A method, system, computer and storage medium for reconciling accounts

Abstract

The application relates to the technical field of data processing, and provides an account checking method, an account checking system, a computer and a storage medium. The account checking method comprises the following steps: obtaining a to-be-checked data set, cleaning the to-be-checked data set to obtain standard upstream billing data and standard downstream billing data; obtaining a field mapping rule, and respectively mapping the standard upstream billing data and the standard downstream billing data into to-be-checked upstream data and to-be-checked downstream data according to the field mapping rule; obtaining account checking scene parameters, constructing a multi-dimensional difference rule network, comparing the to-be-checked upstream data and the to-be-checked downstream data based on the account checking scene parameters and the multi-dimensional difference rule network to obtain a difference identification result; performing an abnormal category prediction on the difference identification result to generate an abnormal identification result, and generating an account checking report based on the abnormal identification result. Through the above method, the data of upstream and downstream business systems can be connected, and the low account checking efficiency and the flexibility and adaptability of the account checking can be improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a reconciliation method, system, computer, and storage medium. Background Technology

[0002] With the acceleration of enterprise digitalization, especially the rapid development of the fintech sector, the importance of internal data settlement, reconciliation, and financial auditing is becoming increasingly prominent.

[0003] Currently, many enterprises have introduced various digital management systems, which have improved the efficiency of verifying massive amounts of data compared to the traditional manual reconciliation model. These management systems can extract data from various business modules, perform data calculations and business verifications according to certain rules, and some management systems also support data interoperability with banks and third-party payment systems.

[0004] Existing automated reconciliation methods verify data through classification and matching and storage structure optimization. For discrepancies, intelligent model methods are introduced for identification. However, data breakpoints still exist when reconciling between different systems. The data sources of upstream and downstream business systems are inconsistent, and the data fields are designed according to their own different specifications, making it difficult to verify massive amounts of data from different systems. Furthermore, the reconciliation rules and strategies in the verification process lack flexibility and adaptability to different business systems. When the amount of data from different systems is very large, the efficiency of processing reconciliation data is also low. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a reconciliation method, system, computer, and storage medium. This invention seeks to solve the technical problems in existing technologies where differing data sources and data standards between upstream and downstream business systems lead to difficulties in data integration, low reconciliation efficiency, and a lack of flexibility.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solution: One reconciliation method includes the following steps: Obtain the dataset to be reconciled, which includes upstream consumption bill data and downstream product bill data; clean the dataset to be reconciled to obtain standard upstream bill data and standard downstream bill data. Obtain field mapping rules, and map the standard upstream billing data and the standard downstream billing data to upstream data to be verified and downstream data to be verified, respectively, according to the field mapping rules; Obtain reconciliation scenario parameters and construct a multidimensional difference rule network. The reconciliation scenario parameters include reconciliation range, difference threshold, and comparison dimension identifier group. Based on the reconciliation scenario parameters and the multidimensional difference rule network, compare the upstream data to be verified and the downstream data to be verified to obtain the difference identification result. Anomaly category prediction is performed on the difference identification results to generate anomaly identification results, and an account reconciliation report is generated based on the anomaly identification results.

[0007] Furthermore, the field mapping rules include several field positioning sub-rules and several transformation sub-rules. The transformation sub-rules are direct mapping, type conversion, format conversion, encoding mapping, or formula calculation. The step of mapping the standard upstream billing data and the standard downstream billing data to the upstream data to be verified and the downstream data to be verified, respectively, according to the field mapping rules includes: Both the standard upstream billing data and the standard downstream billing data are divided into several source fields; Based on several transformation sub-rules corresponding to the source fields, several source fields are transformed into several target fields; Based on several field location sub-rules corresponding to the source fields, several target fields are combined to form upstream data to be verified and downstream data to be verified.

[0008] Furthermore, the steps for constructing the multidimensional difference rule network include: Obtain the rule base, select several rule templates from the rule base based on the reconciliation scenario parameters, and convert the several rule templates into several rule use cases; Obtain priority configuration parameters, and sort several rule test cases into several rule chains according to the priority configuration parameters; Traverse the rule chain and determine in turn whether there is a data reference relationship between two adjacent rule use cases in the rule chain. If there is no data reference relationship between two adjacent rule use cases, then the two rule use cases are grouped into a parallel execution group. Based on the number of parallel execution groups, the execution mode of the rule chain is determined, which is serial execution, parallel execution, or hybrid execution. Computational resource configuration parameters are obtained. Based on the execution mode, the computational resource configuration parameters, and several rule chains, a multidimensional differential rule network is constructed.

