Economic responsibility auditing portrait method and system for expressway enterprises

Abstract

The invention relates to the field of audit management, and discloses an economic responsibility audit portrait method and system for a highway enterprise, and the method comprises the steps: building a multi-layer label model of the highway enterprise, and the multi-layer label model comprises a target layer label, a criterion layer label and an index layer label; wherein the target layer label comprises an economic responsibility auditing label, the criterion layer label is a lower layer refining label of the target layer label, and the index layer label is a lower layer refining label of the criterion layer label; determining a label index weight corresponding to each label in the multi-layer label model, determining a score value corresponding to each label according to an internal assessment management method and external audit data of the expressway enterprise, and determining a product of the score value corresponding to each label and the label index weight as a score of each label, and the score of each label is displayed. According to the method, the auditing portrait can be efficiently constructed for the expressway enterprise, and the auditing effect of the expressway enterprise is improved.

Description

Technical Field

[0001] The present application relates to the technical field of audit management, and more specifically, to an economic responsibility audit portrait method and system for highway enterprises. Background Art

[0002] User portrait technology is a method of creating virtual representatives of user attributes and behaviors by collecting and analyzing user information. The core of user portrait technology is to simplify complex user data into tags for easy understanding and operation. Currently, there is an audit portrait technology similar to user portrait technology to improve audit efficiency. By drawing on the theory and methods of user portrait, audit portrait identifies and selects characteristic indicators for describing the audit object from a large amount of data, quantifies them with data, and visually displays the audit object information. Both user portrait and audit portrait are means of feature extraction and description of specific objects. Audit portrait is mainly used to assist audit work, including assisting in audit project establishment, supporting pre-audit investigation, determining audit focuses, optimizing audit resource allocation, etc., so as to achieve the comprehensiveness, accuracy and efficiency of audit, and discover potential risks and problems.

[0003] Currently, there is no efficient method for establishing an audit portrait of highway enterprises. In the data system of highway enterprises, cross-system data mainly has the following characteristics: (1) Hybrid data: including structured, unstructured and semi-structured data. During the operation and related business activities of highways, a lot of data generated by enterprises, such as project progress records, highway operation toll information, highway maintenance information, etc., exist in unstructured and semi-structured forms such as text, pictures, logs, videos, and audios. (2) Data dispersion: For current highway enterprises, business systems, official websites, etc. are all important sources of enterprise big data, and these data are scattered in different systems or platforms. (3) Huge data volume: The data volume on each business system is very large, and the data capacity in the database storing measurement-type data increases rapidly. Therefore, it is currently difficult to efficiently construct an audit portrait for highway enterprises. Summary of the Invention

[0004] The purpose of the present application is to provide an economic responsibility audit portrait method and system for highway enterprises, which solves the technical problem of difficultly and efficiently constructing an audit portrait for highway enterprises, and achieves the technical effect of efficiently constructing an audit portrait for highway enterprises.

[0005] An economic responsibility audit portrait method for highway enterprises provided by an embodiment of the present application, the method includes: establishing a multi-layer label model for highway enterprises, the multi-layer label model includes target layer labels, criterion layer labels, and index layer labels; wherein, the target layer labels include economic responsibility audit labels, the criterion layer labels are the lower-level refined labels of the target layer labels, and the index layer labels are the lower-level refined labels of the criterion layer labels; determining the label index weights corresponding to each label in the multi-layer label model, and determining the scoring values corresponding to each label according to the internal assessment management method of the highway enterprise and external audit data, and determining the product of the scoring value corresponding to each label and the label index weight as the score of each label, and displaying the scores of each item of label to obtain the economic responsibility audit portrait of the highway enterprise.

[0006] In a possible implementation manner, the method further includes: determining index layer abnormal labels through an index layer abnormal identification unit; determining the criterion layer labels including the index layer abnormal labels, and determining the target layer labels including the criterion layer labels; when the score corresponding to the criterion layer abnormal label is less than or equal to the preset criterion layer abnormal label score, determining the criterion layer label as the criterion layer abnormal label; when the score corresponding to the target layer abnormal label is less than or equal to the preset target layer abnormal label score, determining the target layer label as the target layer abnormal label; sending out audit clue prompt information corresponding to the index layer abnormal label, the criterion layer abnormal label, and the target layer abnormal label.

[0007] In another possible implementation manner, the method further includes: when there are multiple index layer abnormal labels under the same criterion layer abnormal label, obtaining the external condition influence factors corresponding to the multiple index layer abnormal labels respectively; wherein, the external condition influence factor is used to characterize the degree of abnormal contribution of the external condition to the index layer label, and the external condition influence factor includes a construction period stage factor; determining the target index layer abnormal label with the largest external condition influence factor corresponding to the multiple index layer abnormal labels respectively, and sending out audit clue prompt information corresponding to the target index layer abnormal label, the criterion layer abnormal label, and the target layer abnormal label.

[0008] In another possible implementation, the method further includes: when there are multiple abnormal labels at the index layer under multiple abnormal labels at the criterion layer, obtaining the causal influence relationship between the multiple abnormal labels at the index layer and the causal influence factor corresponding to the causal influence relationship; when there are multiple abnormal labels at the criterion layer under the same abnormal label at the target layer, obtaining the causal influence relationship between the multiple abnormal labels at the criterion layer and the causal influence factor corresponding to the causal influence relationship; according to the causal influence relationship between the multiple abnormal labels at the index layer, determining the root abnormal label at the index layer among the multiple abnormal labels at the index layer according to the influence relationship corresponding to the maximum causal influence factor, determining the upper-level criterion layer abnormal label corresponding to the root abnormal label at the index layer, and sending the audit clue prompt information corresponding to the root abnormal label at the index layer, the upper-level criterion layer abnormal label, and the abnormal label at the target layer; according to the causal influence relationship between the multiple abnormal labels at the criterion layer, determining the root criterion layer abnormal label among the multiple abnormal labels at the criterion layer according to the influence relationship corresponding to the maximum causal influence factor, and sending the audit clue prompt information corresponding to the root criterion layer abnormal label and the abnormal label at the target layer.

[0009] In another possible implementation, the method further includes: obtaining the historical audit data of the root abnormal label at the index layer and determining the causal influence relationship confirmation rate corresponding to the historical audit data of the root abnormal label at the index layer, where the causal influence relationship confirmation rate represents the audit confirmation ratio of the causal influence relationship of the historical audit data on the root abnormal label at the index layer; when the causal influence relationship confirmation rate corresponding to the root abnormal label at the index layer is less than the preset causal influence relationship confirmation rate, multiplying the causal influence factor corresponding to the causal influence relationship of the root abnormal label at the index layer by a preset causal influence adjustment factor to adjust the causal influence relationship of the root abnormal label at the index layer.

[0010] In another possible implementation, the method further includes: when there are multiple anomaly labels at the metric layer under multiple anomaly labels at the criterion layer, obtaining the co-influence relationship between the multiple anomaly labels at the metric layer and the co-influence factor corresponding to the co-influence relationship; when there are multiple anomaly labels at the criterion layer under the same anomaly label at the target layer, obtaining the co-influence relationship between the multiple anomaly labels at the criterion layer and the co-influence factor corresponding to the co-influence relationship; according to the co-influence relationship between the multiple anomaly labels at the metric layer, sorting according to the number of co-influence relationships between the multiple anomaly labels at the metric layer, determining the multi-factor anomaly label at the metric layer affected by the most anomaly labels at the metric layer among the multiple anomaly labels at the metric layer, and determining the upper-level criterion layer anomaly label corresponding to the multi-factor anomaly label at the metric layer, and sending out the audit trail prompt information corresponding to the multi-factor anomaly label at the metric layer, the upper-level criterion layer anomaly label, and the target layer anomaly label; according to the co-influence relationship between the multiple anomaly labels at the criterion layer, sorting according to the number of co-influence relationships between the multiple anomaly labels at the metric layer, determining the multi-factor criterion layer anomaly label affected by the most anomaly labels at the criterion layer among the multiple anomaly labels at the criterion layer, and sending out the audit trail prompt information corresponding to the multi-factor criterion layer anomaly label and the target layer anomaly label.

[0011] In another possible implementation, the method further includes: obtaining the historical audit data of the multi-factor anomaly label at the metric layer, and determining the co-influence relationship confirmation rate corresponding to the historical audit data of the multi-factor anomaly label at the metric layer, where the co-influence relationship confirmation rate represents the audit confirmation ratio of the historical audit data for each co-influence relationship of the multi-factor anomaly label at the metric layer; when the first co-influence relationship confirmation rate corresponding to the multi-factor anomaly label at the metric layer is less than the preset first co-influence relationship confirmation rate, multiplying the first co-influence relationship factor corresponding to the first co-influence relationship of the multi-factor anomaly label at the metric layer by the preset co-influence adjustment factor to adjust the co-influence relationship of the multi-factor anomaly label at the metric layer.

