Reconciliation difference analysis method and device, medium and electronic equipment

Abstract

The invention provides an account checking difference analysis method and device, a medium and electronic equipment, and the method comprises the steps: decoupling an accounting rule in a hosting accounting system from an accounting engine, and obtaining pre-accounting information which comprises product information, an accounting date and a product accounting parameter; calling an accounting rule to carry out trial calculation processing on the pre-accounting information to obtain a trial calculation result; establishing a model by using a mixed feature type factor screening algorithm in machine learning, and training the model by using historical accounting transaction data to obtain an optimization algorithm model; processing the trial calculation result by adopting an optimization algorithm model to obtain a product accounting parameter combination; and comparing the product accounting parameter combination with the existing product parameter combination of the trusteeship accounting system, analyzing the difference, and determining the reason of the accounting difference. Therefore, the problems that the account checking difference cannot be analyzed and the parameter setting error of the accounting system possibly exists in the prior art are solved.

Description

Technical Field

[0001] This application relates to the field of financial data processing technology, and more specifically, to a reconciliation discrepancy analysis method, a reconciliation discrepancy analysis device, a computer-readable storage medium, and an electronic device. Background Technology

[0002] Related technologies involve acquiring system reconciliation results data, identifying discrepancies, and then performing a second reconciliation to analyze whether the discrepancy is due to data source delays or a malfunction in the financial audit system. Other technical solutions involve manual analysis of the causes of reconciliation discrepancies, including adjusting product accounting parameter combinations and rerunning the reconciliation process using Excel.

[0003] Ultimately, there are only two possible reasons for reconciliation discrepancies: data source delays or system failures. Therefore, if technical analysis fails to identify a discrepancy, it could also be due to incorrect parameter settings in the accounting system. Manually analyzing reconciliation discrepancies through rerunning records or using Excel for derivation is slow and prone to errors. Furthermore, rerunning records generates redundant vouchers and balance data.

[0004] In other words, the inability of relevant technologies to analyze reconciliation discrepancies may also be due to incorrect parameter settings in the accounting system. Summary of the Invention

[0005] The main objective of this application is to provide a method, apparatus, computer-readable storage medium, and electronic device for analyzing reconciliation discrepancies, so as to at least solve the problem that related technologies may fail to analyze reconciliation discrepancies due to incorrect parameter settings in the accounting system.

[0006] To achieve the above objectives, according to one aspect of this application, a method for analyzing reconciliation discrepancies is provided, comprising: decoupling the accounting rules in the escrow accounting system from the accounting engine, and obtaining pre-accounting information, the pre-accounting information including product information, accounting date, and product accounting parameters; calling the accounting rules to perform trial calculations on the pre-accounting information to obtain trial calculation results; establishing a model using a hybrid feature factor screening algorithm in machine learning, and training the model using historical accounting transaction data to obtain an optimized algorithm model; processing the trial calculation results using the optimized algorithm model to obtain a combination of product accounting parameters; comparing the combination of product accounting parameters with the existing product parameter combinations in the escrow accounting system and analyzing the differences to determine the causes of the accounting discrepancies.

[0007] Optionally, the accounting rules in the managed accounting system are decoupled from the accounting engine, including: using interface technology to separate the accounting rules from the core accounting engine of the managed accounting system; encapsulating the execution logic of the accounting rules with an independent service layer; and updating and applying new accounting rules in real time, at least according to the execution logic of the accounting rules, without restarting the system.

[0008] Optionally, the model is trained using historical accounting transaction data to obtain an optimized algorithm model, including: using a feature selection algorithm to filter and process the historical accounting transaction data to obtain a first accounting parameter, wherein the feature selection algorithm is one of the following: LASSO regression, random forest, and gradient boosting tree; using a feature dimensionality reduction method to reduce the dimensionality of the first accounting parameter to obtain a second accounting parameter; and using the second accounting parameter to train the model to obtain the optimized algorithm model.

