the second feature extraction convolution layer, the first time series prediction layer and the second time series prediction layer may also include: obtaining a time series data set of historical transaction behavior within a reference time period to obtain a first training sample set; using the time series data set of historical transaction behavior in the target time period as a first label; training the first feature extraction convolution layer, the second feature extraction convolution layer, the first time series prediction layer and the second time series prediction layer based on the first label and the first training sample set; and outputting the trained first feature extraction convolution layer, the second feature extraction convolution layer, the first time series prediction layer and the second time series prediction layer after the model converges.

[0101] Correspondingly, training the second feature extraction convolution layer, the second feature extraction convolution layer, the second time series prediction layer and the second time series prediction layer can also include: obtaining a historical capital change time series data set within a reference time period to obtain a second training sample set; using the historical capital change time series data set for the target time period as a second label; training the second feature extraction convolution layer, the second feature extraction convolution layer, the second time series prediction layer and the second time series prediction layer based on the second label and the second training sample set; and outputting the trained second feature extraction convolution layer, the second feature extraction convolution layer, the second time series prediction layer and the second time series prediction layer after the model converges.

[0102] According to an embodiment of the present disclosure, outputting N first time series feature vectors includes: performing a first mapping conversion process on N types of historical transaction behavior time series data sets, respectively, to obtain N historical transaction behavior feature data sets corresponding one-to-one to the N types of historical transaction behavior time series data sets; inputting the N historical transaction behavior feature data sets into N first feature extraction convolution layers one-to-one, and outputting N first time series feature vectors.

[0103] According to an embodiment of the present disclosure, the first mapping conversion process may include: performing mapping conversion processing on N types of historical transaction behavior time series data sets respectively through N first shallow feature extraction layers, thereby outputting N historical transaction behavior feature data sets that correspond one-to-one to the N types of historical transaction behavior time series data sets.

[0104] According to the embodiments of the present disclosure, a first shallow feature extraction layer corresponds to only one type of historical transaction behavior feature data set. By inputting N historical transaction behavior feature data sets into N first feature extraction convolution layers one by one, rather than directly inputting all historical transaction behavior feature data sets into one feature extraction layer after merging and processing, the category features of various types of historical transaction behavior feature data sets can be further accurately extracted.

[0105] According to an embodiment of the present disclosure, outputting M second time series feature vectors includes: performing a second mapping conversion process on M types of historical funds change time series data sets, respectively, to obtain M historical funds change feature data sets corresponding one-to-one to the M types of historical funds change time series data sets; inputting the M historical funds change feature data sets into M second feature extraction convolution layers one-to-one, and outputting M second time series feature vectors.

[0106] According to an embodiment of the present disclosure, the second mapping conversion process may include: performing the second mapping conversion process on M types of historical fund change time series data sets respectively through M second shallow feature extraction layers, and obtaining M historical fund change feature data sets corresponding to the M types of historical fund change time series data sets one by one. Compared with directly inputting all historical fund change time series data sets into one feature extraction layer after merging and processing, the category features of various types of historical transaction behavior feature data sets can be accurately extracted through one second shallow feature extraction layer corresponding only to one type of historical fund change time series data set.

[0107] According to an embodiment of the present disclosure, based on N first time series feature vectors, through the first time series prediction layer, outputting the transaction behavior prediction result includes: performing a first splicing operation on the N first time series feature vectors to obtain a first target feature vector; inputting the first target feature vector into the first time series prediction layer, and outputting the transaction behavior prediction result.

[0108] Figure 4 The schematic diagram of outputting the prediction result of transaction behavior according to the embodiment of the present disclosure is schematically shown. Figure 4 As shown, the historical transaction behavior time series data set 1~historical transaction behavior time series data set n can be input into the first shallow feature extraction layer 1~n one by one, so as to output n historical transaction behavior feature data sets; after the n historical transaction behavior feature data sets are input into the first feature extraction convolution layer 1~n one by one, n first time series feature vectors can be output. The first target feature vector can be obtained by performing a first concatenation operation on the first time series feature vector. After the first target feature vector is output to the first time series prediction layer and a full connection operation is performed, the transaction behavior prediction result can be output.

