Budget auditing method and system based on robust support vector machine

Abstract

The invention relates to the technical field of budget auditing, and discloses a budget auditing method and system based on a robust support vector machine, and the method comprises the steps: obtaining historical budget approval data, building an approval model formula according to the historical budget approval data, and building an approval model according to each approval model formula; obtaining budget information to be approved, extracting information parameters in the budget information to be approved, determining whether the budget information to be approved is passed or not based on the relationship between the information parameters and the approval model set, and if the budget information to be approved is not passed, reporting the budget information to be approved to an authority person. If the judgment result is successful, the budget information to be approved is passed, and if the judgment result is successful, the budget information to be approved is passed; and if the judgment result is that the budget information does not pass, refusing to pass the budget information to be approved. The to-be-approved budget is automatically judged by establishing the approval model, so that the approval efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of budget auditing, and in particular to a method and system for budget auditing based on a robust support vector machine. Background Art

[0002] Comprehensive budgeting is an applied method for the coordinated management of business indicators, expenses, and asset status. Its purpose is to improve client control over projects and adjust resource and personnel costs based on the progress of enterprise projects to provide more reasonable guidance for enterprises.

[0003] With the advancement of society and the expansion of businesses, management is becoming increasingly complex and cumbersome. Comprehensive budgeting, based on a company's development strategy, helps companies define specific operational objectives by detailing future operating activities and financial results. These objectives can be quantified into key indicators such as sales, costs, and profits, providing employees with a clear direction and goals to strive for. However, while this improved management convenience also comes with a corresponding increase in the number of reviewers and the number of management decision-making levels. With the increasing number of corporate layers, daily compilation and review tasks consume significant time. Furthermore, compilation often relies solely on personal experience and lacks the use of historical data. This is detrimental to the company's future direction. Furthermore, the results of this approach lack scientific validity and can lead to errors.

[0004] Therefore, there is an urgent need to invent a scientific and rapid budget audit technology to solve the current technology problems. Due to the complexity of enterprise management and the increase in decision-making levels, budget preparation and auditing work takes a lot of time, relies on personal experience, and lacks effective use of historical data, resulting in the lack of scientificity and accuracy of audit results. Summary of the Invention

[0005] In view of this, the present invention proposes a method and system for budget review based on a robust support vector machine, aiming to solve the problem in current technology that due to the complexity of enterprise management and the increase in decision-making levels, budget preparation and review work takes a lot of time, relies on personal experience, lacks effective use of historical data, and leads to the lack of scientificity and accuracy of audit results.

[0006] The present invention proposes a method for budget review based on a robust support vector machine, comprising:

[0007] Acquire each historical budget approval data, establish an approval model based on the historical budget approval data, and establish an approval model based on each approval model;

[0008] Obtaining budget information to be approved, extracting information parameters from the budget information to be approved, and determining whether to approve the budget information to be approved based on a relationship between the information parameters and the approval model set, wherein:

[0009] If it is determined that the budget information to be approved is not passed, the budget information to be approved is reported to the authority, and the authority's judgment result on the budget information to be approved is obtained, wherein:

[0010] If the judgment result is passed, the budget information to be approved is approved;

[0011] If the judgment result is failure, the budget information to be approved is rejected.

[0012] Furthermore, obtaining each historical budget approval data, establishing an approval model according to the historical budget approval data, and establishing an approval model according to each approval model includes:

[0013] Obtaining consistent duplicate data from each of the historical budget approval data, and removing the duplicate data;

[0014] The approval model formulas are established based on the historical budget approval data after removing duplicate data, and the approval model is established based on the approval model formulas.

[0015] Furthermore, when establishing each of the approval models based on each of the historical budget approval data after removing duplicate data, the method includes:

[0016] Obtain information attributes of each component in historical preset approval data;

[0017] Acquire the historical preset approval data with the same information attribute, establish a linear axis of the information attribute based on the historical preset approval data with the same information attribute, and establish the approval model formula based on the linear axis;

[0018] Obtaining parameters of each component of any of the historical preset approval data in the historical preset approval data with the same information attributes;

[0019] Substituting each component parameter into the approval model formula, and obtaining a preset approval result output by the approval model formula;

[0020] According to the relationship between the preset approval result and the approval result of the historical preset approval data, it is determined whether to modify the approval model.