[0012] In another possible implementation, determining the anomaly label at the metric layer through the metric layer anomaly recognition unit includes: through the metric layer anomaly recognition unit, clustering the multiple historical label data corresponding to the first label under the first metric layer to obtain multiple historical label data categories and multiple historical label data cluster centers corresponding to the multiple historical label data categories respectively; obtaining the multiple first label data updated by the first label under the first metric layer, respectively determining the multiple Euclidean distances between the multiple first label data and the multiple historical label data cluster centers, and determining the variance value of the multiple Euclidean distances. When the variance value of the multiple Euclidean distances is greater than the preset Euclidean distance variance value, taking the first label as the anomaly label at the metric layer.

[0013] In another possible implementation, to determine the anomaly labels at the metric layer through the metric layer anomaly recognition unit, the method further includes: obtaining a first time window corresponding to multiple historical label data under a historical label data category, and obtaining a first external condition influence factor corresponding to the first time window; when the first external condition influence factor is greater than or equal to a preset external condition influence factor, through the metric layer anomaly recognition unit, obtaining multiple historical label data corresponding to the first label under the first metric layer in a second time window following the first time window, clustering the multiple historical label data corresponding to the first label under the first metric layer in the second time window, and updating the multiple historical label data categories and the clustering centers of the multiple historical label data corresponding to the multiple historical label data categories respectively.

[0014] The embodiment of the present application further provides an economic responsibility audit portrait system for highway enterprises, including a unit for executing the method described in any one of the above.

[0015] The beneficial effects of the embodiment of the present application compared with the prior art are as follows:

[0016] The embodiment of the present application provides an economic responsibility audit portrait method for highway enterprises. The method includes: establishing a multi-layer label model for highway enterprises, where the multi-layer label model includes target layer labels, criterion layer labels, and metric layer labels; among them, the target layer labels include economic responsibility audit labels, the criterion layer labels are the lower-level refined labels of the target layer labels, and the metric layer labels are the lower-level refined labels of the criterion layer labels; determining the label index weights corresponding to each label in the multi-layer label model, determining the scoring value corresponding to each label according to the internal assessment management method and external audit data of highway enterprises, and determining the product of the scoring value corresponding to each label and the label index weight as the score of each label, and displaying the scores of each label to obtain the economic responsibility audit portrait of highway enterprises. The embodiment of the present application utilizes the internal and external data of highway enterprises, combines data mining technology, and constructs indicators covering across systems and specialties, which can efficiently construct an audit portrait for highway enterprises and improve the audit effect of highway enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 It is a schematic flowchart of the first economic responsibility audit portrait method for highway enterprises provided by the embodiment of the present application;

[0019] Figure 2 It is a schematic diagram of the first economic responsibility audit portrait for highway enterprises provided by the embodiments of this application;

[0020] Figure 3 It is a schematic diagram of the first economic responsibility audit portrait display system for highway enterprises provided by the embodiments of this application;

[0021] Figure 4 It is a schematic flowchart of the second economic responsibility audit portrait method for highway enterprises provided by the embodiments of this application;

[0022] Figure 5 It is a schematic flowchart of the third economic responsibility audit portrait method for highway enterprises provided by the embodiments of this application;

[0023] Figure 6 It is a schematic flowchart of the fourth economic responsibility audit portrait method for highway enterprises provided by the embodiments of this application;

[0024] Figure 7 It is a schematic flowchart of the fifth economic responsibility audit portrait method for highway enterprises provided by the embodiments of this application;

[0025] Figure 8 It is a schematic diagram of the logical structure of an economic responsibility audit portrait system for highway enterprises provided by the embodiments of this application. Detailed implementation manners

[0026] It should be understood that when used in the specification and appended claims of this application, the term "comprising" indicates the presence of the described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their combinations.

[0027] It should also be understood that the term "and / or" as used in the specification and appended claims of this application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0028] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when...", "once", "in response to determining" or "in response to detecting" according to the context. Similarly, the phrase "if determined" or "if [the described condition or event] is detected" can be interpreted as meaning "once determined", "in response to determining", "once [the described condition or event] is detected" or "in response to detecting [the described condition or event]" according to the context.

[0029] In addition, in the description of the specification and the appended claims of the present application, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be construed as indicating or implying relative importance.

[0030] The reference to "one embodiment" or "some embodiments" etc. described in the specification of the present application means that in one or more embodiments of the present application, the specific features, structures or characteristics described in connection with that embodiment are included. Thus, the statements "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments", etc. that appear in different places in this specification do not necessarily all refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "comprising", "including", "having" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0031] In the expressway enterprise data system, the cross-system data mainly has the characteristics of mixed data, data dispersion and large data volume. The data volume on each business system is very large, and the data capacity in the database storing measurement data increases rapidly. Therefore, it is currently difficult to efficiently construct an audit portrait for expressway enterprises.

[0032] For the above reasons, the embodiments of the present application provide an economic responsibility audit portrait method for expressway enterprises. This method includes: establishing a multi-layer label model for expressway enterprises, where the multi-layer label model includes target layer labels, criterion layer labels and index layer labels; among them, the target layer labels include economic responsibility audit labels, the criterion layer labels are the lower-level refined labels of the target layer labels, and the index layer labels are the lower-level refined labels of the criterion layer labels; determining the label index weights corresponding to each label in the multi-layer label model, and determining the scoring values corresponding to each label according to the internal assessment management method and external audit data of the expressway enterprise, and determining the product of the scoring value corresponding to each label and the label index weight as the score of each label, and displaying the scores of each item of label to obtain the economic responsibility audit portrait of the expressway enterprise. The embodiments of the present application utilize the internal and external data of expressway enterprises, combined with data mining technology, to construct a cross-system and cross-professional covering index, which can efficiently construct an audit portrait for expressway enterprises and improve the audit effect of expressway enterprises.

[0033] In some scenarios, an economic responsibility audit portrait method for expressway enterprises in the embodiments of the present application can be applied to the construction of the audit portrait of expressway enterprises, which can improve the construction efficiency of the audit portrait of expressway enterprises.

[0034] The following specifically describes a method for drawing an economic responsibility audit portrait for highway enterprises provided by the embodiments of the present application in combination with specific examples.

[0035] Figure 1 FIG. is a schematic flowchart of the first method for drawing an economic responsibility audit portrait for highway enterprises provided by the embodiments of the present application. As Figure 1 shown, the above method includes S110 to S120, and the following specifically describes S110 to S120.

[0036] S110. Establish a multi-layer label model for highway enterprises. The multi-layer label model includes target layer labels, criterion layer labels, and index layer labels. Among them, the target layer labels include economic responsibility audit labels, the criterion layer labels are the lower-level refinement labels of the target layer labels, and the index layer labels are the lower-level refinement labels of the criterion layer labels.

[0037] In the embodiments of the present application, a multi-layer label model for highway enterprises can be established first. By constructing a hierarchical label system, a structured expression of the economic responsibility audit dimension can be realized. The multi-layer label model includes target layer labels, criterion layer labels, and index layer labels. The multi-layer label model adopts a three-level architecture of "target layer - criterion layer - index layer" and can refine audit evaluation elements layer by layer.

[0038] When constructing the target layer labels, criterion layer labels, and index layer labels, the target layer labels can include economic responsibility audit labels, the criterion layer labels can be the lower-level refinement labels of the target layer labels, and the index layer labels can be the lower-level refinement labels of the criterion layer labels. When constructing the hierarchical relationship of the target layer labels, criterion layer labels, and index layer labels, the semantic relationship between the labels can be defined, and an association map of "fund flow - business process - responsible entity" can be established.

[0039] It should be noted that the design of the target layer labels can take "economic responsibility audit" as the core goal and be divided into four major dimensions: economic decision-making responsibility, economic execution responsibility, economic management responsibility, and economic supervision responsibility. For example, the economic decision-making responsibility focuses on the compliance of major investment and financing projects and covers core directions such as project decision-making processes and debt risk assessments. The economic execution responsibility focuses on the effectiveness of budget execution and includes labels in key areas such as the in-place rate of maintenance funds and cost control of reconstruction and expansion projects.

[0040] Exemplarily, Figure 2 FIG. is a schematic diagram of the first economic responsibility audit portrait for highway enterprises provided by the embodiments of the present application. As Figure 2 shown, the portrait related to economic responsibility audit can include target layer labels and criterion layer labels in multiple aspects. The criterion layer labels can also include multiple index layer labels ( Figure 2 not shown in the figure).

[0041] As Figure 2 shown, the target layer labels may include strategic development and corporate governance labels, the criterion layer labels may include implementing the guidelines and policies of superiors and decision-making arrangements, and the indicator layer labels may include labels on compliance with relevant laws and regulations, implementing the guidelines and policies of the Party, the state, and superior management agencies regarding economic work, decision-making arrangements, and implementation effects.

[0042] It should be noted that the construction of the criterion layer labels can be carried out by decomposing the target layer through expert evaluation to form multiple secondary labels. Specifically, the criterion layer may include criterion layer labels such as toll operation management (ETC fund collection timeliness rate, success rate of toll evasion recovery), asset preservation and appreciation (road property insurance coverage rate, deviation degree of franchise value assessment), debt risk control (interest-bearing debt ratio, accuracy rate of identifying hidden debts), and project compliance management (tender process compliance rate, integrity of project change approval).