[0009] Furthermore, the comparison dimension identifier group includes at least one of time dimension identifier, product dimension identifier, and account dimension identifier. The step of comparing the upstream data to be verified and the downstream data to be verified based on the reconciliation scenario parameters and the multidimensional difference rule network to obtain the difference identification result includes: Based on the upstream data to be verified and the downstream data to be verified, combined with the reconciliation scope and the comparison dimension identifier group, a multidimensional aggregation sequence group is constructed. The multidimensional aggregation sequence group includes at least one of time dimension aggregation sequence, product dimension aggregation sequence and account dimension aggregation sequence. The multidimensional aggregated sequence group is input into the multidimensional difference rule network to obtain several difference values. The difference values ​​are compared with the difference threshold to obtain several difference location points. The several difference location points constitute the difference recognition result.

[0010] Furthermore, the step of constructing a multidimensional aggregated sequence group based on the upstream data to be verified and the downstream data to be verified, combined with the reconciliation range and the comparison dimension identifier group, includes: When the comparison dimension identifier group includes the time dimension identifier, the time range is extracted from the reconciliation range, and several time grouping keys are extracted from the timestamp fields of the upstream data to be verified and the downstream data to be verified based on the time range. Data is extracted from the upstream data to be verified and the downstream data to be verified based on the time grouping keys, and the upstream total amount, upstream total number of transactions, downstream total amount and downstream total number of transactions are calculated to form a time dimension subsequence. Several time dimension subsequences are combined into a time dimension aggregation sequence. When the comparison dimension identifier group includes the product dimension identifier, the product range is extracted from the reconciliation range. Based on the product range, several product grouping keys are extracted from the product number field in the upstream data to be verified and the downstream data to be verified. Based on the product grouping keys, data is extracted from the upstream data to be verified and the downstream data to be verified. The upstream total amount, the upstream total number of transactions, the downstream total amount, and the downstream total number of transactions are calculated to form a product dimension subsequence. Several product dimension subsequences are combined into a product dimension aggregation sequence. When the comparison dimension identifier group includes the account dimension identifier, the account range is extracted from the reconciliation range. Based on the account range, several account grouping keys are extracted from the account code field in the upstream data to be verified and the downstream data to be verified. Based on the account grouping keys, data is extracted from the upstream data to be verified and the downstream data to be verified. The upstream total amount, the upstream total number of transactions, the downstream total amount, and the downstream total number of transactions are calculated to form an account dimension subsequence. Several account dimension subsequences are combined to form an account dimension aggregation sequence.

[0011] Furthermore, the step of predicting the anomaly category from the difference identification results to generate anomaly identification results includes: Extract upstream difference data and downstream difference data corresponding to the difference identification results from the upstream data to be verified and the downstream data to be verified; Obtain an anomaly classification model, input the upstream difference data and the downstream difference data into the anomaly classification model to predict the anomaly category, so as to obtain the anomaly category; If the anomaly category is a caliber conflict, then the subtype of the caliber conflict is determined and a correction suggestion is generated. The anomaly category and the correction suggestion constitute the anomaly identification result. If the anomaly category is a missing or duplicate bill, a data tracing report is generated. The anomaly category and the data tracing report together constitute the anomaly identification result.

[0012] Furthermore, the step of determining the subtype of caliber conflict also includes: Obtain a preset tax rate, and based on the preset tax rate, the upstream difference data, and the downstream difference data, determine whether the subtype of the tax rate conflict is a tax rate conflict. If the subtype of the conflicting caliber is not a tax caliber conflict, then obtain the field mapping status to determine whether the subtype of the conflicting caliber is a mapping caliber conflict.

[0013] A reconciliation system, employing the reconciliation method described in the above technical solution, the system comprising: The acquisition module is used to acquire the dataset to be reconciled, which includes upstream consumption bill data and downstream product bill data. The dataset to be reconciled is cleaned to obtain standard upstream bill data and standard downstream bill data. The mapping module is used to obtain field mapping rules and map the standard upstream billing data and the standard downstream billing data to upstream data to be verified and downstream data to be verified, respectively, according to the field mapping rules. The difference module is used to obtain reconciliation scenario parameters and construct a multi-dimensional difference rule network. The reconciliation scenario parameters include reconciliation range, difference threshold and comparison dimension identifier group. Based on the reconciliation scenario parameters and the multi-dimensional difference rule network, the upstream data to be verified and the downstream data to be verified are compared to obtain the difference identification result. The category module is used to predict the anomaly category of the difference identification results, generate anomaly identification results, and generate a reconciliation report based on the anomaly identification results.

[0014] A computer includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the reconciliation method as described above.

