Cross-industry auditing template dynamic collaboration method and system

Abstract

The invention discloses a cross-industry auditing template dynamic collaboration method and system, and relates to the technical field of auditing, and the method comprises the steps: obtaining multi-format policy original data through responding to an auditing template updating request, carrying out the structural processing, and constructing policy-auditing influence mapping data through semantic mapping; performing three-dimensional feature extraction and correlation weight quantization on the mapping items, coupling subjective weights of the sequence relation analysis method and objective weights of the grey correlation degree analysis method, generating three-dimensional comprehensive correlation weights, and judging item priorities; retrieving a preset cross-industry basic auditing template library by taking the applicable industry as a retrieval key, and generating an initial industry exclusive auditing template in combination with a priority matching directional adjustment strategy; template structure uniformization and integration compliance are completed based on the preset industry knowledge graph, the cross-industry auditing template is finally determined, and the problems that auditing template updating lacks pertinence and cross-industry adaptation efficiency is low in the prior art are solved.

Description

Technical Field

[0001] This invention relates to the technical field of auditing, and in particular to a dynamic collaborative method and system for cross-industry audit templates. Background Technology

[0002] Driven by the digital wave, cross-industry business models are becoming increasingly common, and the frequency of policy and regulation updates is accelerating. The audit industry faces an urgent need to quickly adapt to diverse regulatory requirements and optimize audit processes and templates. Dynamic collaboration of audit templates in cross-industry scenarios has become a core issue for improving audit compliance and efficiency.

[0003] Currently, while automated auditing tools have made breakthroughs in basic aspects such as data processing and anomaly detection, and can replace some repetitive auditing work, most are limited to the fixed updating of templates in a single industry or the superficial analysis of policy texts. They cannot achieve dynamic retrieval and adaptation of audit templates and structural integration for compliance based on the business characteristics of different industries. This results in insufficient targeting of audit template updates, low efficiency in cross-industry adaptation, and difficulty in quickly responding to the differentiated compliance needs of various industries. This not only increases the time and cost of manual intervention, but also poses the risk of overlooking compliance risks. There is an urgent need for a technical solution that can accurately quantify the impact of policies and enable intelligent collaboration of templates across industries. Summary of the Invention

[0004] This invention provides a method and system for dynamic collaboration of cross-industry audit templates, which solves the technical problems of lack of targeted updates and low efficiency of cross-industry adaptation in the prior art.

[0005] The first aspect of this invention provides a cross-industry audit template dynamic collaboration method, comprising: In response to audit template update requests, the system obtains raw policy data in multiple different formats and performs structured processing to obtain structured policy data. Semantic mapping is performed on the structured policy data to obtain policy-audit impact mapping data; Three-dimensional correlation weight quantification and item priority determination are performed on multiple policy-audit impact mapping items within the policy-audit impact mapping data to obtain the corresponding item priorities; The policy and audit impact mapping data are used to retrieve industry templates, and the initial industry-specific audit templates for each applicable industry are determined by combining the item priority. Based on the pre-built industry knowledge graph, the initial industry-specific audit templates are integrated and made compliant across industries to determine the cross-industry audit templates corresponding to each applicable industry.

[0006] Optionally, in response to the audit template update request, multiple original policy data in different formats are obtained and structured to obtain structured policy data, including: Responding to audit template update requests, it retrieves raw policy data in multiple different formats; The original data of each policy were standardized to obtain multiple plain text data. All the plain text data are encoded to obtain standardized policy text data; The standardized policy text data is subjected to regular expression extraction to obtain core metadata; Key clauses were obtained by performing keyword retrieval and clause classification on the standardized policy text data; The core metadata is mapped to the key terms to obtain structured policy data.

[0007] Optionally, the step of semantically mapping the structured policy data to obtain policy-audit impact mapping data includes: The structured policy data is filtered to obtain core corpus data; The core corpus data is normalized using synonyms to obtain standardized core corpus data; The standardized core corpus data is semantically parsed using a pre-set fine-tuned large language model to obtain the initial mapping results between policy and audit. A confidence assessment was conducted on the initial mapping results between the stated policy and the audit. The initial mapping results between policies and audits with a confidence level greater than or equal to a pre-set confidence threshold are selected as the target confidence level mapping results; The initial mapping results between policies and audits with confidence levels lower than the preset confidence threshold are corrected to obtain complete mapping results; The target confidence mapping result and the complete mapping result are merged and deduplicated to obtain policy and audit impact mapping data.

[0008] Optionally, the step of performing three-dimensional correlation weight quantification and item priority determination on multiple policy-audit impact mapping items within the policy-audit impact mapping data to obtain the corresponding item priority includes: Three-dimensional feature extraction is performed on multiple policy-audit impact mapping entries within the policy-audit impact mapping data to obtain the dimensional feature vectors corresponding to the three dimensions. The three-dimensional dimensions are weighted using the ordinal relation analysis method to obtain the subjective weight vectors of the three-dimensional dimensions corresponding to the multiple policy-audit impact mapping items. Based on the grey relational analysis method, grey relational analysis is performed using all the dimensional feature vectors and the associated preset ideal feature vectors to obtain the three-dimensional objective weight vectors corresponding to multiple policy-audit impact mapping items. Using the fuzzy comprehensive evaluation method, the subjective weight vector and the objective weight vector of the three-dimensional dimension are weighted and coupled to obtain the three-dimensional comprehensive correlation weights corresponding to multiple policy-audit impact mapping items; Based on the three-dimensional comprehensive correlation weight, the multiple policy-audit impact mapping items are sorted in descending order, and the item priority corresponding to each policy-audit impact mapping item is determined according to the preset item priority partitioning interval to which the three-dimensional comprehensive correlation weight belongs.

[0009] Optionally, the step of extracting three-dimensional features from multiple policy-audit impact mapping entries within the policy-audit impact mapping data to obtain dimensional feature vectors corresponding to the three dimensions includes: Policy dimension features are extracted from multiple policy-audit impact mapping entries within the policy-audit impact mapping data to construct a policy dimension feature vector; Industry-dimensional features are extracted from multiple policy-audit impact mapping entries within the policy-audit impact mapping data to construct an industry-dimensional feature vector; Audit node dimension features are extracted from multiple policy-audit impact mapping entries within the policy-audit impact mapping data to construct an audit node dimension feature vector.

[0010] Optionally, the step of using the policy and audit impact mapping data to perform industry template retrieval and determining the initial industry-specific audit template for each applicable industry based on the item priority includes: Target field data is extracted from the policy and audit impact mapping data and used as the search key; the target field data includes applicable industries; The search key is used to retrieve a pre-set cross-industry basic audit template library to obtain the benchmark audit templates corresponding to each applicable industry. Perform integrity verification on each of the aforementioned benchmark audit templates to obtain the verification benchmark audit templates corresponding to each of the aforementioned applicable industries; The corresponding adjustment strategy is matched according to the priority of the item, and the audit templates of each verification benchmark are adjusted according to the adjustment strategy to obtain the initial industry-specific audit templates corresponding to each applicable industry.

[0011] Optionally, the step of integrating and standardizing the initial industry-specific audit templates based on a pre-set industry knowledge graph to determine the cross-industry audit templates corresponding to each applicable industry includes: Based on the pre-built industry knowledge graph, the structure of each initial industry-specific audit template is made consistent to obtain the intermediate industry-specific audit template corresponding to each applicable industry. The audit templates for each of the aforementioned intermediate industries are annotated to obtain the target industry-specific audit templates for each of the applicable industries. Each target industry-specific audit template undergoes compliance verification, and the target industry-specific audit templates that pass the compliance verification are used as cross-industry audit templates.

[0012] Optionally, it also includes: Analyze the compliance issue list for the target industry-specific audit template that failed the compliance verification, and then proceed to execute the step of matching the corresponding adjustment strategy according to the priority of the item.