[0021] Furthermore, when determining whether to modify the approval model according to the relationship between the preset approval result and the approval result of the historical preset approval data, the method includes:

[0022] Comparing the preset approval result with the approval result of the historical preset approval data, and determining whether to proofread the approval model according to the comparison result, wherein:

[0023] If the preset approval result is consistent with the approval result of the historical preset approval data, it is determined that there is no need to modify the approval model;

[0024] If the preset approval result is inconsistent with the approval result of the historical preset approval data, the approval model is iteratively corrected.

[0025] Furthermore, after establishing the approval model, the following steps are also included:

[0026] Determining a regularization term and a loss function term based on the historical budget approval data after removing duplicate data;

[0027] The approval model is trained according to the regularization term and the loss function term, and the trained approval model is obtained.

[0028] Compared to existing technologies, the present invention offers the following advantages: by effectively utilizing historical budget approval data, a robust support vector machine model is constructed that automatically extracts implicit patterns and rules from past approval data. This approach shifts the budget review process from relying solely on human judgment to a data-driven, intelligent review process. This significantly improves review efficiency, enabling rapid and accurate review of budget information, particularly in complex enterprise management environments with large data volumes, while avoiding the time-consuming and subjective errors associated with manual review. Furthermore, when budget information fails automated review, it is automatically escalated to a reviewer with higher authority. This mechanism ensures that human judgment can still be incorporated into key budget decisions, avoiding the potential misjudgments or omissions that can result from relying solely on machine decision-making. The intervention of authorized personnel makes review results more reliable, combining the efficiency of automated review with the discretion of manual review, achieving an effective balance between the two. Furthermore, this method provides further flexibility after authorized personnel intervene in the review process. If the authorized personnel approve the budget, the approval result is automatically updated, ensuring smooth approval. Otherwise, the budget that does not meet the requirements is rejected. Such a process not only improves the scientificity and rationality of budget review, but also can be dynamically adjusted according to actual business needs, ensuring the flexibility and accuracy of decision-making, and avoiding unreasonable rejections or erroneous approvals caused by the rigidity of automated review.

[0029] On the other hand, the present application also provides a system for budget review based on a robust support vector machine, comprising:

[0030] Data acquisition module, used to obtain historical budget approval data;

[0031] A model generation module, electrically connected to the data acquisition module, the model generation module is used to establish an approval model formula based on the historical budget approval data, and establish an approval model based on each approval model formula;

[0032] an approval module electrically connected to the model generation module, the approval module being configured to obtain the budget information to be approved, extract information parameters from the budget information to be approved, and determine whether to approve the budget information to be approved based on a relationship between the information parameters and the approval model set;

[0033] An inquiry module is electrically connected to the approval module, and is used to obtain the budget information that has not passed the approval, report the budget information to be approved to the authority, and obtain the authority's judgment result on the budget information to be approved, wherein:

[0034] If the judgment result is passed, the inquiry module passes the budget information to be approved;

[0035] If the judgment result is failure, the inquiry module rejects the budget information to be approved.

[0036] Furthermore, the model generation module is used to establish an approval model formula based on the historical budget approval data, and establish an approval model based on each approval model formula, including:

[0037] The model generation module is further configured to obtain consistent duplicate data from each of the historical budget approval data and remove the duplicate data;

[0038] The model generation module is further configured to establish each of the approval model formulas based on each of the historical budget approval data after removing duplicate data, and to establish an approval model based on each of the approval model formulas.

