Method and system for monitoring abnormal behavior of enterprises based on difference between purchase and sales invoices

Abstract

The application discloses a kind of based on the method and system for monitoring the abnormal behavior of enterprise based on input-output invoice difference, comprising: respectively obtaining the input invoice data and output invoice data of enterprise in preset time period, and according to the filtering condition of pre-set, the input invoice data and output invoice data are filtered processing;Based on the name of taxpayer and the name of article, respectively, the input invoice data and output invoice data after filtering processing are aggregated according to pre time period, to obtain the input article aggregation list and output article aggregation list corresponding to each taxpayer;Based on the input article aggregation list and output article aggregation list corresponding to each taxpayer, the difference degree of each taxpayer is determined according to the difference degree determination strategy of pre-set;When determining the input-output difference degree of a certain taxpayer is greater than the input-output difference degree threshold value of pre-set, determine that the taxpayer is abnormal taxpayer subject, and send the alarm information related to the abnormal taxpayer subject to regulatory department.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tax monitoring, and more particularly, to a method and system for monitoring abnormal behavior of enterprises based on differences between input and output invoices. BACKGROUND

[0002] Now, tax big data is used to carry out tax risk analysis. For enterprises obviously suspected of tax evasion (evading tax payment), evading the recovery of tax arrears, tax fraud, tax resistance, and fictitious invoicing, tax inspection is carried out in accordance with the law to effectively maintain national tax security. At present, the existing fictitious invoicing inspection method mainly relies on expert experience, supplemented by some big data analysis and prediction methods. Because the invoice chain is relatively long, data screening based on expert experience cannot fully cover the industry or regional characteristics of fictitious invoicing, resulting in an analysis method only suitable for one type of fictitious invoicing inspection. SUMMARY

[0003] The present application provides a method and system for monitoring abnormal behavior of enterprises based on differences between input and output invoices to solve the problem of how to monitor abnormal behavior of enterprises.

[0004] In order to solve the above problems, according to one aspect of the present application, a method for monitoring abnormal behavior of enterprises based on differences between input and output invoices is provided, the method comprising:

[0005] Respectively acquiring input invoice data and output invoice data of an enterprise within a preset time period, and filtering the input invoice data and the output invoice data according to a preset filtering condition;

[0006] Based on the names of taxpayers and the names of items, the filtered input invoice data and the filtered output invoice data are aggregated according to a preset time period to obtain an input item aggregation list and an output item aggregation list corresponding to each taxpayer;

[0007] Based on the input item aggregation list and the output item aggregation list corresponding to each taxpayer, the difference degree of input and output of each taxpayer is determined according to a preset difference degree determination strategy; the difference degree of input and output of each taxpayer is compared with a preset difference degree threshold of input and output,

[0008] and when it is determined that the difference degree of input and output of a certain taxpayer is greater than the preset difference degree threshold of input and output, the taxpayer is determined to be an abnormal taxpayer subject, and alarm information related to the abnormal taxpayer subject is sent to a regulatory department.

[0009] Preferably, wherein the filtering of the input invoice data and the output invoice data according to the preset filtering condition comprises:

[0010] The input invoice data and the output invoice data whose total invoice amount is less than a preset amount threshold are removed. Preferably, the input-output difference degree of each taxpayer is determined according to a preset difference degree determination strategy based on the input item aggregation list and the output item aggregation list corresponding to each taxpayer, including: when the input / output of an item in the input item aggregation list and / or the output item list corresponding to a certain taxpayer is empty, determining that the item difference degree of the item corresponding to the taxpayer is a preset difference degree threshold; when the input item aggregation list and the output item list corresponding to the taxpayer are not empty, performing joint query based on the input aggregation list and the output aggregation list, removing the data with the same item name and quantity, and using spark to traverse each taxpayer, and comparing all items in the input item aggregation list and the output item aggregation list to determine the difference degree of the input item and the output item corresponding to each taxpayer, forming a two-dimensional matrix; a matching list is obtained based on the Hungarian algorithm according to the two-dimensional matrix, and the item difference degree of each item relative to the entire item list is determined according to the matching list; and the item difference degrees of all items corresponding to each taxpayer are accumulated to obtain the input-output difference degree corresponding to each taxpayer.