[0009] Optionally, the optimization algorithm model is used to process the trial calculation results to obtain a product accounting parameter combination, including: processing the trial calculation results using the optimization algorithm model to obtain multiple initial product accounting parameter combinations and corresponding confidence levels; and determining the initial product accounting parameter combination corresponding to the maximum value of the confidence level as the product accounting parameter combination.

[0010] Optionally, after comparing and analyzing the differences between the product accounting parameter combination and the existing product parameter combination in the managed accounting system to determine the cause of the accounting difference, the method includes: comparing the calculated product accounting parameter combination with the existing system parameter configuration to generate a difference report; and displaying the accounting parameters and accounting rules related to the reconciliation difference in a visual manner based on the difference report.

[0011] Optionally, before establishing a model using a hybrid feature-based factor screening algorithm in machine learning and training the model with historical accounting transaction data to obtain an optimized algorithm model, the method further includes: using statistical or machine learning techniques to identify abnormal accounting parameters and rules, and generating corresponding early warning information.

[0012] Optionally, after establishing a model using a hybrid feature factor selection algorithm in machine learning, the method further includes: upon receiving a parameter correction instruction, correcting the corresponding parameters in the model according to the parameter correction instruction.

[0013] According to another aspect of this application, a reconciliation discrepancy analysis device is provided, comprising: a first processing unit, configured to decouple the accounting rules in the managed accounting system from the accounting engine and obtain pre-accounting information, the pre-accounting information including product information, accounting date, and product accounting parameters; a second processing unit, configured to call the accounting rules to perform trial calculations on the pre-accounting information to obtain trial calculation results; a third processing unit, configured to establish a model using a hybrid feature factor screening algorithm in machine learning and train the model using historical accounting transaction data to obtain an optimized algorithm model; a fourth processing unit, configured to process the trial calculation results using the optimized algorithm model to obtain a combination of product accounting parameters; and a fifth processing unit, configured to compare the combination of product accounting parameters with the existing combination of product parameters in the managed accounting system and analyze the differences to determine the cause of the accounting discrepancy.

[0014] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform any of the methods described.

[0015] According to another aspect of this application, an electronic device is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including methods for performing any one of the methods described.

[0016] By applying the technical solution of this application, decoupling allows accounting rules to run independently of the accounting engine. This means that rules can be modified and added more flexibly without affecting the stability and efficiency of the entire system. This independence also facilitates future expansion to more accounting scenarios, improving the system's adaptability. The hybrid feature-based factor screening algorithm, through model optimization, can filter out the most critical parameter set from a large number of product accounting parameter combinations, thereby more accurately identifying the specific rules causing accounting discrepancies. This is more scientific and efficient than traditional manual analysis or simple statistical methods, reducing the possibility of misjudgment. This invention can automatically call accounting rules for trial calculation and automatically analyze discrepancies through an optimized algorithm model, thus avoiding the tedious process of manually adjusting parameters and repeating accounting, greatly saving human resources and reducing operating costs. Unlike traditional methods that require re-running accounts, this invention does not generate additional redundant data such as vouchers and balances during the reconciliation discrepancy analysis process. This is of great significance for maintaining data cleanliness and reducing storage costs. Using the optimized model to quickly locate the source of accounting errors greatly shortens reconciliation time and improves work efficiency. At the same time, since the model is trained on a large amount of historical data, the accuracy of its analysis results is also guaranteed. Accurately identifying the causes of reconciliation discrepancies helps financial institutions better understand potential problems in their internal control systems and processes, enabling them to make more informed decisions, optimize business processes, and improve service quality. This addresses the issue that some technologies may fail to analyze as causing reconciliation discrepancies due to incorrect parameter settings in the accounting system. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 A flowchart illustrating a reconciliation discrepancy analysis method according to an embodiment of this application is shown;

[0019] Figure 2 A structural block diagram of an account reconciliation discrepancy analysis device provided according to an embodiment of this application is shown. Detailed Implementation

[0020] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application 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 for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises 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 apparatus.

