Data processing method, data processing apparatus, electronic device, and medium

Abstract

The present disclosure provides a data processing method, a data processing device, equipment and a medium, which can be applied to the field of big data processing and financial technology. The method comprises: in response to a data processing request, obtaining a plurality of to-be-processed data from a plurality of target databases of different application platforms, wherein the to-be-processed data represents transaction data generated by a transaction behavior; performing analysis and standardization processing on the plurality of to-be-processed data to obtain transaction feature values; processing the transaction feature values by using a classification recognition model to obtain transaction recognition results, wherein the transaction recognition results include business type data representing whether an initial application log corresponding to the transaction behavior needs to be updated and transaction account information; generating application change data according to the business type data and the transaction account information; and updating the initial application log by using the application change data to obtain a target application log.

Description

Technical Field

[0001] This disclosure relates to the fields of big data processing and financial technology, specifically to a data processing method, data processing device, electronic device, computer storage medium, and program product for self-balancing accounting scenarios. Background Technology

[0002] During the transformation of large-scale banking system architecture from centralized mainframes to distributed platforms, the debit and credit entries in a single accounting transaction no longer belong to the same atomic transaction. This makes it impossible to guarantee consistency in posting dates, etc. Although the debit and credit entries are within the same main transaction, they are actually posted by two separate sub-transactions. This type of accounting scenario leads to debit and credit imbalances, which are self-balancing accounting issues. Therefore, the banking system provides a self-balancing netting mechanism to address these imbalances.

[0003] During the adaptive testing phase of the self-balancing accounting scenario, data on the debit and credit sides of a transaction, occurring before and after the day cutoff, is required; that is, debits occur on day T and credits on day T+1. Due to factors such as short data preparation time, inability to control the timing of debits and credits, and difficulty in clearly identifying the product applications to which debits and credits belong, it is difficult to confirm the business attributes of debits and credits in each scenario. Changes need to be made one by one based on the transaction logic, which can lead to errors in the self-balancing accounting data. Therefore, efficiently and quickly preparing and uniformly changing accounting data for self-balancing scenarios is a pressing problem that needs to be solved. Summary of the Invention

[0004] In view of the above problems, this disclosure provides a data processing method, data processing device, electronic device, computer storage medium and program product for self-balancing accounting scenarios.

[0005] According to the first aspect of this disclosure, a data processing method for a self-balancing accounting scenario is provided, including:

[0006] In response to data processing requests, multiple data to be processed are obtained from multiple target databases of different application platforms, wherein the data to be processed represents the transaction data generated by the transaction behavior.

[0007] Multiple datasets to be processed are parsed and standardized to obtain transaction feature values;

[0008] The transaction feature values ​​are processed using a classification and recognition model to obtain the transaction recognition result. The transaction recognition result includes business type data and transaction account information that indicate whether the initial application log corresponding to the transaction behavior needs to be updated.

[0009] Based on business type data and transaction account information, generate application change data;

[0010] The initial application logs are updated using application change data to obtain the target application logs.

[0011] According to embodiments of this disclosure, application change data is generated based on business type data and transaction account information, including:

[0012] Determine the application address where the initial application log that needs to be updated is located based on the business type data;

[0013] Based on the application address, generate switch configuration data according to the business type data;

[0014] Based on the transaction account information, target change data is generated, including application change data, which includes switch configuration data and target change data.

[0015] According to embodiments of this disclosure, updating the initial application log using application change data to obtain the target application log includes:

[0016] The application change data is formatted to obtain the application change script;

[0017] Execute the application change script to update the initial application logs and obtain the target application logs.

[0018] According to embodiments of this disclosure, the application change script includes multiple change sub-scripts;

[0019] The above data processing methods also include:

[0020] When multiple change sub-scripts are executed, the execution result is generated based on the execution result of each change sub-script;

[0021] Display the execution results in a visual format.

[0022] According to embodiments of this disclosure, when executing multiple change sub-scripts, an execution result is generated based on the execution result of each change sub-script, including:

[0023] If any one of the multiple change sub-scripts fails during sequential execution, the update process will be stopped.

[0024] Generate execution results based on the successfully modified sub-scripts and the modified sub-scripts that have stopped updating.

[0025] According to embodiments of this disclosure, the transaction feature values ​​are processed using a classification and recognition model to obtain transaction recognition results, including:

[0026] The transaction feature values ​​are processed by convolutional layers to obtain sequence features;

[0027] The sequence features are processed by global pooling using a pooling layer to obtain the feature vector;

[0028] The feature vectors are processed using a fully connected layer to obtain the transaction recognition results.

[0029] According to embodiments of this disclosure, the transaction feature value is a matrix sequence of m*n, where m is the number of transaction data for a transaction behavior, and n is the number of transaction attributes of the transaction data;

[0030] Specifically, convolutional layers are used to process the transaction feature values ​​to obtain sequence features, including:

[0031] The first convolutional layer uses a convolutional window with a preset stride to traverse the transaction feature values, resulting in a convolutional sequence.

[0032] The second convolutional layer uses a convolutional window with a preset stride to traverse the convolutional sequence and obtain sequence features.

[0033] According to embodiments of this disclosure, the sequence features are globally pooled using a pooling layer to obtain a feature vector, including:

[0034] The sequence features are scanned using a first pooling window of a preset length by the first pooling layer, so as to generate a first window feature corresponding to each first pooling window based on the data in the first pooling window;

[0035] Based on the first window features corresponding to each first pooling window, generate intermediate features corresponding to the first pooling layer;

[0036] The intermediate features are scanned using a second pooling layer based on a second pooling window of a preset length, so as to generate a second window feature corresponding to each second pooling window based on the data in the second pooling window;

[0037] A feature vector is generated based on the second window features corresponding to each second pooling window.

[0038] According to embodiments of this disclosure, the process of using a fully connected layer to process feature vectors and obtain transaction identification results includes:

[0039] Based on multiple preset weight matrices, bias vectors, and feature vectors, classification probabilities corresponding to different preset weight matrices are generated. Each preset weight matrix corresponds to a classification attribute, which includes business type attributes and transaction account attributes.

[0040] The classification attribute corresponding to the maximum classification probability is determined as the transaction identification result.