[0011] Preferably, the preset difference degree threshold is 1.

[0012] Preferably, the method further comprises:

[0013] Before the input invoice data and the output invoice data are aggregated, special ticket deduplication processing is performed on the input invoice data and the output invoice data; wherein the special ticket includes: red ticket and waste ticket.

[0014] According to another aspect of the present application, a system for monitoring abnormal behavior of enterprises based on input-output invoice difference is provided, the system comprising:

[0015] A data acquisition unit is configured to acquire input invoice data and output invoice data of an enterprise within a preset time period, and perform filtering processing on the input invoice data and the output invoice data according to a preset filtering condition;

[0016] An aggregation unit is configured to aggregate the filtered input invoice data and the filtered output invoice data according to a preset time period based on the taxpayer name and the item name, to obtain an input item aggregation list and an output item aggregation list corresponding to each taxpayer;

[0017] An input-output difference degree determination unit is configured to determine the input-output difference degree of each taxpayer based on the input item aggregation list and the output item aggregation list corresponding to each taxpayer according to a preset difference degree determination strategy;

[0018] An abnormality monitoring unit is configured to compare the input-output difference of each taxpayer with a preset input-output difference threshold, and determine that a taxpayer is an abnormal taxpayer when the input-output difference of the taxpayer is greater than the preset input-output difference threshold, and send alarm information related to the abnormal taxpayer to a regulatory department.

[0019] Preferably, the data acquisition unit filters the input invoice data and the output invoice data according to a preset filtering condition, including:

[0020] The input invoice data and the output invoice data whose total amount is less than a preset amount threshold are removed.

[0021] When the input / output of an item in the input item aggregation list and / or the output item list corresponding to a taxpayer is empty, the item difference degree of the item corresponding to the taxpayer is determined to be a preset difference threshold; when the input item aggregation list and the output item list corresponding to the taxpayer are both not empty, the data with the same item name and quantity are removed based on the joint query of the input aggregation list and the output aggregation list, and the spark is used to traverse each taxpayer to compare all items in the input item aggregation list and the output item aggregation list to determine the difference degree of the input item and the output item corresponding to each taxpayer, forming a two-dimensional matrix; a matching list is obtained based on the Hungarian algorithm according to the two-dimensional matrix, and the item difference degree of each item relative to the entire item list is determined according to the matching list;

[0022] The item difference degrees of all items corresponding to each taxpayer are accumulated to obtain the input-output difference of each taxpayer.

[0023] Preferably, the preset difference threshold is 1.

[0024] Preferably, the data acquisition unit further includes:

[0025] The input invoice data and the output invoice data are processed for special ticket deduplication before being aggregated, wherein the special ticket includes a red ticket and a waste ticket.

[0026] The application provides a method and system for monitoring abnormal behavior of enterprises based on differences between input and output invoices, which analyzes input invoices and output invoices of taxpayers within a preset time period, determines the difference between input and output invoices of the taxpayers, and determines abnormal taxpayers according to the difference between input and output invoices and a preset threshold of the difference between input and output invoices, so as to monitor the abnormal behavior of enterprises; wherein, the parameters of the analysis of the difference between input and output invoices can be flexibly set, and the difference between input and output invoices of different taxpayers can be sorted according to the requirements of analysts, so that the analysts can measure the rationality between input and output commodities of taxpayers according to the sorting results; through the method, after a period of training, the parameters and sorting methods are fixed, and automatic identification can also be started, and a list of problem taxpayers is sent to other supervision departments at regular intervals, to assist in identifying abnormal behavior. BRIEF DESCRIPTION OF DRAWINGS

[0027] The exemplary embodiments of the present application can be more fully understood with reference to the accompanying drawings, in which:

[0028] Figure 1 A flow chart of the method 100 for monitoring abnormal behavior of enterprises based on differences between input and output invoices according to the embodiments of the present application;

[0029] Figure 2 A structural schematic diagram of the system 200 for monitoring abnormal behavior of enterprises based on differences between input and output invoices according to the embodiments of the present application.