[0023] As described in the background section, reconciliation discrepancies can ultimately only be caused by two reasons: data source delay or system failure. Therefore, the inability of relevant technologies to analyze reconciliation discrepancies may also be due to incorrect parameter settings in the accounting system. Manual analysis of reconciliation discrepancies, through rerunning accounts or derivation from Excel, is slow and prone to errors. Furthermore, rerunning accounts generates redundant vouchers and balance data. To address the issue that the inability of relevant technologies to analyze reconciliation discrepancies may also be due to incorrect parameter settings in the accounting system, embodiments of this application provide a reconciliation discrepancy analysis method, a reconciliation discrepancy analysis device, a computer-readable storage medium, and an electronic device.

[0024] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0025] This embodiment provides a method for analyzing reconciliation discrepancies. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] Figure 1 This is a flowchart of the reconciliation discrepancy analysis method according to an embodiment of this application. For example... Figure 1 As shown, the method includes the following steps:

[0027] Step S101: Decouple the accounting rules in the managed accounting system from the accounting engine and obtain pre-calculation information, including product information, accounting date and product accounting parameters;

[0028] Step S102: Call the above accounting rules to perform trial calculations on the above pre-calculation information to obtain trial calculation results;

[0029] Step S103: Build a model using the hybrid feature factor screening algorithm in machine learning, and train the model using historical accounting transaction data to obtain an optimized algorithm model.

[0030] Step S104: The above-mentioned optimization algorithm model is used to process the above-mentioned trial calculation results to obtain the product accounting parameter combination;

[0031] Step S105: Compare the above product accounting parameter combination with the existing product parameter combination in the above managed accounting system and analyze the differences to determine the reasons for the accounting differences.

[0032] In the above steps, decoupling allows the accounting rules to run independently of the accounting engine. This means that rules can be modified and added more flexibly without affecting the stability and efficiency of the entire system. This independence also facilitates future expansion to more accounting scenarios, improving the system's adaptability. The hybrid feature-based factor screening algorithm, through model optimization, can filter out the most critical parameter set from a large number of product accounting parameter combinations, thereby more accurately identifying the specific rules causing accounting discrepancies. This is more scientific and efficient than traditional manual analysis or simple statistical methods, reducing the possibility of misjudgment. This invention can automatically call accounting rules for trial calculations and automatically analyze discrepancies through optimized algorithm models, thus avoiding the tedious process of manually adjusting parameters and repeating calculations, greatly saving human resources and reducing operating costs. Unlike traditional methods that require re-running accounts, this invention does not generate additional redundant data such as vouchers and balances during the reconciliation discrepancy analysis process. This is of great significance for maintaining data cleanliness and reducing storage costs. Using the optimized model to quickly locate the source of accounting errors greatly shortens reconciliation time and improves work efficiency. At the same time, since the model is trained on a large amount of historical data, the accuracy of its analysis results is also guaranteed. Accurately identifying the causes of reconciliation discrepancies helps financial institutions better understand potential problems in their internal control systems and processes, enabling them to make more informed decisions, optimize business processes, and improve service quality. This addresses the issue that some technologies may fail to analyze as causing reconciliation discrepancies due to incorrect parameter settings in the accounting system.

[0033] The specific implementation process of building a model using the hybrid feature factor selection algorithm in machine learning includes: Data preparation: First, historical accounting data (obtained with client authorization), product parameter combination data, and fund valuation results are collected. This data should cover different fund products, accounting dates, and accounting scenarios to ensure the model's generalization ability. The data preparation stage should also include data cleaning, such as removing outliers, filling in missing data, and format standardization.

[0034] Feature engineering is a crucial step in machine learning projects, involving the extraction of useful features from raw data. In this application, features may include various product parameters and their combinations, accounting dates, asset types, etc. Feature engineering also includes: Feature selection: Identifying which parameters have the greatest impact on the valuation results, using statistical methods (such as ANOVA, chi-square test) or model-based methods (such as importance scoring in random forests). Feature encoding: For non-numerical features, such as the categories of accounting rules, encoding transformation is performed; common methods are one-hot encoding or label encoding. Feature scaling: Ensuring all features are of the same magnitude to avoid some features having an excessively large impact on the model due to their numerical size.

