Document auditing sequence determination method and device, computer equipment and program product

Abstract

The invention relates to the technical field of data auditing, and discloses a receipt auditing sequence determination method and device, computer equipment and a program product. As the auditing difference values of the plurality of historical auditing receipts of the target applicant under the target auditing rule are associated with the auditing performance values of the plurality of historical auditing receipts of the to-be-audited receipts, the to-be-audited receipts are audited according to the auditing difference values of the plurality of historical auditing receipts of the target applicant under the target auditing rule; and determining that the to-be-audited sequence of the to-be-audited document under the target audit specification is associated with the audit performance values of the plurality of historical audit documents. According to the method, the to-be-audited sequence is determined under the target audit specification, and the audit performance of the target applicant under the plurality of historical audit receipts under the target audit specification is considered, so that the to-be-audited sequence of the to-be-audited receipts of different target applicants under each target audit specification can be flexibly adapted; and thus, the to-be-audited documents with relatively high risks can be audited preferentially, and the flexibility and efficiency of audit of the to-be-audited documents are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data auditing, and in particular to a bill auditing sequence determination method and device, a computer device and a program product. BACKGROUND

[0002] Bill auditing is further auditing and confirmation of the accuracy of bills after the bills are saved. Bill auditing, as a core business link in the financial, tax, supply chain and other industries, directly affects the enterprise operating cost and risk control ability.

[0003] In related technologies, in the process of auditing bills, the auditing sequence of the bills is generally determined in a unified and fixed manner. Since the auditing sequence does not take into account the historical performance differences of different applicants under various auditing regulations, the determination manner of the auditing sequence is not flexible enough, and thus the bill to be audited with higher risk cannot be audited preferentially. SUMMARY

[0004] Therefore, the present application provides a bill auditing sequence determination method and device, a computer device and a program product to solve the problem that the determination manner of the auditing sequence is not flexible enough in the process of auditing bills, and thus the bill to be audited with higher risk cannot be audited preferentially.

[0005] In a first aspect, the present application provides a bill auditing sequence determination method, which comprises:

[0006] obtaining a bill to be audited and a plurality of historical auditing bills of the bill to be audited of a target applicant under a target auditing regulation, wherein the target auditing regulation is each auditing regulation of a plurality of auditing regulations;

[0007] obtaining an auditing performance value of the target applicant for the plurality of historical auditing bills under the target auditing regulation;

[0008] determining an auditing difference value of the plurality of historical auditing bills of the target applicant under the target auditing regulation based on a plurality of coordinate points of the auditing performance values of the plurality of historical auditing bills in a preset coordinate system, wherein the horizontal coordinate value of each coordinate point represents the historical auditing sequence of each historical auditing bill, the vertical coordinate value of each coordinate point represents the auditing performance value of each historical auditing bill, and the historical auditing sequence of each historical auditing bill is determined by the auditing performance value of the historical auditing bill;

[0009] determining a to-be-audited sequence of the bill to be audited under the target auditing regulation according to the auditing difference value of the plurality of historical auditing bills of the target applicant under the target auditing regulation.

[0010] According to the audit distinctness value of the multiple historical audit documents of the target applicant under the target audit regulation, the audit order of the to-be-audited document under the target audit regulation is determined in association with the audit performance value of the multiple historical audit documents. Finally, the audit order determined by the embodiment under the target audit regulation can flexibly adapt to the audit order of the to-be-audited document of different target applicants under different target audit regulations, and thus is beneficial to preferentially auditing the to-be-audited document with higher risk and improving the flexibility and efficiency of the to-be-audited document audit.

[0011] In some optional embodiments, the audit distinctness value of the multiple historical audit documents of the target applicant under the target audit regulation is determined based on multiple coordinate points of the audit performance value of the multiple historical audit documents in a preset coordinate system, and the method comprises the following steps:

[0012] The number of the multiple coordinate points in the preset coordinate system is marked as n;

[0013] n-1 audit performance change line segments are determined from the n coordinate points in the preset coordinate system, and each audit performance change line segment is connected by two adjacent coordinate points;

[0014] An audit performance change curve is formed by the n-1 audit performance change line segments, and a starting endpoint and a terminal endpoint are determined from the audit performance change curve;

[0015] A closed figure is constructed in the preset coordinate system according to the audit performance change curve, the starting endpoint and the terminal endpoint;

[0016] The total area of the closed figure is calculated;

[0017] The comprehensive change slope of the n-1 audit performance change line segments is calculated, and the comprehensive change slope is calculated by summing the target slope of each audit performance change line segment;

[0018] The audit distinctness value of the multiple historical audit documents of the target applicant under the target audit regulation is calculated according to the comprehensive change slope and the total area of the closed figure.

[0019] According to the embodiment, the closed figure is constructed by the coordinate points in the preset coordinate system, the total area of the closed figure is calculated, and the comprehensive change slope of the audit performance change curve formed by the multiple coordinate points is calculated, so that the audit distinctness value of the multiple historical audit documents of the target applicant under the target audit regulation is finally obtained, which is beneficial to further accurately obtaining the audit order of the to-be-audited document under the target audit regulation.

[0020] In some optional embodiments, according to the audit performance change curve, the starting endpoint and the ending endpoint, a closed figure is constructed in the preset coordinate system, comprising:

[0021] a first perpendicular line segment is drawn from the starting endpoint to the horizontal coordinate axis in the preset coordinate system, and a second perpendicular line segment is drawn from the ending endpoint to the vertical coordinate axis in the preset coordinate system;

[0022] In the preset coordinate system, a closed figure formed by the first perpendicular line segment, the second perpendicular line segment, the horizontal coordinate axis and the audit performance change curve is constructed.