[0030] DETAILED DESCRIPTION Reference will now be made to the drawings, wherein the exemplary embodiments of the present application will be more fully described by reference to the drawings.

[0031] Unless otherwise defined, the terms (including technical terms) used herein have meanings commonly understood by one of ordinary skill in the art. In addition, it is to be understood that the terms defined by dictionaries and the like are to be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art, and are not to be interpreted in an idealized or overly formal sense unless expressly so defined herein.

[0032] Figure 1 A flow chart of the method 100 for monitoring abnormal behavior of enterprises based on differences between input and output invoices according to the embodiments of the present application. As Figure 1As shown, the method for monitoring enterprise abnormal behavior based on input-output invoice difference provided by the embodiment of the application analyzes input invoices and output invoices of taxpayers within a preset time period to determine input-output difference degree of the taxpayers, and determines abnormal taxpayer subject according to the input-output difference degree and a preset input-output difference degree threshold, so as to monitor enterprise abnormal behavior; wherein, parameters of input-output difference degree analysis can be flexibly set, and input-output difference degrees of different taxpayers can also be sorted according to requirements of an analyst, so that the analyst can measure rationality between input and output commodities of the taxpayers according to the sorting result. Through the method, after a period of training and parameter and sorting mode solidification, automatic identification can also be started, and a list of problem taxpayers can be sent to other supervision departments at regular time, to assist in identifying abnormal behavior. The method 100 for monitoring enterprise abnormal behavior based on input-output invoice difference provided by the embodiment of the application starts from step 101, and input invoice data and output invoice data of an enterprise within a preset time period are acquired in step 101, and the input invoice data and the output invoice data are filtered according to a preset filtering condition.

[0033] Preferably, wherein the filtering processing of the input invoice data and the output invoice data according to the preset filtering condition comprises:

[0034] The input invoice data and the output invoice data with invoice total amount less than a preset amount threshold are removed. Preferably, wherein the method further comprises: before the input invoice data and the output invoice data are aggregated, the input invoice data and the output invoice data are processed for special ticket deduplication; wherein, the special ticket comprises: red ticket and waste ticket.

[0035] In the present application, firstly, full ticket information of enterprise input-output invoices is acquired to complete data acquisition, and then the acquired data is filtered and deduplicated. Specifically, filtering is realized by filtering data with invoice amount less than a certain regional value of total invoice amount. For example, data with invoice total amount less than 5000 is deleted. Deduplication is to remove red tickets and waste tickets in the input invoice data and the output invoice data before the input invoice data and the output invoice data are aggregated, to remove interference invoices.

[0036] In step 102, the filtered input invoice data and the filtered output invoice data are aggregated according to a preset time period based on taxpayer name and item name, to acquire input item aggregation list and output item aggregation list corresponding to each taxpayer.

[0037] In the present application, the input invoice data and the output invoice data are analyzed according to the taxpayer information respectively, and the input item aggregation list and the output item aggregation list are obtained. The selected items are aggregated monthly according to the purchaser taxpayer and the item name through input analysis, and the input item aggregation list of each taxpayer is obtained. The selected items are aggregated monthly according to the seller taxpayer and the item name through output analysis, and the output item aggregation list of each taxpayer is obtained.

[0038] In step 103, the input-output difference degree of each taxpayer is determined according to the preset difference determination strategy based on the input item aggregation list and the output item aggregation list corresponding to each taxpayer.