[0013] The second aspect of this invention provides a cross-industry audit template dynamic collaborative system, comprising: The policy collection and parsing module is used to respond to audit template update requests, acquire multiple raw policy data in different formats, and perform structured processing to obtain structured policy data. The policy semantic understanding and impact analysis module is used to perform semantic mapping on the structured policy data to obtain policy and audit impact mapping data; The weight quantification and priority determination module is used to perform three-dimensional correlation weight quantification and item priority determination on multiple policy-audit impact mapping items in the policy-audit impact mapping data to obtain the corresponding item priority. The audit template generation and management module is used to retrieve industry templates using the policy and audit impact mapping data, and determine the initial industry-specific audit templates for each applicable industry based on the item priority. The cross-industry template collaborative configuration module is used to integrate and comply with the initial industry-specific audit templates based on the pre-set industry knowledge graph, and to determine the cross-industry audit templates corresponding to each applicable industry.

[0014] A third aspect of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the cross-industry audit template dynamic collaboration method as described above.

[0015] As can be seen from the above technical solutions, the present invention has the following advantages: This invention provides a method and system for dynamic collaboration of cross-industry audit templates. By responding to audit template update requests, it acquires multi-format policy raw data and performs structured processing. Semantic mapping is used to construct policy-audit impact mapping data. Then, multiple policy-audit impact mapping items in the mapping data undergo three-dimensional feature extraction and association weight quantification based on policy, industry, and audit node dimensions. Subjective weights obtained through ordinal relational analysis and objective weights obtained through grey relational analysis are coupled to generate a three-dimensional comprehensive association weight, which determines the priority of each mapping item. Subsequently, a pre-set cross-industry basic audit template library is retrieved using the applicable industry as the search key. An initial industry-specific audit template is generated by combining item priority matching and targeted adjustment strategies. Finally, template structure consistency and integration compliance are achieved based on a pre-set industry knowledge graph, ultimately determining the cross-industry audit templates corresponding to each applicable industry. This invention overcomes the limitations of traditional audit template updates, which lack precise quantification of policy-audit impact, by employing a three-dimensional correlation weight quantification mechanism. It accurately assesses the importance of each mapping item and prioritizes it based on three core dimensions: policy relevance, industry adaptability, and audit node relevance. This transforms template updates from indiscriminate, all-encompassing adjustments into targeted optimizations focused on high-priority, core impact items, fundamentally solving the problem of untargeted template updates. Furthermore, it achieves rapid and accurate matching of benchmark templates using applicable industries as the retrieval dimension. Combined with priority-based targeted adjustments, it reduces ineffective adaptation operations. Moreover, it utilizes industry knowledge graphs to achieve structural uniformity and differentiated field adaptation across industries, avoiding the cumbersome process of manual industry-by-industry adjustments and coordination required in traditional cross-industry adaptation. This significantly shortens the template adaptation cycle, effectively improves cross-industry adaptation efficiency, and further ensures consistency between templates and policy requirements in the compliance process, reducing compliance risks while improving targeting and efficiency. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating the steps of a cross-industry audit template dynamic collaboration method provided in this embodiment of the invention; Figure 2 A structural block diagram of a cross-industry audit template dynamic collaborative system provided in an embodiment of the present invention; Figure 3 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0018] This invention provides a method and system for dynamic collaboration of cross-industry audit templates, which addresses the technical problems of lack of targeted updates and low efficiency in cross-industry adaptation in existing technologies.

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. It should be noted that in the optional embodiments of the present invention, the object information and other related data involved require the permission or consent of the object when the embodiments of the present invention are applied to specific products or technologies, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. That is to say, if the embodiments of the present invention involve data related to the object, it needs to be obtained with the authorization and consent of the object, the authorization and consent of the relevant departments, and in compliance with the relevant laws, regulations, and standards of the country and region. If personal information is involved in the embodiments, the acquisition of all personal information requires the consent of the individual. If sensitive information is involved, the separate consent of the information subject is required, and the embodiments also need to be implemented with the authorization and consent of the object.

[0020] Please see Figure 1 , Figure 1 The flowchart illustrates the steps of a cross-industry audit template dynamic collaboration method provided in this embodiment of the invention.

[0021] This invention provides a cross-industry audit template dynamic collaboration method, comprising: Step 101: Respond to the audit template update request, obtain multiple original policy data in different formats, and perform structured processing to obtain structured policy data.

[0022] Further, step 101 may include the following sub-steps: S11. Respond to the audit template update request and obtain policy raw data in multiple different formats.

[0023] Audit template update request: refers to the instruction initiated by the user or related business system to trigger changes in the audit template adaptation policy. The core includes configuration parameters such as the target industry scope, the policy update time interval, and the specific audit scenario requirements. It is the trigger condition for starting the entire data collection process.

[0024] Raw policy data refers to basic policy-related data collected directly from authoritative sources without any processing. It covers types such as cross-industry general policies and single-industry specific policies, and is the core information carrier supporting the dynamic updating of audit templates.

[0025] Multi-format policy raw data: refers to the diverse document presentation formats of the collected policy raw data, including but not limited to PDF, HTML, scanned documents, structured data returned by API interfaces, etc., which need to be converted into a unified parsable format through subsequent format standardization processing.

[0026] Dual-mode data collection: This refers to a combined data collection mode that integrates distributed web crawlers (suitable for collecting policies from web pages) with API integration with authoritative databases (suitable for retrieving structured policy data), balancing the comprehensiveness of the collection scope with the compliance of the data sources.

[0027] In this embodiment of the invention, in response to an audit template update request (which includes key parameters such as the target industry scope, policy update time interval, and audit scenario requirements), a targeted collection strategy is configured based on the request parameters. Through a dual-mode collection method that connects distributed web crawlers with authoritative database APIs, original policy data in multiple different formats covering cross-industry general policies and single-industry specific policies are obtained from legitimate and compliant channels such as compliance release platforms and industry regulatory data centers.

[0028] S12. Standardize the format of the original data of each policy to obtain multiple plain text data.

[0029] Format standardization: refers to the process of removing non-textual redundant information and unifying the data presentation format for policy raw data in different file formats (such as PDF, HTML, scanned documents, etc.) through specific technical means. The core purpose is to eliminate the impact of format differences on subsequent data parsing.

[0030] Plain text data: refers to structured data that has been standardized in format, retaining only the core textual information of the policy and removing all formatting tags (such as font styles, layout marks, etc.).

[0031] In this embodiment of the invention, the collected policy raw data in multiple formats, such as PDF, HTML, scanned documents, and API interface returned data, are processed by optical character recognition technology to identify text information in scanned documents, parse format tags in PDF and HTML files and remove non-textual redundant content, and perform a unified format conversion operation on different format data to complete the format standardization process, ultimately obtaining multiple plain text data without format interference that can be directly used for subsequent parsing.

[0032] S13. Encode all plain text data to obtain standardized policy text data.

[0033] Encoding: refers to the process of converting plain text data into a unified character encoding format. The core purpose is to eliminate encoding differences between data from different sources and ensure the consistency and compatibility of data parsing.

[0034] Standardized policy text data: refers to policy text data that has undergone encoding normalization, has unified character encoding, no parsing conflicts, and can be directly used for subsequent structured processing (such as metadata extraction and clause classification).

[0035] In this embodiment of the invention, considering the potential encoding inconsistencies that may arise after different format conversions, all plain text data are uniformly encoded using the UTF-8 encoding format to eliminate parsing anomalies caused by encoding differences, ultimately resulting in standardized policy text data with a unified format that can be stably used for subsequent core metadata extraction and key clause classification.

[0036] S14. Perform regular expression extraction on the standardized policy text data to obtain core metadata.

[0037] Regular expression extraction: refers to the processing method that accurately filters and extracts key target information from standardized policy text data based on preset regular expression rules. Its core advantage is the efficient matching of structured information in a specific format.

[0038] Core metadata refers to a set of key information that reflects the core attributes of a policy, including but not limited to policy number, issuing entity, effective date, and scope of application. It is a core component of structured policy analysis.

[0039] In this embodiment of the invention, a targeted regular expression is preset based on the inherent structural features of the policy text to accurately extract key information reflecting the core attributes of the policy, such as the policy number, issuing entity, effective date, and scope of application, thereby completing the regular expression extraction operation of the core metadata.