[0023] The above embodiments are constructed in the preset coordinate system to construct a closed figure, which is beneficial to further calculate the audit difference values of the multiple historical audit documents of the target applicant under the target audit regulation.

[0024] In some optional embodiments, the audit performance values of the multiple historical audit documents of the target applicant under the target audit regulation are obtained, comprising:

[0025] At least one audit feature corresponding to each historical audit document is extracted, and the audit feature includes: a field integrity audit feature, a format compliance audit feature and a cross-field logic consistency audit feature;

[0026] The at least one audit feature is input into a target audit model for prediction to obtain the audit performance values of the multiple historical audit documents of the target applicant, wherein the target audit model is a pre-trained neural network model.

[0027] The above embodiments are trained by the target audit model on the at least one audit feature, which is beneficial to further accurately obtain the audit performance values of the multiple historical audit documents of the target applicant.

[0028] In some optional embodiments, according to the audit difference values of the multiple historical audit documents of the target applicant under the target audit regulation, the audit order of the to-be-audited document under the target audit regulation is determined, comprising:

[0029] The audit difference values of the multiple historical audit documents of the target applicant under the target audit regulation are sorted in descending order to obtain the historical audit order of the multiple historical audit documents under the target audit regulation;

[0030] According to the historical audit order of the multiple historical audit documents under the target audit regulation, the audit order of the to-be-audited document is determined.

[0031] The embodiment sorts the audit difference values of the plurality of historical audit documents of the target applicant under the target audit regulation in descending order according to the above-mentioned embodiment, and sorts the historical audit documents with higher risk levels in front, so that the to-be-audited documents with higher risk levels are also sorted in front due to the same types of the to-be-audited documents and the historical audit documents, and finally, the to-be-audited documents with higher risk levels are audited preferentially.

[0032] In some optional embodiments, the to-be-audited document of the target applicant under the target audit regulation and the plurality of historical audit documents of the to-be-audited document are acquired, including:

[0033] The request instruction package of the to-be-audited document initiated by the target applicant is received, and the identification information of the target applicant is loaded in the request instruction package;

[0034] The plurality of historical audit documents of the to-be-audited document are acquired according to the identification information of the target applicant.

[0035] The embodiment has the advantage that the plurality of historical audit documents of the to-be-audited document are quickly queried according to the above-mentioned embodiment.

[0036] In a second aspect, the embodiment provides a document audit order determination device, and the device includes:

[0037] The audit document acquisition module is configured to acquire the to-be-audited document of the target applicant under the target audit regulation and the plurality of historical audit documents of the to-be-audited document, wherein the target audit regulation is each audit regulation of a plurality of audit regulations, and the target audit regulation is each audit regulation of the plurality of audit regulations;

[0038] The audit performance acquisition module is configured to acquire the audit performance value of the target applicant to the plurality of historical audit documents under the target audit regulation;

[0039] The audit difference determination module is configured to determine the audit difference value of the plurality of historical audit documents of the target applicant under the target audit regulation based on the plurality of coordinate points of the audit performance value of the plurality of historical audit documents in a preset coordinate system, wherein the horizontal coordinate value of each coordinate point represents the historical audit order of each historical audit document, the vertical coordinate value of each coordinate point represents the audit performance value of each historical audit document, and the historical audit order of each historical audit document is determined by the audit performance value of the historical audit document;

[0040] The audit order determination module is configured to determine the to-be-audited order of the to-be-audited document under the target audit regulation according to the audit difference value of the plurality of historical audit documents of the target applicant under the target audit regulation.

[0041] In a third aspect, the embodiment provides a computer device, and the device includes:

[0042] The memory and the processor are connected in communication with each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the document review sequence determination method of the first aspect or any of the corresponding embodiments.

[0043] In a fourth aspect, the embodiment provides a computer readable storage medium, which stores computer instructions, and the computer instructions are used to perform the document review sequence determination method of the first aspect or any of the corresponding embodiments.

[0044] In a fifth aspect, the embodiment provides a computer program product, which comprises computer instructions, and the computer instructions are used to make a computer perform the document review sequence determination method of the first aspect or any of the corresponding embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings needed to be used in the specific embodiments or prior art description. Obviously, the drawings described below are some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0046] Figure 1 FIG. 1 is a flowchart of a document review sequence determination method according to an embodiment of the present application;

[0047] Figure 2 FIG. 2 is a flowchart of another document review sequence determination method according to an embodiment of the present application;

[0048] Figure 3 FIG. 3 is a schematic diagram of forming a closed figure based on a coordinate point in a preset coordinate system according to an embodiment of the present application;

[0049] Figure 4 FIG. 4 is a flowchart of a document review sequence determination method according to an embodiment of the present application;

[0050] Figure 5 FIG. 5 is a structural block diagram of a document review sequence determination device according to an embodiment of the present application;

[0051] Figure 6 FIG. 6 is a hardware structure schematic diagram of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION

[0052] To make the objects, technical solutions and advantages of the embodiments of the present application clearer, the following will be combined with the drawings of the embodiments of the present application to make a clear and complete description of the technical solutions in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0053] In the embodiment, a bill auditing sequence determination method is provided, which can be used for computer equipment such as mobile phones, tablet computers, desktop computers, portable notebooks, servers and the like. Figure 1 The flowchart of the bill auditing sequence determination method according to the embodiment of the present application is shown in FIG. 1, which includes the following steps: Figure 1

[0054] In step S101, the target applicant's to-be-audited bill under the target auditing regulation and the target applicant's to-be-audited bill's multiple historical auditing bills are obtained, wherein the target auditing regulation is each auditing regulation of multiple auditing regulations.

[0055] The target applicant can be any applicant applying for the to-be-audited bill, for example, the target applicant includes but is not limited to: an applicant applying for an invoice, an applicant applying for a billing, an applicant applying for a registered account, etc. The multiple auditing regulations include: auditing regulation A, auditing regulation B, auditing regulation C, auditing regulation D, wherein the target auditing regulation can be A or B or C or D.