[0039] Furthermore, the model generation module is further configured to establish each of the approval models based on each of the historical budget approval data after removing duplicate data, including:

[0040] The model generation module is also used to obtain information attributes of each component in the historical preset approval data;

[0041] The model generation module is further configured to obtain the historical preset approval data with the same information attribute, establish a linear axis of the information attribute based on the historical preset approval data with the same information attribute, and establish the approval model formula based on the linear axis;

[0042] The model generation module is further configured to obtain parameters of each component of any of the historical preset approval data in the historical preset approval data having the same information attribute;

[0043] The model generation module is further used to substitute the parameters of each component into the approval model formula and obtain the preset approval result output by the approval model formula;

[0044] The model generation module is further configured to determine whether to modify the approval model formula according to a relationship between the preset approval result and the approval result of the historical preset approval data.

[0045] Furthermore, the model generation module is further configured to determine whether to modify the approval model according to the relationship between the preset approval result and the approval result of the historical preset approval data, including:

[0046] The model generation module is further configured to compare the preset approval result with the approval result of the historical preset approval data, and determine whether to proofread the approval model according to the comparison result, wherein:

[0047] If the preset approval result is consistent with the approval result of the historical preset approval data, the model generation module determines that there is no need to modify the approval model;

[0048] If the preset approval result is inconsistent with the approval result of the historical preset approval data, the model generation module iteratively amends the approval model.

[0049] Furthermore, after establishing the approval model, the model generation module further includes:

[0050] The model generation module is further configured to determine a regularization term and a loss function term based on each of the historical budget approval data after removing duplicate data;

[0051] The model generation module is also used to train the approval model according to the regularization term and the loss function term, and obtain the trained approval model.

[0052] It can be understood that the method and system for budget review based on a robust support vector machine in the above embodiments of the present invention have the same beneficial effects and will not be described in detail. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0054] Figure 1 A flowchart of a method for budget review based on a robust support vector machine provided in an embodiment of the present invention;

[0055] Figure 2 This is a functional block diagram of a system for budget review based on a robust support vector machine provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, unless there is a conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0057] like Figure 1 In some embodiments of the present application, a method for budget review based on a robust support vector machine is provided, comprising:

[0058] Step S100: Acquire historical budget approval data, establish an approval model based on the historical budget approval data, and establish an approval model based on each approval model.

[0059] Specifically, obtaining each historical budget approval data, establishing an approval model based on the historical budget approval data, and establishing an approval model based on each approval model includes: obtaining consistent duplicate data from each historical budget approval data and removing the duplicate data; establishing each approval model based on each historical budget approval data after removing the duplicate data; and establishing an approval model based on each approval model.

[0060] Specifically, establishing each approval model based on each historical budget approval data after removing duplicate data includes: obtaining the information attributes of each component in the historical preset approval data. Obtaining each historical preset approval data with the same information attributes, establishing a linear axis of the information attributes based on each historical preset approval data with the same information attributes, and establishing an approval model based on the linear axis. Obtaining the parameters of each component of any historical preset approval data with the same information attributes. Substituting each component parameter into the approval model, and obtaining the preset approval result output by the approval model. Determining whether to modify the approval model based on the relationship between the preset approval result and the approval result of the historical preset approval data.

[0061] Specifically, determining whether to revise the approval model based on the relationship between the preset approval result and the approval results of the historical preset approval data includes: comparing the preset approval result with the approval results of the historical preset approval data, and determining whether to proofread the approval model based on the comparison result, wherein: if the preset approval result and the approval results of the historical preset approval data are consistent, then determining that the approval model does not need to be revised. If the preset approval result and the approval results of the historical preset approval data are inconsistent, then iteratively revising the approval model.