[0043] It should be noted that the extraction of the indicator layer labels can be based on knowledge graph technology to extract atomic-level indicators from the enterprise's ERP system, project management system, and audit problem database. Taking debt risk control as an example, the indicator layer label 1 can be the financing cost deviation degree = (actual financing interest rate - industry benchmark interest rate) / industry benchmark interest rate × 100%, the indicator layer label 2 can be the debt maturity mismatch rate = short-term debt balance / long-term debt due within one year × 100%, and the indicator layer label 3 can be the debt warning value = amount of external guarantees / net assets

[0044] × 100%.

[0045] S120. Determine the label index weights corresponding to each label in the multi-layer label model, determine the scoring value corresponding to each label according to the internal assessment management method of the highway enterprise and external audit data, determine the product of the scoring value corresponding to each label and the label index weight as the score of each label, and display the scores of each label to obtain the economic responsibility audit portrait of the highway enterprise.

[0046] In order to construct the economic responsibility audit portrait, the mapping of audit evaluation can be further quantified through a model, which specifically includes three major modules: weight assignment, data fusion, and visual presentation.

[0047] When assigning weights, the label index weights corresponding to each label in the multi-layer label model can be determined, and specifically, the expert subjective weights can be obtained through the AHP (Analytic Hierarchy Process).

[0048] When determining the scoring value corresponding to each tag based on the internal assessment management method of highway enterprises and external audit data, various types of data such as budget execution rate and cost accounting deviation can be obtained in real time by connecting internal data to the financial shared center and the internal assessment management method, and the KNN imputation method is used to handle missing values; for external data, the audit report can be parsed through NLP technology to construct a triple of "violation problem - responsible entity - capital impact", and the LSTM model is applied to quantify the severity level of the problem.

[0049] After obtaining the scoring value corresponding to each tag and the tag index weight, the product of the scoring value corresponding to each tag and the tag index weight can be determined as the score of each tag, and the score of each tag is the final tag value.

[0050] Figure 3 This is a schematic diagram of the first economic responsibility audit portrait display system for highway enterprises provided by the embodiments of the present application. As Figure 3 shown, after obtaining the score of each tag as the final tag value, the economic responsibility audit portrait of the highway enterprise can be obtained by displaying the scores of each tag. Specifically, a three-dimensional audit portrait platform can be developed to achieve the display of the economic responsibility audit of the highway enterprise.

[0051] As Figure 3 shown, the data mining in this economic responsibility audit portrait display system can be used in combination with technologies such as data collection, tag design, logical inference, and visualization. Through different business scenarios, a comprehensive and three-dimensional portrait display of the audit object or business description can be formed.

[0052] The beneficial effect of the above implementation method is that by using the internal and external data of highway enterprises and combining data mining technology, a cross-system and cross-professional covering index can be constructed, which can efficiently construct an audit portrait of highway enterprises and improve the audit effect of highway enterprises.

[0053] Figure 4 This is a schematic flowchart of the second economic responsibility audit portrait method for highway enterprises provided by the embodiments of the present application. As Figure 4 shown, the above method further includes S210 to S220, and the following is a specific description of S210 to S220.

[0054] S210: Determine the abnormal tags at the index layer through the index layer anomaly recognition unit. Determine the criterion layer tags including the abnormal tags at the index layer, and determine the target layer tags including the criterion layer tags.

[0055] In an embodiment of the present application, the indicator layer corresponds to specific audit indicators (such as accounts receivable turnover rate, debt-to-asset ratio, etc.), the criterion layer corresponds to industry compliance criteria (such as Financial Auditing Standard No. X), and the target layer corresponds to corporate strategic goals (such as annual revenue target achievement rate), thereby realizing the construction of a hierarchical labeling system. In order to identify potential risks from multiple levels and dimensions, the embodiment of the present application can first identify abnormal labels at the indicator layer, and provide feedback on abnormal labels at the criterion layer and abnormal labels at the target layer corresponding to the abnormal labels at the indicator layer, thereby realizing comprehensive feedback at each layer on the abnormal labels at the indicator layer, the abnormal labels at the criterion layer, and the abnormal labels at the target layer, thereby realizing a layer-by-layer mapping from micro indicator abnormalities to macro strategic risks, and providing audit clues for the audit process.

[0056] Exemplarily, the method for determining abnormal labels of the indicator layer may detect the degree to which the scores of the indicator layer labels deviate from the normal interval based on the isolation forest algorithm. If the degree to which the scores of the indicator layer labels deviate from the normal interval exceeds a preset threshold, the indicator layer labels are determined to be abnormal labels.

[0057] S220. When the score corresponding to the criterion layer abnormal label is less than or equal to the preset criterion layer abnormal label score, the criterion layer label is determined to be the criterion layer abnormal label; when the score corresponding to the target layer abnormal label is less than or equal to the preset target layer abnormal label score, the target layer label is determined to be the target layer abnormal label; and audit clue prompt information corresponding to the indicator layer abnormal label, the criterion layer abnormal label and the target layer abnormal label is issued.

[0058] After obtaining the indicator layer abnormal label, it can be determined whether the criterion layer label including the indicator layer abnormal label is abnormal. When the score corresponding to the criterion layer abnormal label is less than or equal to the preset criterion layer abnormal label score, the criterion layer label can be determined to be the criterion layer abnormal label, and then the criterion layer abnormal label including the indicator layer abnormal label can be determined.

[0059] After obtaining the criterion layer abnormal label, it is possible to determine whether the target layer label including the indicator layer abnormal label is abnormal. When the score corresponding to the target layer abnormal label is less than or equal to the preset target layer abnormal label score, the target layer label can be determined to be the target layer abnormal label, and then the target layer abnormal label including the criterion layer abnormal label can be determined.

[0060] When it is determined that there are indicator layer abnormal labels, criterion layer abnormal labels and target layer abnormal labels, audit clue prompt information corresponding to the indicator layer abnormal labels, criterion layer abnormal labels and target layer abnormal labels can be issued, thereby achieving comprehensive feedback of the audit clue prompt information corresponding to the indicator layer abnormal labels, criterion layer abnormal labels and target layer abnormal labels at each level, and improving the feedback effect of the audit clue prompt information.

[0061] In an embodiment of the present application, the generation of audit clue prompt information can adopt natural language generation technology, build an audit prompt template based on a text generation model, and automatically generate a structured report with abnormal labels, related evidence chains (abnormal indicator trend chart, original text of standard clauses) and impact analysis.

[0062] Exemplarily, audit clue prompt information can be pushed through multiple channels, including red warning pop-up windows on the audit workbench and to-do reminders in the enterprise OA system.

[0063] The beneficial effect of the above-mentioned implementation method is that by constructing a hierarchical label system, potential risks can be identified from multiple levels and dimensions, and the indicator layer abnormal labels can be first identified, and the criterion layer abnormal labels and target layer abnormal labels corresponding to the indicator layer abnormal labels can be identified and feedback can be provided, thereby achieving comprehensive feedback of the audit clue prompt information of the abnormal indicator layer abnormal labels, criterion layer abnormal labels and target layer abnormal labels at each level, thereby improving the prompt effect of the audit clue prompt information and being able to improve the audit effect.

[0064] Figure 5 A flow chart of a third method for economic responsibility audit profiling of highway enterprises provided in the embodiment of the present application is as follows: Figure 5 As shown, the above method also includes S310 to S320, and S310 to S320 are described in detail below.

[0065] S310: When there are multiple indicator layer abnormal labels under the same criterion layer abnormal label, obtain the external condition impact factors corresponding to the multiple indicator layer abnormal labels. The external condition impact factor is used to characterize the degree of contribution of the external condition to the abnormality of the indicator layer label, and the external condition impact factor includes the construction period stage factor.

[0066] In the process of economic responsibility audit of highway enterprises, multiple indicator-level abnormal labels may be associated with the same criterion-level abnormal label. For example, under the criterion-level label of "cost control abnormality", there may be multiple indicator-level abnormal labels such as "material procurement cost overrun", "labor cost overrun", "machinery use efficiency low", etc. In order to accurately distinguish the actual responsibility of each abnormal indicator, the embodiment of the present application introduces a quantitative analysis method of external condition influencing factors.

[0067] Specifically, quantitative evaluation parameters can be established through multidimensional data analysis as external condition influencing factors. The core function of external condition influencing factors is to objectively measure the abnormal contribution of the external environment to each audit indicator.

[0068] Exemplarily, the external condition impact factor may include the construction period stage factor. In the construction of expressways, there are significant differences in the impacts of different construction stages (such as subgrade construction period, pavement laying period, and traffic safety facility installation period) on various cost indicators: Conducting subgrade construction during the rainy season will significantly increase the drainage cost, while laying asphalt in winter may lead to an abnormal increase in material heating costs. By establishing a correlation model between the construction period stage and cost indicators, the precise proportion of each abnormal indicator affected by the construction period can be calculated.