[0109] According to an embodiment of the present disclosure, the first target feature vector obtained after performing a first concatenation operation on N first time series feature vectors is input into the first time series prediction layer, which can extract the global information of the historical transaction behavior data and capture the relationship between various types of historical transaction behavior data, thereby improving the accuracy of the transaction prediction results.

[0110] According to an embodiment of the present disclosure, based on M second time series feature vectors, through the second time series prediction layer, outputting the prediction result of capital changes includes: performing a second splicing operation on the M second time series feature vectors to obtain a second target feature vector; inputting the second target feature vector into the second time series prediction layer, and outputting the prediction result of capital changes.

[0111] Figure 5 The schematic diagram of outputting the prediction result of capital change according to the embodiment of the present disclosure is schematically shown. Figure 5 As shown, after inputting the historical capital change time series data set 1 to the historical capital change time series data set n into the second shallow feature extraction layer 1~n in a one-to-one correspondence, n historical capital change feature data sets can be output; by inputting the n historical capital change feature data sets into the second feature extraction convolution layer 1~n in a two-to-two correspondence, n second time series feature vectors can be output. Performing a second splicing operation on the above second time series feature vectors, a second target feature vector can be obtained. The above second target feature vector is output to the second time series prediction layer, and after a full connection operation is performed, the capital change prediction result is output.

[0112] Based on the above data processing method, the present disclosure also provides a data processing device. Figure 6 The device is described in detail.

[0113] Figure 6 The structure block diagram of a data processing device according to an embodiment of the present disclosure is schematically shown.

[0114] like Figure 6As shown, the data processing device 600 of this embodiment includes a first output module 610 , a second output module 620 , a third output module 630 , a fourth output module 640 and a generating module 650 .

[0115] The first output module 610 is used to output N first time series feature vectors based on the pre-acquired N types of historical transaction behavior time series data sets through N first feature extraction convolution layers corresponding to the N types of historical transaction behavior time series data sets. In one embodiment, the first output module 610 can be used to perform the operation S210 described above, which will not be repeated here.

[0116] The second output module 620 is used to output the transaction behavior prediction result through the first time series prediction layer based on the N first time series feature vectors. In one embodiment, the second output module 620 can be used to perform the operation S220 described above, which will not be repeated here.

[0117] The third output module is used to output M second time series feature vectors based on the pre-acquired M types of historical fund change time series data sets through M second feature extraction convolution layers corresponding to the M types of historical fund change time series data sets. In one embodiment, the third output module 630 can be used to perform the operation S230 described above, which will not be repeated here.

[0118] The fourth output module is used to output the fund change prediction result through the second time series prediction layer based on the M second time series feature vectors. In one embodiment, the fourth output module 640 can be used to perform the operation S240 described above, which will not be repeated here.

[0119] The generation module is used to generate data processing results for characterizing the degree of transaction risk based on the transaction behavior prediction results and the capital change prediction results. In one embodiment, the generation module 650 can be used to perform the operation S250 described above, which will not be repeated here.

[0120] According to an embodiment of the present disclosure, the generation module includes a calculation submodule, a first generation submodule and a second generation submodule.

[0121] The calculation submodule is used to calculate the deviation of the transaction behavior prediction results relative to the capital change prediction results; the first generation submodule is used to generate the first type of data processing results when the deviation meets the preset numerical conditions; the second generation submodule is used to generate the second type of data processing results when the deviation does not meet the preset numerical conditions.

[0122] According to an embodiment of the present disclosure, the generation module also includes a third generation submodule.

[0123] The third generating submodule is used to generate warning information based on the first type of data processing results.

[0124] According to an embodiment of the present disclosure, the first output module includes a first mapping conversion processing submodule and a first output submodule.