[0040] S15. Perform keyword retrieval and clause classification on the standardized policy text data to obtain key clauses.

[0041] Keyword retrieval: This refers to a process that uses a pre-defined keyword library based on audit business scenarios to match and filter policy content directly related to audit work from standardized policy text data. The core is to accurately locate audit-related information.

[0042] Key clauses: These refer to the core policy content that, after keyword retrieval and classification, can directly constrain the audit process, clarify compliance requirements, or define audit responsibilities. They are the core carrier for building the relationship between policies and audits.

[0043] In this embodiment of the invention, by combining a keyword library of compliance requirements, audit processes, and liability definition pre-set for audit business scenarios, policy content highly relevant to audit work is retrieved through precise keyword matching. The clauses are then classified according to dimensions such as applicable scenarios, audit stages, and constraint types, ultimately resulting in key clauses that focus on the core needs of auditing.

[0044] S16. Map the core metadata to the key clauses to obtain structured policy data.

[0045] Mapping: refers to the process of establishing a one-to-one correspondence between core metadata and key clauses based on the inherent relevance of policy content. The core is to achieve the structured binding of policy attribute information and core content.

[0046] Structured policy data refers to policy data that has complete attributes, clear relationships, and conforms to a unified data structure standard after being mapped with core metadata and key clauses. It is the core input data for subsequent semantic analysis and impact analysis.

[0047] In this embodiment of the invention, a one-to-one correspondence is established based on the inherent correlation of policy content. The core attribute information of the same policy (such as policy number, scope of application, etc.) is bound and associated with key clauses that focus on audit needs. The mapping process between core metadata and key clauses is completed, and finally, structured policy data with complete attributes, clear correlation, and can be directly used for subsequent policy semantic understanding and impact analysis is formed.

[0048] Step 102: Perform semantic mapping on the structured policy data to obtain policy and audit impact mapping data.

[0049] Furthermore, step 102 may include the following sub-steps: S21. Filter the structured policy data to obtain the core corpus data.

[0050] Filtering: This refers to the process of removing redundant information irrelevant to auditing from structured policy data based on the core needs of auditing business, using rules such as keyword matching and scope of application filtering. The core is to improve the accuracy and efficiency of subsequent semantic parsing.

[0051] Core corpus data: refers to a subset of policy data that has been filtered and processed to focus on the relevance of audit business and remove redundant information. It is the core input data used for semantic mapping between policies and audits.

[0052] In this embodiment of the invention, considering that the structured policy data may contain redundant information unrelated to audit business, non-audit-related content in the structured policy data is screened and removed based on a preset audit business keyword library and policy application scope matching rules, thus completing the data filtering process and finally obtaining core corpus data that focuses on the core audit needs and is free of redundant interference.

[0053] S22. Normalize the core corpus data using synonyms to obtain standardized core corpus data.

[0054] Synonym normalization: refers to the process of replacing different expressions (synonyms, near-synonyms) of the same meaning in the core corpus with standard expressions based on a thesaurus specific to the auditing field. The core is to eliminate semantic ambiguity and unify data expression standards.

[0055] Standardized core corpus data refers to corpus data that has been processed by synonym normalization, has consistent semantics and unified expression, and can be directly used for semantic analysis of large language models. It is a key input to improve the accuracy of semantic mapping of policies and audits.

[0056] In this embodiment of the invention, based on the obtained core corpus data that focuses on the core audit requirements and is free from redundant interference, and considering that there may be different expressions of the same semantics, the synonyms and near-synonyms in the core corpus data are uniformly replaced based on a preset thesaurus of audit domains. This completes the synonym normalization process, eliminates semantic ambiguity, and unifies the data expression standards, ultimately resulting in standardized core corpus data with consistent semantics and unified expression.

[0057] S23. Use a pre-set fine-tuned large language model to perform semantic analysis on standardized core corpus data to obtain the initial mapping results between policy and audit.

[0058] Pre-set fine-tuned large language model: refers to a model based on a general large language model (such as GPT-3.5, LLaMA2, ERNIE3.0, etc.), which is then trained and optimized by inputting audit-specific corpus to adapt it to the needs of policy semantic parsing and association recognition in audit scenarios. It has stronger industry relevance and parsing accuracy. This is a conventional general large language model, which will not be discussed further here.

[0059] Semantic parsing refers to the process of using large language models to perform deep semantic mining on standardized core corpora, understand the constraints and compliance requirements of policy provisions, and identify their potential associations with audit-related elements.

[0060] Initial mapping results between policies and audits: This refers to the set of relationships between policy provisions and audit processes, indicators, compliance points, and other elements initially formed after semantic parsing. It includes basic information such as the relationship dimension and the degree of initial matching, and is the core object of subsequent confidence assessment.

[0061] In this embodiment of the invention, based on the obtained standardized core corpus data with consistent semantics and unified expression, a large language model that has been pre-tuned by professional corpus data in the audit field (such as audit process specifications, historical policy-audit adaptation cases, compliance standards, etc.) is invoked to perform deep semantic understanding and correlation mining on the standardized core corpus data, accurately identify the correspondence between policy clauses and audit processes, audit indicators, and compliance requirements, and finally obtain the initial mapping result of policies and audits that includes the direction of policy impact, audit-related nodes, and preliminary matching degree.

[0062] S24. Conduct a confidence assessment of the initial mapping results between policy and audit.

[0063] Confidence assessment: refers to the process of quantitatively judging the reliability and accuracy of the initial mapping results between policies and audits based on preset multi-dimensional assessment indicators. The core is to screen effective mapping relationships and identify low-confidence mappings that need to be optimized.

[0064] Confidence score and level: refers to the quantitative value (e.g., 0-100 points) and corresponding level (e.g., high, medium, low) that reflects the reliability of the mapping relationship after evaluation. It is the core basis for subsequent classification of the mapping result processing method.

[0065] In this embodiment of the invention, based on the obtained initial mapping results of policies and audits, which include the direction of policy impact, audit-related nodes, and preliminary matching degree, and based on the preset confidence assessment rules (covering core indicators such as semantic matching completeness, industry adaptability, and audit node relevance), a quantitative scoring method is used to determine the reliability of each mapping relationship, output the corresponding confidence score and level, clearly distinguish between high-confidence effective mappings and low-confidence mappings that need to be corrected, and provide a basis for the targeted correction of subsequent low-confidence results.

[0066] S25. Select the policy and audit initial mapping results with a confidence level greater than or equal to the preset confidence threshold as the target confidence level mapping results.

[0067] Preset confidence threshold: This refers to the critical score (which can be dynamically adjusted) set based on factors such as audit business compliance standards and the accuracy of historical adaptation data. It is the core criterion for screening high-confidence effective mappings.

[0068] Target confidence mapping results: These refer to the set of policy and audit correlations that, after screening, reach a pre-set confidence threshold and possess high reliability. They are a core component in constructing the final policy and audit impact mapping data.

[0069] In this embodiment of the invention, based on the compliance requirements of audit business and the historical policy-audit adaptation accuracy, a reasonable confidence threshold (such as 80 points, which can be dynamically adjusted according to the actual scenario) is preset. The policy and audit initial mapping relationship with a confidence score greater than or equal to the threshold is selected and determined as the target confidence mapping result, so as to ensure that the subsequent association relationship used to build the policy and audit impact mapping data has high reliability.

[0070] S26. Correct the initial mapping results between policies and audits with confidence levels lower than the preset confidence threshold to obtain complete mapping results.

[0071] Correction: This refers to the process of performing targeted calibration, semantic completion, or related node optimization based on audit domain rules, historical cases, and semantic association logic for initial mapping results whose confidence level does not reach the preset confidence threshold. The core is to improve the accuracy and effectiveness of low-confidence mapping.

[0072] Complete mapping result: refers to the policy-audit mapping set that integrates the target confidence mapping result of S25 with the corrected low confidence mapping result, which covers all valid correlations and meets the requirements for subsequent processing. It is the complete basis for the policy and audit impact mapping data.