[0015] A storage medium storing a computer program that, when executed by a processor, implements the reconciliation method as described in the above technical solution.

[0016] Compared with existing technologies, the advantages of this invention are as follows: By cleaning and mapping data, it solves the problem of inconsistent field naming, encoding, and business definitions between upstream and downstream systems. When the electronic data format of upstream and downstream systems changes or a new system is connected, adaptation can be achieved simply by modifying the field mapping rules. By dynamically constructing the multi-dimensional difference rule network, it can adapt to the reconciliation needs of different reconciliation business scenarios. When business rules change, the priority configuration parameters can be changed to adjust the rule chain. When computing resources or computing requirements change, the concurrent market data within the engine can be adjusted through the computing resource configuration parameters. Furthermore, the multidimensional difference rule network can quickly quantify the differences in reconciliation data between upstream and downstream, improving reconciliation efficiency. By dynamically constructing the multidimensional aggregation sequence group, different dimensions can be flexibly selected for data aggregation according to reconciliation needs. A large amount of detailed data can be grouped and summarized according to different grouping keys before comparison, avoiding the need to compare a large amount of data line by line, thus improving comparison efficiency. For the difference value that exceeds the difference threshold, its grouping key and other related field information can be quickly located according to its aggregation sequence, efficiently locating the source of reconciliation difference. The predictive model can quickly classify the cause of the anomaly, completing intelligent anomaly identification and processing, greatly reducing manual intervention and improving reconciliation efficiency. Attached Figure Description

[0017] Figure 1 This is a flowchart of the reconciliation method in the first embodiment of the present invention; Figure 2 This is a structural block diagram of the reconciliation system in the second embodiment of the present invention; The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0018] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0019] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0021] Please see Figure 1 The reconciliation method in the first embodiment of the present invention includes the following steps: Step S10: Obtain the reconciliation dataset, which includes upstream consumption bill data and downstream product bill data. Clean the reconciliation dataset to obtain standard upstream bill data and standard downstream bill data. Preferably, the upstream consumer billing data records the user's consumption behavior, such as a transaction record of a payment, including transaction serial number, transaction time, transaction amount, merchant code, payment method, handling fee, etc. It is highly real-time, comes from multiple different business systems and has different data formats. The downstream product billing data refers to the settlement billing data generated by the settlement platform according to the settlement rules, including settlement batch number, settlement cycle, total transaction amount, total number of transactions, deduction details, etc. It has strong periodicity, summary rows, is affected by settlement rules, and is financial settlement oriented. Furthermore, data cleaning includes format verification, removal of data with incorrect format, outlier detection, removal of outliers, and normalization processes, which are beneficial to improving the quality of data processing and avoiding reconciliation errors.

[0022] Step S20: Obtain field mapping rules, and map the standard upstream billing data and the standard downstream billing data to upstream data to be verified and downstream data to be verified respectively according to the field mapping rules; Preferably, the standard upstream billing data and the standard downstream billing data are automatically converted into the upstream data to be verified and the downstream data to be verified with unified field definitions through the field mapping rules. This solves the problems of inconsistent field naming, inconsistent encoding systems, and inconsistent data formats between different systems, and lays a data foundation for subsequent multi-dimensional comparisons.

[0023] In step S20, the field mapping rule includes several field positioning sub-rules and several transformation sub-rules. The transformation sub-rules are direct mapping, type conversion, format conversion, encoding mapping, or formula calculation. Step S20 includes: S210: Divide both the standard upstream billing data and the standard downstream billing data into several source fields; Preferably, the data fields are clearly defined. For example, for financial service products, the source fields defined in the standard upstream billing data include "merchant_code", "trans_amount", "trans_date", "product_type", etc., and the source fields defined in the standard downstream billing data include "partner_id", "fee_amount", "billing_date", "service_id", etc.

[0024] S220: Based on several transformation sub-rules corresponding to the source fields, convert several source fields into several target fields; Preferably, the direct mapping refers to directly assigning the value of the source field to the target field without any conversion; the type conversion refers to converting the data type of the source field to the data type specified by the target field, such as converting String to BigDecimal; the format conversion refers to converting the data format of the source field to the format specified by the target field, such as converting the date format from "20240115" to "2024-01-15"; the encoding mapping refers to converting the encoding value of the source field to the target encoding value according to a preset encoding mapping table, such as converting the upstream merchant code "M001" to the downstream unified merchant code "100001"; and the formula calculation refers to calculating the source field according to a preset calculation formula and then assigning the calculation result to the target field, such as calculating the handling fee based on the transaction amount and fee rate.