[0125] The first mapping conversion processing submodule is used to perform the first mapping conversion processing on N types of historical transaction behavior time series data sets, respectively, to obtain N historical transaction behavior feature data sets corresponding one-to-one to the N types of historical transaction behavior time series data sets; the first output submodule is used to input the N historical transaction behavior feature data sets into N first feature extraction convolution layers one-to-one, and output N first time series feature vectors.

[0126] According to an embodiment of the present disclosure, the third output module includes a second mapping conversion processing submodule and a second output submodule.

[0127] The second mapping conversion processing submodule is used to perform second mapping conversion processing on M types of historical capital change time series data sets, respectively, to obtain M historical capital change feature data sets that correspond one-to-one to the M types of historical capital change time series data sets; the second output submodule is used to input the M historical capital change feature data sets into M second feature extraction convolution layers one-to-one, and output M second time series feature vectors.

[0128] According to an embodiment of the present disclosure, the second output module includes a first splicing operation submodule and a third output submodule.

[0129] The first splicing operation submodule is used to perform a first splicing operation on N first time series feature vectors to obtain a first target feature vector; the third output submodule is used to input the first target feature vector into the first time series prediction layer and output the transaction behavior prediction result.

[0130] According to an embodiment of the present disclosure, the fourth output module includes a second splicing operation submodule and a fourth output submodule.

[0131] The second splicing operation submodule is used to perform a second splicing operation on the M second time series feature vectors to obtain a second target feature vector; the fourth output submodule is used to input the second target feature vector into the second time series prediction layer and output the fund change prediction result.

[0132] According to an embodiment of the present disclosure, any multiple modules of the first output module 610, the second output module 620, the third output module 630, the fourth output module 640 and the generation module 650 can be combined into one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the first output module 610, the second output module 620, the third output module 630, the fourth output module 640 and the generation module 650 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in a suitable combination of any of them. Alternatively, at least one of the first output module 610 , the second output module 620 , the third output module 630 , the fourth output module 640 and the generating module 650 may be at least partially implemented as a computer program module, which may perform a corresponding function when executed.

[0133] Figure 7 A block diagram of an electronic device suitable for implementing a data processing method according to an embodiment of the present disclosure is schematically shown.

[0134] like Figure 7 As shown, the electronic device 700 according to an embodiment of the present disclosure includes a processor 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 to a random access memory (RAM) 703. The processor 701 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 701 may also include an onboard memory for caching purposes. The processor 701 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0135] In RAM 703, various programs and data required for the operation of electronic device 700 are stored. Processor 701, ROM 702 and RAM 703 are connected to each other via bus 704. Processor 701 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 702 and / or RAM 703. It should be noted that the program can also be stored in one or more memories other than ROM 702 and RAM 703. Processor 701 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in one or more memories.

[0136] According to an embodiment of the present disclosure, the electronic device 700 may further include an input / output (I / O) interface 705, which is also connected to the bus 704. The electronic device 700 may further include one or more of the following components connected to the input / output (I / O) interface 705: an input portion 706 including a keyboard, a mouse, etc.; an output portion 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 708 including a hard disk, etc.; and a communication portion 709 including a network interface card such as a LAN card, a modem, etc. The communication portion 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the input / output (I / O) interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage portion 708 as needed.

[0137] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0138] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 702 and / or RAM 703 described above and / or one or more memories other than ROM 702 and RAM 703.

[0139] The embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the data processing method provided by the embodiment of the present disclosure.