[0056] The to-be-audited bill is a bill requested to be audited by the applicant. The historical auditing bill is multiple historical auditing bills of the to-be-audited bill completed by the target applicant before each auditing regulation. Therefore, the to-be-audited bill and the historical auditing bill are of the same type. The type of the to-be-audited bill and the historical auditing bill includes but is not limited to historical invoicing bill, historical registered account bill, historical registered housing bill, historical reimbursement bill, etc.

[0057] In a specific example, the target applicant's to-be-audited bill under the target auditing regulation and the target applicant's to-be-audited bill's multiple historical auditing bills are obtained, including:

[0058] In step a1, a request instruction package of the to-be-audited bill initiated by the target applicant is received, and the request instruction package is loaded with the identification information of the target applicant.

[0059] In step a2, according to the identification information of the target applicant, the multiple historical auditing bills of the to-be-audited bill are obtained.

[0060] In step S102, the target applicant's auditing performance value for the multiple historical auditing bills under the target auditing regulation is obtained.

[0061] ​The audit performance value is different audit performance of the target applicant under the target audit regulation on field integrity, format compliance, cross-field logical consistency, etc.

[0062] The embodiment can preferentially use the pre-trained target audit model to further obtain the audit performance values of the target applicant under the target audit regulation on the plurality of historical audit documents, and thus achieve the purpose of accurately extracting the audit performance values of the plurality of historical audit documents.

[0063] In step S103, based on the plurality of coordinate points of the audit performance values of the plurality of historical audit documents in the preset coordinate system, the audit distinct values of the plurality of historical audit documents of the target applicant under the target audit regulation are determined, wherein the horizontal coordinate value of each coordinate point represents the historical audit order of each historical audit document, the vertical coordinate value of each coordinate point represents the audit performance value of each historical audit document, and the historical audit order of each historical audit document is determined by the audit performance value of the historical audit document.

[0064] In a specific example, the audit performance values of the plurality of historical audit documents of the target applicant under the target audit regulation are sorted in order from large to small to obtain the historical audit orders of the plurality of historical audit documents under the target audit regulation; and the historical audit order of each historical audit document is sorted in order from early to late in the preset coordinate system.

[0065] The greater the audit performance value of the plurality of historical audit documents of the target applicant under the target audit regulation, the more unstable the historical performance of the target applicant on the document, and the earlier the sorting, the historical audit order of the historical audit document corresponding to the audit performance value of the historical audit document sorted earlier is sorted earlier in the preset coordinate system, which means that it can be audited preferentially.

[0066] Specifically, the preset coordinate system can be a rectangular coordinate system. In the rectangular coordinate system, the horizontal coordinate value of each coordinate point represents the historical audit order of each historical audit document. The historical audit order of each historical audit document is labeled in order from early to late on the X-axis of the rectangular coordinate system, and the audit performance value corresponding to each label is taken as the parameter value of the Y-axis. The audit performance value corresponding to each label is marked in the form of a coordinate point in the rectangular coordinate system, and a plurality of coordinate points are marked, and the number of the plurality of coordinate points is n.

[0067] Further, the audit distinct values of the plurality of historical audit documents of the target applicant under the target audit regulation are determined based on the n coordinate points of the audit performance values of the plurality of historical audit documents in the preset coordinate system.

[0068] In step S104, according to the audit distinct values of the plurality of historical audit documents of the target applicant under the target audit regulation, the audit order of the to-be-audited document under the target audit regulation is determined.

[0069] In a specific example, the order of the historical audit documents under the target audit regulation is determined according to the audit difference values of the historical audit documents of the target applicant under the target audit regulation, comprising:

[0070] Step b1, the audit difference values of the historical audit documents of the target applicant under the target audit regulation are sorted in descending order to obtain the order of the audit documents under the target audit regulation.

[0071] Step b2, the order of the audit documents is determined according to the historical audit order of the historical audit documents under the target audit regulation.

[0072] For example, the audit difference value of the historical billing document of the target applicant 1 under the target audit regulation A is 0.33, the audit difference value of the historical registration document of the target applicant 1 under the target audit regulation A is 0.55, and the audit difference value of the historical reimbursement document of the target applicant 1 under the target audit regulation A is 0.57.

[0073] For example, the audit difference values 0.57, 0.55, 0.33 of the historical audit documents of the target applicant 1 under the target audit regulation A are sorted in descending order to determine the order of the historical audit documents under the target audit regulation A as follows: the audit order of the historical reimbursement document is first, the audit order of the historical registration document is second, and the audit order of the historical billing document is third.

[0074] Since the historical audit document is a historical audit document of the audit document, the type of the historical audit document is the same as the type of the audit document. Based on the above example, when the type of the audit document is a reimbursement document, the order of the audit document is first; when the type of the audit document is a registration document, the order of the audit document is second; when the type of the audit document is a billing document, the order of the audit document is third.

[0075] The larger the audit difference value, the greater the fluctuation of the target applicant under the target regulation, which means that the risk of the audit document is greater. Therefore, the audit difference values of the historical audit documents of the target applicant under the target audit regulation are sorted in descending order, and the historical audit documents with higher risk levels are sorted in front. Since the audit document and the historical audit document are of the same type, the audit document with a higher risk level is also sorted in front, which is beneficial to the priority audit of the audit document with a higher risk.

