Method, device and equipment for checking quality of submitted data

Abstract

The embodiment of the invention relates to a submitted data quality checking method, device and equipment. The method comprises the steps of capturing submitted data quality requirements of an announcement, and determining check elements of at least one submitted data type according to submitted data quality historical requirements and the captured submitted data quality requirements; according to the checking elements of the submitted data type, a checking rule set corresponding to the submitted data type in a checking rule base is updated; and when to-be-checked submitted data is received, checking the to-be-checked submitted data according to the target submitted data type of the to-be-checked submitted data and the updated checking rule set corresponding to the target submitted data type in the checking rule base. Through the method provided by the embodiment of the invention, automatic monitoring of submitted data requirements, automatic maintenance of checking rules, automatic checking of data quality and automatic analysis of problems are realized, and the checking efficiency and precision of the submitted data quality are improved, so that the quality of the submitted data is improved.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of human data governance technology, and in particular to a method, apparatus and equipment for checking the quality of reported data. Background Technology

[0002] Commercial banks are required to regularly submit data to the regulatory system in accordance with regulatory requirements. Currently, there are many challenges for commercial banks in terms of the quality of data submitted to the regulatory authorities, such as incomplete, inaccurate, and inconsistent data.

[0003] Existing technologies rely on manual verification of the quality of reported data, which requires manual monitoring of changes in regulatory requirements. This not only consumes a lot of manpower but is also prone to omissions or errors due to human factors. At the same time, due to the numerous internal systems of commercial banks, the complexity of data types, and the frequent data exchanges between systems, traditional methods are difficult to achieve comprehensive verification of reported data. Summary of the Invention

[0004] To address the problems existing in the prior art, embodiments of this specification provide a method, apparatus, and equipment for checking the quality of reported data. These methods and equipment enable automatic monitoring of reported data requirements, automatic maintenance of check rules, automatic checking of data quality, and automatic analysis of problems, thereby improving the efficiency and accuracy of reported data quality checks and ultimately enhancing the quality of the reported data.

[0005] The specific technical solutions of the embodiments in this specification are as follows:

[0006] On the one hand, embodiments of this specification provide a method for checking the quality of reported data, the method comprising:

[0007] The quality requirements for the submitted data in the captured announcements include regulatory data regulations, internal business regulations, internal data standards and specifications, and metadata of internal business systems.

[0008] Based on the historical requirements for the quality of the reported data and the captured quality requirements for the reported data, at least one check element for the type of reported data is determined.

[0009] Based on the verification elements of the reported data type, update the verification rule set corresponding to the reported data type in the verification rule base;

[0010] When data to be inspected is received, the data to be inspected is inspected according to the target data type of the data to be inspected and the updated set of inspection rules corresponding to the target data type in the inspection rule base.

[0011] Furthermore, after verifying the data to be verified according to the target data type of the data to be verified and the updated set of verification rules corresponding to the target data type in the verification rule base, the method further includes:

[0012] Determine whether the inspection result of the data to be inspected is "inspection failed";

[0013] If so, then obtain the multidimensional correlation data corresponding to the data to be inspected and reported;

[0014] The multidimensional associated data is subjected to feature extraction according to its dimensions to obtain a multidimensional feature matrix;

[0015] The multidimensional feature matrix is ​​analyzed to determine the error type of the data to be inspected and reported.

[0016] Furthermore, the multidimensional associated data includes the inspection log of the data to be inspected and submitted, the metadata of the data to be inspected and submitted, the original data of the data to be inspected and submitted, and the in-line data standard specifications of the data to be inspected and submitted.

[0017] The multidimensional feature matrix obtained by extracting features from the multidimensional associated data according to its dimensions further includes:

[0018] Feature extraction is performed on the check log to obtain check error features;

[0019] Feature extraction is performed on the metadata to obtain metadata features;

[0020] Feature extraction is performed on the raw data to obtain the raw data features;

[0021] The in-line data standard specifications of the data to be inspected and reported are feature-extracted to obtain data standard specification features;

[0022] The multidimensional feature matrix is ​​formed based on the error characteristics, metadata characteristics, raw data characteristics, and data standard specification characteristics.

[0023] Furthermore, analyzing the multidimensional feature matrix to determine the error type of the data to be inspected and reported further includes:

[0024] Cluster the aforementioned error features with historical error features of historical reported data to determine the target category to which the error features belong;

[0025] Determine the target root cause analysis model corresponding to the target category;

[0026] The error type is obtained by analyzing the metadata features, raw data features, and data standard specification features using the target root cause analysis model.

[0027] Furthermore, the steps for training the root cause analysis model include:

[0028] Obtain the historical multidimensional correlation data corresponding to the historical reported data;

[0029] Feature extraction is performed on the historical multidimensional associated data to obtain a historical multidimensional feature matrix, which includes historical check error features, historical metadata features, historical raw data features, and historical data standard specification features.

[0030] The historical check error features are clustered according to a predetermined number of categories to obtain multiple categories;

[0031] For each category, the historical metadata features, historical raw data features, and historical data standard specification features corresponding to each historical check error feature within that category are used as training data. The training data is then labeled according to the actual error type of the historical reporting data corresponding to that historical check error feature, and finally, the labeled training dataset for each category is obtained.

[0032] For each category, the initial root cause analysis model for that category is trained using the labeled training dataset to obtain the target root cause analysis model for that category.