[0073] In this embodiment of the invention, the remaining policy and audit initial mapping results with confidence levels lower than the preset confidence threshold after screening are combined with the pre-set association rule base in the audit field, historical high-confidence adaptation cases, and policy-audit semantic association logic to perform targeted calibration and supplementation to address issues such as semantic ambiguity, missing association nodes, and incomplete matching logic in the low-confidence mapping. After the correction process is completed, the results are integrated and summarized with the target confidence mapping results obtained in S25 to finally form a complete mapping result that covers all valid policy-audit associations and meets the standards for accuracy and completeness.

[0074] S27. Merge and deduplicate the target confidence mapping results and the complete mapping results to obtain the policy and audit impact mapping data.

[0075] Merging and deduplication: This refers to the process of matching and comparing core identifiers of a multi-source mapping result set to eliminate duplicate associations and retain unique and valid mappings. The core is to ensure the uniqueness and conciseness of the mapping data.

[0076] Policy and audit impact mapping data: refers to structured data formed after merging and deduplication, which contains all valid and unique policy provisions and audit elements related to each other. It is the core input for subsequent weight quantification and priority determination.

[0077] In this embodiment of the invention, the complete mapping result covering all effective policy-audit relationships (integrating the target confidence mapping result of S25 and the corrected low confidence mapping result) is used to accurately match and compare possible duplicate relationships in the set based on core identifiers such as policy number, audit node, and relationship dimension. Redundant duplicates are eliminated and unique valid mapping relationships are retained. This completes the merging and deduplication process, ultimately yielding policy and audit impact mapping data that is clearly related, free of redundancy, and can directly support subsequent three-dimensional relationship weight quantification and priority determination.

[0078] Step 103: Perform three-dimensional correlation weight quantification and item priority determination on multiple policy-audit impact mapping items within the policy-audit impact mapping data to obtain the corresponding item priorities.

[0079] Furthermore, step 103 may include the following sub-steps: S31. Perform three-dimensional feature extraction on multiple policy-audit impact mapping entries within the policy-audit impact mapping data to obtain the dimensional feature vectors corresponding to the three dimensions.

[0080] In this embodiment of the invention, based on the obtained policy and audit impact mapping data that is clearly associated and free of redundancy, key features are extracted from the three core dimensions of policy, industry, and audit node for multiple policy-audit impact mapping entries and transformed into calculable numerical vectors to form a complete three-dimensional feature vector set.

[0081] Three-dimensional feature extraction: This refers to the process of extracting key features from the policy, industry, and audit node dimensions and quantifying them into vectors for policy-audit impact mapping entries. The core is to provide multi-dimensional feature support for weight quantification.

[0082] Policy dimension feature vector: refers to a numerical vector formed by extracting core features such as policy timeliness and constraint strength and quantifying and encoding them, reflecting the impact attributes of the policy itself on auditing.

[0083] Industry-dimensional feature vector: refers to a numerical vector formed by extracting and quantifying core features such as industry matching degree and policy sensitivity, which reflects the compatibility and correlation attributes between policies and specific industries.

[0084] Audit node dimension feature vector: refers to a numerical vector formed by extracting and quantifying core features such as the relevance and risk level of audit links, which represents the impact attributes of policies on audit process nodes.

[0085] Dimensional feature vector: refers to the standardized numerical vector formed by quantizing and encoding key features in a certain dimension. It has the characteristics of being computable and comparable, and is the core input for weight quantization.

[0086] Furthermore, S31 may include the following sub-steps: S311. Extract policy dimension features from multiple policy-audit impact mapping entries within the policy-audit impact mapping data and construct a policy dimension feature vector.

[0087] In this embodiment of the invention, the core policy attributes corresponding to each policy-audit impact mapping item are focused on, and key features such as policy timeliness, constraint strength, scope of application, and close correlation of clauses are extracted. After standardized quantitative encoding, these features are transformed into a numerical vector of a unified dimension, thus completing the extraction of policy dimension features and constructing a policy dimension feature vector.

[0088] S312. Extract industry-dimensional features from multiple policy-audit impact mapping entries within the policy-audit impact mapping data and construct an industry-dimensional feature vector.

[0089] In this embodiment of the invention, based on the applicable industry information corresponding to each mapping entry, core features such as industry matching fit, industry policy sensitivity coefficient, industry compliance requirement complexity, and industry business association depth are extracted. These features are then converted into standardized numerical vectors through feature quantization processing, thereby realizing the extraction of industry-dimensional features and the construction of industry-dimensional feature vectors.

[0090] S313. Extract audit node dimension features from multiple policy-audit impact mapping entries within the policy-audit impact mapping data, and construct an audit node dimension feature vector.

[0091] In this embodiment of the invention, key features such as the correlation degree of audit links, the influence weight of audit indicators, the intensity of audit process change requirements, and the correlation level of audit risks are extracted around the core nodes of the entire audit business process. These features are then quantified into computable numerical vectors through feature encoding, thereby completing the extraction of audit node dimension features and constructing audit node dimension feature vectors.

[0092] S32. Using the order relation analysis method, the three-dimensional dimensions are weighted according to the order relation to obtain the subjective weight vector of the three-dimensional dimensions corresponding to multiple policy-audit impact mapping items.

[0093] Ordinal relation analysis method: This refers to a method that relies on audit professionals to rank the importance of evaluation dimensions and give relative importance coefficients, and then uses mathematical recursion to calculate the weight of each dimension. The core is to combine expert experience to reflect subjective decision-making preferences.

[0094] Order relation weighting: refers to the process of transforming experts' subjective judgments on the importance of the three dimensions into quantitative weights based on the order relation analysis method, highlighting the subjective decision-making orientation in business scenarios.

[0095] The three-dimensional subjective weight vector refers to a numerical vector containing the subjective weight values ​​of the policy dimension, industry dimension, and audit node dimension, obtained by weighting through order relations. It reflects the expert's subjective perception of the importance of each dimension.

[0096] In this embodiment of the invention, based on the obtained dimensional feature vectors corresponding to the policy dimension, industry dimension, and audit node dimension, and based on the actual needs of policy-audit correlation in audit business scenarios, professional and technical personnel in the audit field compare the importance of the three dimensions pairwise and determine the order relationship, such as policy dimension > audit node dimension > industry dimension, or adjust the sorting logic according to the actual business scenario. Then, they combine industry experience to give the relative importance coefficient between adjacent important dimensions, such as the importance of a more important dimension relative to a less important dimension ranging from 1.0 to 1.8. Then, according to the weight calculation logic of the order relationship analysis method, the subjective weight value of each dimension is recursively calculated through the relative importance coefficient, and finally a three-dimensional subjective weight vector is formed, which corresponds one-to-one with each policy-audit impact mapping item and contains the subjective weight of the policy dimension, the subjective weight of the industry dimension, and the subjective weight of the audit node dimension, providing a subjective decision-making basis for subsequent coupling calculation with the objective weight vector.

[0097] S33. Based on the grey relational analysis method, grey relational analysis is performed using all dimension feature vectors and the pre-set ideal feature vectors of the association to obtain the three-dimensional objective weight vectors corresponding to multiple policy-audit impact mapping items.

[0098] Preset ideal feature vector: refers to the reference vector formed by the combination of "optimal correlation strength" features under various dimensions, which is constructed based on the optimal compliance standards of audit business and the core needs of cross-industry adaptation. It is the benchmark reference sequence for grey relational analysis.

[0099] Grey relational analysis refers to a data analysis method that uses a preset ideal feature vector as a reference and calculates the correlation coefficient between the comparison sequence (feature vectors of each dimension) and the reference sequence to quantify the degree of closeness between the two. Its core is to derive objective weights based on data correlation.

[0100] Three-dimensional objective weight vector: refers to a numerical vector obtained through grey relational analysis, which includes the subjective weight values ​​of the policy dimension, industry dimension, and audit node dimension, reflecting the actual correlation importance of each dimension at the data level.