[0041] According to embodiments of this disclosure, multiple pieces of data to be processed are parsed and standardized to obtain transaction feature values, including:

[0042] Multiple data to be processed are parsed to obtain multiple data attributes corresponding to each data to be processed, where the data attributes represent different behavioral attributes of the transaction behavior;

[0043] For each piece of data to be processed, multiple data attributes are preprocessed to obtain intermediate data;

[0044] The transaction feature values ​​are obtained by standardizing multiple intermediate data.

[0045] According to embodiments of this disclosure, the standardization process of multiple intermediate data to obtain transaction feature values ​​includes:

[0046] Multiple intermediate data points are grouped to obtain multiple data groups;

[0047] Feature extraction is performed on multiple data groups to obtain multiple target feature values;

[0048] The transaction feature values ​​are obtained by standardizing and transforming multiple target feature values.

[0049] According to embodiments of this disclosure, the target feature value includes at least one of text-type feature values ​​and Boolean-type feature values;

[0050] This involves format conversion of multiple target feature values ​​to obtain transaction feature values, including:

[0051] When the target feature value is a text-type feature value, the text-type feature value is format-converted based on the lexical hypothesis method to obtain the first intermediate feature value;

[0052] When the target feature value is a Boolean feature value, perform a binary conversion on the Boolean feature value to obtain the second intermediate feature value;

[0053] The transaction feature values ​​are obtained by standardizing and transforming multiple intermediate feature values.

[0054] According to embodiments of this disclosure, preprocessing multiple data attributes to obtain intermediate data includes:

[0055] If a data attribute is irrelevant, the data attribute is filtered out to obtain multiple filtered data attributes.

[0056] For each filtered data attribute, the format of the filtered data attribute is corrected to obtain an initial data;

[0057] Based on multiple initial data points of the data to be processed, intermediate data corresponding to the data to be processed is generated.

[0058] A second aspect of this disclosure provides a data processing apparatus, comprising:

[0059] The acquisition module is used to respond to data processing requests and acquire multiple data to be processed from multiple target databases of different application platforms. The data to be processed represents the transaction data generated by the transaction behavior.

[0060] The first processing module is used to parse and standardize multiple data to be processed to obtain transaction feature values.

[0061] The second processing module is used to process the transaction feature values ​​using a classification and recognition model to obtain the transaction recognition result. The transaction recognition result includes business type data and transaction account information that indicate whether the initial application log corresponding to the transaction behavior needs to be updated.

[0062] The generation module is used to generate application change data based on business type data and transaction account information;

[0063] The update module is used to update the initial application logs using application change data to obtain the target application logs.

[0064] A third aspect of this disclosure provides an electronic device comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the data processing method described above.

[0065] A fourth aspect of this disclosure also provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to perform the data processing method described above.

[0066] The fifth aspect of this disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described data processing method.

[0067] Based on the data processing methods, apparatus, equipment, media, and program products provided in this disclosure, data to be processed on different platforms is standardized. The resulting transaction feature values ​​are input into a classification and identification model to determine the corresponding business type data and transaction account information. Application change data is generated based on the business type data and transaction account information, and the initial application log is updated using the application change data to obtain the target application log. The classification and identification model can quickly and intelligently identify attributes such as business type, loan account type, and affiliated application, significantly shortening the data change plan generation time. Based on the identification results, the change plan is automatically updated and executed, eliminating the need for manual execution by technical personnel. The entire process of automatic change execution solves the problems of preparing self-balancing accounting data and the difficulty of unified change. Attached Figure Description

[0068] The foregoing contents, as well as other objects, features, and advantages of this disclosure, will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0069] Figure 1 This diagram illustrates an application scenario of the data processing method according to an embodiment of the present disclosure.

[0070] Figure 2 A flowchart illustrating a data processing method according to an embodiment of the present disclosure is shown schematically.

[0071] Figure 3 A flowchart illustrating a method for generating transaction identification results according to an embodiment of the present disclosure is shown schematically.

[0072] Figure 4 A schematic block diagram of a data processing apparatus according to embodiments of the present disclosure is shown; and

[0073] Figure 5 A block diagram schematically illustrates an electronic device suitable for implementing a data processing method according to an embodiment of the present disclosure. Detailed Implementation

[0074] The embodiments of the present disclosure will now 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 disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0075] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0076] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0077] When using expressions such as "at least one of A, B, and C", they should generally be interpreted in accordance with the meaning that is commonly understood by a person skilled in the art (e.g., "a system having at least one of A, B, and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B, and C, etc.).

[0078] In the process of developing this disclosure, it was discovered that the adaptability testing phase of a self-balancing accounting scenario requires data on the debit and credit sides of a transaction, both before and after the daily cutoff. Due to factors such as short data preparation time, inability to control the timing of debit and credit transactions, and difficulty in clearly identifying the product applications to which the debit and credit sides belong, changes need to be confirmed and verified one by one based on the transaction logic. This also leads to errors in the self-balancing accounting data, making data preparation and unified changes quite difficult. Therefore, there is an urgent need for data processing methods specifically for self-balancing accounting scenarios.

[0079] In view of the above, embodiments of this disclosure provide a data processing method, a data processing apparatus, an electronic device, and a medium. The method includes: in response to a data processing request, acquiring multiple data to be processed from multiple target databases of different application platforms, wherein the data to be processed represents transaction data generated by a transaction behavior; parsing and standardizing the multiple data to be processed to obtain transaction feature values; processing the transaction feature values ​​using a classification and recognition model to obtain a transaction recognition result, wherein the transaction recognition result includes business type data and transaction account information indicating whether the initial application log corresponding to the transaction behavior needs to be updated; generating application change data based on the business type data and transaction account information; and updating the initial application log using the application change data to obtain a target application log.

[0080] It should be noted that the data processing methods and devices provided in this disclosure can be used in the financial field, such as financial institutions like banks, or in any field other than the financial field. Therefore, the application areas of the data processing methods and devices provided in this disclosure are not limited.

[0081] In the technical solution of this invention, the user information (including but not limited to user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, and displayed data) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with the relevant laws, regulations, and standards of the relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse.

[0082] Figure 1 The diagram illustrates an application scenario of the data processing method according to an embodiment of the present disclosure.

[0083] like Figure 1 As shown, 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 serves as a medium for providing communication links between the terminal devices 101, 102, 103, and the server 105. The network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0084] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0085] Terminal devices 101, 102, and 103 can be various electronic devices with displays and web browsing capabilities, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0086] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using terminal devices 101, 102, and 103 (for example only). The backend management server can analyze and process data such as received 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 devices.