[0069] Exemplarily, when determining the construction period stage factor, the construction period progress data of the enterprise project management system, the cost fluctuation database of each stage in historical projects, and environmental monitoring data such as meteorology and geology can be integrated, and an impact factor prediction model can be trained through machine learning algorithms (such as random forest regression), and finally, the construction period stage factor corresponding to each abnormal label of the indicator layer can be output.

[0070] S320. Determine the target indicator layer abnormal label corresponding to the largest external condition impact factor among the abnormal labels of multiple indicator layers, and send out the audit clue prompt information corresponding to the target indicator layer abnormal label, the criterion layer abnormal label, and the target layer abnormal label.

[0071] In the embodiment of the present application, after obtaining the external condition impact factors of the abnormal labels of each indicator layer, the audit portrait system can automatically execute the impact degree sorting algorithm. The impact degree sorting algorithm first performs normalization processing on the impact factors of all indicators under the same criterion layer to eliminate the influence of different dimensions; then it can determine the target indicator layer abnormal label with the largest external condition impact factor, and finally identify the target indicator layer abnormal label most significantly affected by the external conditions, so as to be able to conduct audits starting from the label with the greatest external influence degree, improving the audit efficiency.

[0072] After obtaining the target indicator layer abnormal label, the audit clue prompt information corresponding to the target indicator layer abnormal label, the criterion layer abnormal label, and the target layer abnormal label can be sent out to provide audit clue prompts for the audit process.

[0073] Taking the audit of a certain expressway reconstruction and expansion project as an example, when it is found that the construction period stage factor of the "material cost" target layer abnormal label reaches 0.82 (range 0 - 1), which is significantly higher than 0.35 of the "labor cost" target layer abnormal label in the same period, the system will automatically trigger a three - level linkage audit clue prompt: 1) The target layer prompts the corresponding target layer abnormal label "abnormal operation and management responsibility"; 2) The criterion layer locates the "improper cost control" criterion layer abnormal label corresponding to the target indicator layer abnormal label; 3) The indicator layer highlights the target indicator layer abnormal label "material procurement affected by construction during the flood season reaches 82%". This hierarchical prompt mechanism not only ensures the accurate positioning of the audit focus but also retains the complete logical chain of audit economic responsibility determination.

[0074] The beneficial effects of the above implementation method are that after obtaining the external condition influence factors of the abnormal labels at each index layer, the audit portrait system can determine the target index layer abnormal label with the largest external condition influence factor, and give feedback on the target index layer abnormal label, so as to be able to audit from the abnormal label of the index layer with the greatest external influence degree, improving the audit efficiency.

[0075] Another beneficial effect of the above implementation method is that a dynamic correlation model between external conditions and audit indicators is established, which solves the technical bottleneck that it is difficult to quantify the influence of environmental factors on specific labels in traditional audits. In the audit of highway projects, the accuracy of identifying abnormal responsibilities has been greatly improved.

[0076] Figure 6 FIG. is a schematic flowchart of a fourth economic responsibility audit portrait method for highway enterprises provided by an embodiment of the present application. As Figure 6 shown, the above method further includes S410 to S420, which will be specifically described below.

[0077] S410. When there are multiple index layer abnormal labels under multiple criterion layer abnormal labels, obtain the causal influence relationship between the multiple index layer abnormal labels and the causal influence relationship factor corresponding to the causal influence relationship. When there are multiple criterion layer abnormal labels under the same target layer abnormal label, obtain the causal influence relationship between the multiple criterion layer abnormal labels and the causal influence relationship factor corresponding to the causal influence relationship.

[0078] In the economic responsibility audit of highway enterprises, abnormal situations often show complex multi-level correlation characteristics. In this implementation method, a three-level causal influence analysis system of "target layer - criterion layer - index layer" can be constructed, and then the accurate traceability of audit problems can be realized by quantifying the causal conduction relationship between abnormal labels at each level. Specifically, when there are multiple index layer abnormal labels under multiple criterion layer abnormal labels, the causal influence relationship between the multiple index layer abnormal labels and the causal influence relationship factor corresponding to the causal influence relationship can be obtained, and then the causal influence relationship between the multiple index layer abnormal labels and the causal influence relationship factor corresponding to the causal influence relationship can be analyzed. When there are multiple criterion layer abnormal labels under the same target layer abnormal label, the causal influence relationship between the multiple criterion layer abnormal labels and the causal influence relationship factor corresponding to the causal influence relationship can be obtained, and then the causal influence relationship between the multiple criterion layer abnormal labels and the causal influence relationship factor corresponding to the causal influence relationship can be analyzed.

[0079] In this implementation, a knowledge graph of causal influence relationships can be established. For abnormal labels at the indicator level (such as "material price difference rate exceeds standard", "machine unit fee is abnormal", etc.), causal influence relationship factors between each other are calculated through causal testing and Bayesian network analysis. For example, when the causal factor of "material price increase" and "subcontracting cost increase" reaches 0.78, it indicates that the former has a strong conduction effect on the latter. At the same time, at the criterion level (such as "cost control", "schedule management", etc.), the structural equation model (SEM) can be used to analyze cross-domain causal influences, such as the potential impact path of "schedule delay" on "cost overrun".

[0080] Exemplarily, this implementation method can integrate project management system logs, material procurement flow, financial accounting vouchers, and supervision daily report data to identify causal influence relationships.

[0081] S420. According to the causal influence relationship between multiple indicator-layer abnormal labels and the influence relationship corresponding to the maximum causal influence relationship factor, the root indicator-layer abnormal label among the multiple indicator-layer abnormal labels is determined, and the upper criterion-layer abnormal label corresponding to the root indicator-layer abnormal label is determined, and the audit clue prompt information corresponding to the root indicator-layer abnormal label, the upper criterion-layer abnormal label and the target-layer abnormal label is issued. According to the causal influence relationship between multiple criterion-layer abnormal labels and the influence relationship corresponding to the maximum causal influence relationship factor, the root criterion-layer abnormal label among the multiple criterion-layer abnormal labels is determined, and the audit clue prompt information corresponding to the root criterion-layer abnormal label and the target-layer abnormal label is issued.

[0082] After obtaining the causal influence relationship among multiple indicator-layer abnormal labels, the system can execute the root cause analysis algorithm according to the causal influence relationship among the multiple indicator-layer abnormal labels and based on the established causal influence relationship network, determine the root indicator-layer abnormal label among the multiple indicator-layer abnormal labels according to the influence relationship corresponding to the maximum causal influence relationship factor, and determine the upper-level criterion-layer abnormal label corresponding to the root indicator-layer abnormal label, and issue audit clue prompt information corresponding to the root indicator-layer abnormal label, the upper-level criterion-layer abnormal label and the target-layer abnormal label.

[0083] In this implementation, when determining the abnormal label of the root indicator layer, we can trace back along the causal chain to the node with the largest causal factor to determine it as the root indicator (such as identifying "fluctuation in asphalt purchase price" as the root cause of "pavement cost exceeding standard"); for criterion layer abnormalities, the centrality index of each node is calculated through the influence propagation model to determine the root criterion layer label (for example, identifying "design change management" as the core cause of "double out of control of progress and cost").

[0084] Exemplarily, when determining the root cause indicator layer anomaly label among the indicator layer anomaly label A1, the indicator layer anomaly label A2, and the indicator layer anomaly label A3 according to the influence relationship corresponding to the maximum causal influence relationship factor, the causal influence relationship factor K11 between the indicator layer anomaly label A1 and the indicator layer anomaly label A2 that it affects can be obtained first. The causal influence relationship factor K11 represents the influence degree of the indicator layer anomaly label A1 on the indicator layer anomaly label A2, and the causal influence relationship factor K23 between the indicator layer anomaly label A2 and the indicator layer anomaly label A3 that it affects can be obtained. The causal influence relationship factor K23 represents the influence degree of the indicator layer anomaly label A2 on the indicator layer anomaly label A3, and the causal influence relationship factor K13 between the indicator layer anomaly label A1 and the indicator layer anomaly label A3 that it affects can be obtained. The causal influence relationship factor K13 represents the influence degree of the indicator layer anomaly label A1 on the indicator layer anomaly label A3.

[0085] After obtaining the causal influence relationship factor K11, the causal influence relationship factor K23, and the causal influence relationship factor K13, the magnitude relationship among the causal influence relationship factor K11, the causal influence relationship factor K23, and the causal influence relationship factor K13 can be determined. When the causal influence relationship factor K23 is the largest causal influence relationship factor, the indicator layer anomaly label A2 can be determined as the root cause indicator layer anomaly label.

[0086] After obtaining the root cause indicator layer anomaly label A2, the upper layer criterion layer anomaly label LA2 including the root cause indicator layer anomaly label A2 can be determined, and the audit trail prompt information corresponding to the root cause indicator layer anomaly label A2, the upper layer criterion layer anomaly label LA2, and the target layer anomaly label TA2 can be issued, where the target layer anomaly label TA2 is the target layer anomaly label including the upper layer criterion layer anomaly label LA2.