[0025] S230: Based on several field location sub-rules corresponding to the source fields, several target fields are combined into upstream data to be verified and downstream data to be verified.

[0026] Understandably, the conversion sub-rules support encoding mapping and formula calculation, and can flexibly handle complex caliber conversion needs. When the composition of upstream and downstream business systems changes, such as the addition of multiple consumer sub-systems upstream, the data source and data format change, only the field mapping rules need to be adjusted to adaptably unify the data.

[0027] Step S30: Obtain reconciliation scenario parameters and construct a multidimensional difference rule network. The reconciliation scenario parameters include reconciliation range, difference threshold and comparison dimension identifier group. Based on the reconciliation scenario parameters and the multidimensional difference rule network, compare the upstream data to be verified and the downstream data to be verified to obtain the difference identification result. Preferably, the reconciliation scope refers to the data range limitation conditions that need to be processed in this reconciliation task, used to filter out data records that meet the conditions from the upstream data to be verified and the downstream data to be verified. The difference threshold refers to the critical value for judging whether the amount difference is within the allowable range. Understandably, by dynamically constructing the multi-dimensional difference rule network based on the reconciliation scenario parameters, the system can flexibly adapt to the reconciliation needs of different business scenarios. When business rules change, only the reconciliation scenario parameters need to be updated without modifying the core code, greatly improving the system's adaptability and scalability.

[0028] In step S30, the comparison dimension identifier group includes at least one of time dimension identifier, product dimension identifier, and account dimension identifier, and step S30 includes: S310: Obtain the rule base, select several rule templates from the rule base based on the reconciliation scenario parameters, and convert the several rule templates into several rule use cases; Preferably, the rule base pre-stores templates for amount comparison rules, number of transactions comparison rules, time comparison rules, and existence comparison rules. Each rule template defines the framework of the comparison logic and configurable parameters. For example, the amount comparison rule is used to compare whether the upstream amount and the downstream amount are consistent, and a tolerance range can be set; the number of transactions comparison rule is used to compare whether the number of upstream consumer transactions is consistent with the number of downstream product settlement transactions; the time comparison rule is used to compare whether the transaction time and the settlement cycle match; and the existence comparison rule is used to compare whether the records exist in both upstream and downstream data. The rule use cases are code instances formed based on the rule templates.

[0029] S320: Obtain priority configuration parameters, and sort several rule test cases into several rule chains according to the priority configuration parameters; Preferably, the priority can be configured according to different rule types. For example, the existence rule has the highest priority, and the amount comparison rule is second only to the existence rule.

[0030] S330: Traverse the rule chain and determine in turn whether there is a data reference relationship between two adjacent rule use cases in the rule chain. If there is no data reference relationship between two adjacent rule use cases, then the two rule use cases are grouped into a parallel execution group. S340: Based on the number of parallel execution groups, determine the execution mode of the rule chain, wherein the execution mode is serial execution, parallel execution, or hybrid execution; obtain computing resource configuration parameters; and construct a multidimensional differential rule network based on the execution mode, the computing resource configuration parameters, and several rule chains.

[0031] Preferably, the computing resource configuration parameters include the number of threads and the execution timeout, which can be flexibly configured according to the reconciliation task and the computer equipment resources used, which is beneficial to improving reconciliation efficiency.

[0032] S350: Based on the upstream data to be verified and the downstream data to be verified, combined with the reconciliation scope and the comparison dimension identifier group, a multi-dimensional aggregation sequence group is constructed. The multi-dimensional aggregation sequence group includes at least one of time dimension aggregation sequence, product dimension aggregation sequence and account dimension aggregation sequence. The S350 also includes: S3510: When the comparison dimension identifier group includes the time dimension identifier, extract the time range from the reconciliation range, extract several time grouping keys from the timestamp fields of the upstream data to be verified and the downstream data to be verified based on the time range, extract data from the upstream data to be verified and the downstream data to be verified based on the time grouping keys, and calculate the upstream total amount, the upstream total number of transactions, the downstream total amount and the downstream total number of transactions to form a time dimension subsequence, and combine several time dimension subsequences into a time dimension aggregation sequence; S3520: When the comparison dimension identifier group includes the product dimension identifier, extract the product range from the reconciliation range, extract several product grouping keys from the product number field in the upstream data to be verified and the downstream data to be verified based on the product range, extract data from the upstream data to be verified and the downstream data to be verified based on the product grouping keys, and calculate the upstream total amount, the upstream total number of transactions, the downstream total amount and the downstream total number of transactions to form a product dimension subsequence, and combine several product dimension subsequences into a product dimension aggregation sequence; S3530: When the comparison dimension identifier group includes the account dimension identifier, extract the account range from the reconciliation range, extract several account grouping keys from the account code field in the upstream data to be verified and the downstream data to be verified based on the account range, extract data from the upstream data to be verified and the downstream data to be verified based on the account grouping keys, and calculate the upstream total amount, the upstream total number of transactions, the downstream total amount and the downstream total number of transactions to form an account dimension subsequence, and combine several account dimension subsequences into an account dimension aggregation sequence.