[0087] It should be noted that the data processing method provided in this embodiment can generally be executed by server 105. Correspondingly, the data processing device provided in this embodiment can generally be located in server 105. The data processing method provided in this embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105. Correspondingly, the data processing device provided in this embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105.

[0088] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0089] The following will be based on Figure 1 The described scene, through Figures 2-3 The data processing method of the disclosed embodiments will be described in detail.

[0090] Figure 2 A flowchart illustrating a data processing method according to an embodiment of the present disclosure is shown schematically.

[0091] like Figure 2 As shown, the method 200 includes operations S210 to S250.

[0092] In operation S210, in response to a data processing request, multiple data to be processed are obtained from multiple target databases of different application platforms, wherein the data to be processed represents the transaction data generated by the transaction behavior.

[0093] According to embodiments of this disclosure, the multiple target databases of the application platform may refer to multiple databases involved in the business of the bank's distributed platform.

[0094] According to embodiments of this disclosure, the data to be processed is transaction data information based on clearly defined lending and borrowing information prepared in advance by business testers and belonging to the self-balancing business scenario type.

[0095] In operation S220, multiple data to be processed are parsed and standardized to obtain transaction feature values.

[0096] According to embodiments of this disclosure, the data to be processed is parsed and standardized. The parsing and standardization process includes retaining data attribute features such as business scenario description, business type, debit application, debit account, debit amount, debit date, debit accounting status, credit application, credit account, credit amount, and credit date.

[0097] In operation S230, the transaction feature values ​​are processed using a classification and recognition model to obtain the transaction recognition result. The transaction recognition result includes business type data and transaction account information that indicate whether the initial application log corresponding to the transaction behavior needs to be updated.

[0098] According to embodiments of this disclosure, the database contains multiple initial application logs, which are used to record business transaction operations.

[0099] According to embodiments of this disclosure, a JDBC database connection is established starting from the initial application logs involving both the lender and borrower to obtain the full amount of transaction data.

[0100] According to embodiments of this disclosure, the classification and identification model can determine the type of data belonging to a new transaction scenario in real time. The transaction identification result includes the lender's account type, business type, lender's application name, and other content.

[0101] According to embodiments of this disclosure, the borrower account type includes corporate accounts, personal accounts, internal accounts, etc.; the business type includes provident fund business, loan business, deposit and withdrawal business, etc.; the borrower application name includes the software name used by the borrower and lender or the counter number used for counter services, etc.

[0102] When operating S240, application change data is generated based on business type data and transaction account information.

[0103] According to embodiments of this disclosure, application change data is generated by using the borrower account type, business type, and borrower application name content generated by the classification and identification model.

[0104] According to embodiments of this disclosure, due to different business scenarios, there may be multiple debits or credits, which may involve changes to multiple application log tables, and application change data of each initial application log is generated sequentially.

[0105] In operation S250, the initial application log is updated using application change data to obtain the target application log.

[0106] According to embodiments of this disclosure, changes are made to each initial application log to obtain the target application log, and the correctness and executability of the changes are automatically checked.

[0107] According to embodiments of this disclosure, data to be processed from different platforms is standardized, and the resulting transaction feature values ​​are input into a classification and recognition model to determine the corresponding business type data and transaction account information. Application change data is generated based on the business type data and transaction account information, and the initial application log is updated using this data to obtain the target application log. The execution status is then displayed in real time. The classification and recognition model can quickly and intelligently identify business types, loan account types, and affiliated applications, significantly shortening the data change plan generation time. The change plan is automatically updated based on the identification results, eliminating the need for manual execution by technical personnel. The entire process of change execution is automated, solving the problems of preparing self-balancing accounting data and the difficulty of unified change management.

[0108] According to embodiments of this disclosure, application change data is generated based on business type data and transaction account information, including:

[0109] Determine the application address where the initial application log that needs to be updated is located based on the business type data;

[0110] Based on the application address, generate switch configuration data according to the business type data;

[0111] Based on the transaction account information, target change data is generated, including application change data, which includes switch configuration data and target change data.

[0112] According to embodiments of this disclosure, based on the identified business type data, the business testing requirements are determined, i.e., whether it is a test host process or a platform process, and then the application address where the initial application log that needs to be updated is located is determined.

[0113] According to embodiments of this disclosure, based on the application address and the identified service type data, corresponding service type switch configuration data is generated. The switch configuration data includes host switches and platform switches. For example, "Service type 22: Host switch set to off, platform switch set to on".

[0114] According to embodiments of this disclosure, based on the identified borrower / debtor account type and borrower / debtor application name, corresponding target change data is generated, namely, change application name, change table name, and other field data. Taking the date of changing the internal account type as an example, the workdate field of the credit or debit details in the log table t_kfdtl_rgs of the internal account type is modified according to certain transaction information conditions.

[0115] According to embodiments of this disclosure, before data modification, it is necessary to check whether the self-balancing process has started. If the self-balancing process has not yet started, data modification can be performed, and the self-balancing process will be automatically submitted after the modification is performed; if the self-balancing process has started, the conditions for modification are not met, and no modification is required.

[0116] According to embodiments of this disclosure, due to different business scenarios, there may be multiple debits or credits, which may involve changes to multiple initial application logs, and application change data for each initial application log is generated sequentially.

[0117] According to embodiments of this disclosure, the generated change data are automatically concatenated in sequence to generate overall application change data, and business personnel can make personalized adjustments to ensure the accuracy of change execution.

[0118] According to embodiments of this disclosure, application change data is generated based on the identified business type data and transaction account information, which includes changing transaction date data, changing application name, and changing table name, thereby balancing the transaction records of borrowers and lenders in the same business scenario.

[0119] According to embodiments of this disclosure, an initial application log is updated using application change data to obtain a target application log, including:

[0120] The application change data is formatted to obtain the application change script;

[0121] Execute the application change script to update the initial application logs and obtain the target application logs.

[0122] According to embodiments of this disclosure, based on the test plan, the application change data of each test plan is converted into application change scripts such as HTTP call interfaces and update / select SQL statements, and the correctness and executability of the generated application change scripts are automatically checked.

[0123] According to embodiments of this disclosure, an application change script is executed to update the initial application log, thereby obtaining the target application log.

[0124] According to embodiments of this disclosure, the application change script includes multiple change sub-scripts.