[0076] In summary, since the audit distinct values of the plurality of historical audit documents of the target applicant under the target audit regulation are associated with the audit performance values of the plurality of historical audit documents, the audit order of the to-be-audited document under the target audit regulation is determined according to the audit distinct values of the plurality of historical audit documents of the target applicant under the target audit regulation, and is associated with the audit performance values of the plurality of historical audit documents. Finally, the to-be-audited order determined by the embodiment under the target audit regulation can flexibly adapt to the to-be-audited order of the to-be-audited document of different target applicants under each audit regulation by considering the audit performance of the plurality of historical audit documents of the target applicant under the target audit regulation, thereby facilitating the priority audit of the to-be-audited document with higher risk and improving the flexibility and efficiency of the to-be-audited document audit.

[0077] In the embodiment, a document audit order determination method is provided, which can be used for computer equipment such as mobile phones, tablet computers, desktop computers, portable notebooks, servers, etc. Figure 2 is a flowchart of the document audit order determination method according to the embodiment of the application, as shown in Figure 2 S103, based on the plurality of coordinate points of the audit performance values of the plurality of historical audit documents in the preset coordinate system, determining the audit distinct values of the plurality of historical audit documents of the target applicant under the target audit regulation, comprising:

[0078] S1031, marking the number of the plurality of coordinate points in the preset coordinate system as n.

[0079] S1032, determining n-1 audit performance change line segments from the n coordinate points in the preset coordinate system, and each audit performance change line segment is connected by two adjacent coordinate points.

[0080] As shown in Figure 3 , the preset coordinate system is a rectangular coordinate system, n coordinate points are marked in the rectangular coordinate system, and n-1 audit performance change line segments are determined, and the audit performance change line segment OC is connected by two adjacent coordinate points.

[0081] S1033, constructing an audit performance change curve through the n-1 audit performance change line segments, and determining a starting endpoint and a terminating endpoint from the audit performance change curve.

[0082] In Figure 3 , the n-1 audit performance change line segments constitute an audit performance change curve L, and the starting endpoint E and the terminating endpoint N are determined from the audit performance change curve L.

[0083] S1034, constructing a closed figure in the preset coordinate system according to the audit performance change curve, the starting endpoint and the terminating endpoint.

[0084] In a specific example, the step S1034 of constructing the closed figure in the preset coordinate system according to the audit performance change curve, the starting endpoint and the ending endpoint comprises:

[0085] The step c1 comprises: making a first perpendicular segment from the starting endpoint to a horizontal coordinate axis in the preset coordinate system, and making a second perpendicular segment from the ending endpoint to a vertical coordinate axis in the preset coordinate system.

[0086] In the step c1, Figure 3 a first perpendicular segment m1 is made from the starting endpoint to a horizontal coordinate axis (X-axis) in the rectangular coordinate system, and a second perpendicular segment m2 is made from the ending endpoint to a vertical coordinate axis (Y-axis) in the rectangular coordinate system.

[0087] The step c2 comprises: constructing, in the preset coordinate system, a closed figure formed by the first perpendicular segment, the second perpendicular segment, the horizontal coordinate axis and the audit performance change curve.

[0088] In the step c2, Figure 3 a closed figure M is constructed in the rectangular coordinate system by the first perpendicular segment m1, the second perpendicular segment m2, the horizontal coordinate axis (X-axis), the vertical coordinate axis (Y-axis) and the audit performance change curve.

[0089] The step S1035 comprises: calculating a total area of the closed figure.

[0090] The total area of the closed figure is determined by the shape of the closed figure. For example, when the closed figure is a trapezoid, the total area of the closed trapezoid is calculated according to the calculation formula of the area of the trapezoid; when the closed figure is a rectangle, the total area of the rectangle is calculated according to the calculation formula of the area of the rectangle; when the closed figure can be divided into multiple regular figures, the total sum of each regular figure is calculated.

[0091] The step S1036 comprises: calculating a comprehensive change slope of the n-1 audit performance change line segments, the comprehensive change slope being calculated by summing up the target slopes of each audit performance change line segment.

[0092] For example, in the step S1036, Figure 3 the target slope of the audit change line segment OC is represented by OC k .

[0093] wherein (x2, y2) represents the coordinates of the point C, and (x1, y1) represents the coordinates of the point O.

[0094] The total sum of the target slopes of the n-1 audit performance change line segments is calculated to obtain the comprehensive change slope Slcoy of the n-1 audit performance change line segments.

[0095] Step S1037, according to the comprehensive change slope and the total area of the closed figure, the audit difference value of the target applicant under the target audit regulation is calculated.

[0096] The comprehensive change slope is denoted by SICOY, and the total area of the closed figure is denoted by HKBm. The audit difference value of the target applicant under the target audit regulation is calculated by the following formula:

[0097]

[0098] In the embodiment, the audit difference value of the target applicant under the target audit regulation is determined by the above-mentioned manner, which is beneficial to further accurately and flexibly determine the to-be-audited order of the to-be-audited document under the target audit regulation, and expose the to-be-audited document with higher risk in advance, so as to finally achieve the priority audit of the to-be-audited document with higher risk.

[0099] In some specific embodiments, as shown in Figure 4 The step S102 of obtaining the audit performance value of the target applicant under the target audit regulation includes:

[0100] Step S1021, at least one audit feature corresponding to each historical audit document is extracted, and the audit feature includes: field integrity audit feature, format compliance audit feature, and cross-field logic consistency audit feature.

[0101] Step S1022, at least one audit feature is input into the target audit model for prediction to obtain the audit performance value of the target applicant under the target audit regulation, wherein the target audit model is a pre-trained neural network model.

[0102] For example, the target applicant 1 has multiple historical audit documents under the target audit regulation A, which are historical billing document 11, historical billing document 12, historical registration document 21, and historical reimbursement document 31. The historical audit documents in this embodiment are pre-processed historical audit documents.

[0103] In a specific example, step S1022, the field integrity audit feature corresponding to each historical audit document is extracted, including:

[0104] Step d1, the mandatory field attribute list of each historical audit document is obtained, and the mandatory field attribute list includes a plurality of mandatory field attributes.