[0035] Model building: Selecting a suitable machine learning model. Hybrid feature factor selection requires combining multiple types of machine learning methods, such as linear regression, logistic regression, decision trees, and neural networks. The specific choice depends on the characteristics of the data and the nature of the problem.

[0036] In one embodiment of this application, the accounting rules in the managed accounting system are decoupled from the accounting engine, including: using interface technology to separate the accounting rules from the core accounting engine of the managed accounting system (ensuring that the updating and maintenance of the rules will not affect the stability and performance of the entire system); encapsulating the execution logic of the accounting rules with an independent service layer (so that these rules can run independently in different environments without being constrained by other system components); and updating and applying new accounting rules in real time, at least according to the execution logic of the accounting rules, without restarting the system.

[0037] Separate the calculation rules from the core calculation engine using interface technology: Interface Design: Develop a clearly defined set of interfaces that describe how the core calculation engine interacts with the calculation rules. For example, interfaces might include functions such as retrieving parameters, performing specific calculation operations, and returning calculation results. Engine Refactoring: Modify the core calculation engine so that it no longer directly holds the calculation rules, but instead communicates with the independent calculation rule service by calling the aforementioned interfaces. In this way, the core calculation engine becomes a caller, while the calculation rules are moved to a separate service. Independent Calculation Rule Service: Create an independent calculation rule service that contains the logic of all calculation rules and implements the interfaces defined above. The service design should follow good software design principles, such as the Single Responsibility Principle, where each rule is responsible for a specific calculation task.

[0038] The execution logic of the above calculation rules is encapsulated in an independent service layer: Service Layer Design: An independent service layer is constructed. This layer is responsible for receiving requests from the core calculation engine and calling the corresponding calculation rule service to execute the specific calculation logic. Parameter Passing: When executing the calculation task, the service layer receives necessary parameters from the core calculation engine and passes them to the calculation rule service. This may include product information, calculation date, specific calculation parameters, etc. Result Feedback: After the calculation rule service completes the calculation logic, it returns the result to the service layer, which then packages the result and sends it back to the core calculation engine.

[0039] Real-time updates and application of new accounting rules: Dynamic loading mechanism: Design a mechanism that allows new accounting rules to be loaded and updated without restarting the entire system. This typically involves technologies such as hot-plugging, dynamic link libraries, or containerized service updates in a microservice architecture. Version control: Implement version control for each accounting rule to ensure that even during updates, it is possible to roll back to previous versions, preventing update risks from impacting system stability. Rule configuration center: Establish a rule configuration center that centrally manages all accounting rules, provides an interface for administrators to modify rules in real time, and synchronizes the changes to all service instances.

[0040] After decoupling, accounting rules can be developed, tested, and deployed independently of the core accounting engine, improving system flexibility. This means that changes to accounting rules will not affect the stability of the entire system and can be quickly adjusted according to business needs. The independent encapsulation of accounting rules makes them easier to understand and maintain. For complex financial accounting scenarios, this decoupling can significantly reduce dependencies between codes and lower maintenance costs. The ability to update and apply new accounting rules in real time allows the system to react quickly to market changes or regulatory updates, avoiding business interruptions caused by system restarts. Due to the independence of accounting rules and the service layer, it is easier to add new accounting rules, and even different asset types can have different service layers, increasing system scalability.

[0041] In one embodiment of this application, the above model is trained using historical accounting transaction data to obtain an optimized algorithm model, including: using a feature selection algorithm to filter and process the historical accounting transaction data to obtain a first accounting parameter, wherein the feature selection algorithm is one of the following: LASSO regression, random forest, and gradient boosting tree; using a feature dimensionality reduction method to reduce the dimensionality of the first accounting parameter to obtain a second accounting parameter; and using the second accounting parameter to train the above model to obtain the optimized algorithm model.