[0125] According to embodiments of this disclosure, the data processing method further includes:

[0126] When multiple change sub-scripts are executed, the execution result is generated based on the execution result of each change sub-script;

[0127] Display the execution results in a visual format.

[0128] According to embodiments of this disclosure, corresponding change sub-scripts are generated for each application change data. The automatic execution of each change sub-script is triggered according to the sequential relationship of the application change data, and the execution result of the current script is returned in real time. The results are then displayed to staff using visual means such as displays and voice devices, allowing staff to promptly understand any problems encountered during the change process.

[0129] According to embodiments of this disclosure, when executing multiple change sub-scripts, an execution result is generated based on the execution result of each change sub-script, including:

[0130] If any one of the multiple change sub-scripts fails during sequential execution, the update process will be stopped.

[0131] Generate execution results based on the successfully modified sub-scripts and the modified sub-scripts that have stopped updating.

[0132] According to the embodiments of this disclosure, the execution of the next change sub-script is automatically triggered after the expected execution result is met. If the execution of a change sub-script fails, the update is stopped and the successfully executed changes are rolled back.

[0133] According to embodiments of this disclosure, execution results are generated based on successfully modified sub-scripts and modified sub-scripts that have stopped updating.

[0134] According to embodiments of this disclosure, the final test analysis result for this test scenario, i.e., the execution result, is obtained based on the execution status of each change sub-script, and the analysis result is displayed in real time. For example, "Test data change for batch date in scenario XX was successful" or "Test data change for batch date in scenario XX failed, execution was interrupted at step XXX".

[0135] According to the embodiments of this disclosure, in the entire self-balancing accounting scenario, switch settings, change plan formulation, data changes, etc., do not need to be manually executed by technical personnel. The execution of change plans is more automated, and the real-time display of the entire process execution status is conducive to mastering the automatic execution of changes throughout the entire process.

[0136] Figure 3 A flowchart illustrating a method for generating transaction identification results according to an embodiment of the present disclosure is shown.

[0137] like Figure 3 As shown, the transaction feature values ​​are processed using a classification and recognition model to obtain the transaction recognition result, including operations S310 to S380.

[0138] In steps S310 to S320, the transaction feature values ​​are processed by convolutional layers to obtain sequence features.

[0139] In steps S330 to S360, the sequence features are globally pooled using a pooling layer to obtain feature vectors.

[0140] In steps S370 to S380, the feature vector is processed using a fully connected layer to obtain the transaction identification result.

[0141] According to embodiments of this disclosure, the convolutional layer learns local features from transaction feature values ​​composed of word vectors. A convolutional window traverses the sequence with a fixed stride, and the elements within the window are multiplied by multiple convolutional kernels. After each convolutional kernel completes its operation with the entire sequence, a new sequence feature is generated, whose value reflects the features exhibited by the input data under that convolutional kernel.

[0142] According to embodiments of this disclosure, a pooling layer is used to sample sequence features. It scans the data using a window, replacing the entire window's data with the maximum or average value of the elements within the window, thereby shortening the data length. The pooling layer uses a max-pooling algorithm to process the submatrix composed of each row of data, calculating the maximum value of each column in the submatrix to obtain a vector composed of the maximum values ​​of each column. Max-pooling is then performed on all submatrices. After global max-pooling, the matrix generates a processable one-dimensional feature vector, where the feature vector contains high-order features of the original input data.

[0143] According to embodiments of this disclosure, fully connected layers are used to learn the mapping relationship between feature vectors and recognition types.

[0144] According to embodiments of this disclosure, the identification type includes lender account type, business type, lender application name, etc.

[0145] According to embodiments of this disclosure, the number of convolutional layers can be at least two, the number of pooling layers can be at least two, and the number of fully connected layers can be at least one. It should be noted that the number of convolutional layers, pooling layers, and fully connected layers can be set according to actual needs.

[0146] According to embodiments of this disclosure, the convolutional layer extracts transaction feature information from the training transaction data, thereby outputting corresponding sequence feature information; the pooling layer extracts the output sequence feature information, thereby outputting a corresponding one-dimensional feature vector; and the fully connected layer processes the output feature vector information to obtain the corresponding transaction recognition result data.

[0147] According to embodiments of this disclosure, the classification and recognition model is trained using known transaction recognition result data. The trained and optimized model supports the import of new transaction scenario data. Cross-entropy loss is used as the loss function, and the root mean square propagation (RMSprop) algorithm is used as the optimizer. For newly added unknown transaction data, the classification model determines the type of the new transaction scenario data in real time and outputs the transaction recognition result.

[0148] According to embodiments of this disclosure, the classification and recognition model can intelligently identify borrower and lender data information in business scenarios. By using a convolutional neural network algorithm, it can quickly and intelligently identify business types, loan account types, and affiliated applications, significantly shortening the data change scheme generation time and effectively solving problems such as numerous self-balancing business scenario types, complex transaction logic, and unclear borrowers and lenders.

[0149] According to embodiments of this disclosure, the transaction feature value is a matrix sequence of m*n, where m is the number of transaction data for a transaction behavior, and n is the number of transaction attributes of the transaction data.

[0150] Specifically, in operation S310, the first convolutional layer uses a convolutional window with a preset stride to traverse the transaction feature values ​​and obtain the convolutional sequence.

[0151] In operation S320, the second convolutional layer uses a convolutional window based on a preset stride to traverse the convolutional sequence and obtain sequence features.

[0152] According to embodiments of this disclosure, convolutional layer C1 learns local features from transaction feature values ​​composed of word vectors. A convolutional window is used to traverse the transaction feature values ​​with a fixed stride, and the elements within the window are subjected to dot product operations with multiple convolutional kernels. After each convolutional kernel completes its operation with the entire sequence, a new convolutional sequence is generated, representing the features exhibited by the transaction data under that convolutional kernel.

[0153] In one embodiment, the convolutional layer is operated using formula (1):

[0154] s = f(W c ·d+b) (1)

[0155] Among them, W c It is the c-th convolutional kernel among n convolutional kernels of length p in convolutional layer C1, d represents the submatrix composed of each row of data; b∈R represents the preset bias value; f is the activation function; s is the convolutional kernel W. c The result of the dot product with the i-th submatrix.

[0156] According to embodiments of this disclosure, a second convolutional layer C2 uses a convolutional window to traverse the convolutional sequence with a fixed stride. Elements within the window are subjected to dot product operations with multiple convolutional kernels. After each convolutional kernel completes its operation with the entire sequence, sequence features are generated, characterizing the features exhibited by the transaction data at a higher span under that convolutional kernel.