[0033] Furthermore, the inspection elements include inspection data items and inspection standards for the inspection data items;

[0034] Based on the historical requirements for the quality of the reported data and the captured quality requirements for the reported data, the verification elements for at least one type of reported data further include:

[0035] The trained large model is used to analyze the captured regulatory data, internal business regulations, internal data standards and specifications, and internal business system metadata to determine at least one type of data to be submitted in the quality requirements of the captured data, as well as the latest verification data item and the latest verification standard for the data to be submitted.

[0036] The large model is used to analyze the historical regulatory data provisions, historical internal business provisions, historical internal data standard specifications, and historical internal business system metadata in the historical requirements for the quality of reported data, and to determine at least one historical reporting data type in the historical requirements for the quality of reported data, the historical inspection data item corresponding to the historical reporting data type, and the historical inspection standard corresponding to the historical inspection data item.

[0037] Find historical reporting data types that are the same as the reported data type. If found, compare the latest check data items, the latest check standards, and the historical check data items and historical check standards of the reported data type to obtain the check data items, the check standards, and the update form of the check data items corresponding to the reported data type. The update form includes addition, modification, deletion, or duplication.

[0038] If not found, the latest check data item of the reporting data type is taken as a check data item of the reporting data type, the latest check standard of the latest check data item is taken as the check standard of the check data item, and the update form of the check data item is set to new.

[0039] Furthermore, comparing the latest check data items, the latest check standards, and the historical check data items and historical check standards for the reported data type, the updated forms of the check data items, the check standards, and the check data items corresponding to the reported data type further include:

[0040] The latest check data item of the reporting data type is taken as one of the check data items of the reporting data type;

[0041] Determine whether there is a historical check data item that is identical to the check data item;

[0042] If they exist, compare whether the historical inspection standard corresponding to the historical inspection data item is the same as the latest inspection standard of the inspection data item. If they are the same, set the update form of the inspection data item to duplicate. If they are different, use the latest inspection standard as the inspection standard of the inspection data item and set the update form of the inspection data item to modified.

[0043] If it does not exist, the latest verification standard will be used as the verification standard for the verification data item, and the update format of the verification data item will be set to "new".

[0044] Furthermore, by comparing the latest check data items, the latest check standards, and historical check data items and standards for the reported data type, the updated forms of the check data items, the check standards, and the check data items corresponding to the reported data type are obtained, which also includes:

[0045] Determine if there are any historical check data items that are different from all the latest check data items;

[0046] If it exists, then set the update format of that historical check data item to delete.

[0047] Furthermore, updating the check rule set corresponding to the reported data type in the check rule base according to the check elements of the reported data type further includes:

[0048] The inspection rules for the inspection data item are generated based on the inspection criteria of the inspection data item.

[0049] If the update format of the check data item is set to "add", then the generated check rule is added to the check rule set.

[0050] If the update form of the inspection data item is modification, then the inspection rule of the historical inspection data item corresponding to the inspection data item in the inspection rule set is replaced with the generated inspection rule;

[0051] If the update form of the inspection data item is deletion, then the inspection rules of the historical inspection data items corresponding to the inspection data item in the inspection rule set will be deleted.

[0052] If the update form of the inspection data item is duplicated, then the inspection rules for the historical inspection data items corresponding to the inspection data item in the inspection rule set remain unchanged.

[0053] Furthermore, the verification of the data to be verified, based on the target data type of the data to be verified and the updated set of verification rules corresponding to the target data type in the verification rule base, further includes:

[0054] Identify the data items to be inspected in the submitted data to be inspected;

[0055] Search the updated set of inspection rules corresponding to the target reporting data type from the inspection rule base;

[0056] Search for the target check rule corresponding to the data item to be checked from the updated check rule set;

[0057] The target verification rules are used to verify the submitted data to be verified.

[0058] On the other hand, embodiments of this specification also provide a data quality checking device, the device comprising:

[0059] The capture unit is used to capture the quality requirements for the submitted data of the announcement, which include regulatory data regulations, internal business regulations, internal data standards and specifications, and metadata of internal business systems.

[0060] The verification element determination unit is used to determine at least one verification element of the data type of the submitted data based on the historical requirements for the quality of the submitted data and the captured data quality requirements.

[0061] The check rule set update unit is used to update the check rule set corresponding to the reported data type in the check rule base according to the check elements of the reported data type.

[0062] The inspection unit is used to inspect the data to be inspected when it receives the data to be inspected, based on the target data type of the data to be inspected and the updated inspection rule set corresponding to the target data type in the inspection rule base.

[0063] On the other hand, embodiments of this specification also provide a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the above-described method.

[0064] On the other hand, embodiments of this specification also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0065] On the other hand, embodiments of this specification also provide a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method.

[0066] The method described in this specification captures and analyzes data reporting quality requirements to obtain the check elements of the reporting data type. Then, based on these check elements, the check rules corresponding to the reporting data type in the check rule base are updated. This achieves automatic updating of the check rule base according to changes in the reporting data quality requirements, ensuring that the check rules in the check rule base are consistent with the latest reporting data quality requirements. This improves the accuracy of the data verification and can accurately identify data in the reporting data that does not meet the reporting data quality requirements. Compared to existing methods that rely on manual verification of the reporting data, the method described in this specification significantly improves both verification accuracy and efficiency. Attached Figure Description

[0067] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0068] Figure 1 The diagram shown is a schematic representation of an implementation system for a data quality verification method according to an embodiment of this specification.

[0069] Figure 2 The diagram shown is a flowchart illustrating how, in an embodiment of this specification, at least one check element for a data type of the submitted data is determined based on historical data quality requirements and the captured data quality requirements.