[0039] Preferably, the input-output difference degree of each taxpayer is determined based on the input item aggregation list and the output item aggregation list corresponding to each taxpayer according to the preset difference determination strategy, which includes: when the input / output of a certain item in the input item aggregation list and / or the output item list corresponding to a certain taxpayer is empty, the item difference degree of the item corresponding to the taxpayer is determined as a preset difference threshold; when the input item aggregation list and the output item list corresponding to the taxpayer are not empty, a joint query is performed based on the input aggregation list and the output aggregation list, data with equal item names and quantities are removed, and spark is used to traverse each taxpayer, and all items in the input item aggregation list and the output item aggregation list are compared in a loop to determine the difference degree of the input item and the output item corresponding to each taxpayer, forming a two-dimensional matrix; a matching list is obtained based on the Hungarian algorithm according to the two-dimensional matrix, and the item difference degree of each item relative to the entire item list is determined according to the matching list;

[0040] The item difference degrees of all items corresponding to each taxpayer are accumulated to obtain the input-output difference degree corresponding to each taxpayer.

[0041] Preferably, the preset difference threshold is 1.

[0042] In the present application, the calculation of the input-output difference degree is divided into three categories: the input of the item is empty, the output of the item is empty, and the input and output of the item are not empty. For the difference degree analysis of certain specific items, the experience parameters of the item difference degree can be adjusted.

[0043] For the item with empty input, each taxpayer with empty input is traversed to set the item difference degree of the item corresponding to each taxpayer with empty input as the preset difference degree threshold 1. Similarly, for the item with empty output, each taxpayer with empty output is traversed to set the item difference degree of the item corresponding to each taxpayer with empty output as the preset difference degree threshold 1. For the item with non-empty input and output, the process of determining the difference degree includes: obtaining the input and output item list of each taxpayer by jointly querying the input item aggregation list and the output item aggregation list; directly taking out the items with equal item names and quantities and all the items with difference degrees less than the experience parameter, and not calculating the difference degrees, and the difference degree is 0; performing the Hungarian algorithm on the remaining items, traversing each taxpayer using spark, comparing all the items in the input item list and the output item list, calculating the difference degrees of the input item and the output item, forming a two-dimensional matrix, using the Hungarian algorithm to obtain a matching list, and obtaining the difference degree of each item relative to the entire input and output item list based on the result obtained by matching. Finally, the item difference degrees of all the items corresponding to each taxpayer are accumulated to obtain the input and output difference degree corresponding to each taxpayer.

[0044] In step 104, the input and output difference degree of each taxpayer is compared with the preset input and output difference degree threshold, and when it is determined that the input and output difference degree of a certain taxpayer is greater than the preset input and output difference degree threshold, the taxpayer is determined to be an abnormal taxpayer subject, and alarm information related to the abnormal taxpayer subject is sent to the supervisory department.

[0045] In the present application, enterprises are sorted according to the size of the input and output difference degree value, the user can set the input and output difference degree threshold, and the enterprises exceeding the preset input and output difference degree threshold will be screened out as abnormal taxpayer subjects, and the alarm information related to the abnormal taxpayer subjects is sent to the supervisory department. Based on the method of the present application, the sampling invoice range and invoice type can be flexibly configured, and the reuse range of the method can be improved by using parameter configuration. When the successful application cases of the method increase, some parameters can be fixed as the difference monitoring model and independently applied. In addition, the big data analysis method used by the system improves the calculation speed.

[0046] Figure 2 A structure schematic diagram of the system 200 for monitoring the abnormal behavior of enterprises based on the input and output invoice difference according to the embodiment of the present application is shown in FIG. 1. Figure 2 As shown in FIG. 1, the system 200 for monitoring the abnormal behavior of enterprises based on the input and output invoice difference provided by the embodiment of the present application includes a data acquisition unit 201, an aggregation unit 202, an input and output difference degree determination unit 203, and an abnormal monitoring unit 204.

[0047] Preferably, the data acquisition unit 201 is configured to acquire the input invoice data and the output invoice data of the enterprise in a preset time period, and perform filtering processing on the input invoice data and the output invoice data according to a preset filtering condition.

[0048] Preferably, the data acquisition unit 201 performs filtering processing on the input invoice data and the output invoice data according to a preset filtering condition, including:

[0049] The input invoice data and the output invoice data with a total invoice amount less than a preset amount threshold are removed.