[0062] As can be seen, model quality is ensured by obtaining and removing duplicate historical budget approval data. During data processing, duplicate data can introduce noise and bias, leading to overfitting of the model or inaccurate predictions for new data. Removing this duplicate data ensures the uniqueness and diversity of the dataset used, allowing the model training to better reflect the true trends and patterns of the budget data. This data cleaning process improves the model's generalization and reduces interference from invalid data. Next, based on the deduplicated historical budget approval data, the system extracts specific information attributes for each component. Each attribute represents a different dimension of the budget approval process, such as department, project category, and budget amount. By categorizing and analyzing data with the same information attribute, the system creates linear axes based on each attribute, generating a preliminary approval model. This step simplifies the complexity of the multi-dimensional data by linearizing it, facilitating the construction of a mathematical model suitable for budget approval. Based on this, the model is further optimized by obtaining specific parameters from historical budget approval data with the same information attributes and substituting them into the preliminary approval model to generate a pre-determined approval result. This pre-determined result is compared with the actual historical approval result to assess the model's accuracy. If the preset results align with historical results, the model is valid and requires no revision. However, if the results are inconsistent, the model requires further optimization. Subsequently, when the preset approval results differ from historical results, the model is iteratively revised. This feedback mechanism, similar to supervised learning in machine learning, automatically adjusts model parameters by continuously comparing model predictions with actual values, ensuring that the final approval model output better reflects actual conditions. This iterative revision process leverages feedback from historical data, enhancing the model's robustness and adaptability, ensuring more accurate and efficient performance in future budget approvals. Finally, budget approval data constantly changes with business development and management hierarchies, so a fixed model may not be able to adapt to complex budget review needs in the long term. This iterative revision mechanism allows the model to automatically adjust to new data inputs, maintaining its validity and accuracy. This provides greater flexibility and reliability for enterprise budget review.

[0063] As you can see, by obtaining historical budget approval data and removing duplicate data, the dataset used is unique and representative. This process reduces data redundancy and improves data purity, thus avoiding bias caused by duplicate data during model training. Duplicate data often causes models to overfit specific patterns, but removing this data helps the model better capture the true patterns in the data and improve its adaptability to new data. This data cleansing process ensures that the model is trained on more authentic and diverse historical data, making it more robust. Secondly, after removing duplicate data, the information attributes in the historical budget approval data are analyzed. Extracting information attributes is a key step in generating an approval model. By classifying data with the same information attributes, linear axes can be created based on these attributes, and the approval model formula can be generated based on these attributes. This method, through the classification and linearization of information attributes, effectively simplifies data complexity and makes the model building process more efficient. The linear axes of information attributes help the system accurately model each dimension of the budget approval data, thereby forming budget approval models suitable for different scenarios. This step simplifies the multidimensional data into an operational linear model, facilitating rapid decision-making during the budget review process. Next, by substituting the parameters of each component into the preliminary approval model, a pre-defined approval outcome is generated. This process utilizes specific parameters with the same information attributes in historical data to simulate possible future budget approval outcomes. These pre-defined outcomes can be used to predict model performance and compared with actual historical approval results to assess model accuracy. If the pre-defined outcomes are consistent with historical results, the model is accurate and effective and requires no further revision; if they are inconsistent, the model requires adjustment. This comparison of pre-defined and actual results provides an efficient feedback mechanism, enabling timely identification and correction of model deviations. More importantly, when a discrepancy is detected between the pre-defined and actual approval outcomes, an iterative correction mechanism is initiated. This dynamic adjustment allows the approval model to be continuously optimized based on the latest feedback data, ensuring that it adapts to the ever-changing budget review environment. This iterative correction mechanism not only enhances the model's robustness but also improves its adaptability to new data, avoiding the potential misjudgments that static models can encounter when faced with novel situations. The model's dynamic correction capabilities enable the system to adjust itself at any time, maintaining efficient and accurate budget review capabilities. Furthermore, this feedback-based iterative correction mechanism effectively improves the long-term stability and scalability of the approval model. As a company's business expands, the budget approval process will become more complex and diverse. Fixed models may not be able to continuously adapt to this complex environment. However, through iterative correction, accurate analysis and judgment of budget data can be maintained under constantly changing conditions.Regardless of whether the company expands in size or the types of business increase, this system can adjust model parameters in real time through continuous data feedback to ensure the efficiency and stability of the budget review process.

[0064] Specifically, after establishing the approval model, the process also includes: determining a regularization term and a loss function term based on the historical budget approval data after removing duplicate data, training the approval model based on the regularization term and the loss function term, and obtaining the trained approval model.