[0042] Data Preprocessing: First, the collected historical accounting transaction data is cleaned and formatted to ensure data quality and consistency, preparing it for subsequent algorithm applications. Feature Selection Algorithm Application: One of three algorithms (LASSO Regression, Random Forest, or Gradient Boosting Tree) is selected to perform feature selection on the preprocessed data to determine the primary accounting parameters. These three algorithms each have their own characteristics: LASSO Regression: This is a linear regression model that narrows down some parameters by imposing a penalty term, or even setting them to zero, thereby achieving feature selection. In this invention, LASSO Regression can be used to exclude accounting parameters with less impact on the accounting trial results, retaining key parameters. Random Forest: This is an ensemble learning algorithm based on decision trees, which improves the stability and accuracy of the model by constructing multiple decision trees and combining their predictions. Random Forest can assess the relative importance of each accounting parameter, helping us determine the features that have the greatest impact on the accounting trial results. Gradient Boosting Tree: This is another ensemble learning method that iteratively improves the model by adding decision trees, particularly suitable for handling large amounts of data and highly nonlinear relationships. Gradient Boosting Tree can also provide feature importance ranking for filtering key accounting parameters. Feature Reduction: After determining the first set of accounting parameters, the next step is to use feature reduction methods, such as Principal Component Analysis (PCA), t-SNE, or Linear Discriminant Analysis (LDA), to reduce the dimensionality of these parameters and obtain the second set of accounting parameters. Dimensionality reduction helps remove redundant information between parameters, reduces computational burden, and improves model training efficiency. Model Training: Finally, the model is trained using the second set of accounting parameters after feature reduction. By continuously adjusting the model parameters, it can more accurately predict the results of accounting reconciliation, thus obtaining an optimized algorithm model. This model will be used to quickly identify and analyze reconciliation discrepancies, improving reconciliation efficiency.

[0043] By employing feature selection and dimensionality reduction, the model can operate with fewer data dimensions, significantly improving the speed and efficiency of analyzing reconciliation discrepancies. Feature dimensionality reduction reduces the computational resources required for model training, lowering operating costs. Feature selection ensures the model focuses only on the most relevant parameters, simplifying the model structure and enhancing its generalization ability, maintaining high accuracy across different fund products and accounting scenarios. It also reduces the risk of human error: automated processes minimize manual intervention, avoiding potential human errors in manual analysis and improving the reliability of reconciliation analysis.

[0044] In one embodiment of this application, the above-mentioned optimization algorithm model is used to process the above-mentioned trial calculation results to obtain a product accounting parameter combination, including: processing the above-mentioned trial calculation results using the above-mentioned optimization algorithm model to obtain multiple initial product accounting parameter combinations and corresponding confidence levels; determining the above-mentioned initial product accounting parameter combination corresponding to the maximum value of the above-mentioned confidence level as the above-mentioned product accounting parameter combination.

[0045] Confidence level reflects the model's confidence in the correctness of the predicted parameter combination. Selecting the parameter combination with the highest confidence level means that this combination is most likely to accurately reflect the fund manager's actual accounting rules, thereby improving the accuracy of reconciliation discrepancy analysis. The optimized algorithm model, through learning and analyzing historical data, can quickly filter out the most likely correct combination from numerous possible product accounting parameter combinations. This method avoids the tedious process of manually comparing parameter combinations one by one, significantly improving the efficiency of reconciliation discrepancy analysis. The parameter combinations and confidence levels derived through the algorithm model reduce the influence of human factors and lower the risk of misjudgment due to individual analytical abilities and biases. This makes reconciliation discrepancy analysis more objective and reliable.

[0046] In one embodiment of this application, after comparing and analyzing the differences between the above-mentioned product accounting parameter combination and the existing product parameter combination of the above-mentioned managed accounting system to determine the cause of the accounting difference, the above method includes: comparing the calculated product accounting parameter combination with the existing system parameter configuration to generate a difference report; and displaying the accounting parameters and accounting rules related to the reconciliation difference in a visual manner based on the above-mentioned difference report.