[0070] Figure 3 The diagram shown is a flowchart illustrating the process of checking the data to be checked based on the target data type of the data to be checked and the updated set of check rules corresponding to the target data type in the check rule base in an embodiment of this specification.

[0071] Figure 4 The diagram shown is a flowchart illustrating the analysis of the causes of verification errors in an embodiment of this specification.

[0072] Figure 5 The diagram shown is a structural schematic of a data quality verification device according to an embodiment of this specification.

[0073] Figure 6 The diagram shown is a structural schematic of the computer device in an embodiment of this specification.

[0074] [Explanation of Figure Markers]:

[0075] 501. Grabbing Unit;

[0076] 502. Unit for determining inspection elements;

[0077] 503. Check rule set update unit;

[0078] 504. Inspection Unit;

[0079] 602. Computer equipment;

[0080] 604, Processor;

[0081] 606. Memory;

[0082] 608. Drive mechanism;

[0083] 610. Input / output module;

[0084] 612. Input devices;

[0085] 614. Output devices;

[0086] 616. Presentation equipment;

[0087] 618. Graphical User Interface;

[0088] 620. Network interface;

[0089] 622. Communication link;

[0090] 624. Communication bus. Detailed Implementation

[0091] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the embodiments of this specification.

[0092] It should be noted that the terms "first," "second," etc., in the description, claims, and accompanying drawings of the embodiments herein are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, apparatus, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0093] It should be noted that the acquisition, storage, use, and processing of data in the technical solutions of the embodiments in this specification all comply with the relevant provisions of national laws and regulations.

[0094] It should be noted that in the embodiments of this specification, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, they do not mean that the applicant has used or necessarily used the solution.

[0095] To address the problems existing in the prior art, this specification provides an embodiment of a method for checking the quality of reported data. Figure 1 The diagram illustrates a flowchart of a data quality verification method according to an embodiment of this specification. The diagram depicts the process of analyzing data quality requirements to update the verification rule base, and using the rule base to verify the submitted data. However, based on conventional or non-creative work, this may include more or fewer steps. The order of steps listed in the embodiment is merely one possible execution order among many and does not represent the only possible order. In actual system or device products, the methods shown in the embodiment or the accompanying drawings can be executed sequentially or in parallel. Specifically, as shown... Figure 1As shown, the method can be executed by a server within the bank and may include:

[0096] Step 101: Capture the data quality requirements for the announcement, which include regulatory data regulations, internal business regulations, internal data standards and specifications, and metadata of internal business systems;

[0097] Step 102: Determine at least one check element for the data type of the submitted data based on the historical requirements for the quality of the submitted data and the captured data quality requirements;

[0098] Step 103: Update the check rule set corresponding to the reported data type in the check rule base according to the check elements of the reported data type;

[0099] Step 104: When receiving the data to be inspected, inspect the data to be inspected according to the target data type of the data to be inspected and the updated set of inspection rules corresponding to the target data type in the inspection rule base.

[0100] The method described in this specification captures and analyzes data reporting quality requirements to obtain the check elements of the reporting data type. Then, based on these check elements, the check rules corresponding to the reporting data type in the check rule base are updated. This achieves automatic updating of the check rule base according to changes in the reporting data quality requirements, ensuring that the check rules in the check rule base are consistent with the latest reporting data quality requirements. This improves the accuracy of the data verification and can accurately identify data in the reporting data that does not meet the reporting data quality requirements. Compared to existing methods that rely on manual verification of the reporting data, the method described in this specification significantly improves both verification accuracy and efficiency.

[0101] In the embodiments of this specification, the reported data quality requirements include regulatory data regulations, internal business regulations, internal data standard specifications, and internal business system metadata, wherein the above data can be obtained by four intelligent agents.

[0102] For example, agent A, responsible for monitoring regulatory provisions and penalty records, obtains regulatory provisions and penalty records from official websites of regulatory agencies through web crawling technology. It then performs structured processing and semantic analysis on the obtained data to extract requirements and penalty cases related to the regulatory reporting data, and promptly detects changes in regulatory requirements.

[0103] The intelligent agent B, responsible for monitoring internal business regulations, regularly scans the business rules and regulations texts stored in the internal document management system, business systems, etc., uses natural language processing technology to parse the texts, identify information such as business processes, data definitions, and operating procedures, monitors the updates and changes to business rules and regulations, and analyzes the impact of these changes on the quality of regulatory reporting data.

[0104] The intelligent agent C, responsible for monitoring the industry's data standards and specifications, interfaces with the industry's data standards management system to obtain real-time information on version changes of data standards and specifications, including data element definitions, code sets, data formats, etc., to track the modification of data standards and specifications and assess their impact on the rules for checking the quality of data submitted to the regulatory authorities.

[0105] The intelligent agent D, responsible for monitoring the metadata of the business system, extracts metadata information from the metadata database of the business system, such as data structure, data type, data length, relationships, data source and destination, monitors changes in metadata, and promptly detects the potential impact of metadata changes caused by system upgrades, module adjustments, etc. on data quality.

[0106] Then, based on the historical requirements for the quality of the reported data and the captured reported data quality requirements, at least one check element for the type of reported data is determined.

[0107] Specifically, the inspection elements include inspection data items and inspection standards for the inspection data items.