[0105] For example, for the historical billing document, the mandatory field attribute list of the historical billing document is obtained, and the mandatory field attribute list includes but is not limited to: invoice code, billing date, amount, and purchaser name.

[0106] Step d2, extract the Boolean vector of each mandatory field attribute in the mandatory field attribute list of each historical audit document.

[0107] For example, for each historical audit document, if there is a mandatory field attribute, the field is marked as 1, otherwise, it is marked as 0, and then a Boolean vector is generated.

[0108] Step d3, encode the Boolean vector of each mandatory field attribute in the mandatory field attribute list of each historical audit document to obtain the field encoding result of the historical audit document.

[0109] When the historical audit document is a historical billing document, the historical billing document is encoded by the One-Hot encoding network or the Embedding encoding network, and the encoding result obtained is: for example, the historical billing document = (1, 0, 0). When the historical audit document is a historical reimbursement document, the historical reimbursement document = (0, 1, 0). Therefore, by encoding each historical audit document, the embodiment can solve the problem of differences in mandatory fields of different types of documents.

[0110] Step d4, according to the business rules, the weights of the plurality of mandatory field attributes are assigned, and the field integrity audit features corresponding to each historical audit document are obtained.

[0111] When the historical audit document is a historical billing document, the mandatory field attribute is a billing code, and the weight of the "billing code" is 10. The mandatory field attribute is a billing address, and the weight of the "billing address" is 5. When the mandatory field attribute is missing, the influence degree of the optional field attribute can be calculated according to the weight, and the influence degree is used for audit result scoring.

[0112] In another specific example, step S1022, the format compliance audit features corresponding to each historical audit document are extracted, including:

[0113] Step e1, identify a plurality of text field attributes of each historical audit document.

[0114] The plurality of text attributes include but are not limited to date, mailbox, code, etc. In the specific identification of the plurality of text field attributes, the plurality of text field attributes can be obtained by regular matching.

[0115] For example, regular expressions such as date format ^\d{4}-\d{2}-\d{2}$, email format ^[\w.-]+@[\w.-]+\.\w+$, etc. are written according to the format rule; regular matching is performed on the field value, and matching is successful = 1, and failure = 0, and then the format compliance text features are generated.

[0116] For example, the field value "2025 / 04 / 22" matches the date regular expression fails, and the match fails = 0; "2025-04-22" matches successfully, and the match fails = 1.

[0117] Step e2, define the label of each text field attribute of each historical audit document.

[0118] For example, "date" = 01, "mailbox" = 02, "invoice code" = 03.

[0119] Step e3, encode the label of each text field attribute of each historical audit document to obtain the label encoding result of the historical audit document.

[0120] By using One-Hot encoding network or Embedding encoding network, the label of each text field attribute of each historical audit document is converted into a vector, and then the cross-field format rule difference is solved.

[0121] Step e4, according to the label encoding result of each historical audit document, judge whether the current data amount of each historical audit document is equal to the preset data amount.

[0122] Step e5, mark the historical audit document whose current data amount is not equal to the preset data amount as abnormal data;

[0123] Step e6, the historical audit document whose current data amount is equal to the preset data amount is taken as the format compliance audit feature corresponding to each historical audit document.

[0124] For example, when the historical audit document is a historical invoice document, the preset data amount is 12 characters, and the preset data amount can be flexibly set according to the actual situation. If the current data amount of the historical invoice document is not equal to 12 characters, the historical audit document is marked as abnormal data.

[0125] In another specific example, step S1012, extracting the format compliance audit feature corresponding to each historical audit document further includes:

[0126] Step f1, identifying a plurality of numerical field attributes of each historical audit document, the plurality of numerical field attributes including: integer field attributes and decimal field attributes.

[0127] Step f2, according to the decimal field attribute of each historical audit document, judging whether the number of digits after the decimal point in the decimal field meets the requirements.

[0128] Step f3, when the number of digits after the decimal point of each historical audit document does not meet the requirements, marking the historical audit document as a first vector;

[0129] Step f4, when the number of decimal places of each historical audit document meets the requirements, mark the historical audit document as the second vector;

[0130] The first vector is represented by 1, and the second vector is represented by 0. For example, judging the decimal compliance, whether the number of decimal places meets the requirements, when the number of decimal places meets the requirements, mark 0, otherwise, mark 1.

[0131] Step f5, detecting whether there are abnormal symbols in the multiple numerical field attributes of each historical audit document;

[0132] Step f6, when there are abnormal symbols in the multiple numerical field attributes of each historical audit document, mark the historical audit document as the first vector;

[0133] Step f7, when there are no abnormal symbols in the multiple numerical field attributes of each historical audit document, mark the historical audit document as the second vector;

[0134] For example, when the historical audit document is a historical billing document, when the historical billing document appears “￥” “$” and other non-numeric characters, mark 1, indicating that there are illegal symbols, otherwise, mark 0.

[0135] In another specific example, step S1012, extracting the cross-field logical consistency audit features corresponding to each historical audit document, including:

[0136] Step g1, obtaining multiple numerical fields of each historical audit document;

[0137] Step g2, determining a target numerical field and an associated numerical field of the target numerical field from the multiple numerical fields;

[0138] Step g3, calculating the numerical difference or numerical ratio between the target numerical field and the associated numerical field;

[0139] For example, the historical audit document is a historical billing document, the target numerical field is the total amount, and the sub-item amount is the associated numerical field. The numerical difference between the two is represented as total amount-sub-item amount. When the target numerical field is tax amount and the associated numerical field is amount, the numerical ratio between the two is represented as “tax amount / amount”.