[0033] Preferably, the account scope can be a specified list of merchants, and the product scope can be used to define financial product types such as payment service products and credit service products.

[0034] S360: Input the multidimensional aggregated sequence group into the multidimensional difference rule network to obtain several difference values, compare the difference values ​​with the difference threshold to obtain several difference location points, and the several difference location points constitute the difference recognition result.

[0035] Preferably, if the comparison dimension identifier group simultaneously includes the time dimension identifier, the product dimension identifier, and the account dimension identifier, then a three-dimensional cross-comparison is performed. A first dimension is established based on the time grouping key, a second dimension is established based on the product grouping key, and a third dimension is established based on the account grouping key, thus constructing a three-dimensional cross-comparison matrix. The three-dimensional cross-comparison matrix includes several cells. Data matching the time, product, and account subject of the cell is extracted from the upstream data to be verified and the downstream data to be verified. The upstream aggregation value and the downstream aggregation value are calculated respectively through the multi-dimensional difference rule network to obtain the difference value corresponding to the cell. If the difference value is greater than the difference threshold, then the cell is established as the difference location point.

[0036] Understandably, by dynamically constructing the multidimensional aggregated sequence group based on the comparison dimension identifier group, massive amounts of detailed data are compressed into a small amount of summary data for comparison, significantly improving comparison efficiency. By inputting the multidimensional aggregated sequence group into the multidimensional difference rule engine for automated comparison, the tedious process of manual verification is avoided. By comparing the difference value with the difference threshold, small-range calculation errors can be tolerated, avoiding false alarms in reconciliation difference results. By obtaining the difference location point, the specific date, product type, or account entity where the difference exists can be quickly located, enabling rapid identification of the source of the difference.

[0037] Step S40: Perform anomaly category prediction on the difference identification results to generate anomaly identification results, and generate a reconciliation report based on the anomaly identification results.

[0038] Preferably, the difference recognition results are quickly classified using deep learning technology to perform intelligent diagnosis and generate corresponding report results, providing a basis for subsequent processing.

[0039] Step S40 includes: S410: Extract upstream difference data and downstream difference data corresponding to the difference identification result from the upstream data to be verified and the downstream data to be verified; S420: Obtain the anomaly classification model, input the upstream difference data and the downstream difference data into the anomaly classification model to predict the anomaly category, so as to obtain the anomaly category; Preferably, the anomaly classification model is trained using an MLP architecture, and the anomaly category is used as the training label. The training label can be a conflict of standards, missing bills, inconsistent amounts, duplicate records, etc.

[0040] S430: If the anomaly category is a caliber conflict, then determine the subtype of the caliber conflict and generate a correction suggestion. The anomaly category and the correction suggestion constitute the anomaly identification result. Preferably, a discrepancy in accounting standards indicates that the upstream consumer billing data and the downstream product billing data are inconsistent in terms of data definition, calculation rules, and statistical standards. For example, the upstream amount includes tax while the downstream amount does not; the upstream is based on transaction time while the downstream is based on settlement time; or the upstream and downstream use different coding systems.

[0041] The S430 also includes: S4310: Obtain a preset tax rate, and based on the preset tax rate, the upstream difference data, and the downstream difference data, determine whether the subtype of the tax rate conflict is a tax rate conflict. Preferably, the preset tax rate includes a preset tax rate and a preset fee rate. When checking differences in tax rates, the difference rate between the upstream difference data and the downstream difference data is calculated. If the difference rate is close to the preset tax rate, it indicates a conflict in the tax rate scope. The ratio of the upstream difference data to the downstream difference data is calculated and compared with the preset fee rate to identify whether there is a conflict in the product fee rate scope. If a conflict in the tax rate scope is confirmed, it is helpful to prompt the modification of the mapping rules, add the tax calculation formula, and avoid batch reconciliation differences.

[0042] S4320: If the subtype of the conflicting caliber is not a tax caliber conflict, then obtain the field mapping status to determine whether the subtype of the conflicting caliber is a mapping caliber conflict.

[0043] Preferably, the main focus is on obtaining the status of the encoding mapping and verifying whether the encodings in different business systems have been fully mapped.