[0157] According to embodiments of this disclosure, transaction feature values ​​are repeatedly processed by two convolutional layers to extract sequence features of transaction information from a higher span.

[0158] According to embodiments of this disclosure, a pooling layer is used to perform global pooling processing on sequence features to obtain a feature vector, including operations S330 to S360.

[0159] In operation S330, the sequence features are scanned using the first pooling layer based on a first pooling window of a preset length, so as to generate a first window feature corresponding to each first pooling window according to the data in the first pooling window.

[0160] In operation S340, intermediate features corresponding to the first pooling layer are generated based on the first window features corresponding to each first pooling window.

[0161] In operation S350, the intermediate features are scanned using the second pooling layer based on a second pooling window of a preset length, so as to generate a second window feature corresponding to each second pooling window according to the data in the second pooling window.

[0162] In operation S360, feature vectors are generated based on the second window features corresponding to each second pooling window.

[0163] According to an embodiment of this disclosure, the first pooling layer P1 scans the data using a first pooling window of length q, and replaces the data of the entire window with the maximum or average value of the elements within the window to generate a first window feature corresponding to each first pooling window.

[0164] According to an embodiment of this disclosure, the first pooling layer P1 uses the max pooling algorithm to process the submatrix composed of each row of data of the first window feature, calculates the maximum value of each column in the submatrix, obtains a vector composed of the maximum values ​​of each column of the submatrix, performs max pooling operation on all submatrixes, and obtains the intermediate feature composed of the maximum values ​​of each column of all submatrixes.

[0165] According to an embodiment of this disclosure, the second pooling layer P2 uses a second pooling window of length ρ to scan the intermediate feature data, and replaces the data of the entire window with the maximum or average value of the elements in the window to generate a second window feature corresponding to each second pooling window, where ρ is equal to the number of rows of the intermediate feature data.

[0166] According to an embodiment of this disclosure, the second pooling layer P2 uses a global max pooling algorithm to process the matrix composed of each row of data of the second window feature, calculates the maximum value of each column in the matrix, obtains a vector composed of the maximum values ​​of each column of the matrix, and performs max pooling operation on all submatrices to obtain a one-dimensional feature vector.

[0167] According to embodiments of this disclosure, a pooling layer is added after the convolutional layer. The operation of the two pooling layers effectively reduces the size of the feature vector matrix, thereby reducing the number of weights in the fully connected layer, while also speeding up the model operation and preventing overfitting.

[0168] According to embodiments of this disclosure, a fully connected layer is used to process feature vectors to obtain transaction identification results, including operations S370 to S380.

[0169] In operation S370, classification probabilities corresponding to different preset weight matrices are generated based on multiple preset weight matrices, bias vectors, and feature vectors. Each preset weight matrix corresponds to a classification attribute, which includes business type attributes and transaction account attributes.

[0170] In operation S380, the classification attribute corresponding to the maximum classification probability is determined as the transaction recognition result.

[0171] According to embodiments of this disclosure, a fully connected layer is used to learn the mapping relationship between feature vector z and k recognition categories.

[0172] In one embodiment, the mapping function used for the mapping relationship is shown in formula (2):

[0173] y=f(vz+B) (2)

[0174] Where v is a preset weight matrix; B is a preset bias vector; f is an activation function; and the output classification probability y represents the probability that transaction data is classified into a certain classification attribute, that is, the probability that transaction data is classified into the business type attribute and the transaction account attribute respectively.

[0175] According to embodiments of this disclosure, the weight values ​​of each node in the convolutional network are randomized. During the training of the classification and recognition model, the backpropagation algorithm and the mini-batch gradient descent algorithm are used to gradually update the weight values ​​of the network. After multiple rounds of training and testing, the optimal weight values ​​are selected as the network parameters output of the model.

[0176] According to embodiments of this disclosure, the classification attribute corresponding to the maximum classification probability is determined as the transaction identification result.

[0177] According to embodiments of this disclosure, multiple pieces of data to be processed are parsed and standardized to obtain transaction feature values, including:

[0178] Multiple data to be processed are parsed to obtain multiple data attributes corresponding to each data to be processed, where the data attributes represent different behavioral attributes of the transaction behavior;

[0179] For each piece of data to be processed, multiple data attributes are preprocessed to obtain intermediate data;

[0180] The transaction feature values ​​are obtained by standardizing multiple intermediate data.

[0181] According to embodiments of this disclosure, a JDBC database connection is established starting from the initial application logs involving both the lender and borrower to obtain the full amount of transaction data, i.e., the data to be processed.

[0182] According to embodiments of this disclosure, preprocessing operations such as data dimensionality reduction and data cleaning are performed on multiple datasets to be processed, irrelevant attribute fields and symbols are removed, key fields are retained, the data dimensionality is reduced, and thus the accuracy and efficiency of the classification and recognition model are improved.

[0183] According to embodiments of this disclosure, multiple intermediate data are standardized to obtain transaction feature values, including:

[0184] Multiple intermediate data points are grouped to obtain multiple data groups;

[0185] Feature extraction is performed on multiple data groups to obtain multiple target feature values;

[0186] The transaction feature values ​​are obtained by standardizing and transforming multiple target feature values.

[0187] According to embodiments of this disclosure, multiple intermediate data are grouped according to a business scenario description field. For example, if there are h business scenario description types, the data is divided into h groups. Business scenarios can refer to deposit and withdrawal services, loan services, insurance services, etc.

[0188] According to embodiments of this disclosure, g transaction information fields are extracted from each data group as target feature value fields to obtain multiple target feature values.

[0189] According to embodiments of this disclosure, each target feature value field is standardized to obtain transaction feature values ​​that the classification and recognition model can recognize.

[0190] According to embodiments of this disclosure, the target feature value includes at least one of text-type feature values ​​and Boolean-type feature values;

[0191] This involves format conversion of multiple target feature values ​​to obtain transaction feature values, including:

[0192] When the target feature value is a text-type feature value, the text-type feature value is format-converted based on the lexical hypothesis method to obtain the first intermediate feature value;

[0193] When the target feature value is a Boolean feature value, perform a binary conversion on the Boolean feature value to obtain the second intermediate feature value;

[0194] The transaction feature values ​​are obtained by standardizing and transforming multiple intermediate feature values.