[0087] In this implementation manner, similarly, according to the causal influence relationship among multiple criterion layer anomaly labels, the root cause criterion layer anomaly label among the multiple criterion layer anomaly labels can be determined according to the influence relationship corresponding to the maximum causal influence relationship factor, and the audit trail prompt information corresponding to the root cause criterion layer anomaly label and the target layer anomaly label can be issued.

[0088] Exemplarily, the audit trail prompt can be presented in a dynamically expandable manner. The root cause anomaly label (red warning) can be displayed at the first level, the complete causal path diagram (including all causal factor values) can be expanded at the second level, and the specific evidence chain (associated contract terms, meeting minutes, etc.) can be provided at the third level. In particular, for cross-level complex causal relationships (such as the indicator layer "stone shortage" leading to the criterion layer "schedule delay" and then affecting the target layer "economic benefits"), the system will automatically generate a three-dimensional association matrix for visual display.

[0089] The beneficial effects of the above implementation methods are as follows: in terms of audit accuracy, the introduction of the causal impact relationship factor improves the accuracy rate of root cause problem identification; in terms of liability determination, the multi-level causal analysis effectively distinguishes direct liability and transmitted liability, improving the audit efficiency.

[0090] In some implementation methods, according to the causal impact relationships between abnormal labels of multiple indicator layers, and in accordance with the impact relationship corresponding to the maximum causal impact relationship factor, the root cause indicator layer abnormal label among the abnormal labels of multiple indicator layers is determined, including: obtaining the attenuation factor of the causal impact relationship between abnormal labels of multiple indicator layers, determining the product of the causal impact relationship factor and the attenuation factor of the causal impact relationship between abnormal labels of multiple indicator layers, and determining the sum of the products of the causal impact relationship factor and the attenuation factor of the causal impact relationship for each indicator layer abnormal label as the cumulative causal impact relationship factor, and determining the indicator layer abnormal label corresponding to the maximum value of the cumulative causal impact relationship factors corresponding to the abnormal labels of multiple indicator layers as the root cause indicator layer abnormal label.

[0091] In order to accurately determine the root cause indicator layer abnormal label, the direct causal impact relationship and the indirect causal impact relationship between abnormal labels of multiple indicator layers can also be statistically analyzed. Specifically, the attenuation factor of the causal impact relationship between abnormal labels of multiple indicator layers can be obtained, and the product of the causal impact relationship factor and the attenuation factor of the causal impact relationship between abnormal labels of multiple indicator layers can be determined to realize the statistical analysis of the indirect causal impact relationship between abnormal labels of multiple indicator layers.

[0092] Exemplarily, the indirect causal impact relationship between abnormal labels of multiple indicator layers can be a chain conduction indirect causal impact relationship where the abnormal label A4 of the indicator layer affects the abnormal label A5 of the indicator layer, and the abnormal label A5 of the indicator layer affects the abnormal label A6 of the indicator layer. By obtaining the attenuation factors of the causal impact relationships where the abnormal label A4 of the indicator layer affects the abnormal label A5 of the indicator layer and the abnormal label A5 of the indicator layer affects the abnormal label A6 of the indicator layer, the product of the causal impact relationship factor and the attenuation factor of the causal impact relationship between the abnormal label A4 of the indicator layer affecting the abnormal label A5 of the indicator layer and the abnormal label A5 of the indicator layer affecting the abnormal label A6 of the indicator layer can be determined, and the sum of the products of the causal impact relationship factor and the attenuation factor of the causal impact relationship between the abnormal label A4 of the indicator layer affecting the abnormal label A5 of the indicator layer and the abnormal label A5 of the indicator layer affecting the abnormal label A6 of the indicator layer is determined as the cumulative causal impact relationship factor corresponding to the abnormal label A6 of the indicator layer, and the indicator layer abnormal label corresponding to the maximum value of the cumulative causal impact relationship factors corresponding to the abnormal label A4 of the indicator layer affecting the abnormal label A5 of the indicator layer and the abnormal label A5 of the indicator layer respectively is determined as the root cause indicator layer abnormal label, realizing the accurate identification of the root cause indicator layer abnormal label.

[0093] Exemplarily, the attenuation factor of the causal influence relationship between multiple indicator layer abnormal labels can be dynamically determined according to external conditions such as market environment conditions and time conditions, so that the causal influence relationship between multiple indicator layer abnormal labels can adapt to the current external conditions to improve the accuracy of discovering audit clues.

[0094] The beneficial effect of the above implementation method is that adding an attenuation coefficient to the influencing factor of the indirect causal relationship between multiple indicator layer abnormal labels of a long path can improve the accuracy of determining the root indicator layer abnormal label and improve the recognition accuracy of the root indicator layer abnormal label.

[0095] In some implementations, the above method further includes S510 to S520, and S510 to S520 are described in detail below.

[0096] S510. Obtain historical audit data of abnormal labels at the root indicator layer, and determine the causal impact relationship confirmation rate corresponding to the historical audit data of abnormal labels at the root indicator layer. The causal impact relationship confirmation rate represents the audit confirmation ratio of the causal impact relationship of the historical audit data to the abnormal labels at the root indicator layer.

[0097] In an embodiment of the present application, a quantitative indicator of causal influence relationship confirmation rate can be further introduced into the economic responsibility audit of highway enterprises to evaluate the support strength of historical audit data for current root cause anomaly labels.

[0098] When the causal influence relationship confirmation rate is introduced, the audit files of similar projects of the same audit target in the past three years can be automatically retrieved, and a historical audit case knowledge base can be constructed. Specifically, natural language processing technology (such as the BERT-BiLSTM model) can be used to perform structured analysis on historical audit reports to extract the "abnormal label-causal relationship-audit conclusion" triples contained therein. For example, the complete logical chain of "material price difference rate exceeds the standard → supplier qualifications are incomplete (causal factor 0.85) → procurement management failure" can be parsed from a project audit report.

[0099] In this implementation, a similarity matching algorithm based on case-based reasoning (CBR) can also be developed to automatically retrieve audit files of similar projects of the same audit target in the past three years. Specifically, the abnormal labels of the current root indicator layer (such as "abnormal machine unit fees") and historical cases can be compared in multiple dimensions to calculate the matching score. Specific matching dimensions include: project characteristics (such as mountain / plain highways), abnormal numerical characteristics (such as deviation amplitude, duration), environmental factors (such as climate conditions, geological conditions), etc. If the causal relationship of each matching case is consistent with the current analysis, it can be counted as a confirmed case.

[0100] Exemplarily, the calculation formula for the causal impact relationship confirmation rate can be: confirmation rate = Σ (matched case weight × audit conclusion confidence level) / Σ matched case weights, where the case weight is jointly determined by the time decay factor (higher weight for recent cases) and project similarity, and the audit conclusion confidence level is rated based on the sufficiency of audit evidence at that time.

[0101] S520. When the causal impact relationship confirmation rate corresponding to the abnormal label in the root cause indicator layer is less than the preset causal impact relationship confirmation rate, multiply the causal impact relationship factor corresponding to the causal impact relationship of the abnormal label in the root cause indicator layer by a preset causal impact adjustment factor to adjust the causal impact relationship of the abnormal label in the root cause indicator layer.

[0102] In this implementation, when the system detects that the confirmation rate of an abnormal label in a certain root cause indicator layer is lower than the preset threshold (e.g., 60%), it can automatically trigger the dynamic adjustment program of the causal impact factor, multiply the causal impact relationship factor corresponding to the causal impact relationship of the abnormal label in the root cause indicator layer by a preset causal impact adjustment factor to adjust the causal impact relationship of the abnormal label in the root cause indicator layer.

[0103] Exemplarily, the preset confirmation rate can be a non-fixed value, but is intelligently adjusted according to the type of abnormal label in the root cause indicator layer: for abnormal types such as "material cost", due to large market price fluctuations, the threshold is set to 50%; while for rigid expenditures such as "work safety fees", the threshold is set to 75%, and the threshold rule base can be continuously optimized through machine learning.

[0104] Exemplarily, the preset causal impact adjustment factor can adopt the form of a logarithmic function: when the confirmation rate is near the threshold (e.g., 55% - 65%), a mild adjustment is adopted (the preset causal impact adjustment factor can be 1.2); when the confirmation rate is extremely low (< 30%), a radical adjustment is implemented (the preset causal impact adjustment factor can be 2.0), and at the same time, an industry expert verification link can be introduced to require manual review for major adjustments.

[0105] It should be noted that the causal impact relationship of the abnormal label in the root cause indicator layer after each adjustment will be stored in the knowledge base as a new case and marked with an "artificial correction" label. The system will regularly evaluate the adjustment effect and continuously optimize the adjustment algorithm by comparing the predicted causal relationship with the actual audit conclusion.

[0106] The beneficial effect of the above implementation is that it can improve the reliability of audit conclusions. In the pilot of a certain transportation investment group, the reconsideration rate of audit conclusions after applying this technology has decreased. Especially for abnormal types such as price increases of new materials and special processes, which used to have greater disputes, the confirmation accuracy of audit conclusions has been greatly improved.