[0050] Preferably, the data acquisition unit 201 further includes:

[0051] The input invoice data and the output invoice data are subjected to special ticket deduplication processing before being aggregated; wherein the special ticket includes a red ticket and a waste ticket.

[0052] Preferably, the aggregation unit 202 is configured to aggregate the filtered input invoice data and the filtered output invoice data according to a preset time period based on the taxpayer name and the item name, to obtain an input item aggregation list and an output item aggregation list corresponding to each taxpayer.

[0053] Preferably, the input-output difference determination unit 203 is configured to determine the input-output difference of each taxpayer based on the input item aggregation list and the output item aggregation list corresponding to each taxpayer according to a preset difference determination strategy.

[0054] Preferably, the input-output difference determination unit 203 determines the input-output difference of each taxpayer based on the input item aggregation list and the output item aggregation list corresponding to each taxpayer according to a preset difference determination strategy, including: when the input / output of a certain item in the input item aggregation list and / or the output item list corresponding to a certain taxpayer is empty, determining that the item difference of the item corresponding to the taxpayer is a preset difference threshold;

[0055] When the input item aggregation list and the output item list corresponding to a taxpayer are both not empty, joint query is performed based on the input aggregation list and the output aggregation list, data with equal item names and quantities are removed, and spark is used to traverse each taxpayer, and all items in the input item aggregation list and the output item aggregation list are compared in a loop to determine the difference of the input item and the output item corresponding to each taxpayer, to form a two-dimensional matrix; a matching list is obtained based on the Hungarian algorithm according to the two-dimensional matrix, and the item difference of each item relative to the entire item list is determined according to the matching list;

[0056] The item difference degree of all items corresponding to each taxpayer is accumulated to obtain the input and output item difference degree corresponding to each taxpayer.

[0057] Preferably, the preset difference degree threshold is 1.

[0058] Preferably, the abnormality monitoring unit 204 is configured to compare the input and output item difference degree of each taxpayer with a preset input and output item difference degree threshold, and determine a taxpayer as an abnormal taxpayer when it is determined that the input and output item difference degree of the taxpayer is greater than the preset input and output item difference degree threshold, and send alarm information related to the abnormal taxpayer to a regulatory department.

[0059] The system 200 for monitoring abnormal behavior of enterprises based on input and output invoice difference of the embodiment of the present application corresponds to the method 100 for monitoring abnormal behavior of enterprises based on input and output invoice difference of another embodiment of the present application, which will not be described here.

[0060] The present application has been described with reference to a few embodiments. Obviously, a person skilled in the art would find that for the present application the disclosed embodiments as well as some others not particularly described herein are equivalents in spirit within the scope of the present application, as defined by the appended patent claims.

[0061] Generally, all terms used in the claims are to be interpreted according to their ordinary meaning in the technical field, unless explicitly defined otherwise herein. All references to "a / an / the [device, component, etc.]” are to be interpreted openly as referring to at least one instance of said device, component, etc., unless explicitly stated otherwise. As it is explicitly stated above that any method disclosed herein can be performed in any suitable order, the order of steps or actions of methods disclosed herein shall not be interpreted in a way contrary to the ordinary meaning of the steps or actions, unless explicitly stated otherwise. It is understood that the embodiments of the present application can be provided as a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, and the like) embodying computer program code thereon for use by a computer or processor.

[0062] 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 flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the computer or other programmable data processing apparatus produce a device that implements the functions specified in the flowcharts and / or block diagrams. These computer program instructions can also be stored in a computer-usable storage medium that can guide a computer or other programmable data processing apparatus to work in a specific manner, so that the computer program instructions stored in the computer-usable storage medium produce a device that implements the functions specified in the flowcharts and / or block diagrams.Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0063] 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.