[0195] According to embodiments of this disclosure, target feature values ​​can be divided into text-type feature values ​​and Boolean-type feature values.

[0196] According to embodiments of this disclosure, when the target feature value is a text-type feature value, a first intermediate feature value is obtained by performing binary format conversion using the bag of words method.

[0197] According to embodiments of this disclosure, when the target feature value is a Boolean feature value, a binary value format conversion is directly performed to obtain a second intermediate feature value.

[0198] According to embodiments of this disclosure, standardization transformation is used to regularize the output of data after feature value construction.

[0199] In one embodiment, the characteristic value fields of each intermediate feature are standardized to obtain the transaction characteristic value. The standardization transformation formula is shown in formula (3):

[0200]

[0201] Where x represents a specific intermediate eigenvalue, and μ and δ are the expected value and standard deviation of the intermediate eigenvalue, respectively. Calculate scores for the standardized data.

[0202] According to embodiments of this disclosure, multiple intermediate feature value fields are standardized and transformed to generate data that can be recognized by the classification and recognition model, which also improves the efficiency of the classification and recognition model algorithm.

[0203] According to embodiments of this disclosure, multiple data attributes are preprocessed to obtain intermediate data, including:

[0204] If a data attribute is irrelevant, the data attribute is filtered out to obtain multiple filtered data attributes.

[0205] For each filtered data attribute, the format of the filtered data attribute is corrected to obtain an initial data;

[0206] Based on multiple initial data points of the data to be processed, intermediate data corresponding to the data to be processed is generated.

[0207] According to embodiments of this disclosure, irrelevant attributes of multiple data to be processed are filtered out. For example, irrelevant attribute fields include transaction code, region code, branch number, teller number, channel type, account type, and remarks.

[0208] According to the embodiments of this disclosure, key attribute fields such as business scenario description, business type, debit application, debit account, debit amount, debit date, debit accounting status, credit application, credit account, credit amount, and credit date are retained.

[0209] According to embodiments of this disclosure, the remaining key attribute fields are formatted by removing irrelevant information such as spaces and punctuation marks, simplifying the field format, generating multiple initial data sets corresponding to the data to be processed, and thus obtaining intermediate data.

[0210] Based on the above data processing method, this disclosure also provides a data processing apparatus. The following will be combined with... Figure 4 The device is described in detail.

[0211] Figure 4 A schematic block diagram of a data processing apparatus according to an embodiment of the present disclosure is shown.

[0212] like Figure 4 As shown, the data processing device 400 in this embodiment includes an acquisition module 410, a first processing module 420, a second processing module 430, a generation module 440, and an update module 450.

[0213] The acquisition module 410 is used to acquire multiple data to be processed from multiple target databases of different application platforms in response to a data processing request. The data to be processed represents transaction data generated by a transaction. In one embodiment, the acquisition module 410 can be used to perform the operation S210 described above, which will not be repeated here.

[0214] The first processing module 420 is used to parse and standardize multiple pieces of data to obtain transaction feature values. In one embodiment, the first processing module 420 can be used to perform the operation S220 described above, which will not be repeated here.

[0215] The second processing module 430 is used to process the transaction feature values ​​using a classification and recognition model to obtain a transaction recognition result. The transaction recognition result includes business type data and transaction account information indicating whether the initial application log corresponding to the transaction behavior needs to be updated. In one embodiment, the second processing module 430 can be used to execute the operation S230 described above, which will not be repeated here.

[0216] The generation module 440 is used to generate application change data based on business type data and transaction account information. In one embodiment, the generation module 440 can be used to perform the operation S240 described above, which will not be repeated here.

[0217] The update module 450 is used to update the initial application log using application change data to obtain the target application log. In one embodiment, the update module 450 can be used to perform the operation S250 described above, which will not be repeated here.

[0218] According to embodiments of this disclosure, data to be processed from different platforms is standardized, and the resulting transaction feature values ​​are input into a classification and recognition model to determine the corresponding business type data and transaction account information. Application change data is generated based on the business type data and transaction account information, and the initial application log is updated using this application change data to obtain the target application log. The classification and recognition model can quickly and intelligently identify attributes such as business type, loan account type, and affiliated application, significantly shortening the data change plan generation time. Based on the identification results, the change plan is automatically updated and executed, eliminating the need for manual execution by technical personnel. The entire process of automatic change execution solves the problems of preparing self-balancing accounting data and the difficulty of unified change management.

[0219] According to embodiments of this disclosure, the generation module 440 includes a first generation submodule, a second generation submodule, and a third generation submodule.

[0220] The first generation submodule is used to determine the application address where the initial application log that needs to be updated is located based on the business type data.

[0221] The second generation submodule is used to generate switch configuration data based on the application address and business type data.

[0222] The third generation submodule is used to generate target change data based on transaction account information. The application change data includes switch configuration data and target change data.

[0223] According to embodiments of this disclosure, the update module 450 includes a first update submodule and a second update submodule.

[0224] The first update submodule is used to convert the application change data into a format to obtain the application change script.

[0225] The second update submodule is used to execute the application change script to update the initial application log and obtain the target application log.

[0226] According to embodiments of this disclosure, the first update submodule includes a first update unit and a second update unit.

[0227] The first update unit is used to generate an execution result based on the execution result of each change sub-script when multiple change sub-scripts are executed.

[0228] The second update unit is used to display the execution results in a visual format.

[0229] According to embodiments of this disclosure, the first update unit includes a first update subunit and a second update subunit.

[0230] The first update subunit is used to stop the update process if any one of the multiple change subscripts fails during the sequential execution of multiple change subscripts.

[0231] The second update subunit is used to generate execution results based on the successfully modified subscripts and the modified subscripts that have stopped updating.

[0232] According to embodiments of this disclosure, the second processing module 430 includes a sequence feature generation submodule, a feature vector generation submodule, and a transaction identification result generation submodule.

[0233] The sequence feature generation submodule is used to perform convolution processing on the transaction feature values ​​using convolutional layers to obtain sequence features.

[0234] The feature vector generation submodule is used to perform global pooling on sequence features using pooling layers to obtain feature vectors.

[0235] The transaction recognition result generation submodule is used to process feature vectors using a fully connected layer to obtain transaction recognition results.

[0236] According to embodiments of this disclosure, the sequence feature generation submodule includes a first processing unit and a second processing unit.

[0237] The first processing unit is used to traverse the transaction feature values ​​using the convolution window of the first convolutional layer based on a preset stride to obtain the convolutional sequence.