[0101] In this embodiment of the invention, based on the obtained dimensional feature vectors corresponding to the policy dimension, industry dimension, and audit node dimension, a preset ideal feature vector is first constructed based on the optimal compliance standards for audit business and the core requirements for cross-industry template adaptation. This is a combination of features reflecting the "optimal correlation strength" under each dimension (e.g., for the policy dimension, feature values ​​that are the most timely, have the highest constraint strength, and whose applicable scope and audit scenario are perfectly matched; for the industry dimension, feature values ​​that have the highest matching degree, the best policy sensitivity coefficient, and are adapted to compliance complexity; and for the audit node dimension, feature values ​​that have the strongest correlation, the largest impact weight, and are adapted to risk level). This forms an ideal reference benchmark of the same dimension as the feature vectors of each dimension. Subsequently, grey relational analysis is used. The degree analysis method uses a preset ideal feature vector as a reference sequence. It takes the three-dimensional feature vector of each policy-audit impact mapping item as a comparison sequence, eliminates the difference in feature dimensions through normalization, calculates the gray correlation coefficient between each comparison sequence and the reference sequence (quantifying the closeness of a single feature to the ideal feature), and then derives the objective weight value of each dimension based on the average proportion of the correlation coefficient under each dimension. Finally, it forms a three-dimensional objective weight vector that corresponds one-to-one with each policy-audit impact mapping item and includes objective weights of the policy dimension, industry dimension, and audit node dimension. This provides a data-driven objective basis for subsequent coupling calculation with the subjective weight vector.

[0102] S34. Using the fuzzy comprehensive evaluation method, the subjective weight vector and the objective weight vector of the three-dimensional dimension are weighted and coupled to obtain the three-dimensional comprehensive correlation weights corresponding to multiple policy-audit impact mapping items.

[0103] Fuzzy comprehensive evaluation method: Based on fuzzy mathematics theory, it is an evaluation method that integrates multi-dimensional subjective weights and objective weights according to preset fusion rules to achieve the normalization and coupling of multi-source weight information. Its core is to take into account the dual value of subjective decision-making and objective data.

[0104] Weighted coupling: refers to the process of linearly weighting subjective weight vectors and objective weight vectors by setting a fusion coefficient, so as to achieve complementary fusion of the two types of weight information. The core is to balance the influence weight of subjective experience and objective data.

[0105] Three-dimensional comprehensive correlation weight: refers to the vector set that includes normalized weights of three dimensions: policy, industry, and audit node, obtained after weighted coupling and normalization. It comprehensively quantifies the comprehensive correlation importance of the mapped items in multiple dimensions.

[0106] In this embodiment of the invention, based on the obtained three-dimensional subjective weight vector (denoted as...) ),in Subjective weighting of policy dimensions Subjective weighting for industry dimensions The subjective weights of the audit node dimension and the objective weight vector of the three dimensions obtained by S33 (denoted as...) (corresponding to the objective weights of the three dimensions mentioned above), based on the dual requirements of auditing operations for subjective decision-making preferences and data objectivity, a subjective weight fusion coefficient is first preset. Integration coefficient with objective weights (Constraints are satisfied) , and In this embodiment , (This can be dynamically adjusted according to the actual business scenario); then, a weighted coupling model is constructed using the fuzzy comprehensive evaluation method. The first step is to calculate the preliminary comprehensive weight of each dimension, that is, the weight of each dimension is obtained by weighting the subjective weight and the objective weight according to the fusion coefficient, specifically:

[0107] in, Corresponding policy dimensions Corresponding industry dimensions Corresponding audit node dimensions.

[0108] The second step is to normalize the preliminary combined weights of the three dimensions to eliminate differences in units and ensure the comparability and superposition of the weights. The normalization formula is as follows:

[0109] Finally, the three-dimensional comprehensive correlation weights corresponding to each policy-audit impact mapping item are obtained. This weighting system comprehensively integrates subjective decision-making experience with the correlation characteristics of objective data, providing a quantitative basis for subsequent item priority determination.

[0110] S35. Based on the three-dimensional comprehensive correlation weight, sort the multiple policy-audit impact mapping items in descending order, and determine the item priority corresponding to each policy-audit impact mapping item according to the preset item priority partitioning interval to which the three-dimensional comprehensive correlation weight belongs.

[0111] In this embodiment of the invention, based on the three-dimensional comprehensive correlation weights corresponding to each policy-audit impact mapping item, the comprehensive score of each item is first calculated. Based on the actual impact weight of each dimension on the audit template update, a preset dimension impact coefficient is established. (Constraints are satisfied) In this embodiment (This can be adjusted according to the business scenario). The comprehensive score formula is:

[0112] Subsequently, all policy-audit impact mapping items are sorted in descending order according to their overall scores. Items with higher scores correspond to policies with more critical impacts on the audit template and are ranked higher. Simultaneously, priority zones for these items are preset (this embodiment uses three priority levels, which can be expanded as needed): high priority zone... Medium priority range Low priority interval Finally, the overall score of each item is matched with the preset range to determine the priority of each item (e.g., a score of 0.82 corresponds to high priority, and a score of 0.55 corresponds to medium priority), providing a priority basis for the targeted updates and collaborative configuration of the audit template in the future.

[0113] Step 104: Use policy and audit impact mapping data to retrieve industry templates and determine the initial industry-specific audit templates for each applicable industry based on the item priority.

[0114] Furthermore, step 104 may include the following sub-steps: S41. Extract target field data from the policy and audit impact mapping data as the search key; the target field data includes applicable industries.

[0115] Target field data: refers to the set of key information extracted from the policy and audit impact mapping data, used to retrieve and match audit templates. The core includes the applicable industry, and related fields such as policy number and audit node type can be added as needed.

[0116] Search key: refers to a structured search identifier formed by combining target field data in a preset format. It has industry relevance and core feature recognition, and is used to quickly match the benchmark template in the cross-industry audit template library.

[0117] In this embodiment of the invention, after determining the priority of each policy-audit impact mapping item, the policy-audit impact mapping data (including information such as correlation, three-dimensional comprehensive correlation weight, and item priority) is used. Based on the accuracy requirements of audit template matching, according to the preset field extraction rules, key information including the applicable industry is selected and extracted from each policy-audit impact mapping item as target field data. At the same time, fields strongly related to template matching, such as policy number, audit core node type, and compliance constraint type, can be added (flexibly adjusted according to actual retrieval needs). These extracted target field data are combined in a preset format to form a structured retrieval key, ensuring that the retrieval key not only clearly identifies the applicable industry attributes corresponding to the associated policy, but also accurately carries the core correlation characteristics between the policy and the audit.

[0118] S42. Use the search key to retrieve the pre-set cross-industry basic audit template library to obtain the benchmark audit templates corresponding to each applicable industry.

[0119] Pre-built cross-industry basic audit template library: refers to a database that stores standardized basic audit templates for various industries according to an industry classification system. Each template is associated with index information such as industry identifier, audit node, and applicable policy type, and is the core data source for providing benchmark audit templates.

[0120] Benchmark audit template: refers to the standardized basic audit template for the corresponding applicable industry obtained through retrieval and matching. It includes the industry's routine audit process, indicators and basic compliance requirements. It is the original blueprint for subsequent adjustments and optimizations based on policy impacts.

[0121] In this embodiment of the invention, a pre-built cross-industry basic audit template library is invoked based on a structured search key containing associated fields such as applicable industry and policy number, and audit core node type. This template library stores standardized basic audit templates for each industry in an organized manner according to an industry classification system. Each template is associated with index information such as industry identifier, core audit node, and applicable policy type, supporting accurate matching of multiple fields. A search logic of "industry identifier-priority accurate matching + auxiliary field secondary verification" is adopted. First, the template category of the corresponding industry in the template library is quickly located using the applicable industry field in the search key. Then, the templates under the category are screened for adaptability by combining other target fields such as policy number and compliance constraint type, eliminating templates that do not conform to the core policy-related characteristics. Finally, a benchmark audit template corresponding to each applicable industry that can cover the industry's conventional audit processes, indicators, and basic compliance requirements is obtained.

[0122] S43. Perform integrity verification on each benchmark audit template to obtain the verification benchmark audit template corresponding to each applicable industry.

[0123] Completeness verification process: This refers to the process of verifying the content coverage, completeness of core elements, and format standardization of the benchmark audit template based on relevant policy provisions, industry compliance standards, and a list of core audit elements, and then supplementing and standardizing it. The core is to ensure that the template meets the needs of policy adaptation and audit practice.

