Data mining method and system for group standard compliance assessment

Abstract

This invention discloses a data mining method and system for compliance assessment of group standards, relating to the field of data mining technology. The method includes: collecting related data of group standards, performing format governance and clause location, and generating a basic dataset of evidence for group standard clauses; extracting the clause normative strength and clause evidence requirements from the basic dataset of evidence for group standard clauses, and differentially configuring the clause evidence requirements according to the clause normative strength to generate a clause normative strength evidence requirement table; extracting corresponding evidence from the basic dataset of evidence for group standard clauses based on the clause normative strength evidence requirement table, and verifying the coverage status, evidence validity status, and object consistency status between the corresponding evidence and the clause evidence requirements to generate a clause evidence coverage matching set. This invention improves the traceability, stability, and relevance of dynamic compliance assessment reports for group standards.

Description

Technical Field

[0001] This invention relates to the field of data mining technology, and in particular to data mining methods and systems for assessing compliance with group standards. Background Technology

[0002] As the number of group standards continues to grow, standard text parsing, evidence data governance, compliance risk assessment, and data mining technologies are gradually merging. These technologies have evolved from manual retrieval to structured processing of clauses, evidence association modeling, and dynamic report generation, providing a technological foundation for the digital assessment of the implementation status of group standards.

[0003] In the relevant compliance assessment process, the evidence corresponding to the clauses is usually based on positive matching. The evaluation focuses on whether there are supporting materials. There is little reverse backtracking on evidence of abnormal detection, time conflict, inconsistency of objects, and unclosed rectification under the same clause. This can easily lead to insufficient basis for judging the risk status of the clauses and the credibility of the evidence. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a data mining method for assessing the compliance of group standards to address the problem of insufficient credibility of compliance conclusions caused by one-way matching of evidence for group standard clauses.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: Firstly, this invention provides a data mining method for assessing the compliance of group standards. The method includes: collecting related data of group standards, performing format governance and clause location, and generating a basic dataset of evidence for group standard clauses; extracting the clause normative strength and clause evidence requirements from the basic dataset of evidence for group standard clauses, and configuring the clause evidence requirements differently according to the clause normative strength, generating a table of clause normative strength evidence requirements; extracting corresponding evidence from the basic dataset of evidence for group standard clauses based on the table of clause normative strength evidence requirements, and verifying the coverage status, validity status, and object consistency status between the corresponding evidence and the clause evidence requirements, generating a clause evidence coverage matching set; performing reverse evidence retrieval on the clause evidence coverage matching set, extracting evidence content that has abnormal indicators, time conflicts, object inconsistencies, and unresolved rectification issues with the corresponding evidence, and performing conflict judgment between the corresponding evidence and the reverse evidence, generating a positive and negative evidence conflict verification chain; determining the clause risk status and evidence credibility status based on the positive and negative evidence conflict verification chain, and classifying rectification levels and review requirements according to the degree of evidence conflict in the positive and negative evidence conflict verification chain, generating a dynamic assessment report of group standard compliance.

[0007] As a preferred embodiment of the data mining method for assessing compliance with group standards as described in this invention, the generation of the basic dataset of evidence for group standard clauses specifically includes: Collect group standard text files, compilation instructions, referenced document status information, implementation evidence files, testing record files, and rectification record files, and uniformly mark them to generate group standard related data; The data associated with group standards is converted in format, restored in layout, cleaned in fields, and located in clauses. The standard number, standard name, chapter title, clause number, clause text position, reference document position, implementation evidence position, test record position, and rectification record position are marked to generate group standard clause location data. The clause numbers, clause text locations, reference document locations, implementation evidence locations, inspection record locations, and rectification record locations in the group standard clause location data are grouped together using the same clause number. Based on each location, the clause text content, reference document content, implementation evidence content, inspection record content, and rectification record content are extracted to generate the group standard clause evidence base dataset.

[0008] As a preferred embodiment of the data mining method for assessing the compliance of group standards described in this invention, the generation of the clause specification strength evidence requirement table specifically includes, Extract the text content, source of evidence, time of evidence, and object of evidence fields of the same clause number from the group standard clause evidence base dataset, and perform normative word marking and obligation action extraction on the text content of the clause to obtain the normative strength of the clause; Based on the strength of the clauses, the content of the clause text, and the source of evidence under the same clause number, the testing requirements, supporting material requirements, implementation record requirements, and review requirements are correspondingly categorized to obtain the clause evidence requirements. Based on the strength of the clause's normative provisions, the evidentiary requirements for each clause are configured differently, generating a table of evidentiary requirements for the strength of the clause's normative provisions.