[0047] By visually displaying reconciliation discrepancies, even managers without technical backgrounds can intuitively understand the differences, enhancing decision-makers' comprehension, reducing communication costs, and improving team collaboration efficiency. The discrepancy report directly points out deviations from existing system parameter configurations, avoiding blind searches and quickly identifying problematic parameters and rules, thus accelerating problem resolution. Automated comparison and report generation reduce the possibility of manual operation, avoiding false or missed reports due to human factors, and improving the accuracy and reliability of the reports. Detailed discrepancy reports provide historical records, facilitating subsequent tracking and auditing, helping to identify potential system improvement points, and also providing a basis for compliance checks.

[0048] In one embodiment of this application, before establishing a model using a hybrid feature-based factor screening algorithm in machine learning and training the model with historical accounting transaction data to obtain an optimized algorithm model, the method further includes: using statistical or machine learning techniques to identify abnormal accounting parameters and rules, and generating corresponding early warning information.

[0049] Statistical and machine learning models can analyze massive amounts of transaction data in real time, quickly identifying accounting parameters and rules that deviate from normal ranges. This provides immediate alerts, allowing managers to take immediate action to reduce potential risks and losses. Automated monitoring replaces traditional manual review, significantly reducing the time and resources required for human audits; intelligent algorithms can process massive amounts of data in a short time, greatly improving the efficiency of problem detection. Once fully trained, machine learning models can predict future anomalies based on historical patterns, preventing problems before they occur. Models can learn from past data, identifying early signs of abnormal behavior, thereby preventing major errors. Anomaly detection systems can increase the transparency of reconciliation and accounting processes, ensuring all operations are under regulatory oversight. This helps in complying with the stringent regulations of the financial industry, promptly correcting non-compliant accounting parameters, and reducing compliance risks.

[0050] In one embodiment of this application, after establishing a model using a hybrid feature factor screening algorithm in machine learning, the method further includes: upon receiving a parameter correction instruction, correcting the corresponding parameters in the model according to the parameter correction instruction.

[0051] Specifically, this dynamic adjustment capability enables models to quickly respond to new information or environmental changes, thereby continuously optimizing their predictive and analytical abilities. For example, in the reconciliation discrepancy analysis of a fund custody accounting system, if market conditions change or new accounting rules are discovered, the model can update its internal logic based on the latest parameter correction instructions, improving the accuracy of future reconciliation discrepancy analysis. By continuously correcting model parameters, key factors affecting reconciliation discrepancies can be captured more accurately. In the field of machine learning, this typically means that the model can gradually reduce errors during training, improving the accuracy of prediction or classification. Automatic parameter correction reduces the need for human intervention and avoids errors that may be introduced by human operation. For complex scenarios such as fund custody, manual parameter adjustment is time-consuming and labor-intensive, and it is difficult to cover all possible influencing factors, while automated parameter correction can efficiently handle large amounts of data and complex rules.

[0052] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0053] This application also provides a reconciliation discrepancy analysis device. It should be noted that this device can be used to execute the reconciliation discrepancy analysis method provided in this application. This device is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0054] The following describes the reconciliation discrepancy analysis device provided in the embodiments of this application.

[0055] Figure 2 This is a schematic diagram of a reconciliation discrepancy analysis device according to an embodiment of this application. Figure 2 As shown, the device includes:

[0056] The first processing unit 21 is used to decouple the accounting rules in the managed accounting system from the accounting engine and obtain pre-accounting information, including product information, accounting date, and product accounting parameters. The second processing unit 22 is used to call the accounting rules to perform trial calculations on the pre-accounting information and obtain trial calculation results. The third processing unit 23 is used to build a model using a hybrid feature factor screening algorithm in machine learning and train the model using historical accounting transaction data to obtain an optimized algorithm model. The fourth processing unit 24 is used to process the trial calculation results using the optimized algorithm model to obtain a combination of product accounting parameters. The fifth processing unit 25 is used to compare the combination of product accounting parameters with the existing combination of product parameters in the managed accounting system and analyze the differences to determine the reasons for the accounting differences.