[0124] Verification benchmark audit template: refers to the benchmark audit template that has been completed and standardized in terms of completeness verification, format, and adaptability to policy impacts. It is the basic version for subsequent targeted optimization.

[0125] In this embodiment of the invention, based on the core related clauses, industry compliance standards, and core elements of the audit process in the policy-audit impact mapping data, a completeness verification process is carried out. This involves checking whether the benchmark audit template covers the audit requirements corresponding to all high / medium priority policy-audit impact mapping items, and whether it lacks key audit nodes, indicator definitions, compliance judgment rules, and other core content. Simultaneously, it verifies whether the template format conforms to the unified cross-industry template standard. For templates with missing content, the corresponding policy-required audit elements are automatically added. Templates with inconsistent formats are standardized and adjusted. The result is a benchmark audit template that covers policy impact requirements, has complete elements, conforms to the format, and can be directly used for subsequent optimization in various applicable industries. This provides a high-quality basic template for targeted updates based on priority items.

[0126] S44. Match the corresponding adjustment strategy according to the item priority, and adjust the audit template of each verification benchmark according to the adjustment strategy to obtain the initial industry-specific audit template corresponding to each applicable industry.

[0127] Adjustment strategy: refers to template optimization rules (including mandatory adjustment, on-demand adjustment, optional adjustment, etc.) that are tied to the priority of items and contain specific operational details. It clarifies the adjustment actions, the target objects, and the adaptation standards to ensure that the adjustments can be implemented.

[0128] Initial industry-specific audit template: This refers to an audit template that has been differentiated and adjusted to meet industry policy requirements, incorporate policy impacts, retain the core industry audit framework, and whose adjustment trajectory is traceable. It is a basic and optimized version of the industry-specific template.

[0129] In this embodiment of the invention, based on the determined priority (high, medium, and low levels) of each policy-audit impact mapping item and the complete and formatted verification benchmark audit template obtained in S43, the preset priority-adjustment strategy-operation details three-dimensional mapping rule library is first invoked (clearly defining the specific adjustment actions, operation objects, and adaptation standards corresponding to different priorities): among which high-priority items are bound to the "forced adjustment + priority adaptation" strategy, and the operation details are to directly add / modify the core audit elements of the template—for audit-related nodes in the policy-audit impact mapping items, embedding the special audit steps required by the policy in the corresponding process links of the template, and synchronously updating the definition, calculation logic, and compliance judgment threshold of the core audit indicators (e.g., if the policy clearly states that "the R&D expense deduction ratio is increased to 75%", then the corresponding indicator is directly modified). The template is divided into several parts: 1) Calculation rules, with policy clause numbers and effective dates marked; 2) Medium-priority items are bound to the "adjust as needed + optimization and adaptation" strategy, with the operational details being to optimize the adaptability of existing elements in the template—adjusting the weighting of non-core audit indicators based on industry audit practice scenarios (e.g., if the policy strengthens "environmental compliance" requirements, the scoring proportion of environmentally related indicators can be increased), and supplementing the detailed audit guidelines corresponding to the policy (e.g., clarifying the scope of material verification and evidence retention requirements), without changing the core framework of the template; 3) Low-priority items are bound to the "optional adjustment + supplementary adaptation" strategy, with the operational details being to only supplement the policy association explanation in the template's notes column—listing the policy name, applicable scenarios, and weak correlation points with the audit template (e.g., if the policy adds "small and micro enterprise support clauses," only indicate the reference significance of this clause for industry audits at the end of the template, without adjusting the audit process and indicators). Subsequently, following the order of "high-priority items are processed first → medium-priority items are adapted second best → low-priority items are supplemented with explanations", the specific related clauses of each policy-audit impact mapping item are matched one by one. Targeted adjustments are made to the verification benchmark audit templates of each applicable industry, and each adjustment action is associated with a unique identifier of the corresponding mapping item (for easy traceability). Finally, initial industry-specific audit templates are obtained for each applicable industry that are adapted to the corresponding industry policy needs, integrate the impact of policies of different priorities, retain the core framework of industry audit, and have traceable adjustment trajectories.

[0130] Step 105: Based on the pre-built industry knowledge graph, integrate and comply with the cross-industry templates of each initial industry-specific audit template, and determine the cross-industry audit templates corresponding to each applicable industry.

[0131] Furthermore, step 105 may include the following sub-steps: S51. Based on the pre-set industry knowledge graph, the structure of each initial industry-specific audit template is made consistent to obtain the intermediate industry-specific audit template corresponding to each applicable industry.

[0132] The pre-built industry knowledge graph is a structured semantic knowledge base built to support the standardized and differentiated adaptation of cross-industry audit templates. Its core stores standardized and specific information related to the audit domain and multiple industries in the form of "entity-relationship-attribute" triples, specifically as follows: Core entity types: covering national economic industry classifications (such as manufacturing, finance, information technology, etc.), core audit modules (audit preparation, implementation, reporting, archiving), cross-industry unified field standards (such as audit indicator naming conventions, process node coding rules), industry-specific audit elements (such as "capital adequacy audit items" for the financial industry and "production process compliance audit items" for the manufacturing industry), policy adaptation types (such as tax, environmental protection, and regulatory policies), and template format specifications (such as table styles and mandatory field rules). Key relationships include: the correspondence between "industry-specific audit elements" (e.g., retail industry → merchandise inventory audit module), the mapping between "cross-industry unified fields and industry-specific fields" (e.g., "compliance judgment threshold" → "capital adequacy ratio threshold" in the financial industry / "environmental emission threshold" in the manufacturing industry), the relationship between "audit core modules and industry adaptation priority" (e.g., the "risk assessment module" in the financial industry has higher priority than ordinary industries), and the adaptation between "policy type and audit elements" (e.g., environmental protection policy → pollution discharge compliance audit item in the manufacturing industry). Core attribute information: Each entity is accompanied by specific attributes (such as industry entities including policy level and audit process complexity attributes, and audit module entities including core field list and format specification version attributes).

[0133] Structural Consistency: This refers to the process of completing modules, unifying fields, streamlining processes, and standardizing formats of initial industry-specific audit templates based on cross-industry unified audit structure specifications in industry knowledge graphs. The core is to achieve cross-industry universality and standardization of the template structure while retaining industry-specific core elements.

[0134] Intermediate Industry-Specific Audit Template: This refers to an audit template that has been standardized in structure, with complete modules, unified fields, and standardized processes. It takes into account both industry characteristics and cross-industry applicability and serves as a standardized basic version for subsequent cross-industry collaborative configuration.

[0135] In this embodiment of the invention, a pre-built industry knowledge graph is invoked. This pre-built industry knowledge graph includes cross-industry unified audit structure specifications (such as core module division, field naming standards, process node sorting rules, format uniformity requirements, etc.) and mapping relationships of industry-specific elements. First, the existing structure of each initial industry-specific audit template is compared with the unified specifications in the knowledge graph to check whether the template has problems such as missing core modules (such as audit preparation, implementation, reporting, and archiving), inconsistent field naming, disordered process node order, and format inconsistencies with cross-industry standards. Then, targeted adjustments are made according to the unified specifications: missing core audit modules are supplemented, industry-specific fields are named according to the specifications and mapped to the cross-industry unified field system, and the process node order is adjusted to conform to general audit logic, while retaining the core business elements such as industry-specific audit indicators and judgment rules (without changing policy adaptation and industry adaptability). Finally, the structural consistency processing is completed to obtain intermediate industry-specific audit templates for each applicable industry that are complete in modules, unified in fields, standardized in processes, and take into account both industry characteristics and cross-industry universality.

[0136] S52. Mark the audit templates for each intermediate industry to obtain the target industry-specific audit templates for each applicable industry.

[0137] Labeling: This refers to the process of adding core attribute tags such as unique identifiers, industry information, policy associations, and version numbers to the audit templates for intermediate industries according to the system's preset rules. The core is to make the template attributes clear and traceable.