[0140] Step g4, according to whether the numerical difference or numerical ratio between the target numerical field and the associated numerical field meets the preset requirements;

[0141] Step g5, if not, mark the compliance between the target numerical field and the associated numerical field as the first vector.

[0142] Step g6, if yes, mark the compliance between the target value field and the associated value field as the second vector, and perform cross-value field extreme value checking on the target value field to obtain the cross-value field logical consistency audit feature corresponding to each historical audit document;

[0143] For example, the first vector is 1 and the second vector is 0. If the value difference between the target value field and the associated value field is equal to 0, the compliance between the target value field and the associated value field is marked as 0, otherwise, it is marked as 1. If the value ratio between the target value field and the associated value field is within the allowed preset error range, and the value ratio is less than or equal to ≤0.1%, the compliance between the target value field and the associated value field is marked as 0, otherwise, it is marked as 1.

[0144] For example, the total amount of the purchase order is 1000 yuan, and the item sum is 999 yuan. The difference is 1 yuan. If the allowed error is 5 yuan, it is marked as 0, indicating that the difference between the total amount and the item sum is compliant.

[0145] Further, when the value difference or value ratio between the target value field and the associated value field meets the preset requirement, cross-value field extreme value checking is performed on the target value field.

[0146] The target value field is identified to obtain the cross-value field extreme value in the target value field, and then the value of the target value field is further checked for compliance in combination with the cross-value field extreme value. The compliance is marked as the second vector 0, and then the cross-value field logical consistency audit feature corresponding to each historical audit document is obtained. Otherwise, it is marked as the first vector 1, indicating that the value in the target value field exceeds the cross-value field extreme value.

[0147] For example, when the historical audit document is a historical billing document, when the current unit price value is less than or equal to the historical maximum unit price value, it indicates that the target value field cross-value field extreme value checking result meets the requirement.

[0148] In another specific example, step S1012, the cross-field logical consistency audit feature corresponding to each historical audit document is extracted, including:

[0149] Step h1, obtaining a plurality of time fields of each historical audit document.

[0150] Step h2, calculating the time difference between each target time in each historical audit document and the preset time;

[0151] Step h3, determining whether the time difference is less than or equal to the preset difference;

[0152] Step h4, if no, determining that the time of each historical audit document is not compliant, and marking it as the first vector.

[0153] Step h5, if yes, determine the time compliance of each historical audit document, and mark as the second vector.

[0154] For example, if the delivery date in the historical audit document is less than or equal to the signing date, mark it as the second vector, and determine that the historical audit document is compliant, otherwise, mark it as the first vector, and determine that the historical audit document is not compliant.

[0155] In another specific example, step S1012, extract the cross-field logical consistency audit features corresponding to each historical audit document, including:

[0156] Step i1, extract a plurality of text field features of each historical audit document.

[0157] Step i2, determine the cross-field logical consistency audit features corresponding to each historical audit document according to the text similarity of each text field feature of each historical audit document.

[0158] Wherein, the text similarity of each text field feature is determined by the text similarity between the text field feature and another text field feature.

[0159] Text field features with a text similarity greater than or equal to a preset threshold are taken as cross-field logical consistency audit features of each text field feature.

[0160] Step i3, discard text field features with a text similarity less than a preset threshold.

[0161] For example, if the similarity between "Contractor A name" and "Supplier name" is greater than or equal to a preset threshold (90%), it means that the two text features are compliant, otherwise, they are not compliant.

[0162] The target audit model in this embodiment is a pre-trained neural network model, and each historical audit document is a historical audit document collected for each audit regulation, therefore, each audit regulation corresponds to an independently trained deep learning model, and each deep learning model is trained independently to be the corresponding target audit model.

[0163] During the training process, a deep learning model is built based on the field integrity audit features, format compliance audit features, and cross-field logical consistency audit features of the training samples. The deep learning model is then trained. During this process, each audit feature corresponding to the training sample is assigned an audit performance value, ranging from 1 to 5. It is important to note that the audit performance value has a specific meaning: a higher value indicates greater compliance with the audit standards. The audit features corresponding to the training samples are then divided into training, validation, and test sets according to specific ratios: 60%, 20%, and 20%. This division facilitates comprehensive and scientific model training, validation, and final performance testing, ensuring that the target audit model accurately assesses compliance with the audit standards based on the input audit features. After training is complete and the dataset is divided, the target audit model is repeatedly trained using the training set, allowing it to learn and grasp the inherent correlation between the audit features and audit performance values ​​corresponding to each historical audit document. During the training process, the validation set was used to timely verify the performance of the target audit model during the training phase. Based on the verification results, the target audit model's parameters were adjusted promptly, optimizing its structure and moving it towards greater accuracy and stability. Once the target audit model achieved the expected performance on the training and validation sets, a final performance test was conducted on the test set to verify its generalization and accuracy on documents not included in the training. Ultimately, a target audit model was successfully constructed specifically for auditing field completeness, format compliance, and cross-field logic.

[0164] This embodiment also provides a document review order determination device, which is used to implement the above-mentioned embodiments and preferred embodiments. Details already described will not be repeated here. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0165] This embodiment provides a document review order determination device, such as Figure 5 As shown, including:

[0166] The audit document acquisition module 501 is used to acquire the documents to be audited and multiple historical audit documents of the documents to be audited under the target audit regulations of the target applicant, wherein the target audit regulations are each audit regulation of the multiple audit regulations; wherein the target audit regulations are each audit regulation of the multiple audit regulations;

[0167] The auditing performance obtaining module 502 is configured to obtain an auditing performance value of the target applicant under the target auditing regulation for a plurality of historical auditing documents;

[0168] The auditing difference determining module 503 is configured to determine an auditing difference value of the target applicant under the target auditing regulation for the plurality of historical auditing documents based on a plurality of coordinate points of the auditing performance values of the plurality of historical auditing documents in a preset coordinate system, wherein a horizontal coordinate value of each coordinate point represents a historical auditing sequence of each historical auditing document, a vertical coordinate value of each coordinate point represents the auditing performance value of each historical auditing document, and the historical auditing sequence of each historical auditing document is determined by the auditing performance value of the historical auditing document.