[0057] In one embodiment of this application, the first processing unit includes: a first processing module for separating the accounting rules from the core accounting engine of the aforementioned managed accounting system using interface technology; a second processing module for encapsulating the execution logic of the aforementioned accounting rules using an independent service layer; and a third processing module for updating and applying new accounting rules in real time, at least according to the execution logic of the aforementioned accounting rules, without restarting the system.

[0058] In one embodiment of this application, the third processing unit includes: a fourth processing module for filtering the historical accounting transaction data using a feature selection algorithm to obtain a first accounting parameter, wherein the feature selection algorithm is one of the following: LASSO regression, random forest, and gradient boosting tree; a fifth processing module for reducing the dimensionality of the first accounting parameter using a feature dimensionality reduction method to obtain a second accounting parameter; and a sixth processing module for training the model using the second accounting parameter to obtain the optimized algorithm model.

[0059] In one embodiment of this application, the fourth processing unit includes: a seventh processing module for processing the trial calculation results using the above-mentioned optimization algorithm model to obtain multiple initial product accounting parameter combinations and corresponding confidence levels; and an eighth processing module for determining the initial product accounting parameter combination corresponding to the maximum value of the above-mentioned confidence level as the product accounting parameter combination.

[0060] In one embodiment of this application, the above-mentioned apparatus further includes: a sixth processing unit configured to compare and analyze the differences between the above-mentioned product accounting parameter combination and the existing product parameter combination of the above-mentioned managed accounting system, and determine the cause of the accounting difference, and then compare the calculated product accounting parameter combination with the existing system parameter configuration to generate a difference report; and a seventh processing unit configured to display the accounting parameters and accounting rules related to the reconciliation difference in a visual manner based on the above-mentioned difference report.

[0061] In one embodiment of this application, the above-mentioned apparatus further includes: an eighth processing unit for identifying abnormal accounting parameters and rules using statistical or machine learning techniques and generating corresponding early warning information before establishing a model using a hybrid feature factor screening algorithm in machine learning and training the model with historical accounting transaction data to obtain an optimized algorithm model.

[0062] In one embodiment of this application, the above-mentioned apparatus further includes: a ninth processing unit configured to, after establishing a model using a hybrid feature factor screening algorithm in machine learning, correct the corresponding parameters in the model according to the parameter correction instruction received.

[0063] The aforementioned reconciliation discrepancy analysis device includes a processor and a memory. The first, second, third, fourth, and fifth processing units are all stored as program units in the memory, and the processor executes these program units to achieve their respective functions. All of the above modules are located in the same processor; alternatively, the modules may be located in different processors in any combination.

[0064] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured; adjusting kernel parameters can resolve issues that might arise from incorrect parameter settings in the accounting system, even if related technologies cannot analyze reconciliation discrepancies.

[0065] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0066] This invention provides a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform the reconciliation discrepancy analysis method.

[0067] This invention provides a processor for running a program, wherein the program executes the reconciliation discrepancy analysis method.

[0068] This invention provides a device including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs at least the following steps: decoupling the accounting rules in the managed accounting system from the accounting engine, and obtaining pre-accounting information, including product information, accounting date, and product accounting parameters; calling the accounting rules to perform trial calculations on the pre-accounting information to obtain trial calculation results; establishing a model using a hybrid feature factor screening algorithm in machine learning, and training the model using historical accounting transaction data to obtain an optimized algorithm model; processing the trial calculation results using the optimized algorithm model to obtain a combination of product accounting parameters; comparing the combination of product accounting parameters with existing product parameter combinations in the managed accounting system and analyzing the differences to determine the cause of the accounting discrepancies. The device in this document can be a server, PC, PAD, mobile phone, etc.

[0069] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program having at least the following method steps: decoupling the accounting rules in the managed accounting system from the accounting engine and obtaining pre-accounting information, the pre-accounting information including product information, accounting date, and product accounting parameters; calling the accounting rules to perform trial calculations on the pre-accounting information to obtain trial calculation results; establishing a model using a hybrid feature factor screening algorithm in machine learning, and training the model using historical accounting transaction data to obtain an optimized algorithm model; processing the trial calculation results using the optimized algorithm model to obtain a combination of product accounting parameters; comparing the combination of product accounting parameters with the existing product parameter combinations in the managed accounting system and analyzing the differences to determine the cause of the accounting differences.