[0108] like Figure 2 As shown, determining at least one check element for the data type of a reported data according to the historical requirements for reported data quality and the captured reported data quality requirements further includes:

[0109] Step 201: Analyze the captured regulatory data regulations, internal business regulations, internal data standards and specifications, and internal business system metadata using the trained large model to determine at least one type of data to be submitted in the quality requirements of the captured data, as well as the latest verification data item corresponding to the data type and the latest verification standard for the latest verification data item.

[0110] Step 202: Analyze the historical regulatory data provisions, historical internal business provisions, historical internal data standard specifications, and historical internal business system metadata in the historical requirements for data quality of the submitted data using the large model, and determine at least one historical data type in the historical requirements for data quality of the submitted data, the historical inspection data item corresponding to the historical data type, and the historical inspection standard corresponding to the historical inspection data item.

[0111] Step 203: Locate historical reporting data types that are the same as the reported data type. If found, compare the latest check data items, the latest check standards, and the historical check data items and historical check standards of the reported data type to obtain the check data items, the check standards, and the update form of the check data items corresponding to the reported data type. The update form includes addition, modification, deletion, or duplication.

[0112] Step 204: If not found, the latest check data item of the reporting data type is taken as a check data item of the reporting data type, the latest check standard of the latest check data item is taken as the check standard of the check data item, and the update form of the check data item is set to new.

[0113] In the embodiments of this specification, the trained large model can be a general large model. The general large model is used to analyze the regulatory data regulations, internal business regulations, internal data standard specifications and internal business system metadata collected by the above four agents to determine at least one reporting data type in the quality requirements of the reported data, as well as the latest verification data item corresponding to the reporting data type and the latest verification standard of the latest verification data item.

[0114] By using a general large model, we analyze the historical regulatory data provisions, historical internal business provisions, historical internal data standard specifications, and historical internal business system metadata in the historical requirements for the quality of historical data submitted, and determine at least one historical data type in the historical requirements for the quality of the submitted data, the historical inspection data item corresponding to the historical data type, and the historical inspection standard corresponding to the historical inspection data item.

[0115] Then, the latest check data items and their latest check labels are compared with the historical check data items and their corresponding historical check standards to determine the check data items that need to be updated and the corresponding check standards.

[0116] In the embodiments of this specification, the forms of update include adding, modifying, deleting, or repeating.

[0117] The verification rules for the historical verification data items corresponding to each reporting data type have been stored in the verification rule base. If the latest verification data item of the reporting data type is different from all the historical verification data items of the reporting data type, it means that there are no verification rules for the latest verification data item in the verification rule base. In this case, it is necessary to generate a verification rule for the latest verification data item, set its update form to "add", and store it in the verification rule base.

[0118] The check rules in the check rule base can be searched by the data type of the report and the check data item, which will not be described in detail in the embodiments of this specification.

[0119] If a latest check data item of the analyzed reporting data type is the same as a historical check data item of the same reporting data type, it means that a check rule already exists for the latest check data item in the check rule base. In this case, a check rule for the latest check data item needs to be generated and updated in the check rule base. If a check rule corresponding to the historical check data item already exists in the check rule base and is the same as the check rule corresponding to the latest check data item, it means that the latest reported data quality requirements have not modified the reported data quality requirements for this check data item. Therefore, the update format for this check data item is set to repeat, keeping the check rules for the historical check data items corresponding to the check data items in the check rule set corresponding to this reporting data type unchanged.

[0120] If a latest check data item of the reported data type is identical to a historical check data item of the same data type, it indicates that a check rule already exists for the latest check data item in the check rule base. In this case, a check rule for the latest check data item needs to be generated and updated in the check rule base. If the check rule corresponding to the historical check data item already exists in the check rule base is different from the check rule corresponding to the latest check data item, it indicates that the latest reported data quality requirements have modified the reported data quality requirements for this check data item. In this case, the update format should be set to "modify," and the check rule for the historical check data item in the check rule set corresponding to this reported data type should be replaced with the generated check rule.

[0121] If all the latest check data items of the reported data type are different from one or more historical check data items, it means that the latest data reporting quality requirements no longer require the reporting of this check data item, or no longer specify the reporting format of this check data item. In this case, the update format of the historical check data item is set to delete, thereby deleting the check rules of the historical check data item corresponding to the check data item in the check rule set of the reported data type.

[0122] In the embodiments of this specification, the data type of the report is specified by the regulatory authority, such as report type, detail type, etc. The data items to be checked in different data types can be the same, but the check rules can be different. The data items to be checked represent the data items that commercial banks are required to report according to the prescribed data type. The check standard indicates the format, value range, and total consistency of the data item as specified by the regulatory authority. The function of the check rule is to determine whether the format, value, total data, detail data, etc. before the data items to be checked in the report conform to the format, value range and / or total consistency specified in the check standard.

[0123] In the embodiments of this specification, the logical expression for the inspection of the inspection criteria of the inspection data items can be generated through a general large model, thereby obtaining the inspection rules. The embodiments of this specification will not be elaborated further.

[0124] In some other embodiments of this specification, if the update mode of a historical check data item is set to deletion, the check rule for that historical check data item may not be immediately deleted from the check rule base. Instead, the check rule for that historical check data item is marked as invalid, and when the data to be checked is checked, the invalid check rule will not be matched. If the data quality requirements of the crawling announcement for a specified number of rounds have been exceeded, and the invalid check data item is no longer present in the check data items corresponding to the data type of the report, then the invalid check rule is deleted from the check rule base. This avoids the problem of increasing the workload of updating the check rule base due to temporary adjustments in the data quality requirements.