[0138] Industry-specific audit templates: These are industry-specific audit templates that have been annotated, have complete annotation information, can be quickly identified as core attributes, and are easy to trace and manage. They are the final template versions output by the system that can be directly used for auditing practices.

[0139] In this embodiment of the invention, each intermediate industry-specific audit template is standardized and annotated according to the system's preset annotation rules. The core annotation content includes the template's unique identifier ID, the full name and code of the applicable industry, the associated policy number and effective date, the priority distribution of the corresponding policy-audit impact mapping items (e.g., "3 high-priority items, 5 medium-priority items"), the template version number (associated with the version control logic of the audit template generation and version management module), the structure consistency processing identifier, and the annotation timestamp. The annotation position is uniformly placed at the top of the template's homepage and at the beginning of the core module, using a standardized label format to ensure that the annotation information is intuitively visible and does not interfere with the actual use of auditing. After the annotation is completed, the target industry-specific audit templates corresponding to each applicable industry are obtained with complete annotation information, quick identification of core attributes, and ease of version traceability and compliance review, providing clear attribute identification support for subsequent template push, feedback collection, and cross-industry collaborative management.

[0140] S53. Conduct compliance verification on the industry-specific audit templates for each target industry, and use the industry-specific audit templates that pass the compliance verification as cross-industry audit templates.

[0141] Compliance verification: This refers to the process of verifying the policy adaptability, compliance, and traceability of changes of industry-specific audit templates by combining policy compliance standard library, industry regulatory norms, general auditing standards, and template change traceability records. The core is to ensure that the templates meet regulatory requirements and audit practice norms.

[0142] Cross-industry audit templates: These are standardized templates that have passed compliance verification, possess industry adaptability, compliance, and traceability, and can be directly used for audit practices in the corresponding industry. They are the final cross-industry applicable audit results output by the system.

[0143] In this embodiment of the invention, the entire process record of template changes (including policy-related basis, adjustment trajectory, annotation information, etc.) stored in the log recording and audit trail module is invoked. Combined with the pre-set policy compliance standard library (covering the original clauses of related policies, effective time and interpretation details), regulatory norms of various industries and general auditing standards, compliance verification is carried out on the audit templates of each target industry. The key checks are whether the audit requirements in the template are fully consistent with the related policy clauses, whether they meet the compliance bottom line of the corresponding industry, and whether the audit indicator settings and judgment logic violate general auditing standards. At the same time, it is verified whether the template change content has a clear policy basis and the change trajectory is traceable, with no compliance conflicts or policy adaptation omissions. The target industry-specific audit templates that have been verified to be fully compliant with policy requirements, industry norms and auditing standards, and whose changes are traceable, are identified as cross-industry audit templates, ensuring that they can be directly used for audit practice in the corresponding industry and meet compliance review requirements.

[0144] S54. Analyze the compliance issue list for the target industry-specific audit template that failed the compliance verification, and then proceed to execute the steps of matching the corresponding adjustment strategy according to the priority of the items.

[0145] Compliance Issues Checklist: This refers to a structured checklist compiled for templates that failed compliance verification. It includes the issue type, compliance basis, related item identifiers and priorities, and the cause of the issue. It serves as the core basis for subsequent targeted adjustments and rectification.

[0146] Jump execution: This refers to the system triggering a process linkage mechanism, using the compliance issue list as supplementary basis, returning to the "matching the corresponding adjustment strategy according to the priority of the item" step to re-execute the template adjustment process. The core is to achieve closed-loop rectification of compliance issues.

[0147] In this embodiment of the invention, for the industry-specific audit template that fails the compliance verification, a structured compliance issue list is first compiled by combining the verification logs recorded during the compliance verification process (including specific issues such as policy adaptation deviations, conflicts with industry regulatory norms, violations of audit standards, and untraceable change trajectories). The list clearly marks the type of each compliance issue (such as inconsistencies in policy clause adaptation, audit indicators not meeting industry regulatory requirements, etc.), the corresponding related policy clauses / regulatory norms, the policy-audit impact mapping item identifiers and priorities involved, and briefly explains the cause of the problem (such as high-priority items not fully conforming to the original policy when adjusted). (The system addresses issues such as wording discrepancies, conflicts between optimization logic for medium-priority items and industry standards). Subsequently, the system automatically triggers a process redirection mechanism, using the compliance issue list as supplementary adjustment basis, and redirects to step S44, "Matching the corresponding adjustment strategy based on item priority." When matching the adjustment strategy, compliance issue rectification requirements are simultaneously integrated (e.g., for high-priority items with policy adaptation deviations, strengthening the "word-by-word verification of policy clauses" step in the adjustment strategy; for medium-priority items with industry standard conflicts, optimizing the industry standard adaptation verification logic of the adaptation strategy). This ensures that subsequent adjustments to the verification benchmark audit template accurately resolve compliance issues and promote the template's compliance verification.

[0148] Please see Figure 2 , Figure 2 This is a structural block diagram of a cross-industry audit template dynamic collaborative system provided in an embodiment of the present invention.

[0149] This invention provides a cross-industry audit template dynamic collaborative system, comprising: The policy collection and parsing module 201 is used to respond to audit template update requests, obtain multiple raw policy data in different formats, and perform structured processing to obtain structured policy data. The policy semantic understanding and impact analysis module 202 is used to perform semantic mapping on structured policy data to obtain policy and audit impact mapping data; The weight quantification and priority determination module 203 is used to perform three-dimensional correlation weight quantification and item priority determination on multiple policy-audit impact mapping items in the policy-audit impact mapping data to obtain the corresponding item priority. The audit template generation and management module 204 is used to retrieve industry templates using policy and audit impact mapping data, and determine the initial industry-specific audit templates for each applicable industry based on the priority of the entries. The cross-industry template collaborative configuration module 205 is used to integrate and comply with the initial industry-specific audit templates based on the pre-set industry knowledge graph, and to determine the cross-industry audit templates corresponding to each applicable industry.

[0150] Structured processing refers to the process of extracting key elements (such as policy clauses, metadata, audit nodes, etc.) from unstructured / semi-structured policy texts, audit-related information, etc., according to preset specifications, and transforming them into a structured data format with clear modules, unified fields, and computability and reusability. It is mainly used for policy parsing and template standardization to support subsequent automated processing by the system.

[0151] Semantic mapping: refers to the process of establishing a precise correspondence between the semantics of policy provisions and audit elements (audit processes, indicators, compliance rules, etc.) based on the policy semantic understanding capability of Large Language Model (LLM). The core is to identify the specific impact of policies on audits and provide a semantic matching basis for template adjustment.

[0152] Three-dimensional correlation weight quantification and item priority determination: For policy-audit impact mapping items, features are extracted and quantified into dimensional feature vectors from policy, industry, and audit node dimensions. The three-dimensional comprehensive correlation weight is calculated by coupling the order relation analysis method (subjective weight) and the grey relational analysis method (objective weight). Then, based on the comprehensive score ranking and preset interval, the importance level (high / medium / low) of the item is determined. The core is to realize the quantitative classification of policy impact and guide the priority of template adjustment.

[0153] Industry template retrieval: This refers to the process of locating and filtering the corresponding industry benchmark audit template from a pre-built cross-industry basic audit template library based on a structured search key containing key information such as applicable industry, policy number, and audit node type, using the logic of "industry identifier priority matching + auxiliary field verification". The core is to quickly obtain standardized template blueprints that meet the basic needs of the industry.

[0154] Cross-industry template integration and compliance: This refers to the process of unifying the structure of benchmark audit templates from different industries based on industry knowledge graphs (unifying modules, fields, and process specifications while retaining industry characteristics), making differentiated adjustments based on policy impact items, and ensuring the templates are compliant and usable through compliance verification (checking policy adaptability, industry standard compliance, and change traceability). The core is to achieve unified management and compliance adaptation of cross-industry templates.

[0155] Since the above is a system corresponding to a cross-industry audit template dynamic collaboration method, its implementation principle is the same as that of a cross-industry audit template dynamic collaboration method. For the sake of convenience and brevity, those skilled in the art can clearly understand that the specific working process of the system and modules described above can be referred to the corresponding process in the aforementioned method embodiments, and will not be repeated here.