[0044] S440: If the anomaly category is a missing or duplicate record in a bill, a data tracing report is generated, and the anomaly category and the data tracing report constitute the anomaly identification result.

[0045] Preferably, missing bills indicate possible data synchronization failures or abnormal bill generation; duplicate records indicate possible duplicate data pushes or system duplicate processing. Data tracing can be performed on the data flow path of multiple business systems based on key identifiers such as transaction serial number, transaction time, and transaction amount to find the location of data errors. The reconciliation report includes a summary of differences, details of differences, analysis of the causes of differences, and handling suggestions, which helps users to intuitively understand the reconciliation results and make quick decisions and take action.

[0046] Please see Figure 2 The reconciliation system provided in the second embodiment of the present invention applies the reconciliation method described in the first embodiment above, and the system includes: The acquisition module 10 is used to acquire the dataset to be reconciled, which includes upstream consumption bill data and downstream product bill data. The dataset to be reconciled is cleaned to obtain standard upstream bill data and standard downstream bill data. Mapping module 20 is used to obtain field mapping rules and map the standard upstream billing data and the standard downstream billing data to upstream data to be verified and downstream data to be verified, respectively, according to the field mapping rules. In the mapping module 20, the field mapping rules include several field positioning sub-rules and several transformation sub-rules. The transformation sub-rules are direct mapping, type conversion, format conversion, encoding mapping, or formula calculation. The mapping module 20 includes: The first unit is used to divide both the standard upstream billing data and the standard downstream billing data into several source fields; The second unit is used to convert several source fields into several target fields according to several transformation sub-rules corresponding to the source fields; The third unit is used to assemble several target fields into upstream data to be verified and downstream data to be verified, based on several field location sub-rules corresponding to the source fields.

[0047] The difference module 30 is used to obtain reconciliation scenario parameters and construct a multi-dimensional difference rule network. The reconciliation scenario parameters include reconciliation range, difference threshold and comparison dimension identifier group. Based on the reconciliation scenario parameters and the multi-dimensional difference rule network, the upstream data to be verified and the downstream data to be verified are compared to obtain the difference identification result. In the difference module 30, the comparison dimension identifier group includes at least one of time dimension identifier, product dimension identifier, and account dimension identifier, and the difference module 30 includes: The fourth unit is used to obtain a rule base, select several rule templates from the rule base based on the reconciliation scenario parameters, and convert the several rule templates into several rule use cases; The fifth unit is used to obtain priority configuration parameters and sort several rule test cases into several rule chains according to the priority configuration parameters. The sixth unit is used to traverse the rule chain and determine whether there is a data reference relationship between two adjacent rule use cases in the rule chain. If there is no data reference relationship between two adjacent rule use cases, the two rule use cases are grouped into a parallel execution group. The seventh unit is used to determine the execution mode of the rule chain based on the number of parallel execution groups, wherein the execution mode is serial execution, parallel execution, or hybrid execution; obtain computing resource configuration parameters; and construct a multidimensional differential rule network based on the execution mode, the computing resource configuration parameters, and several rule chains.

[0048] The eighth unit is used to construct a multi-dimensional aggregation sequence group based on the upstream data to be verified and the downstream data to be verified, combined with the reconciliation range and the comparison dimension identifier group. The multi-dimensional aggregation sequence group includes at least one of time dimension aggregation sequence, product dimension aggregation sequence and account dimension aggregation sequence. The eighth unit is specifically used for: When the comparison dimension identifier group includes the time dimension identifier, the time range is extracted from the reconciliation range, and several time grouping keys are extracted from the timestamp fields of the upstream data to be verified and the downstream data to be verified based on the time range. Data is extracted from the upstream data to be verified and the downstream data to be verified based on the time grouping keys, and the upstream total amount, upstream total number of transactions, downstream total amount and downstream total number of transactions are calculated to form a time dimension subsequence. Several time dimension subsequences are combined into a time dimension aggregation sequence. When the comparison dimension identifier group includes the product dimension identifier, the product range is extracted from the reconciliation range. Based on the product range, several product grouping keys are extracted from the product number field in the upstream data to be verified and the downstream data to be verified. Based on the product grouping keys, data is extracted from the upstream data to be verified and the downstream data to be verified. The upstream total amount, the upstream total number of transactions, the downstream total amount, and the downstream total number of transactions are calculated to form a product dimension subsequence. Several product dimension subsequences are combined into a product dimension aggregation sequence. When the comparison dimension identifier group includes the account dimension identifier, the account range is extracted from the reconciliation range. Based on the account range, several account grouping keys are extracted from the account code field in the upstream data to be verified and the downstream data to be verified. Based on the account grouping keys, data is extracted from the upstream data to be verified and the downstream data to be verified. The upstream total amount, the upstream total number of transactions, the downstream total amount, and the downstream total number of transactions are calculated to form an account dimension subsequence. Several account dimension subsequences are combined to form an account dimension aggregation sequence.