[0125] In the embodiments of this specification, the generation of verification rules and the updating of the verification rule base can also be performed by an intelligent agent. For example, an intelligent agent E that sets and maintains data quality monitoring rules can perform the generation of verification rules and the updating of the verification rule base.

[0126] When data to be inspected is received, it is inspected according to the target data type of the data and the updated set of inspection rules corresponding to the target data type in the inspection rule base. Specifically, an agent F responsible for deploying and running data quality inspection rules can be set to deploy the maintained inspection rules to the data inspection engine at preset time intervals (such as daily, weekly, or monthly), automatically run the data quality inspection rules, perform a comprehensive inspection of the regulatory data, and record the inspection results in detail, including data items that pass the inspection and data items that fail the inspection, and generate inspection logs.

[0127] like Figure 3 As shown, the verification of the data to be verified, based on the target data type of the data to be verified and the updated set of verification rules corresponding to the target data type in the verification rule base, further includes:

[0128] Step 301: Determine the data items to be inspected in the submitted data to be inspected;

[0129] Step 302: Search the updated set of inspection rules corresponding to the target reporting data type in the inspection rule base;

[0130] Step 303: Search for the target check rule corresponding to the data item to be checked from the updated check rule set;

[0131] Step 304: Use the target verification rules to verify the data to be verified.

[0132] According to one embodiment of this specification, if the verification result of the submitted data to be checked is "failed," the reason may be a system data error or an input error. Therefore, in order to identify the specific cause of the error, such as... Figure 4 As shown, after verifying the data to be verified according to the target data type of the data to be verified and the updated set of verification rules corresponding to the target data type in the verification rule base, the method further includes:

[0133] Step 401: Determine whether the inspection result of the data to be inspected is "inspection failed";

[0134] Step 402: If yes, then obtain the multidimensional correlation data corresponding to the data to be inspected and reported;

[0135] Step 403: Extract features from the multidimensional associated data according to the dimensions to obtain a multidimensional feature matrix;

[0136] Step 404: Analyze the multidimensional feature matrix to determine the error type of the data to be inspected and reported.

[0137] In the embodiments of this specification, multidimensional associated data represents all related data of the data to be inspected and reported. For example, the multidimensional associated data includes the inspection log of the data to be inspected and reported, the metadata of the data to be inspected and reported, the original data of the data to be inspected and reported, and the in-line data standard specifications of the data to be inspected and reported.

[0138] In the embodiments of this specification, the check log records which data items or formats do not conform to the check standard. For example, the data length of a check data item exceeds the data length range specified in the check standard, or the data format of a check data item does not conform to the data format specified in the check standard. Therefore, the embodiments of this specification use the check log as associated data for one dimension of error type analysis.

[0139] Metadata of data to be inspected and submitted refers to the storage format requirements of the raw data of the data to be inspected and submitted in the system, including data type, data length, etc.

[0140] The raw data of the data to be inspected and submitted refers to the raw data of the inspection data items in the data to be inspected and submitted stored in the system.

[0141] The internal data standard specifications for data to be inspected and submitted refer to the format specifications for storing inspection data items in the system within the bank's internal data to be inspected and submitted, such as storage format and storage data length.

[0142] This specification performs feature extraction on the multidimensional associated data according to its dimensions to obtain a multidimensional feature matrix, which further includes:

[0143] Feature extraction is performed on the check log to obtain check error features;

[0144] Feature extraction is performed on the metadata to obtain metadata features;

[0145] Feature extraction is performed on the raw data to obtain the raw data features;

[0146] The in-line data standard specifications of the data to be inspected and reported are feature-extracted to obtain data standard specification features;

[0147] The multidimensional feature matrix is ​​formed based on the error characteristics, metadata characteristics, raw data characteristics, and data standard specification characteristics.

[0148] In the embodiments described in this specification, a trained general-purpose model can be used to extract key information from check logs, metadata, raw data, and industry standards and specifications to obtain check error features, metadata features, raw data features, and data standard and specification features. These features are then analyzed using a machine learning model to determine the error type.

[0149] According to one embodiment of this specification, because there are many verification standards, in order to improve the accuracy of machine learning models in analyzing error types, this embodiment of the specification obtains historical multidimensional correlation data corresponding to the historical reporting data;

[0150] Feature extraction is performed on the historical multidimensional associated data to obtain a historical multidimensional feature matrix, which includes historical check error features, historical metadata features, historical raw data features, and historical data standard specification features.

[0151] The historical check error features are clustered according to a predetermined number of categories to obtain multiple categories;

[0152] For each category, the historical metadata features, historical raw data features, and historical data standard specification features corresponding to each historical check error feature within that category are used as training data. The training data is then labeled according to the actual error type of the historical reporting data corresponding to that historical check error feature, and finally, the labeled training dataset for each category is obtained.

[0153] For each category, the initial root cause analysis model for that category is trained using the labeled training dataset to obtain the target root cause analysis model for that category.

[0154] This can be understood as follows: the embodiments of this specification classify those with the same or similar error characteristics into a category, and then train a machine learning model specifically for this category, thereby improving the recognition accuracy of the model.

[0155] Therefore, the embodiments of this specification further include analyzing the multidimensional feature matrix to determine the error type of the data to be inspected and reported, including:

[0156] Cluster the aforementioned error features with historical error features of historical reported data to determine the target category to which the error features belong;

[0157] Determine the target root cause analysis model corresponding to the target category;

[0158] The error type is obtained by analyzing the metadata features, raw data features, and data standard specification features using the target root cause analysis model.