[0070] This application also provides an electronic device, including: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include methods for performing any of the above-described methods.

[0071] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0072] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0073] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0074] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0075] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0076] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0077] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0078] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0079] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0080] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0081] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for analyzing reconciliation discrepancies, characterized in that, include: Decouple the accounting rules in the managed accounting system from the accounting engine and obtain pre-accounting information, which includes product information, accounting date and product accounting parameters; The pre-calculation information is processed by invoking the calculation rules to obtain the calculation results; A model was built using a hybrid feature-based factor screening algorithm in machine learning, and the model was trained using historical accounting transaction data to obtain an optimized algorithm model. The optimization algorithm model is used to process the trial calculation results to obtain the product accounting parameter combination; The product accounting parameter combination is compared with the existing product parameter combination in the managed accounting system, and the differences are analyzed to determine the reasons for the accounting differences.

2. The method according to claim 1, characterized in that, Decoupling the accounting rules from the accounting engine in the managed accounting system includes: Interface technology is used to separate the accounting rules from the core accounting engine of the managed accounting system. The execution logic of the accounting rules is encapsulated in an independent service layer; Without restarting the system, new accounting rules should be updated and applied in real time, at least according to the execution logic of the accounting rules.

3. The method according to claim 1, characterized in that, The model is trained using historical transaction data to obtain an optimized algorithm model, including: The historical accounting transaction data is filtered and processed using a feature selection algorithm to obtain the first accounting parameters. The feature selection algorithm is one of the following: LASSO regression, random forest, and gradient boosting tree. The first accounting parameter is reduced in dimensionality using a feature dimensionality reduction method to obtain the second accounting parameter; The model is trained using the second calculation parameters to obtain the optimized algorithm model.

4. The method according to claim 1, characterized in that, The optimization algorithm model is used to process the trial calculation results to obtain a combination of product accounting parameters, including: The optimization algorithm model is used to process the trial calculation results to obtain multiple initial product accounting parameter combinations and corresponding confidence levels; The initial product accounting parameter combination corresponding to the maximum value of the confidence level is determined as the product accounting parameter combination.

5. The method according to claim 1, characterized in that, After comparing and analyzing the differences between the product accounting parameter combination and the existing product parameter combination in the managed accounting system to determine the cause of the accounting difference, the method includes: Compare the calculated product accounting parameter combination with the existing system parameter configuration to generate a difference report; The discrepancy report visually displays the accounting parameters and rules related to reconciliation discrepancies.

6. The method according to claim 1, characterized in that, Before establishing a model using a hybrid eigenvalue factor selection algorithm in machine learning and training the model with historical transaction data to obtain an optimized algorithm model, the method further includes: Utilize statistical or machine learning techniques to identify abnormal accounting parameters and rules, and generate corresponding early warning information.

7. The method according to claim 1, characterized in that, After building a model using a hybrid eigenvalue factor selection algorithm in machine learning, the method further includes: Upon receiving a parameter correction instruction, the corresponding parameters in the model are corrected according to the parameter correction instruction.

8. A reconciliation discrepancy analysis device, characterized in that, include: The first processing unit is used to decouple the accounting rules in the managed accounting system from the accounting engine and obtain pre-calculation information, which includes product information, accounting date and product accounting parameters. The second processing unit is used to call the accounting rules to perform trial calculations on the pre-calculation information and obtain the trial calculation results. The third processing unit is used to build a model using a hybrid feature factor screening algorithm in machine learning, and to train the model using historical accounting transaction data to obtain an optimized algorithm model. The fourth processing unit is used to process the trial calculation results using the optimization algorithm model to obtain the product accounting parameter combination; The fifth processing unit is used to compare the product accounting parameter combination with the existing product parameter combination in the managed accounting system and analyze the differences to determine the reasons for the accounting differences.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs comprising methods for performing any one of claims 1 to 7.

Images