[0238] The second processing unit is used to traverse the convolutional sequence using the convolutional window of the second convolutional layer based on a preset stride to obtain sequence features.

[0239] According to embodiments of this disclosure, the feature vector generation submodule includes a third processing unit, a fourth processing unit, a fifth processing unit, and a sixth processing unit.

[0240] The third processing unit is used to scan sequence features using the first pooling layer based on a first pooling window of a preset length, so as to generate a first window feature corresponding to each first pooling window according to the data in the first pooling window.

[0241] The fourth processing unit is used to generate intermediate features corresponding to the first pooling layer based on the first window features corresponding to each first pooling window.

[0242] The fifth processing unit is used to scan intermediate features using the second pooling layer based on a second pooling window of a preset length, so as to generate a second window feature corresponding to each second pooling window according to the data in the second pooling window.

[0243] The sixth processing unit is used to generate feature vectors based on the second window features corresponding to each second pooling window.

[0244] According to embodiments of this disclosure, the transaction identification result generation submodule includes a seventh processing unit and an eighth processing unit.

[0245] The seventh processing unit is used to generate classification probabilities corresponding to different preset weight matrices based on multiple preset weight matrices, bias vectors, and feature vectors. Each preset weight matrix corresponds to a classification attribute, which includes business type attributes and transaction account attributes.

[0246] The eighth processing unit is used to determine the classification attribute corresponding to the maximum classification probability as the transaction recognition result.

[0247] According to embodiments of this disclosure, the first processing module 420 includes a first processing submodule, a second processing submodule, and a third processing submodule.

[0248] The first processing submodule is used to parse multiple pieces of data to be processed and obtain multiple data attributes corresponding to each piece of data to be processed. The data attributes represent different behavioral attributes of the transaction behavior.

[0249] The second processing submodule is used to preprocess multiple data attributes for each piece of data to be processed, so as to obtain intermediate data.

[0250] The third processing submodule is used to standardize multiple intermediate data to obtain transaction feature values.

[0251] According to embodiments of this disclosure, the third processing submodule includes a ninth processing unit, a tenth processing unit, and an eleventh processing unit.

[0252] The ninth processing unit is used to group multiple intermediate data to obtain multiple data groups.

[0253] The tenth processing unit is used to extract features from multiple data groups to obtain multiple target feature values.

[0254] The eleventh processing unit is used to standardize and transform multiple target feature values ​​to obtain transaction feature values.

[0255] According to embodiments of this disclosure, the eleventh processing unit includes a first processing subunit, a second processing subunit, and a third processing subunit.

[0256] The first processing subunit is used to perform format conversion on the text feature value based on the lexical hypothesis method to obtain the first intermediate feature value when the target feature value is a text feature value.

[0257] The second processing subunit is used to perform binary conversion on the Boolean feature value when the target feature value is a Boolean feature value, so as to obtain the second intermediate feature value.

[0258] The third processing subunit is used to perform standardization transformation based on multiple intermediate feature values ​​to obtain transaction feature values.

[0259] According to embodiments of this disclosure, the second processing submodule includes a twelfth processing unit, a thirteenth processing unit, and a fourteenth processing unit.

[0260] The twelfth processing unit is used to filter out data attributes when the data attributes are irrelevant attributes, and obtain multiple filtered data attributes.

[0261] The thirteenth processing unit is used to perform format correction processing on each filtered data attribute to obtain an initial data.

[0262] The fourteenth processing unit is used to generate intermediate data corresponding to the data to be processed based on multiple initial data of the data to be processed.

[0263] According to embodiments of this disclosure, any multiple modules among the acquisition module 410, the first processing module 420, the second processing module 430, the generation module 440, and the update module 450 can be combined into one module, or any one of these modules can be split into multiple modules. Alternatively, at least some of the functions of one or more of these modules can be combined with at least some of the functions of other modules and implemented in one module. According to embodiments of this disclosure, at least one of the acquisition module 410, the first processing module 420, the second processing module 430, the generation module 440, and the update module 450 can be at least partially implemented as hardware circuitry, 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-package, an application-specific integrated circuit (ASIC), or implemented in hardware or firmware by any other reasonable means of integrating or packaging circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, at least one of the acquisition module 410, the first processing module 420, the second processing module 430, the generation module 440, and the update module 450 may be implemented at least partially as a computer program module, which can perform corresponding functions when the computer program module is run.

[0264] Figure 5 A block diagram schematically illustrates an electronic device suitable for implementing a data processing method according to an embodiment of the present disclosure.

[0265] like Figure 5 As shown, an electronic device 500 according to an embodiment of the present disclosure includes a processor 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage portion 508 into a random access memory (RAM) 503. The processor 501 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 501 may also include onboard memory for caching purposes. The processor 501 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.

[0266] RAM 503 stores various programs and data required for the operation of electronic device 500. Processor 501, ROM 502, and RAM 503 are interconnected via bus 504. Processor 501 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 502 and / or RAM 503. It should be noted that programs may also be stored in one or more memories other than ROM 502 and RAM 503. Processor 501 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in one or more memories.

[0267] According to embodiments of this disclosure, the electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to a bus 504. The electronic device 500 may also include one or more of the following components connected to the I / O interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 510 as needed so that computer programs read from it can be installed into the storage section 508 as needed.

[0268] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.

[0269] According to embodiments of this disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this disclosure, the computer-readable storage medium may include ROM 502 and / or RAM 503 and / or one or more memories other than ROM 502 and RAM 503 described above.

[0270] Embodiments of this disclosure also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to enable the computer system to implement the data processing methods provided in the embodiments of this disclosure.

[0271] When the computer program is executed by the processor 501, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0272] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 509, and / or installed from a removable medium 511. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0273] In such an embodiment, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by processor 501, it performs the functions defined in the system of this disclosure embodiment. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0274] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can execute entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0275] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0276] Those skilled in the art will understand that the features described in the various embodiments and / or claims of this disclosure can be combined or combined in various ways, even if such combinations or combinations are not explicitly described in this disclosure. In particular, the features described in the various embodiments and / or claims of this disclosure can be combined or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.