[0159] In the embodiments of this specification, an intelligent agent G can be set up to analyze the check results and conduct in-depth analysis of the problems that fail the check to obtain the error type.

[0160] Furthermore, an agent H can be set up to coordinate data governance and formulate corresponding governance strategies based on the error type. If the error type is a system data generation error, the agent I responsible for notifying the system to carry out source governance will be coordinated; if the error type is an input error, the agent responsible for notifying the personnel in charge to carry out data governance will be coordinated. At the same time, this agent tracks the governance progress to ensure that the governance work is completed on time.

[0161] The intelligent agent I, which is responsible for source governance, sends a governance notification to the source system through the system interface. The notification includes a description of the problem, the data items involved, the governance requirements, and the completion deadline. After receiving the notification, the staff of the source system make corresponding system adjustments and data corrections, and then feed back the governance results to the intelligent agent H, which is responsible for coordinating the data governance.

[0162] The intelligent agent J, responsible for notifying the personnel to carry out data governance, sends governance notices to the personnel through office automation systems, SMS, or email, specifying the problem, the correction method, and the completion time. After the personnel complete the governance, they feed back the results to the intelligent agent H, which is responsible for coordinating the data governance.

[0163] For example, this embodiment provides a method for governing the quality of commercial bank regulatory reporting data based on AI Agent, and the specific steps are as follows:

[0164] 1. Agent Deployment: Multiple agents are deployed in the commercial bank's data governance platform, namely Agent A, responsible for monitoring regulatory provisions and penalty records; Agent B, responsible for monitoring internal business rules and regulations; Agent C, responsible for monitoring internal data standards and specifications; Agent D, responsible for monitoring business system metadata; Agent E, responsible for maintaining data quality monitoring rules; Agent F, responsible for deploying and running data quality verification rules; Agent G, responsible for analyzing verification results; Agent H, responsible for coordinating data governance; Agent I, responsible for notifying source systems to carry out source governance; and Agent J, responsible for notifying personnel to carry out data governance.

[0165] 2. Data quality requirements monitoring:

[0166] (1) Agent A, which is responsible for monitoring regulatory provisions and penalty records, crawls the latest regulatory provisions and penalty records from the official website of the regulatory agency every day using web crawling technology. It performs structured processing on the crawled PDF, Word and other format files and extracts key information (such as regulatory indicators, reporting frequency, data format, etc.).

[0167] (2) Agent B, which is responsible for monitoring the business rules and regulations of the bank, scans the business rules and regulations in the bank's document management system every week, performs natural language processing on the newly uploaded or modified rules and regulations texts, and parses out the changes in business processes, such as the addition of a data collection item in the loan approval process, and analyzes the impact of the change on the data reported by the regulatory governance.

[0168] (3) The intelligent agent C responsible for monitoring the data standards and specifications within the industry connects with the data standards and management system within the industry in real time. When the definition of a certain data element in the data standards and management system is modified (such as the data type of "customer age" being changed from integer to string), the intelligent agent C responsible for monitoring the data standards and specifications within the industry obtains the information in a timely manner.

[0169] (4) The intelligent agent D, which is responsible for monitoring the metadata of the business system, extracts metadata information from the metadata database of the core business system, credit management system and other business systems every hour. When it finds that the table structure of a certain business system has changed (such as adding a field), it records the change.

[0170] 3. Data quality check rule maintenance: Agent E, responsible for maintaining data quality monitoring rules, collects the monitoring results from agents A, B, C, and D, and generates check rules and updates the check rule base.

[0171] 4. Data quality inspection rule operation: The intelligent agent F, which is responsible for deploying and running the data quality inspection rules, deploys the inspection rules maintained by the intelligent agent E to the data inspection engine once a day to inspect the data reported by the regulator. For example, it checks whether the "customer ID number" meets the format requirements and whether the "loan balance" is consistent with the ledger. The inspection results are recorded. For example, if 10 records of "customer ID number format error" are found.

[0172] 5. Inspection Result Analysis: The agent G responsible for analyzing the inspection results analyzed the 10 records that failed the inspection. It retrieved the customer information entry operation logs corresponding to these records and found that 8 of them were due to the operator missing one digit when entering the data, and 2 were due to errors in the system data generation.

[0173] 6. Governance Coordination and Promotion: Based on the analysis results of agent G, agent H, responsible for coordinating data governance, coordinates with agent J, which is responsible for notifying the responsible personnel to carry out data governance, for 8 records entered incorrectly by the personnel; and coordinates with agent I, which is responsible for notifying the source system to carry out source governance for 2 records caused by system interface failures.

[0174] (1) The intelligent agent J responsible for notifying the staff to carry out data governance sends a notification to the corresponding staff through the office automation system. The content is: "The customer ID number you entered has a format error. Please correct it within 24 hours. Error record ID: XXX".

[0175] (2) The intelligent agent I responsible for coordinating the source system to carry out source governance sends a notification to the corresponding business system through the system interface. The content is "System data generation error, resulting in incorrect customer ID number. Please check and fix within 48 hours. Error record ID: YYY".

[0176] 7. Governance Tracking: Agent H tracks the governance progress. After the responsible personnel correct the errors, they provide feedback to Agent H. After the system interface is repaired, the results are also fed back to Agent H. Agent H confirms that all issues have been resolved.

[0177] Through the above implementation methods, the entire process of automated verification of the quality of data reported by commercial banks for regulatory purposes has been achieved, improving verification efficiency and accuracy.