[0107] The beneficial effects of the above implementation method are also that the system can effectively respond to sudden changes in the market environment (such as sharp price fluctuations caused by international exchange rate fluctuations) through the dynamic adjustment mechanism, and can accurately identify non-management responsibility factors.

[0108] The beneficial effects of the above implementation method are also that it can improve the efficiency of knowledge accumulation. The process of calculating the confirmation rate is essentially the digital precipitation of organizational audit experience, and it can shorten the training cycle of new auditors.

[0109] Figure 7 This is a schematic flowchart of the fifth economic responsibility audit portrait method for highway enterprises provided by the embodiments of the present application. As Figure 7 shown, the above method further includes S610 to S620, and the following is a specific description of S610 to S620.

[0110] S610. When there are multiple index layer anomaly labels under multiple criterion layer anomaly labels, obtain the co-influence relationship between the multiple index layer anomaly labels and the co-influence factor corresponding to the co-influence relationship. When there are multiple criterion layer anomaly labels under the same target layer anomaly label, obtain the co-influence relationship between the multiple criterion layer anomaly labels and the co-influence factor corresponding to the co-influence relationship.

[0111] In the economic responsibility audit of highway enterprises, this implementation method can construct a co-influence relationship analysis model to identify the complex interaction between multiple anomaly labels. Specifically, a co-influence relationship graph can be established first. Using graph neural network (GNN) technology, a three-level (index layer - criterion layer - target layer) association network is constructed. For index layer anomaly labels (such as "material loss rate", "mechanical utilization rate", etc.), through mutual information analysis and conditional probability calculation, the co-influence factors between two can be quantified. For example, when the co-factor between "number of rainy season construction days" and "concrete curing cost" reaches 0.72, it indicates that there is a significant co-effect between the two.

[0112] After obtaining the co-influence relationship analysis model, when there are multiple index layer anomaly labels under multiple criterion layer anomaly labels, obtain the co-influence relationship between the multiple index layer anomaly labels and the co-influence factor corresponding to the co-influence relationship, and then analyze the co-influence relationship between the multiple index layer anomaly labels.

[0113] After obtaining the co-influence relationship analysis model, when there are multiple criterion layer anomaly labels under the same target layer anomaly label, obtain the co-influence relationship between the multiple criterion layer anomaly labels and the co-influence factor corresponding to the co-influence relationship, and then analyze the co-influence relationship between the multiple criterion layer anomaly labels.

[0114] Exemplarily, this implementation can also consider the characteristic differences in different project phases, and the system will automatically adjust the collaborative factor weights based on the project progress: during the subgrade construction phase, the collaborative weight of "earthwork transportation distance" and "fuel consumption" is set to 0.8; while during the pavement laying phase, this weight is automatically adjusted to 0.3. The weight rule base is continuously optimized through reinforcement learning.

[0115] S620. According to the collaborative influence relationships among multiple index layer anomaly labels, sort them according to the number of collaborative influence relationships among multiple index layer anomaly labels, determine the multi-factor index layer anomaly label affected by the most index layer anomaly labels among multiple index layer anomaly labels, determine the upper-level criterion layer anomaly label corresponding to the multi-factor index layer anomaly label, and issue the audit clue prompt information corresponding to the multi-factor index layer anomaly label, the upper-level criterion layer anomaly label, and the target layer anomaly label. According to the collaborative influence relationships among multiple criterion layer anomaly labels, sort them according to the number of collaborative influence relationships among multiple index layer anomaly labels, determine the multi-factor criterion layer anomaly label affected by the most criterion layer anomaly labels among multiple criterion layer anomaly labels, and then issue the audit clue prompt information corresponding to the multi-factor criterion layer anomaly label and the target layer anomaly label.

[0116] In this implementation, based on the collaborative influence relationship network, the system can execute a multi-factor analysis algorithm. For the index layer anomaly labels, the in-degree centrality of each index layer anomaly label can be statistically analyzed. According to the collaborative influence relationships among multiple index layer anomaly labels, sort them according to the number of collaborative influence relationships among multiple index layer anomaly labels, determine the multi-factor index layer anomaly label affected by the most index layer anomaly labels among multiple index layer anomaly labels, determine the index layer anomaly label receiving the most collaborative influence as the multi-factor index layer anomaly label (for example, identifying that "steel bar processing fee" is affected by 5 other indicators), and determine the upper-level criterion layer anomaly label corresponding to the multi-factor index layer anomaly label, issue the audit clue prompt information corresponding to the multi-factor index layer anomaly label, the upper-level criterion layer anomaly label, and the target layer anomaly label, and then conduct an audit prompt for the multi-factor index layer anomaly label and the multi-layer labels corresponding to the multi-factor index layer anomaly label.

[0117] In this implementation, for criterion layer anomalies, a community detection algorithm can be used to identify tightly coupled anomaly clusters. According to the co - influence relationship between multiple criterion layer anomaly labels, sort them according to the number of co - influence relationships between multiple metric layer anomaly labels, and determine the multi - factor criterion layer anomaly label affected by the most criterion layer anomaly labels among multiple criterion layer anomaly labels. Determine the multi - factor criterion layer anomaly for the core node, and send the audit clue prompt information corresponding to the multi - factor criterion layer anomaly label and the target layer anomaly label, so as to realize the audit prompt for the multi - factor criterion layer anomaly label and the multi - layer labels corresponding to the multi - factor criterion layer anomaly label.

[0118] Exemplarily, the audit clue prompt information can adopt a visualization scheme of "fishbone diagram + heat map": the main bone shows the multi - factor anomaly label, the fish bones mark each co - influence factor and its factor value, and the heat map intuitively displays the co - influence strength distribution.

[0119] Exemplarily, for major co - anomalies (such as the linkage of more than three criterion layer anomalies), the system will automatically generate a "Multi - factor Impact Analysis Report", which can specifically include a co - influence path diagram, historical comparison data, risk conduction simulation, etc.

[0120] The beneficial effects of the above - mentioned implementation are that the co - factor model greatly improves the recognition rate of complex problems, and the multi - factor positioning technology greatly shortens the analysis time. Especially when dealing with the "cost - schedule - quality" triangular relationship problem, the co - influence analysis report automatically generated by the system can replace the traditional manual compilation.

[0121] The beneficial effects of the above - mentioned implementation are also that the audit conclusions from a co - influence perspective are more valuable for decision - making, and the adoption rate of management suggestions based on co - influence analysis is greatly improved, which is more effective than traditional audits.

[0122] In some implementations, the above - mentioned method further includes S710 to S720, which will be specifically described below.

[0123] S710. Obtain the historical audit data of the multi - factor metric layer anomaly label, and determine the co - influence relationship confirmation rate corresponding to the historical audit data of the multi - factor metric layer anomaly label. The co - influence relationship confirmation rate represents the audit confirmation ratio of the historical audit data for each co - influence relationship of the multi - factor metric layer anomaly label.

[0124] In the economic responsibility audit of highway enterprises, this implementation also introduces the quantitative index of co - influence relationship confirmation rate to evaluate the verification degree of historical audit data for the current multi - factor anomaly label.

[0125] When evaluating the degree of verification of historical audit data for current multi-factor anomaly labels, it can be completed through a confirmation rate calculation engine based on time series matching. For current multi-factor metric layer anomaly labels (such as "abnormal asphalt mixture cost"), audit cases of the same period historical projects (same quarter, similar section scale) can be retrieved first, and the audit confirmation status of each collaborative relationship can be extracted (for example, the "A→B" relationship is confirmed 3 times in 5 cases). At the same time, the confirmation rate can be calculated as Σ(confirmation times × case weight) / Σ(occurrence times × case weight), where the case weight takes into account factors such as project similarity, data integrity, and audit authority.

[0126] Exemplarily, in view of the construction characteristics of highway projects, the system can distinguish the impacts of different engineering stages: the collaborative confirmation rate during the subgrade construction period and the pavement laying period are calculated separately to avoid the interference of seasonal factors.

[0127] S720. When the confirmation rate of the first collaborative impact relationship corresponding to the multi-factor metric layer anomaly label is less than the preset first collaborative impact relationship confirmation rate, multiply the first collaborative impact relationship factor corresponding to the first collaborative impact relationship of the multi-factor metric layer anomaly label by the preset collaborative impact adjustment factor to adjust the collaborative impact relationship of the multi-factor metric layer anomaly label.

[0128] After obtaining the confirmation rate of the first collaborative impact relationship corresponding to the multi-factor metric layer anomaly label, when the confirmation rate of the first collaborative impact relationship corresponding to the multi-factor metric layer anomaly label is less than the preset first collaborative impact relationship confirmation rate, it indicates that the confirmation rate of the first collaborative impact relationship corresponding to the multi-factor metric layer anomaly label is too low. Furthermore, the first collaborative impact relationship factor corresponding to the first collaborative impact relationship of the multi-factor metric layer anomaly label can be multiplied by the preset collaborative impact adjustment factor to adjust the collaborative impact relationship of the multi-factor metric layer anomaly label.