[0140] The above functions defined in the system / device of the embodiment of the present disclosure are performed when the computer program is executed by the processor 701. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0141] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 709, and / or installed from the removable medium 711. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0142] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the processor 701, the above functions defined in the system of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0143] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0144] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0145] It will be appreciated by those skilled in the art that the features described in the various embodiments of the present disclosure may be combined and / or combined in a variety of ways, even if such combinations or combinations are not explicitly described in the present disclosure. In particular, without departing from the spirit and teachings of the present disclosure, the features described in the various embodiments of the present disclosure may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0146] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A data processing method, characterized in that: The method comprises: Based on the pre-acquired N types of historical transaction behavior time series data sets, outputting N first time series feature vectors through N first feature extraction convolution layers corresponding one-to-one to the N types of historical transaction behavior time series data sets; Based on the N first time series feature vectors, outputting a transaction behavior prediction result through a first time series prediction layer; Based on the pre-acquired M types of historical fund change time series data sets, outputting M second time series feature vectors through M second feature extraction convolution layers corresponding one-to-one to the M types of historical fund change time series data sets; Based on the M second time series feature vectors, outputting a fund change prediction result through a second time series prediction layer; Based on the transaction behavior prediction results and the capital change prediction results, a data processing result for characterizing the degree of transaction risk is generated.

2. The method according to claim 1, characterized in that The N types of historical transaction behavior time series data sets include at least one of the following: a historical income data set, a historical consumption data set, and a historical repayment data set; The M-type historical funds change time series data set includes at least one of the following: a historical single balance change data set, a historical cumulative balance change data set, a historical transfer-in transaction data set, and a historical transfer-out transaction data set.

3. The method according to claim 1, characterized in that The generating of the data processing result for characterizing the degree of transaction risk based on the transaction behavior prediction result and the capital change prediction result comprises: Calculating the deviation of the transaction behavior prediction result relative to the capital change prediction result; When the deviation satisfies a preset numerical condition, a first type of data processing result is generated; When the deviation does not satisfy the preset numerical condition, a second type of data processing result is generated.

4. The method according to claim 3, characterized in that The method further comprises: Based on the first type of data processing results, alarm information is generated.

5. The method according to claim 1, characterized in that The outputting of N first time series feature vectors comprises: Performing a first mapping conversion process on the N types of historical transaction behavior time series data sets respectively to obtain N historical transaction behavior feature data sets corresponding one-to-one to the N types of historical transaction behavior time series data sets; The N historical transaction behavior feature data sets are input into the N first feature extraction convolution layers one by one, and the N first time series feature vectors are output.

6. The method according to claim 1, characterized in that The outputting M second time series feature vectors comprises: Performing a second mapping conversion process on the M types of historical fund change time series data sets respectively, to obtain M historical fund change feature data sets corresponding one-to-one to the M types of historical fund change time series data sets; The M historical funds change feature data sets are input into the M second feature extraction convolution layers one by one, and the M second time series feature vectors are output.

7. The method according to claim 1, characterized in that The outputting of the transaction behavior prediction result based on the N first time series feature vectors through the first time series prediction layer includes: Performing a first concatenation operation on the N first time series feature vectors to obtain a first target feature vector; The first target feature vector is input into the first time series prediction layer, and the transaction behavior prediction result is output.

8. The method according to claim 1, characterized in that The outputting of the fund change prediction result based on the M second time series feature vectors through the second time series prediction layer includes: Performing a second concatenation operation on the M second time series feature vectors to obtain a second target feature vector; The second target feature vector is input into the second time series prediction layer, and the fund change prediction result is output.

9. A data processing device, characterized in that: The device comprises: A first output module is used to output N first time series feature vectors based on N types of historical transaction behavior time series data sets acquired in advance through N first feature extraction convolution layers corresponding one-to-one to the N types of historical transaction behavior time series data sets; A second output module, configured to output a transaction behavior prediction result through a first time series prediction layer based on the N first time series feature vectors; A third output module is used to output M second time series feature vectors based on the pre-acquired M types of historical capital change time series data sets through M second feature extraction convolution layers corresponding to the M types of historical capital change time series data sets one by one; a fourth output module, configured to output a fund change prediction result through a second time series prediction layer based on the M second time series feature vectors; and A generation module is used to generate a data processing result for characterizing the degree of transaction risk based on the transaction behavior prediction result and the capital change prediction result.

10. An electronic device comprising: one or more processors; a memory for storing one or more computer programs, It is characterized in that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images