[0178] Based on the same inventive concept, embodiments of this specification also provide a device for checking the quality of reported data, such as... Figure 5 As shown, the device includes:

[0179] The capture unit 501 is used to capture the quality requirements for the submitted data of the announcement, which include regulatory data regulations, internal business regulations, internal data standard specifications and internal business system metadata;

[0180] The verification element determination unit 502 is used to determine at least one verification element of the reporting data type based on the historical requirements for the quality of the reported data and the captured reporting data quality requirements.

[0181] The check rule set update unit 503 is used to update the check rule set corresponding to the reported data type in the check rule base according to the check elements of the reported data type.

[0182] The inspection unit 504 is used to inspect the data to be inspected when it receives the data to be inspected, according to the target data type of the data to be inspected and the updated inspection rule set corresponding to the target data type in the inspection rule base.

[0183] The beneficial effects obtained by the above-described device are the same as those obtained by the above-described method, and will not be described in detail in the embodiments of this specification.

[0184] like Figure 6 The diagram shown is a structural schematic of a computer device according to an embodiment of this specification. The methods described in this specification can be applied to the computer device of this embodiment.

[0185] Computer device 602 may include one or more processors 604, such as one or more central processing units (CPUs), each of which may implement one or more hardware threads. Computer device 602 may also include any memory 606 for storing information of any kind, such as code, settings, data, etc. Non-limitingly, for example, memory 606 may include any type of RAM, any type of ROM, flash memory, hard disk, optical disk, etc. More generally, any storage resource can be used to store information using any technology.

[0186] Furthermore, any storage resource can provide volatile or non-volatile retention of information.

[0187] Furthermore, any storage resource can represent a fixed or removable component of the computer device 602. In one case, when the processor 604 executes associated instructions stored in any storage resource or combination of storage resources, the computer device 602 can perform any operation of the associated instructions. The computer device 602 also includes one or more drive mechanisms 608 for interacting with any storage resource, such as a hard disk drive system, an optical disk drive system, etc.

[0188] Computer device 602 may also include an input / output module 610 (I / O) for receiving various inputs (via input device 612) and providing various outputs (via output device 614). A specific output mechanism may include a presentation device 616 and an associated graphical user interface (GUI) 618. In other embodiments, the input / output module 610 (I / O), input device 612, and output device 614 may be omitted, and the device may function solely as a computer device within a network. Computer device 602 may also include one or more network interfaces 620 for exchanging data with other devices via one or more communication links 622. One or more communication buses 624 couple the components described above together.

[0189] Communication link 622 can be implemented in any way, such as via a local area network, a wide area network (e.g., the Internet), a point-to-point connection, or any combination thereof. Communication link 622 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0190] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0191] This specification also provides computer-readable instructions, wherein when a processor executes the instructions, the program therein causes the processor to perform the above-described method.

[0192] It should be understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0193] It should also be understood that, in the embodiments of this specification, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the embodiments of this specification, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0194] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the embodiments in this specification.

[0195] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0196] In the embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, devices, or units, or they may be electrical, mechanical, or other forms of connection.

[0197] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments described in this specification, depending on actual needs.

[0198] Furthermore, the functional units in the various embodiments of this specification can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0199] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this specification, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this specification. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0200] This specification describes the principles and implementation methods of the embodiments using specific examples. The above descriptions of the embodiments are only for the purpose of helping to understand the methods and core ideas of the embodiments in this specification. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the embodiments in this specification. Therefore, the content of this specification should not be construed as a limitation on the embodiments in this specification.

Claims

1. A method for checking the quality of reported data, characterized in that, The method includes: The quality requirements for the submitted data in the captured announcements include regulatory data regulations, internal business regulations, internal data standards and specifications, and metadata of internal business systems. Based on the historical requirements for the quality of the reported data and the captured quality requirements for the reported data, at least one check element for the type of reported data is determined. Based on the inspection elements of the reported data type, update the inspection rule set corresponding to the reported data type in the inspection rule base; When data to be inspected is received, the data to be inspected is inspected according to the target data type of the data to be inspected and the updated set of inspection rules corresponding to the target data type in the inspection rule base.

2. The method according to claim 1, characterized in that, After verifying the data to be verified according to the target data type of the data to be verified and the updated set of verification rules corresponding to the target data type in the verification rule base, the method further includes: Determine whether the inspection result of the data to be inspected is "inspection failed"; If so, then obtain the multidimensional correlation data corresponding to the data to be inspected and reported; The multidimensional associated data is subjected to feature extraction according to its dimensions to obtain a multidimensional feature matrix; The multidimensional feature matrix is ​​analyzed to determine the error type of the data to be inspected and reported.

3. The method according to claim 2, characterized in that, The multidimensional associated data includes the inspection log of the data to be inspected and reported, the metadata of the data to be inspected and reported, the original data of the data to be inspected and reported, and the in-line data standard specifications of the data to be inspected and reported. The multidimensional feature matrix obtained by extracting features from the multidimensional associated data according to its dimensions further includes: Feature extraction is performed on the check log to obtain check error features; Feature extraction is performed on the metadata to obtain metadata features; Feature extraction is performed on the raw data to obtain the raw data features; The in-line data standard specifications of the data to be inspected and reported are feature-extracted to obtain data standard specification features; The multidimensional feature matrix is ​​formed based on the error characteristics, metadata characteristics, raw data characteristics, and data standard specification characteristics.