[0065] Specifically, by obtaining the attributes of the components of each dimension, the approval model is trained based on the regularization term and the loss function term, and the model formula I is established:

[0066]

[0067] in, is the regularization term, is the loss function,

[0068] Specifically, the training process involves obtaining the attribute dimensions of each component in the historical budget approval data and the specific values ​​of the attribute dimensions (x1, x2, …, xn, y, -1). Here, x1, x2, …, xn represent the attribute dimensions of each component, such as the organizational dimension, indicator dimension, period dimension, budget dimension, and approval status, and y represents the specific value of the dimension composed of these dimensions. The obtained attribute dimensions of each component and the specific values ​​of the attribute dimensions (x1, x2, …, xn, y, -1) are trained using Formula II to obtain specific parameters w (where w represents the execution data of previous years) and b. Formula II is as follows:

[0069]

[0070] Among them, u>0 is a parameter that can be trained to obtain a specific value through actual execution data of previous years, β=[1-e -η ] -1 , σ is a kernel parameter.

[0071] It can be seen that by introducing regularization terms and loss function terms to train the approval model, the generalization ability and robustness of the model can be effectively improved, and the overfitting problem can be reduced. At the same time, through in-depth analysis and modeling of the attributes of each dimension and their specific values, the system can more accurately capture the complex relationship of the budget approval data and generate optimized model parameters. The use of these optimized parameters (such as w, b and kernel parameters) further improves the accuracy and predictability of the model, making the budget approval process more scientific and reasonable, thereby significantly improving the accuracy and efficiency of budget review.

[0072] Step S200: Obtain budget information to be approved, extract information parameters from the budget information to be approved, and determine whether to approve the budget information to be approved based on the relationship between the information parameters and the approval model set.

[0073] Step S300: If it is determined that the budget information to be approved is not passed, the budget information to be approved is reported to the authority, and the authority's judgment result on the budget information to be approved is obtained. If the judgment result is passed, the budget information to be approved is approved. If the judgment result is not passed, the budget information to be approved is rejected.

[0074] In the above embodiment, by effectively utilizing historical budget approval data, a robust support vector machine model is constructed that can automatically extract implicit patterns and regularities from past approval data. This approach shifts the budget review process from relying solely on human judgment to a data-driven, intelligent review process. This significantly improves review efficiency, particularly in complex enterprise management environments with large data volumes. Budget information can be quickly and accurately reviewed, avoiding the time-consuming and subjective errors associated with manual review. Furthermore, when budget information fails automated review, it is automatically escalated to a reviewer with higher authority. This mechanism ensures that human judgment can still be involved in key budget decisions, avoiding the potential misjudgments or omissions that could result from relying solely on machine decision-making. The intervention of authorized personnel makes the review results more reliable, combining the efficiency of automated review with the discretion of manual review, achieving an effective balance between the two. Furthermore, this approach provides further flexibility after the authorized personnel intervene in the review. If the authorized personnel approve the budget, the approval result is automatically updated, ensuring smooth approval. Otherwise, the budget that does not meet the requirements is rejected. Such a process not only improves the scientificity and rationality of budget review, but also can be dynamically adjusted according to actual business needs, ensuring the flexibility and accuracy of decision-making, and avoiding unreasonable rejections or erroneous approvals caused by the rigidity of automated review.

[0075] In another preferred embodiment based on the above embodiment, Figure 2 This embodiment provides a system for budget review based on a robust support vector machine, including: a data acquisition module, a model generation module, an approval module, and an inquiry module.

[0076] Specifically, the data acquisition module is used to obtain various historical budget approval data. The model generation module is electrically connected to the data acquisition module, and the model generation module is used to establish an approval model formula based on the historical budget approval data, and to establish an approval model based on each approval model formula. The approval module is electrically connected to the model generation module, and the approval module is used to obtain the budget information to be approved, and extract the information parameters in the budget information to be approved, and determine whether the budget information to be approved is passed based on the relationship between the information parameters and the approval model formula set. The inquiry module is electrically connected to the approval module, and the inquiry module is used to obtain the budget information to be approved that has not passed, and report the budget information to be approved to the authority, and obtain the authority's judgment result on the budget information to be approved, wherein: if the judgment result is passed, the inquiry module passes the budget information to be approved. If the judgment result is failed, the inquiry module refuses to pass the budget information to be approved.