[0129] Exemplarily, the preset first collaborative impact relationship confirmation rate can be set dynamically. The base value of the preset first collaborative impact relationship confirmation rate can be 60%, and the preset first collaborative impact relationship confirmation rate can also fluctuate up and down by 10%-15% according to the type of the multi-factor metric layer anomaly label.

[0130] Exemplarily, the preset collaborative impact adjustment factor can also perform a smoothing adjustment on the first collaborative impact relationship factor through a logarithmic function. When the confirmation rate is in the range of 40-60%, the adjustment factor can take values from 1.2 to 1.5; when the confirmation rate is lower than 40%, the adjustment factor can take values from 1.5 to 2.0. At the same time, the first collaborative impact relationship corresponding to the first collaborative impact relationship factor can be marked as the "to be verified" state.

[0131] Exemplarily, after multiplying the first co - influence relationship factor corresponding to the abnormal label of the multi - factor index layer by a preset co - influence adjustment factor, expert review and knowledge update can be carried out by tracking the whole - chain data of material procurement - transportation - use, verifying the vouchers and approval records of relevant projects, and comparing the differences between construction logs and progress plans.

[0132] Exemplarily, major adjustments (such as adjustment factor > 1.8) can be automatically pushed to the expert workbench, and an "adjustment proposal" (including historical comparison, current evidence chain, simulated impact analysis) is provided. The confirmed adjustment results will be updated to the case library, and the adjustment basis and responsible person will be marked.

[0133] The beneficial effects of the above - mentioned implementation method are that it can enhance the scientific nature of audit conclusions, especially the determination of complex co - influence relationships such as "material price - construction technology - climate conditions" is more accurate. The dynamic adjustment mechanism reduces the manual review workload, greatly improves the audit efficiency, and makes the knowledge accumulation in the audit process more systematic.

[0134] In some implementation methods, in the above - mentioned S210, through the index - layer anomaly identification unit, the index - layer abnormal label is determined, including S211 to S212. The following will specifically describe S211 to S212.

[0135] S211: Through the index - layer anomaly identification unit, cluster the multiple historical label data corresponding to the first label under the first index layer to obtain multiple historical label data categories and multiple historical label data cluster centers corresponding to the multiple historical label data categories respectively.

[0136] In this implementation method, when identifying the index - layer abnormal label, in the economic responsibility audit of highway enterprises, dynamic clustering technology can be used to perform intelligent analysis on the index - layer label data to achieve the identification of the index - layer abnormal label.

[0137] When identifying the index - layer abnormal label, an index - layer anomaly identification unit including a multi - dimensional clustering feature space can be constructed to cluster the multiple historical label data corresponding to the first label under the first index layer.

[0138] Specifically, when clustering multiple historical label data (such as "material price difference rate") under the first index layer, multi - dimensional feature vectors in the multiple historical label data can be extracted, which can specifically include time - dimension features (quarterly volatility, annual trend), project - dimension features (section scale, terrain type), environment - dimension features (climate conditions, geological conditions), etc., and the improved K - means++ algorithm is used to cluster the multiple historical label data.

[0139] Exemplarily, each finally determined historical label data clustering center is accompanied by a complete metadata description, specifically including the coverage time period, project type distribution, data quality score, etc.

[0140] S212. Obtain multiple first label data updated by the first label under the first index layer, respectively determine multiple Euclidean distances between the multiple first label data and multiple historical label data clustering centers, and determine the variance value of the multiple Euclidean distances. When the variance value of the multiple Euclidean distances is greater than the preset Euclidean distance variance value, the first label is taken as an abnormal label at the index layer.

[0141] When identifying the subsequently updated first label data, multiple first label data updated by the first label under the first index layer can be obtained. The multiple first label data can be actual project data used to evaluate whether the first label is abnormal. Furthermore, multiple Euclidean distances between the multiple first label data and multiple historical label data clustering centers can be respectively determined, and the variance value of the multiple Euclidean distances can be determined. The variance value of the Euclidean distances represents the fluctuation range of the distances between the multiple first label data and multiple historical label data clustering centers. When the variance value of the multiple Euclidean distances is greater than the preset Euclidean distance variance value, it indicates that the fluctuation range of the distances from the subsequently updated first label data to multiple historical label data clustering centers is relatively large, indicating that the subsequently updated first label data no longer conforms to the clustering characteristics of the multiple historical label data. Therefore, the first label can be taken as an abnormal label at the index layer.

[0142] The beneficial effect of the above implementation method is that it can improve audit sensitivity. In the audit of mountain expressway projects, multiple hidden anomalies caused by geological condition changes have been successfully identified.

[0143] The beneficial effect of the above implementation method is also that the automatic clustering mechanism greatly shortens the data processing time. Especially in the audit of cross-year projects, the clustering analysis report automatically generated by the system can replace the workload of manual analysis and improve the audit efficiency.

[0144] The beneficial effect of the above implementation method is also that in terms of adaptability, it effectively responds to the data distribution changes brought about by policy adjustments. The system can automatically identify the labels that need to be recalibrated, ensuring the timeliness and reliability of audit conclusions.

[0145] In some implementation methods, in the above S210, through the abnormal identification unit at the index layer, to determine the abnormal label at the index layer, S213 to S214 are also included. The following specifically describes S213 to S214.

[0146] S213. Obtain the first time window corresponding to multiple historical label data under the historical label data category, and obtain the first external condition influence factor corresponding to the first time window.

[0147] In this implementation method, in the economic responsibility audit of highway enterprises, a mechanism for analyzing the impact of time-window-sensitive external conditions can also be introduced to further improve the effect of analyzing the impact of external conditions.

[0148] When analyzing the impact of external conditions, a multi-dimensional external condition evaluation system can be established. For each historical label data category (such as the category of "abnormal material costs"), the system can automatically extract the external environment data within its first time window (usually 3 to 6 months), including specific meteorological data (rainfall, temperature), geological data (soil change rate), market data (material price index), etc.

[0149] Exemplarily, an entropy weight-TOPSIS combined algorithm can be used to calculate the comprehensive impact factor, and its calculation formula is:

[0150] F = α×Wweather + β×Wgeology + γ×Wmarket

[0151] Among them, α, β, and γ are dynamic weight coefficients, Wweather represents the impact degree of meteorological data, Wgeology represents the impact degree of geological data, Wmarket represents the impact degree of market data, and α, β, and γ can be regularly adjusted through the LSTM network prediction model.

[0152] S214. When the first external condition impact factor is greater than or equal to the preset external condition impact factor, through the index layer anomaly recognition unit, obtain multiple historical label data corresponding to the first label under the first index layer in the second time window following the first time window, cluster the multiple historical label data corresponding to the first label under the first index layer in the second time window, and update the multiple historical label data categories and the multiple historical label data cluster centers corresponding to the multiple historical label data categories.

[0153] When the first external condition impact factor is greater than or equal to the preset external condition impact factor, through the index layer anomaly recognition unit, obtain multiple historical label data corresponding to the first label under the first index layer in the second time window following the first time window, and cluster the multiple historical label data corresponding to the first label under the first index layer in the second time window to improve the recognition accuracy of the index layer anomaly labels in the second time window.

[0154] As a supplement, in the audit portrait, when significant changes occur in external conditions, the system can initiate an intelligent update process. The intelligent update process can also adaptively adjust the time window, specifically determining the length of the second time window automatically according to the intensity of changes in external conditions. Under moderate influence (0.7 ≤ the first external condition influence factor < 0.8), the first time window can be extended by 50% to obtain the length of the second time window. Under severe influence (the first external condition influence factor ≥ 0.8), the first time window can be extended by 100% to obtain the length of the second time window, and subsequent analysis can be initiated.

[0155] After clustering the multiple historical tag data corresponding to the first tag under the first indicator layer in the second time window, the multiple historical tag data categories and the multiple historical tag data clustering centers respectively corresponding to the multiple historical tag data categories can be updated to improve the subsequent recognition accuracy of abnormal tags in the indicator layer.

[0156] The beneficial effects of the above implementation method are as follows. In terms of audit adaptability, it can accurately distinguish management responsibilities from the influence of natural factors, and the perception of external conditions improves the recognition accuracy of abnormal tags in the indicator layer. For example, in the audit of a mountain highway, the system can automatically identify the abnormal cost ratio caused by the rainy season.

[0157] The beneficial effects of the above implementation method are also as follows. In terms of the timeliness of the audit system, the dynamic update mechanism keeps the clustering model in an optimal state. The response speed of the audit system to material price fluctuations can be shortened from 2 weeks in the traditional method to 3 days, enabling it to promptly capture the impact of market changes.

[0158] The beneficial effects of the above implementation method are also as follows. In the multi-dimensional analysis report, it can provide more comprehensive audit basis. Based on the "external conditions - cost impact" correlation diagram generated by the updated audit system, a targeted cost control plan can be specified.

[0159] The embodiment of the present application also provides an economic responsibility audit portrait system for highway enterprises, including units for executing the method described in any one of the above.