[0009] As a preferred embodiment of the data mining method for assessing compliance with group standards as described in this invention, the generation of the clause evidence coverage matching set specifically includes, Extract the evidence requirements for the same clause number from the table of evidence requirements for the strength of the clause, and organize the evidence extraction conditions according to the evidence type, evidence field, evidence time and evidence object to generate the clause evidence extraction rules. According to the rules for extracting evidence in the clauses, locate the corresponding implementation evidence content, test record content, referenced document content, supporting material content, rectification record content, and review record content in the basic dataset of evidence for the clauses of the group standard, and group the implementation evidence content, test record content, referenced document content, supporting material content, rectification record content, and review record content into the same clause number to generate evidence corresponding to the clause; The evidence requirements for the corresponding clauses are verified against the clause evidence requirements in the clause strength evidence requirement table. The verification includes coverage status, evidence validity status, and object consistency status, generating a clause evidence coverage matching set.

[0010] As a preferred embodiment of the data mining method for compliance assessment of group standards described in this invention, the extraction of evidence that exhibits abnormal indicators, time conflicts, inconsistencies in objects, and incomplete rectification loops specifically includes: Extract the corresponding evidence and verification results of the clauses in the clause evidence coverage matching set, and transform the content that does not meet the verification conditions into reverse retrieval trigger items to generate reverse evidence retrieval conditions. Based on the reverse evidence retrieval criteria, the evidence content under the same clause number is traced back from the group standard clause evidence base dataset, and the associated records of the same evidence object under different clause numbers are extracted to generate candidate reverse evidence data. The candidate reverse evidence data is checked for anomalies in indicators, time conflicts, consistency of objects, and rectification closure loop. The evidence content with anomalies is classified into the same clause number to generate reverse evidence.

[0011] As a preferred embodiment of the data mining method for assessing the compliance of group standards described in this invention, the content that does not meet the verification conditions specifically includes: Extract the corresponding evidence for the same clause number from the clause evidence coverage matching set, as well as the verification content of coverage status verification, evidence validity status verification and object consistency status verification, and generate clause verification content collection data. The corresponding evidence in the data collection of clause verification content is compared with the evidence requirements of the clause in the table of evidence requirements for the strength of clause specifications, and the verification content with insufficient evidence coverage, unsatisfactory evidence validity status and inconsistent evidence objects is marked to generate abnormal verification mark data. The abnormal verification marker data is collected according to the clause number, the corresponding evidence for the clause, the abnormality type, and the location of the abnormality source, and the content that does not meet the verification conditions is generated.

[0012] As a preferred embodiment of the data mining method for assessing the compliance of group standards described in this invention, the generation of the conflict verification chain of positive and negative evidence specifically includes: Using clause number and evidence object as the association key, the evidence corresponding to the clause is paired with the reverse evidence within the same clause and across clauses for the same evidence object, generating comparative data of positive and negative evidence; Conflict verification is conducted on the indicator values, evidence time, evidence objects and rectification closure status in the data comparing positive and negative evidence, and the conflict impact level is divided based on the evidence requirement table of the strength of the clauses and norms, and the degree of evidence conflict is generated. The degree of evidence conflict is chained together with the corresponding clause number, the corresponding evidence for the clause, and the reverse evidence to generate a chain for verifying the conflict between positive and negative evidence.

[0013] As a preferred embodiment of the data mining method for assessing the compliance of group standards described in this invention, the determination of the risk status and the credibility status of the evidence specifically includes: Extract clause number, clause normative strength, degree of evidence conflict, corresponding evidence and reverse evidence from the chain of conflict verification between positive and negative evidence, and merge and organize them according to the same clause number to generate clause conflict merging data; Based on the degree of evidence conflict and the strength of clause norms in the clause conflict consolidation data, the index deviation status, time conflict status, object conflict status and rectification closure status are classified and judged to generate clause risk status. Using the risk status of the clauses as the verification entry point, the source credibility, time validity, object consistency and conflict admissibility of the corresponding evidence and reverse evidence in the clause conflict merging data are verified to generate the credibility status of the evidence.

[0014] As a preferred embodiment of the data mining method for assessing the compliance of group standards according to the present invention, the generation of a dynamic assessment report for the compliance of group standards specifically includes: Based on the degree of conflict of evidence, the risk status of the clause, and the credibility status of the evidence, the clauses to be rectified under the same clause number are merged and sorted to generate clause rectification sorting data. The rectification data of the clauses is sorted into rectification levels, and the frequency of review, review objects, scope of review evidence and review pass conditions are configured according to the rectification level to generate rectification review configuration data; The rectification level and review requirements in the rectification and review configuration data are organized into reports and written with the corresponding clause number, clause risk status, evidence credibility status and evidence conflict degree, generating a dynamic assessment report on the compliance of group standards.