[0077] Specifically, the model generation module is used to establish an approval model formula based on historical budget approval data, and when establishing an approval model based on each approval model formula, the model generation module is further used to obtain consistent duplicate data from each historical budget approval data and eliminate the duplicate data. The model generation module is further used to establish each approval model formula based on each historical budget approval data after eliminating the duplicate data, and establish an approval model based on each approval model formula.

[0078] Specifically, when the model generation module is used to establish each approval model formula based on each historical budget approval data after removing duplicate data, it includes: the model generation module is also used to obtain the information attributes of each component in the historical preset approval data. The model generation module is also used to obtain each historical preset approval data with the same information attributes, and establish a linear axis of the information attributes based on each historical preset approval data with the same information attributes, and establish an approval model formula based on the linear axis. The model generation module is also used to obtain the parameters of each component of any historical preset approval data in each historical preset approval data with the same information attributes. The model generation module is also used to substitute the parameters of each component into the approval model formula, and obtain the preset approval result output by the approval model formula. The model generation module is also used to determine whether to modify the approval model formula based on the relationship between the preset approval result and the approval result of the historical preset approval data.

[0079] Specifically, the model generation module is further configured to determine whether to revise the approval model based on the relationship between the preset approval result and the approval results of the historical preset approval data. The model generation module is further configured to compare the preset approval result with the approval results of the historical preset approval data, and determine whether to proofread the approval model based on the comparison result. If the preset approval result is consistent with the approval results of the historical preset approval data, the model generation module determines that the approval model does not need to be revised. If the preset approval result is inconsistent with the approval results of the historical preset approval data, the model generation module iteratively revises the approval model.

[0080] Specifically, after establishing the approval model, the model generation module further includes: the model generation module is further configured to determine a regularization term and a loss function term based on each historical budget approval data after removing duplicate data. The model generation module is further configured to train the approval model based on the regularization term and the loss function term, and obtain the trained approval model.

[0081] It can be understood that the method and system for budget review based on a robust support vector machine in the above embodiments of the present invention have the same beneficial effects and will not be described in detail.

[0082] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or a combination of software and hardware embodiments. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0083] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0084] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0085] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for budget review based on robust support vector machine, characterized in that: include: Acquire each historical budget approval data, establish an approval model based on the historical budget approval data, and establish an approval model based on each approval model; Obtaining budget information to be approved, extracting information parameters from the budget information to be approved, and determining whether to approve the budget information to be approved based on a relationship between the information parameters and the approval model set, wherein: If it is determined that the budget information to be approved is not passed, the budget information to be approved is reported to the authority, and the authority's judgment result on the budget information to be approved is obtained, wherein: If the judgment result is passed, the budget information to be approved is approved; If the judgment result is failure, the budget information to be approved is rejected.

2. The method for budget review based on robust support vector machine according to claim 1, characterized in that: Acquiring each historical budget approval data, establishing an approval model formula based on the historical budget approval data, and establishing an approval model based on each approval model formula includes: Obtaining consistent duplicate data from each of the historical budget approval data, and removing the duplicate data; The approval model formulas are established based on the historical budget approval data after removing duplicate data, and the approval model is established based on the approval model formulas.

3. The method for budget review based on robust support vector machine according to claim 2, characterized in that: When establishing each of the approval models based on each of the historical budget approval data after removing duplicate data, the method includes: Obtain information attributes of each component in historical preset approval data; Acquire the historical preset approval data with the same information attribute, establish a linear axis of the information attribute based on the historical preset approval data with the same information attribute, and establish the approval model formula based on the linear axis; Obtaining parameters of each component of any of the historical preset approval data in the historical preset approval data with the same information attributes; Substituting each component parameter into the approval model formula, and obtaining a preset approval result output by the approval model formula; According to the relationship between the preset approval result and the approval result of the historical preset approval data, it is determined whether to modify the approval model.