[0049] The ninth unit is used to input the multidimensional aggregated sequence group into the multidimensional difference rule network to obtain several difference values, compare the difference values ​​with the difference threshold to obtain several difference location points, and the several difference location points constitute the difference recognition result.

[0050] Category module 40 is used to predict the abnormal category of the difference identification result and generate an abnormal identification result, so as to generate a reconciliation report based on the abnormal identification result.

[0051] The category module 40 includes: The tenth unit is used to extract upstream difference data and downstream difference data corresponding to the difference identification result from the upstream data to be verified and the downstream data to be verified. The eleventh unit is used to obtain an anomaly classification model. The upstream difference data and the downstream difference data are input into the anomaly classification model to predict the anomaly category and obtain the anomaly category. The twelfth unit is used to determine the subtype of the caliber conflict if the anomaly category is a caliber conflict, and generate a correction suggestion. The anomaly category and the correction suggestion constitute the anomaly identification result. The twelfth unit is specifically used for: Obtain a preset tax rate, and based on the preset tax rate, the upstream difference data, and the downstream difference data, determine whether the subtype of the tax rate conflict is a tax rate conflict. If the subtype of the conflicting caliber is not a tax caliber conflict, then obtain the field mapping status to determine whether the subtype of the conflicting caliber is a mapping caliber conflict.

[0052] If the anomaly category is a missing or duplicate bill, a data tracing report is generated. The anomaly category and the data tracing report together constitute the anomaly identification result.

[0053] A third embodiment of the present invention provides a computer, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the reconciliation method as described in the first embodiment.

[0054] The fourth embodiment of the present invention provides a storage medium on which a computer program is stored, which, when executed by a processor, implements the reconciliation method as described in the first embodiment.

[0055] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0056] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A reconciliation method, characterized in that, Includes the following steps: Obtain the dataset to be reconciled, which includes upstream consumption bill data and downstream product bill data; clean the dataset to be reconciled to obtain standard upstream bill data and standard downstream bill data. Obtain field mapping rules, and map the standard upstream billing data and the standard downstream billing data to upstream data to be verified and downstream data to be verified, respectively, according to the field mapping rules; Obtain reconciliation scenario parameters and construct a multidimensional difference rule network. The reconciliation scenario parameters include reconciliation range, difference threshold, and comparison dimension identifier group. Based on the reconciliation scenario parameters and the multidimensional difference rule network, compare the upstream data to be verified and the downstream data to be verified to obtain the difference identification result. Anomaly category prediction is performed on the difference identification results to generate anomaly identification results, and an account reconciliation report is generated based on the anomaly identification results.

2. The reconciliation method according to claim 1, characterized in that, The field mapping rules include several field location sub-rules and several transformation sub-rules. The transformation sub-rules are direct mapping, type conversion, format conversion, encoding mapping, or formula calculation. The step of mapping the standard upstream billing data and the standard downstream billing data to the upstream data to be verified and the downstream data to be verified, respectively, according to the field mapping rules includes: Both the standard upstream billing data and the standard downstream billing data are divided into several source fields; Based on several transformation sub-rules corresponding to the source fields, several source fields are transformed into several target fields; Based on several field location sub-rules corresponding to the source fields, several target fields are combined to form upstream data to be verified and downstream data to be verified.

3. The reconciliation method according to claim 1, characterized in that, The steps for constructing the multidimensional difference rule network include: Obtain the rule base, select several rule templates from the rule base based on the reconciliation scenario parameters, and convert the several rule templates into several rule use cases; Obtain priority configuration parameters, and sort several rule test cases into several rule chains according to the priority configuration parameters; Traverse the rule chain and determine in turn whether there is a data reference relationship between two adjacent rule use cases in the rule chain. If there is no data reference relationship between two adjacent rule use cases, then the two rule use cases are grouped into a parallel execution group. Based on the number of parallel execution groups, the execution mode of the rule chain is determined, which is serial execution, parallel execution, or hybrid execution. Computational resource configuration parameters are obtained. Based on the execution mode, the computational resource configuration parameters, and several rule chains, a multidimensional differential rule network is constructed.