[0156] Please see Figure 3 , Figure 3 This is a structural block diagram of an electronic device provided in an embodiment of the present invention.

[0157] An electronic device according to an embodiment of the present invention includes: a memory 301 and a processor 302. The memory 301 stores a computer program. When the computer program is executed by the processor 302, the processor 302 executes the cross-industry audit template dynamic collaboration method as described in the above embodiment.

[0158] Memory 301 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. Memory 301 has storage space 303 for program code 313 for performing any of the method steps described above. For example, storage space 303 for program code may include various program codes 313 for implementing the various steps in the methods described above. These program codes may be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, CDs, memory cards, or floppy disks. The program code may be compressed, for example, in a suitable form. When run by a computing processing device, this code causes the computing processing device to perform the various steps in the methods described above. These program codes may be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, CDs, memory cards, or floppy disks. The program code may be compressed, for example, in a suitable form. When this code is run by a computing device, it causes the computing device to execute the various steps in the cross-industry audit template dynamic collaboration method described above.

[0159] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0160] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0161] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0162] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0163] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0164] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions 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 the present invention.

Claims

1. A cross-industry audit template dynamic collaboration method, characterized in that, include: In response to audit template update requests, the system obtains raw policy data in multiple different formats and performs structured processing to obtain structured policy data. Semantic mapping is performed on the structured policy data to obtain policy-audit impact mapping data; Three-dimensional correlation weight quantification and item priority determination are performed on multiple policy-audit impact mapping items within the policy-audit impact mapping data to obtain the corresponding item priorities; The policy and audit impact mapping data are used to retrieve industry templates, and the initial industry-specific audit templates for each applicable industry are determined by combining the item priority. Based on the pre-built industry knowledge graph, the initial industry-specific audit templates are integrated and made compliant across industries to determine the cross-industry audit templates corresponding to each applicable industry.

2. The cross-industry audit template dynamic collaboration method according to claim 1, characterized in that, The response to the audit template update request obtains multiple raw policy data in different formats and performs structured processing to obtain structured policy data, including: Responding to audit template update requests, it retrieves raw policy data in multiple different formats; The original data of each policy were standardized to obtain multiple plain text data. All the plain text data are encoded to obtain standardized policy text data; The standardized policy text data is subjected to regular expression extraction to obtain core metadata; Key clauses were obtained by performing keyword retrieval and clause classification on the standardized policy text data; The core metadata is mapped to the key terms to obtain structured policy data.

3. The cross-industry audit template dynamic collaboration method according to claim 1, characterized in that, The semantic mapping of the structured policy data to obtain policy-audit impact mapping data includes: The structured policy data is filtered to obtain core corpus data; The core corpus data is normalized using synonyms to obtain standardized core corpus data; The standardized core corpus data is semantically parsed using a pre-set fine-tuned large language model to obtain the initial mapping results between policy and audit. A confidence assessment was conducted on the initial mapping results between the stated policy and the audit. The initial mapping results between policies and audits with a confidence level greater than or equal to a pre-set confidence threshold are selected as the target confidence level mapping results; The initial mapping results between policies and audits with confidence levels lower than the preset confidence threshold are corrected to obtain complete mapping results; The target confidence mapping result and the complete mapping result are merged and deduplicated to obtain policy and audit impact mapping data.

4. The cross-industry audit template dynamic collaboration method according to claim 1, characterized in that, The process involves performing three-dimensional correlation weight quantification and item priority determination on multiple policy-audit impact mapping items within the policy-audit impact mapping data to obtain the corresponding item priorities, including: Three-dimensional feature extraction is performed on multiple policy-audit impact mapping entries within the policy-audit impact mapping data to obtain the dimensional feature vectors corresponding to the three dimensions. The three-dimensional dimensions are weighted using the ordinal relation analysis method to obtain the subjective weight vectors of the three-dimensional dimensions corresponding to the multiple policy-audit impact mapping items. Based on the grey relational analysis method, grey relational analysis is performed using all the dimensional feature vectors and the associated preset ideal feature vectors to obtain the three-dimensional objective weight vectors corresponding to multiple policy-audit impact mapping items. Using the fuzzy comprehensive evaluation method, the subjective weight vector and the objective weight vector of the three-dimensional dimension are weighted and coupled to obtain the three-dimensional comprehensive correlation weights corresponding to multiple policy-audit impact mapping items; Based on the three-dimensional comprehensive correlation weight, the multiple policy-audit impact mapping items are sorted in descending order, and the item priority corresponding to each policy-audit impact mapping item is determined according to the preset item priority partitioning interval to which the three-dimensional comprehensive correlation weight belongs.

5. The cross-industry audit template dynamic collaboration method according to claim 4, characterized in that, The step of extracting three-dimensional features from multiple policy-audit impact mapping entries within the policy-audit impact mapping data to obtain dimensional feature vectors corresponding to the three dimensions includes: Policy dimension features are extracted from multiple policy-audit impact mapping entries within the policy-audit impact mapping data to construct a policy dimension feature vector; Industry-dimensional features are extracted from multiple policy-audit impact mapping entries within the policy-audit impact mapping data to construct an industry-dimensional feature vector; Audit node dimension features are extracted from multiple policy-audit impact mapping entries within the policy-audit impact mapping data to construct an audit node dimension feature vector.

6. The cross-industry audit template dynamic collaboration method according to claim 1, characterized in that, The process of using the policy and audit impact mapping data to perform industry template retrieval, and determining the initial industry-specific audit template for each applicable industry based on the item priority, includes: Target field data is extracted from the policy and audit impact mapping data and used as the search key; the target field data includes applicable industries; The search key is used to retrieve a pre-set cross-industry basic audit template library to obtain the benchmark audit templates corresponding to each applicable industry. Perform integrity verification on each of the aforementioned benchmark audit templates to obtain the verification benchmark audit templates corresponding to each of the aforementioned applicable industries; The corresponding adjustment strategy is matched according to the priority of the item, and the audit templates of each verification benchmark are adjusted according to the adjustment strategy to obtain the initial industry-specific audit templates corresponding to each applicable industry.

7. The cross-industry audit template dynamic collaboration method according to any one of claims 1-6, characterized in that, The process involves integrating and standardizing the initial industry-specific audit templates based on a pre-built industry knowledge graph, thereby determining the cross-industry audit templates corresponding to each applicable industry. This includes: Based on the pre-built industry knowledge graph, the structure of each initial industry-specific audit template is made consistent to obtain the intermediate industry-specific audit template corresponding to each applicable industry. The audit templates for each of the aforementioned intermediate industries are annotated to obtain the target industry-specific audit templates for each of the applicable industries. Each target industry-specific audit template undergoes compliance verification, and the target industry-specific audit templates that pass the compliance verification are used as cross-industry audit templates.

8. The cross-industry audit template dynamic collaboration method according to claim 6, characterized in that, Also includes: Analyze the compliance issue list for the target industry-specific audit template that failed the compliance verification, and then proceed to execute the step of matching the corresponding adjustment strategy according to the priority of the item.

9. A cross-industry audit template dynamic collaborative system, characterized in that, include: The policy collection and parsing module is used to respond to audit template update requests, acquire multiple raw policy data in different formats, and perform structured processing to obtain structured policy data. The policy semantic understanding and impact analysis module is used to perform semantic mapping on the structured policy data to obtain policy and audit impact mapping data; The weight quantification and priority determination module is used to perform three-dimensional correlation weight quantification and item priority determination on multiple policy-audit impact mapping items in the policy-audit impact mapping data to obtain the corresponding item priority. The audit template generation and management module is used to retrieve industry templates using the policy and audit impact mapping data, and determine the initial industry-specific audit templates for each applicable industry based on the item priority. The cross-industry template collaborative configuration module is used to integrate and comply with the initial industry-specific audit templates based on the pre-set industry knowledge graph, and to determine the cross-industry audit templates corresponding to each applicable industry.

10. An electronic device, characterized in that, The system includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the cross-industry audit template dynamic collaboration method as described in any one of claims 1-8.

Images