4. The method for budget review based on robust support vector machine according to claim 3, characterized in that: Determining whether to modify the approval model according to the relationship between the preset approval result and the approval result of the historical preset approval data includes: Comparing the preset approval result with the approval result of the historical preset approval data, and determining whether to proofread the approval model according to the comparison result, wherein: If the preset approval result is consistent with the approval result of the historical preset approval data, it is determined that there is no need to modify the approval model; If the preset approval result is inconsistent with the approval result of the historical preset approval data, the approval model is iteratively corrected.

5. The method for budget review based on robust support vector machine according to claim 1, characterized in that: After establishing the approval model, it also includes: Determining a regularization term and a loss function term based on the historical budget approval data after removing duplicate data; The approval model is trained according to the regularization term and the loss function term, and the trained approval model is obtained.

6. A system for budget review based on a robust support vector machine, applicable to the method for budget review based on a robust support vector machine as claimed in any one of claims 1 to 5, characterized in that: include: Data acquisition module, used to obtain historical budget approval data; A model generation module, electrically connected to the data acquisition module, the model generation module is used to establish an approval model formula based on the historical budget approval data, and establish an approval model based on each approval model formula; an approval module electrically connected to the model generation module, the approval module being configured to obtain the budget information to be approved, extract information parameters from the budget information to be approved, and determine whether to approve the budget information to be approved based on a relationship between the information parameters and the approval model set; An inquiry module is electrically connected to the approval module, and is used to obtain the budget information that has not passed the approval, report the budget information to be approved to the authority, and obtain the authority's judgment result on the budget information to be approved, wherein: If the judgment result is passed, the inquiry module passes the budget information to be approved; If the judgment result is failure, the inquiry module rejects the budget information to be approved.

7. The system for budget review based on robust support vector machine according to claim 6, characterized in that: The model generation module is used to establish an approval model formula based on the historical budget approval data, and to establish an approval model based on each approval model formula, including: The model generation module is further configured to obtain consistent duplicate data from each of the historical budget approval data and remove the duplicate data; The model generation module is further configured to establish each of the approval model formulas based on each of the historical budget approval data after removing duplicate data, and to establish an approval model based on each of the approval model formulas.

8. The system for budget review based on robust support vector machine according to claim 7, characterized in that: The model generation module is further configured to establish each of the approval models based on each of the historical budget approval data after removing duplicate data, including: The model generation module is also used to obtain information attributes of each component in the historical preset approval data; The model generation module is further configured to obtain the historical preset approval data with the same information attribute, establish a linear axis of the information attribute based on the historical preset approval data with the same information attribute, and establish the approval model formula based on the linear axis; The model generation module is further configured to obtain parameters of each component of any of the historical preset approval data in the historical preset approval data having the same information attribute; The model generation module is further used to substitute the parameters of each component into the approval model formula and obtain the preset approval result output by the approval model formula; The model generation module is further configured to determine whether to modify the approval model formula according to a relationship between the preset approval result and the approval result of the historical preset approval data.

9. The system for budget review based on robust support vector machine according to claim 8, characterized in that: The model generation module is further configured to determine whether to modify the approval model according to the relationship between the preset approval result and the approval result of the historical preset approval data, including: The model generation module is further configured to compare the preset approval result with the approval result of the historical preset approval data, and determine whether to proofread the approval model according to the comparison result, wherein: If the preset approval result is consistent with the approval result of the historical preset approval data, the model generation module determines that there is no need to modify the approval model; If the preset approval result is inconsistent with the approval result of the historical preset approval data, the model generation module iteratively amends the approval model.

10. The system for budget review based on robust support vector machine according to claim 6, characterized in that: After establishing the approval model, the model generation module further includes: The model generation module is further configured to determine a regularization term and a loss function term based on each of the historical budget approval data after removing duplicate data; The model generation module is also used to train the approval model according to the regularization term and the loss function term, and obtain the trained approval model.

Images