4. The reconciliation method according to claim 1, characterized in that, The comparison dimension identifier group includes at least one of time dimension identifier, product dimension identifier, and account dimension identifier. The step of comparing the upstream data to be verified and the downstream data to be verified based on the reconciliation scenario parameters and the multidimensional difference rule network to obtain the difference identification result includes: Based on the upstream data to be verified and the downstream data to be verified, combined with the reconciliation scope and the comparison dimension identifier group, a multidimensional aggregation sequence group is constructed. The multidimensional aggregation sequence group includes at least one of time dimension aggregation sequence, product dimension aggregation sequence and account dimension aggregation sequence. The multidimensional aggregated sequence group is input into the multidimensional difference rule network to obtain several difference values. The difference values ​​are compared with the difference threshold to obtain several difference location points. The several difference location points constitute the difference recognition result.

5. The reconciliation method according to claim 4, characterized in that, The step of constructing a multidimensional aggregated sequence group based on the upstream data to be verified and the downstream data to be verified, combined with the reconciliation range and the comparison dimension identifier group, includes: When the comparison dimension identifier group includes the time dimension identifier, the time range is extracted from the reconciliation range, and several time grouping keys are extracted from the timestamp fields of the upstream data to be verified and the downstream data to be verified based on the time range. Data is extracted from the upstream data to be verified and the downstream data to be verified based on the time grouping keys, and the upstream total amount, upstream total number of transactions, downstream total amount and downstream total number of transactions are calculated to form a time dimension subsequence. Several time dimension subsequences are combined into a time dimension aggregation sequence. When the comparison dimension identifier group includes the product dimension identifier, the product range is extracted from the reconciliation range. Based on the product range, several product grouping keys are extracted from the product number field in the upstream data to be verified and the downstream data to be verified. Based on the product grouping keys, data is extracted from the upstream data to be verified and the downstream data to be verified. The upstream total amount, the upstream total number of transactions, the downstream total amount, and the downstream total number of transactions are calculated to form a product dimension subsequence. Several product dimension subsequences are combined into a product dimension aggregation sequence. When the comparison dimension identifier group includes the account dimension identifier, the account range is extracted from the reconciliation range. Based on the account range, several account grouping keys are extracted from the account code field in the upstream data to be verified and the downstream data to be verified. Based on the account grouping keys, data is extracted from the upstream data to be verified and the downstream data to be verified. The upstream total amount, the upstream total number of transactions, the downstream total amount, and the downstream total number of transactions are calculated to form an account dimension subsequence. Several account dimension subsequences are combined to form an account dimension aggregation sequence.

6. The reconciliation method according to claim 1, characterized in that, The step of predicting the anomaly category from the difference identification result to generate anomaly identification result includes: Extract upstream difference data and downstream difference data corresponding to the difference identification results from the upstream data to be verified and the downstream data to be verified; Obtain an anomaly classification model, input the upstream difference data and the downstream difference data into the anomaly classification model to predict the anomaly category, so as to obtain the anomaly category; If the anomaly category is a caliber conflict, then the subtype of the caliber conflict is determined and a correction suggestion is generated. The anomaly category and the correction suggestion constitute the anomaly identification result. If the anomaly category is a missing or duplicate bill, a data tracing report is generated. The anomaly category and the data tracing report together constitute the anomaly identification result.

7. The reconciliation method according to claim 6, characterized in that, The steps for determining the subtype of caliber conflict also include: Obtain a preset tax rate, and based on the preset tax rate, the upstream difference data, and the downstream difference data, determine whether the subtype of the tax rate conflict is a tax rate conflict. If the subtype of the conflicting caliber is not a tax caliber conflict, then obtain the field mapping status to determine whether the subtype of the conflicting caliber is a mapping caliber conflict.

8. A reconciliation system, employing the reconciliation method as described in any one of claims 1 to 7, characterized in that, The system includes: The acquisition module is used to acquire the dataset to be reconciled, which includes upstream consumption bill data and downstream product bill data. The dataset to be reconciled is cleaned to obtain standard upstream bill data and standard downstream bill data. The mapping module is used to obtain field mapping rules and map the standard upstream billing data and the standard downstream billing data to upstream data to be verified and downstream data to be verified, respectively, according to the field mapping rules. The difference module is used to obtain reconciliation scenario parameters and construct a multi-dimensional difference rule network. The reconciliation scenario parameters include reconciliation range, difference threshold and comparison dimension identifier group. Based on the reconciliation scenario parameters and the multi-dimensional difference rule network, the upstream data to be verified and the downstream data to be verified are compared to obtain the difference identification result. The category module is used to predict the anomaly category of the difference identification results, generate anomaly identification results, and generate a reconciliation report based on the anomaly identification results.

9. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the reconciliation method as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the reconciliation method as described in any one of claims 1 to 7.

Images