[0160] Figure 8 It is a schematic logical structure diagram of an economic responsibility audit portrait system for highway enterprises provided by an embodiment of the present application. As Figure 8 shown, the system 1 of this embodiment includes a processing unit 11, a storage unit 12, and a transceiver unit 13. The processing unit 11 is used to process data, the storage unit 12 is used to store data, and the transceiver unit 13 is used to send and receive data. The processing unit 11, the storage unit 12, and the transceiver unit 13 cooperate with each other to implement the above method. The beneficial effects of the embodiment of the present application have been described in the above method and will not be elaborated here.

[0161] It should be noted that for the content such as information interaction and execution process between the above-mentioned devices / units, since it is based on the same concept as the method embodiments of the present application, for its specific functions and the technical effects brought, reference can be specifically made to the method embodiment part, and details are not described herein again.

[0162] Those skilled in the art can clearly understand that for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is used as an example. In actual applications, the above-mentioned functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiments can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of the present application. The specific working process of the units and modules in the above system can refer to the corresponding process in the foregoing method embodiments, and details are not described herein again.

[0163] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above method embodiments of the present application, a computer program can be used to instruct the relevant hardware to complete. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the photographing device / terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disc, etc. In some jurisdictions, according to legislation and patent practice, the computer-readable medium cannot be an electrical carrier signal and a telecommunication signal.

[0164] In the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0165] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. A professional technician can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

[0166] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the devices or units can be in electrical, mechanical or other forms.

[0167] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0168] The above-described embodiments are only used to illustrate the technical solutions of this application, rather than to limit them; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included in the protection scope of this application.

Claims

1. An economic responsibility audit portrait method for highway enterprises, characterized in that, The method includes: Establishing a multi-layer label model for highway enterprises, where the multi-layer label model includes target layer labels, criterion layer labels, and index layer labels; among them, the target layer labels include economic responsibility audit labels, the criterion layer labels are the lower-level refined labels of the target layer labels, and the index layer labels are the lower-level refined labels of the criterion layer labels; Determining the label index weights corresponding to each label in the multi-layer label model, determining the scoring values corresponding to each label according to the internal assessment management method and external audit data of highway enterprises, and determining the product of the scoring value and the label index weight corresponding to each label as the score of each label, and displaying the scores of each item of labels to obtain the economic responsibility audit portrait of highway enterprises.

2. The method according to claim 1, wherein The method further includes: Determining the abnormal labels at the index layer through the index layer abnormal identification unit; determining the criterion layer labels including the abnormal labels at the index layer, and determining the target layer labels including the criterion layer labels; When the score corresponding to the criterion layer abnormal label is less than or equal to the preset score of the criterion layer abnormal label, determining the criterion layer label as the criterion layer abnormal label; when the score corresponding to the target layer abnormal label is less than or equal to the preset score of the target layer abnormal label, determining the target layer label as the target layer abnormal label; sending out the audit clue prompt information corresponding to the abnormal labels at the index layer, the criterion layer abnormal labels, and the target layer abnormal labels.

3. The method according to claim 2, wherein The method further includes: When there are multiple abnormal labels at the index layer under the same criterion layer abnormal label, obtaining the external condition influence factors corresponding to each of the multiple abnormal labels at the index layer; among them, the external condition influence factor is used to characterize the degree of abnormal contribution of the external condition to the index layer label, and the external condition influence factor includes the construction period stage factor; Determining the target index layer abnormal label with the largest external condition influence factor corresponding to each of the multiple abnormal labels at the index layer, and sending out the audit clue prompt information corresponding to the target index layer abnormal label, the criterion layer abnormal label, and the target layer abnormal label.

4. The method according to claim 3, wherein The method further includes: When there are multiple abnormal labels at the index layer under multiple criterion layer abnormal labels, obtaining the causal influence relationship between the multiple abnormal labels at the index layer and the causal influence factor corresponding to the causal influence relationship; when there are multiple criterion layer abnormal labels under the same target layer abnormal label, obtaining the causal influence relationship between the multiple criterion layer abnormal labels and the causal influence factor corresponding to the causal influence relationship; According to the causal influence relationship between the multiple abnormal labels at the index layer, determining the root cause index layer abnormal label among the multiple abnormal labels at the index layer according to the influence relationship corresponding to the largest causal influence factor, and determining the upper-level criterion layer abnormal label corresponding to the root cause index layer abnormal label, and sending out the audit clue prompt information corresponding to the root cause index layer abnormal label, the upper-level criterion layer abnormal label, and the target layer abnormal label; according to the causal influence relationship between the multiple criterion layer abnormal labels, determining the root cause criterion layer abnormal label among the multiple criterion layer abnormal labels according to the influence relationship corresponding to the largest causal influence factor, and sending out the audit clue prompt information corresponding to the root cause criterion layer abnormal label and the target layer abnormal label.

5. The method according to claim 4, characterized in that, The method further includes: Obtain the historical audit data of the abnormal labels at the root cause indicator layer, and determine the causal impact relationship confirmation rate corresponding to the historical audit data of the abnormal labels at the root cause indicator layer. The causal impact relationship confirmation rate represents the audit confirmation ratio of the causal impact relationship of the historical audit data on the abnormal labels at the root cause indicator layer; When the causal impact relationship confirmation rate corresponding to the abnormal label at the root cause indicator layer is less than the preset causal impact relationship confirmation rate, multiply the causal impact relationship factor corresponding to the causal impact relationship of the abnormal label at the root cause indicator layer by the preset causal impact adjustment factor to adjust the causal impact relationship of the abnormal label at the root cause indicator layer.

6. The method according to claim 5, wherein The method further includes: When there are multiple abnormal labels at the indicator layer under multiple abnormal labels at the criterion layer, obtain the co-influence relationship between the multiple abnormal labels at the indicator layer and the co-influence factor corresponding to the co-influence relationship; when there are multiple abnormal labels at the criterion layer under the same abnormal label at the target layer, obtain the co-influence relationship between the multiple abnormal labels at the criterion layer and the co-influence factor corresponding to the co-influence relationship; According to the co-influence relationship between the multiple abnormal labels at the indicator layer, sort according to the number of co-influence relationships between the multiple abnormal labels at the indicator layer, determine the multi-factor indicator layer abnormal label affected by the most abnormal labels at the indicator layer among the multiple abnormal labels at the indicator layer, and determine the upper-level criterion layer abnormal label corresponding to the multi-factor indicator layer abnormal label, and issue the audit clue prompt information corresponding to the multi-factor indicator layer abnormal label, the upper-level criterion layer abnormal label and the target layer abnormal label; according to the co-influence relationship between the multiple abnormal labels at the criterion layer, sort according to the number of co-influence relationships between the multiple abnormal labels at the indicator layer, determine the multi-factor criterion layer abnormal label affected by the most abnormal labels at the criterion layer among the multiple abnormal labels at the criterion layer, and issue the audit clue prompt information corresponding to the multi-factor criterion layer abnormal label and the target layer abnormal label.

7. The method according to claim 6, wherein The method further includes: Obtain the historical audit data of the multi-factor indicator layer abnormal label, and determine the co-influence relationship confirmation rate corresponding to the historical audit data of the multi-factor indicator layer abnormal label. The co-influence relationship confirmation rate represents the audit confirmation ratio of each co-influence relationship of the historical audit data on the multi-factor indicator layer abnormal label; When the first co-influence relationship confirmation rate corresponding to the multi-factor indicator layer abnormal label is less than the preset first co-influence relationship confirmation rate, multiply the first co-influence relationship factor corresponding to the first co-influence relationship of the multi-factor indicator layer abnormal label by the preset co-influence adjustment factor to adjust the co-influence relationship of the multi-factor indicator layer abnormal label.

8. The method according to claim 7, wherein Through the indicator layer anomaly recognition unit, determine the abnormal labels at the indicator layer, including: Through the indicator layer anomaly recognition unit, cluster the multiple historical label data corresponding to the first label under the first indicator layer to obtain multiple historical label data categories and multiple historical label data cluster centers corresponding to the multiple historical label data categories respectively; Obtain multiple first tag data updated by the first tag under the first index layer, respectively determine multiple Euclidean distances between the multiple first tag data and multiple historical tag data clustering centers, and determine the variance value of the multiple Euclidean distances. When the variance value of the multiple Euclidean distances is greater than the preset Euclidean distance variance value, the first tag is used as the abnormal tag of the index layer.

9. The method according to claim 8, characterized in that, Determining the abnormal tags of the index layer through the index layer abnormal identification unit further includes: Obtain the first time window corresponding to multiple historical tag data under the historical tag data category, and obtain the first external condition influence factor corresponding to the first time window; When the first external condition influence factor is greater than or equal to the preset external condition influence factor, through the index layer abnormal identification unit, obtain multiple historical tag data corresponding to the first tag under the first index layer in the second time window subsequent to the first time window, cluster the multiple historical tag data corresponding to the first tag under the first index layer in the second time window, and update the multiple historical tag data categories and the multiple historical tag data clustering centers corresponding to the multiple historical tag data categories respectively.

10. An economic responsibility audit portrait system for highway enterprises, characterized in that, Includes a unit for performing the method according to any one of claims 1 to 9.

Images