[0277] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. The scope of this disclosure is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A data processing method, comprising: In response to a data processing request, multiple data to be processed are obtained from multiple target databases of different application platforms, wherein the data to be processed represents transaction data generated by a transaction behavior. The multiple datasets to be processed are parsed and standardized to obtain transaction feature values; The transaction feature values ​​are processed using a classification and recognition model to obtain transaction recognition results, wherein the transaction recognition results include business type data and transaction account information indicating whether the initial application log corresponding to the transaction behavior needs to be updated; Based on the business type data and transaction account information, generate application change data; The initial application log is updated using the application change data to obtain the target application log; Specifically, the initial application log is updated using the specific operation of updating the application change data to obtain the target application log, including: The application change data is converted to a new format to obtain the application change script; The application change script is executed to update the initial application log to obtain the target application log. The application change script includes multiple change sub-scripts. If any one of the multiple change sub-scripts fails to execute when multiple change sub-scripts are executed sequentially, the update process is stopped. Generate execution results based on the successfully modified sub-scripts and the modified sub-scripts that have stopped updating.

2. The method according to claim 1, wherein, Based on the business type data and transaction account information, application change data is generated, including: The application address where the initial application log that needs to be updated is located is determined based on the business type data; Based on the application address, generate switch configuration data according to the service type data; Based on the transaction account information, target change data is generated, wherein the application change data includes the switch configuration data and the target change data.

3. The method according to claim 1, wherein, The method further includes: The execution results are displayed in a visual format.

4. The method according to claim 1, wherein, The transaction feature values ​​are processed using a classification and recognition model to obtain transaction recognition results, including: The transaction feature values ​​are processed by convolutional layers to obtain sequence features; The sequence features are subjected to global pooling using a pooling layer to obtain a feature vector; The feature vector is processed using a fully connected layer to obtain the transaction identification result.

5. The method according to claim 4, wherein the transaction feature value is an m*n matrix sequence, wherein, m is the number of transaction data for the transaction behavior, and n is the number of transaction attributes for the transaction data; Specifically, convolutional layers are used to perform convolution processing on the transaction feature values ​​to obtain sequence features, including: The transaction feature values ​​are traversed using a convolutional window of the first convolutional layer with a preset stride to obtain a convolutional sequence; The second convolutional layer uses the convolutional window based on the preset stride to traverse the convolutional sequence and obtain the sequence features.

6. The method according to claim 4, wherein, The sequence features are subjected to global pooling using a pooling layer. The feature vector is obtained, including: The sequence features are scanned using a first pooling layer based on a first pooling window of a preset length, so as to generate a first window feature corresponding to each first pooling window based on the data in the first pooling window; Based on the first window features corresponding to each first pooling window, generate intermediate features corresponding to the first pooling layer; The intermediate features are scanned using a second pooling layer based on a second pooling window of a preset length, so as to generate a second window feature corresponding to each second pooling window based on the data in the second pooling window; The feature vector is generated based on the second window features corresponding to each second pooling window.

7. The method according to claim 4, wherein, The feature vector is processed using a fully connected layer to obtain the transaction identification result, including: Based on multiple preset weight matrices, bias vectors, and feature vectors, classification probabilities corresponding to different preset weight matrices are generated. Each preset weight matrix corresponds to a classification attribute, which includes a business type attribute and a transaction account attribute. The classification attribute corresponding to the maximum classification probability is determined as the transaction identification result.

8. The method according to claim 1, wherein, The multiple datasets to be processed are parsed and standardized to obtain transaction feature values, including: The multiple data to be processed are parsed to obtain multiple data attributes corresponding to each data to be processed, wherein the data attributes represent different behavioral attributes of the transaction behavior; For each piece of data to be processed, multiple data attributes are preprocessed to obtain intermediate data; The transaction feature values ​​are obtained by standardizing multiple intermediate data sets.

9. The method according to claim 8, wherein, The transaction feature values ​​are obtained by standardizing multiple intermediate data sets, including: The intermediate data are grouped to obtain multiple data groups; Feature extraction is performed on multiple data groups to obtain multiple target feature values; The transaction feature values ​​are obtained by standardizing and transforming multiple target feature values.

10. The method according to claim 9, wherein the target feature value includes at least one of text-type feature values ​​and Boolean-type feature values; in, The transaction feature values ​​are obtained by format conversion of multiple target feature values, including: When the target feature value is the text-type feature value, the text-type feature value is format-converted based on the lexical hypothesis method to obtain the first intermediate feature value; When the target feature value is the Boolean feature value, the Boolean feature value is converted into binary to obtain the second intermediate feature value; The transaction feature value is obtained by standardizing and transforming multiple intermediate feature values.

11. The method according to claim 8, wherein, Preprocessing is performed on multiple data attributes to obtain intermediate data, including: If the data attribute is an irrelevant attribute, the data attribute is filtered out to obtain multiple filtered data attributes. For each of the filtered data attributes, the format of the filtered data attributes is corrected to obtain an initial data; Based on the multiple initial data of the data to be processed, intermediate data corresponding to the data to be processed is generated.

12. A data processing apparatus, comprising: The acquisition module is used to acquire multiple data to be processed from multiple target databases of different application platforms in response to data processing requests, wherein the data to be processed represents transaction data generated by transaction behavior. The first processing module is used to parse and standardize multiple data to be processed to obtain transaction feature values; The second processing module is used to process the transaction feature values ​​using a classification and recognition model to obtain a transaction recognition result, wherein the transaction recognition result includes business type data and transaction account information that characterize whether the initial application log corresponding to the transaction behavior needs to be updated; The generation module is used to generate application change data based on the business type data and transaction account information; The update module is used to update the initial application log using the application change data to obtain the target application log; The update module includes: The first update submodule is used to convert the application change data into a format to obtain the application change script. The second update submodule is used to execute the application change script to update the initial application log and obtain the target application log. The application change script includes multiple change subscripts. The first update submodule includes: The first update unit is used to generate an execution result based on the execution result of each change sub-script when multiple change sub-scripts are executed. The first update unit includes: The first update subunit is used to stop the update process if any one of the multiple change subscripts fails during the sequential execution of multiple change subscripts. The second update subunit is used to generate execution results based on the successfully modified subscripts and the modified subscripts that have stopped updating.

13. An electronic device, comprising: One or more processors; Storage device for storing one or more programs. Wherein, when the one or more programs are executed by the one or more processors, the one or more processors perform the method according to any one of claims 1 to 11.

14. A computer-readable storage medium having stored thereon executable instructions that, when executed by a processor, cause the processor to perform the method according to any one of claims 1 to 11.

Images