[0169] The auditing sequence determining module 504 is configured to determine a to-be-audited sequence of the to-be-audited document under the target auditing regulation according to the auditing difference value of the plurality of historical auditing documents of the target applicant under the target auditing regulation.

[0170] In some optional embodiments, the auditing difference determining module 503 comprises:

[0171] The number marking sub-module is configured to mark a number of the plurality of coordinate points in the preset coordinate system as n.

[0172] The line segment determining sub-module is configured to determine n-1 auditing performance change line segments from the n coordinate points in the preset coordinate system, and each auditing performance change line segment is connected by two adjacent coordinate points.

[0173] The end point determining sub-module is configured to form an auditing performance change curve by the n-1 auditing performance change line segments, and determine a starting end point and a terminal end point from the auditing performance change curve.

[0174] The figure determining sub-module is configured to construct a closed figure in the preset coordinate system according to the auditing performance change curve, the starting end point and the terminal end point.

[0175] The area calculating sub-module is configured to calculate a total area of the closed figure.

[0176] The slope calculating sub-module is configured to calculate a comprehensive change slope of the n-1 auditing performance change line segments, and the comprehensive change slope is calculated by summing a target slope of each auditing performance change line segment.

[0177] The difference determining sub-module is configured to calculate the auditing difference value of the target applicant under the target auditing regulation for the plurality of historical auditing documents according to the comprehensive change slope and the total area of the closed figure.

[0178] In some optional embodiments, the figure determining sub-module comprises:

[0179] The first determining unit is configured to make a first perpendicular line segment from the starting endpoint to a horizontal coordinate axis in a preset coordinate system, and make a second perpendicular line segment from the ending endpoint to a vertical coordinate axis in the preset coordinate system.

[0180] The second determining unit is configured to construct, in the preset coordinate system, a closed figure formed by the first perpendicular line segment, the second perpendicular line segment, the horizontal coordinate axis and the audit performance change curve.

[0181] In some optional embodiments, the audit performance obtaining module 502 comprises:

[0182] The feature extraction sub-module is configured to extract at least one audit feature corresponding to each historical audit document, and the audit feature comprises a field integrity audit feature, a format compliance audit feature and a cross-field logic consistency audit feature.

[0183] The feature training sub-module is configured to input the at least one audit feature into a target audit model for prediction to obtain the audit performance value of the plurality of historical audit documents of the target applicant, wherein the target audit model is a pre-trained neural network model.

[0184] In some optional embodiments, the audit order determining module 504 comprises:

[0185] The first determining sub-module is configured to sequentially sort the audit performance values of the plurality of historical audit documents of the target applicant under the target audit regulation in descending order to obtain a historical audit order of the plurality of historical audit documents under the target audit regulation.

[0186] The second determining sub-module is configured to determine the to-be-audited order of the to-be-audited document according to the historical audit order of the plurality of historical audit documents under the target audit regulation.

[0187] In some optional embodiments, the audit document obtaining module 501 comprises:

[0188] The identifier receiving sub-module is configured to receive a request instruction package of the to-be-audited document initiated by the target applicant, and the request instruction package loads the identifier information of the target applicant.

[0189] The document obtaining sub-module is configured to obtain the plurality of historical audit documents of the to-be-audited document according to the identifier information of the target applicant.

[0190] Further function descriptions of the above various modules and units are the same as those of the above corresponding embodiments, and will not be described herein.

[0191] The document auditing sequence determination apparatus in the embodiment is in the form of functional units, and the units refer to ASIC (Application Specific Integrated Circuit) circuits, processors and memories for executing one or more software or fixed programs, and / or other devices capable of providing the above functions.

[0192] The embodiment of the present application also provides a computer device having the above document auditing sequence determination apparatus.

[0193] Please refer to Figure 6 , Figure 6 is a structural schematic diagram of a computer device provided by an optional embodiment of the present application, as Figure 6 shown, the computer device comprises one or more processors 10, a memory 20, and an interface for connecting components, including a high-speed interface and a low-speed interface. The components are communicatively connected with each other by different buses, and can be installed on a common mainboard or in other manners as needed. The processor can process instructions executed in the computer device, including instructions stored in the memory or on the memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, multiple processors and / or buses can be used with multiple memories and multiple memory, if needed. Similarly, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 6 The processor 10 is taken as an example in the embodiment.

[0194] The processor 10 can be a central processor, a network processor or a combination thereof. The processor 10 can further include a hardware chip. The hardware chip can be an application specific integrated circuit, a programmable logic device or a combination thereof. The programmable logic device can be a complex programmable logic device, a field programmable logic gate array, a generic array logic or any combination thereof.

[0195] The memory 20 stores instructions executable by the at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0196] The memory 20 can include a program storage area and a data storage area. The program storage area can store an operating system, application programs required for at least one function, etc. The data storage area can store data created by the computer device, etc. In addition, the memory 20 can include a high-speed random access memory, and can also include a non-transitory memory such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid state memory device. In some alternative embodiments, the memory 20 can optionally include memory that is remotely located with respect to the processor 10, and which can be connected to the computer device through a network. Examples of such networks include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communications network, and combinations thereof.

[0197] The memory 20 can include a volatile memory, such as a random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid state memory device. The memory 20 can also include an array of multi-state flash memory cells, which can be used to store data and / or instructions to implement the embodiments described herein.