4. The method according to claim 3, characterized in that, Analyzing the multidimensional feature matrix to determine the error type of the data to be inspected and reported further includes: Cluster the aforementioned error features with historical error features of historical reported data to determine the target category to which the error features belong; Determine the target root cause analysis model corresponding to the target category; The error type is obtained by analyzing the metadata features, raw data features, and data standard specification features using the target root cause analysis model.

5. The method according to claim 4, characterized in that, The steps for training the root cause analysis model include: Obtain the historical multidimensional correlation data corresponding to the historical reported data; Feature extraction is performed on the historical multidimensional associated data to obtain a historical multidimensional feature matrix, which includes historical check error features, historical metadata features, historical raw data features, and historical data standard specification features. The historical check error features are clustered according to a predetermined number of categories to obtain multiple categories; For each category, the historical metadata features, historical raw data features, and historical data standard specification features corresponding to each historical check error feature within that category are used as training data. The training data is then labeled according to the actual error type of the historical reporting data corresponding to that historical check error feature, and finally, the labeled training dataset for each category is obtained. For each category, the initial root cause analysis model for that category is trained using the labeled training dataset to obtain the target root cause analysis model for that category.

6. The method according to claim 1, characterized in that, The inspection elements include inspection data items and inspection standards for the inspection data items; Based on the historical requirements for the quality of the reported data and the captured quality requirements for the reported data, the verification elements for at least one type of reported data further include: The trained large model is used to analyze the captured regulatory data, internal business regulations, internal data standards and specifications, and internal business system metadata to determine at least one type of data to be submitted in the quality requirements of the captured data, as well as the latest verification data item and the latest verification standard for the data to be submitted. The large model is used to analyze the historical regulatory data provisions, historical internal business provisions, historical internal data standard specifications, and historical internal business system metadata in the historical requirements for the quality of reported data, and to determine at least one historical reporting data type in the historical requirements for the quality of reported data, the historical inspection data item corresponding to the historical reporting data type, and the historical inspection standard corresponding to the historical inspection data item. Find historical reporting data types that are the same as the reported data type. If found, compare the latest check data items, the latest check standards, and the historical check data items and historical check standards of the reported data type to obtain the check data items, the check standards, and the update form of the check data items corresponding to the reported data type. The update form includes addition, modification, deletion, or duplication. If not found, the latest check data item of the reporting data type is taken as a check data item of the reporting data type, the latest check standard of the latest check data item is taken as the check standard of the check data item, and the update form of the check data item is set to new.

7. The method according to claim 6, characterized in that, Comparing the latest check data items, the latest check standards, and the historical check data items and historical check standards for the reported data type, the updated forms of the check data items, the check standards, and the check data items corresponding to the reported data type further include: The latest check data item of the reporting data type is taken as one of the check data items of the reporting data type; Determine whether there is a historical check data item that is identical to the check data item; If they exist, compare whether the historical inspection standard corresponding to the historical inspection data item is the same as the latest inspection standard of the inspection data item. If they are the same, set the update form of the inspection data item to duplicate. If they are different, use the latest inspection standard as the inspection standard of the inspection data item and set the update form of the inspection data item to modified. If it does not exist, the latest verification standard will be used as the verification standard for the verification data item, and the update format of the verification data item will be set to "new".

8. The method according to claim 7, characterized in that, By comparing the latest check data items, the latest check standards, and the historical check data items and historical check standards for the reported data type, the updated forms of the check data items, the check standards, and the check data items corresponding to the reported data type are obtained, and the method further includes: Determine if there are any historical check data items that are different from all the latest check data items; If it exists, then set the update format of that historical check data item to delete.

9. The method according to claim 6, characterized in that, Updating the check rule set corresponding to the reported data type in the check rule base according to the check elements of the reported data type further includes: The inspection rules for the inspection data item are generated based on the inspection criteria of the inspection data item. If the update format of the check data item is set to "add", then the generated check rule is added to the check rule set. If the update form of the inspection data item is modification, then the inspection rule of the historical inspection data item corresponding to the inspection data item in the inspection rule set is replaced with the generated inspection rule; If the update form of the inspection data item is deletion, then the inspection rules of the historical inspection data items corresponding to the inspection data item in the inspection rule set will be deleted. If the update form of the inspection data item is duplicated, then the inspection rules for the historical inspection data items corresponding to the inspection data item in the inspection rule set remain unchanged.

10. The method according to claim 6, characterized in that, The verification of the data to be verified, based on the target data type of the data to be verified and the updated set of verification rules corresponding to the target data type in the verification rule base, further includes: Identify the data items to be inspected in the submitted data to be inspected; Search the updated set of inspection rules corresponding to the target reporting data type from the inspection rule base; Search for the target check rule corresponding to the data item to be checked from the updated check rule set; The target verification rules are used to verify the submitted data to be verified.

11. A device for checking the quality of reported data, characterized in that, The device includes: The capture unit is used to capture the quality requirements for the submitted data of the announcement, which include regulatory data regulations, internal business regulations, internal data standards and specifications, and metadata of internal business systems. The verification element determination unit is used to determine at least one verification element of the data type of the submitted data based on the historical requirements for the quality of the submitted data and the captured data quality requirements. The check rule set update unit is used to update the check rule set corresponding to the reported data type in the check rule base according to the check elements of the reported data type. The inspection unit is used to inspect the data to be inspected when it receives the data to be inspected, based on the target data type of the data to be inspected and the updated inspection rule set corresponding to the target data type in the inspection rule base.

12. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 10.

14. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 10.

Images