[0064] 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 functions specified in one or more boxes. 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, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for monitoring abnormal enterprise behavior based on discrepancies between purchase and sales invoices, characterized in that, The method includes: The system acquires the enterprise's input invoice data and output invoice data within a preset time period, and filters the input invoice data and output invoice data according to preset filtering conditions. Based on the taxpayer's name and the item's name, the filtered input invoice data and output invoice data are aggregated according to a preset time period to obtain the input item aggregation list and output item aggregation list corresponding to each taxpayer. Based on the aggregated list of input and output goods for each taxpayer, the input-output discrepancy for each taxpayer is determined according to a preset discrepancy determination strategy, including: When the input / output of an item in the input or output item list corresponding to a taxpayer is empty, the item difference degree for that item corresponding to the taxpayer is determined to be a preset difference degree threshold. When neither the input nor output item list corresponding to the taxpayer is empty, a joint query is performed based on the input and output item lists. Input and output invoices with the same item name and quantity are removed. Spark is used to iterate through each taxpayer, comparing all items in the input and output item lists to determine the difference degree of input and output items for each taxpayer, forming a two-dimensional matrix. A matching list is obtained by performing optimal matching based on the Hungarian algorithm on the two-dimensional matrix. The item difference degree of each item relative to the entire item list is determined based on the matching list. The item difference degrees of all items corresponding to each taxpayer are summed to obtain the input and output difference degree for each taxpayer. The preset difference degree threshold is 1. The system compares the input-output discrepancy of each taxpayer with a preset input-output discrepancy threshold. When it is determined that the input-output discrepancy of a taxpayer is greater than the preset threshold, the taxpayer is identified as an abnormal taxpayer and an alarm message related to the abnormal taxpayer is sent to the regulatory authority.

2. The method according to claim 1, characterized in that, The filtering process for the input invoice data and output invoice data according to preset filtering conditions includes: Remove input and output invoices whose total invoice amount is less than a preset threshold.

3. The method according to claim 1, characterized in that, The method further includes: Before aggregating the input and output invoice data, special invoice deduplication processing is performed on the input and output invoice data; the special invoices include: red invoices and void invoices.

4. A system for monitoring abnormal enterprise behavior based on discrepancies between purchase and sales invoices, characterized in that, The system includes: The data acquisition unit is used to acquire the input invoice data and output invoice data of the enterprise within a preset time period, and to filter the input invoice data and output invoice data according to preset filtering conditions. The aggregation unit is used to aggregate the filtered input invoice data and output invoice data according to the taxpayer name and the item name respectively, according to a preset time period, so as to obtain the input item aggregation list and output item aggregation list corresponding to each taxpayer. The input-output discrepancy determination unit is used to determine the input-output discrepancy for each taxpayer based on the taxpayer's corresponding input item aggregation list and output item aggregation list, according to a preset discrepancy determination strategy. This includes: When the input / output of a certain item in the input item aggregation list and / or output item list corresponding to a certain taxpayer is empty, the item difference degree corresponding to that taxpayer is determined to be the preset difference degree threshold. When both the input and output item lists for a taxpayer are not empty, a joint query is performed based on the input and output item lists. Input and output invoices with the same item name and quantity are removed. Spark is then used to iterate through each taxpayer, comparing all items in the input and output item lists to determine the degree of difference between the input and output items for each taxpayer, forming a two-dimensional matrix. Based on the two-dimensional matrix, a best-matching algorithm is used to obtain a matching list. The degree of difference of each item relative to the entire item list is then determined based on the matching list. The item difference scores for all items corresponding to each taxpayer are summed to obtain the input-output difference score for each taxpayer; the preset difference score threshold is 1. The anomaly monitoring unit is used to compare the input-output discrepancy of each taxpayer with a preset input-output discrepancy threshold. When it is determined that the input-output discrepancy of a certain taxpayer is greater than the preset input-output discrepancy threshold, the taxpayer is identified as an abnormal taxpayer and alarm information related to the abnormal taxpayer is sent to the regulatory department.

5. The system according to claim 4, characterized in that, The data acquisition unit filters the input invoice data and output invoice data according to preset filtering conditions, including: Remove input and output invoices whose total invoice amount is less than a preset threshold.

6. The system according to claim 4, characterized in that, The data acquisition unit further includes: Before aggregating the input and output invoice data, special invoice deduplication processing is performed on the input and output invoice data; the special invoices include: red invoices and void invoices.

Images