[0198] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0199] The embodiments of the present application also provide a computer readable storage medium. The above-mentioned method according to the embodiments of the present application can be implemented in hardware, firmware, or recorded in a storage medium, or implemented as computer code to be originally stored in a remote storage medium or a non-transitory machine readable storage medium downloaded through a network and stored in a local storage medium, so that the method described herein can be processed by such software on a storage medium using a general purpose computer, a special purpose processor, or programmable or special purpose hardware. The storage medium can be a magnetic disk, an optical disk, a read-only memory, a random access memory, a flash memory, a hard disk, or a solid state disk, etc. Further, the storage medium can also include a combination of the above-mentioned kinds of storage. It can be understood that the computer, processor, microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code, when the software or computer code is accessed and executed by the computer, processor, or hardware, the method shown in the above embodiments is implemented.

[0200] A portion of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the form in which the computer program instruction exists in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc. Accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium that can be accessed by the computer.

[0201] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for determining the order of document review, characterized in that: The method comprises: Obtaining a target applicant's pending audit documents under a target audit provision and multiple historical audit documents of the pending audit documents, wherein the target audit provision is each audit provision of the multiple audit provisions; Obtain the audit performance values ​​of the target applicant for multiple historical audit documents under the target audit regulations; Determining, based on the audit performance values ​​of the multiple historical audit documents at multiple coordinate points in a preset coordinate system, audit difference values ​​of the multiple historical audit documents of the target applicant under the target audit regulations, wherein the abscissa value of each coordinate point represents the historical audit order of each historical audit document, and the ordinate value of each coordinate point represents the audit performance value of each historical audit document, and the historical audit order of each historical audit document is determined by the audit performance value of the historical audit document; The order of the documents to be reviewed under the target audit regulations is determined according to the audit difference values ​​of multiple historical audit documents of the target applicant under the target audit regulations.

2. The method according to claim 1, characterized in that Based on the audit performance values ​​of the multiple historical audit documents at multiple coordinate points in a preset coordinate system, determining the audit difference values ​​of the multiple historical audit documents of the target applicant under the target audit regulations, including: Mark the number of the plurality of coordinate points in the preset coordinate system as n; Determine n-1 audit performance change line segments from n coordinate points in a preset coordinate system, where each audit performance change line segment is formed by connecting two adjacent coordinate points; Constructing an audit performance change curve through n-1 audit performance change line segments, and determining a starting endpoint and an ending endpoint from the audit performance change curve; Constructing a closed graph in the preset coordinate system according to the audit performance change curve, the starting endpoint and the ending endpoint; Calculating the total area of ​​the closed figure; Calculating the comprehensive change slope of n-1 audit performance change line segments, wherein the comprehensive change slope is calculated by summing the target slope of each audit performance change line segment; According to the comprehensive change slope and the total area of ​​the closed figure, the audit difference values ​​of multiple historical audit documents of the target applicant under the target audit regulations are calculated.

3. The method according to claim 2, characterized in that Constructing a closed graph in the preset coordinate system according to the audit performance change curve, the starting endpoint, and the ending endpoint includes: Draw a first perpendicular line segment from the starting end point to the horizontal coordinate axis of the preset coordinate system, and draw a second perpendicular line segment from the ending end point to the vertical coordinate axis of the preset coordinate system; In the preset coordinate system, a closed figure formed by the first vertical line segment, the second vertical line segment, the horizontal coordinate axis and the audit performance change curve is constructed.

4. The method according to claim 1, wherein Obtain the audit performance values ​​of multiple historical audit documents of the target applicant under the target audit regulations, including: Extract at least one audit feature corresponding to each historical audit document, wherein the audit features include: field integrity audit feature, format compliance audit feature, and cross-field logical consistency audit feature; The at least one audit feature is input into the target audit model for prediction to obtain the audit performance values ​​of multiple historical audit documents of the target applicant, wherein the target audit model is a pre-trained neural network model.

5. The method according to claim 1, wherein Determining the order of the documents to be reviewed under the target audit regulations according to the audit difference values ​​of multiple historical audit documents of the target applicant under the target audit regulations includes: Sort the audit difference values ​​of multiple historical audit documents of the target applicant under the target audit regulations in descending order to obtain the historical audit order of the multiple historical audit documents under the target audit regulations; The order of the documents to be reviewed is determined according to the order of the historical reviews of the multiple historical reviews under the target review regulations.

6. The method according to claim 1, characterized in that Obtain the target applicant's pending documents under the target audit regulations and multiple historical audit documents of the pending documents, including: Receive a request instruction package for documents to be reviewed initiated by a target applicant, where the request instruction package contains identification information of the target applicant; Based on the identification information of the target applicant, multiple historical review documents of the document to be reviewed are obtained.

7. A document review order determination device, characterized in that: The device comprises: An audit document acquisition module is used to acquire a target applicant's pending audit document under a target audit provision and multiple historical audit documents of the pending audit document, wherein the target audit provision is each audit provision of multiple audit provisions; The audit performance acquisition module is used to obtain the audit performance values ​​of the target applicant for multiple historical audit documents under the target audit regulations; an audit difference determination module, configured to determine audit difference values ​​of multiple historical audit documents of a target applicant under the target audit regulations based on multiple coordinate points of the audit performance values ​​of the multiple historical audit documents in a preset coordinate system, wherein the abscissa value of each coordinate point represents the historical audit order of each historical audit document, and the ordinate value of each coordinate point represents the audit performance value of each historical audit document, and the historical audit order of each historical audit document is determined by the audit performance value of the historical audit document; The review order determination module is used to determine the review order of the documents to be reviewed under the target review regulations based on the review difference values ​​of multiple historical review documents of the target applicant under the target review regulations.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the document review order determination method according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the document review order determination method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises computer instructions for causing a computer to execute the method for determining the order of document review according to any one of claims 1 to 6.

Images