[0015] Secondly, this invention provides a data mining system for assessing the compliance of group standards, comprising: a clause database construction module, used to collect relevant data of group standards, perform format governance and clause location, and generate a basic dataset of group standard clause evidence; a strength configuration module, used to extract the clause normative strength and clause evidence requirements from the basic dataset of group standard clause evidence, and to differentiate the clause evidence requirements according to the clause normative strength, generating a clause normative strength evidence requirement table; and an evidence verification module, used to extract the corresponding evidence from the basic dataset of group standard clause evidence according to the clause normative strength evidence requirement table, and to verify the corresponding evidence against the required evidence. The system verifies the coverage status, evidence validity status, and object consistency status of the evidence, generating a clause evidence coverage matching set. The counter-evidence chaining module performs reverse evidence retrieval on the clause evidence coverage matching set, extracting evidence content that has abnormal indicators, time conflicts, object inconsistencies, or unresolved rectification issues with the corresponding clause evidence, and determines the conflict between the corresponding clause evidence and the counter-evidence, generating a positive and negative evidence conflict verification chain. The dynamic evaluation module determines the clause risk status and evidence credibility status based on the positive and negative evidence conflict verification chain, and classifies the rectification level and review requirements according to the degree of evidence conflict in the positive and negative evidence conflict verification chain, generating a group standard compliance dynamic evaluation report.

[0016] The beneficial effects of this invention are as follows: By performing reverse evidence retrieval on the matching set of evidence coverage for clauses and determining the conflict between the corresponding evidence and the reverse evidence, evidence such as abnormal indicators, time conflicts, inconsistent objects, and incomplete rectification can be included in the verification chain under the same clause number. This ensures that the compliance judgment of each clause has both supporting and negative verification basis. The generated positive and negative evidence conflict verification chain can continuously carry the corresponding evidence, reverse evidence, and degree of evidence conflict for the clauses, providing a unified data foundation for determining the risk status of clauses, verifying the credibility of evidence, classifying rectification levels, and configuring review requirements. This improves the traceability, stability, and relevance of the dynamic compliance assessment report for group standards. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. 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.

[0018] Figure 1 A flowchart of a data mining method for assessing compliance with group standards.

[0019] Figure 2 A schematic diagram of a data mining system for assessing compliance with group standards.

[0020] Figure 3 A flowchart for generating the matching set of evidence for the clauses.

[0021] Figure 4 A flowchart for generating a dynamic assessment report on compliance with group standards. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a data mining method for assessing compliance with group standards, comprising the following steps: S1. Collect data related to group standards, perform format governance and clause location, and generate a basic dataset of evidence for group standard clauses.

[0026] S1.1 Collect group standard text files, compilation instructions, referenced document status information, implementation evidence files, testing record files, and rectification record files, and uniformly mark them to generate group standard related data.

[0027] Specifically, the system collects group standard text files, compilation instructions, referenced document status information, implementation evidence files, testing record files, and rectification record files from the National Standards Information Public Service Platform, group standard release platform, social organization public pages, standard compilation database, referenced document status database, implementing unit evidence storage directory, testing institution record interface, and rectification record ledger using web page reading, batch file import, database field extraction, and record interface synchronization methods. Within the same collection batch, the system marks the group standard text files, compilation instructions, and referenced document status information with standard number, standard name, referenced document number, and version status. Through file metadata, table field names, text keywords, and interface field mapping tables, the system extracts evidence source fields, evidence time fields, evidence object fields, indicator fields, evidence conclusion fields, and rectification status fields from the implementation evidence files, testing record files, and rectification record files. The extracted fields are then bound to the corresponding file locations to generate group standard associated data.

[0028] S1.2 Perform format conversion, layout restoration, field cleaning, and clause positioning on the group standard-related data, and mark the standard number, standard name, chapter title, clause number, clause text position, reference document position, implementation evidence position, test record position, and rectification record position to generate group standard clause positioning data.

[0029] Specifically, the group standard text files, compilation instructions, referenced document status information, implementation evidence documents, testing record documents, and rectification record documents in the group standard associated data are uniformly converted into searchable text format. The page structure is restored based on headers and footers, table of contents levels, table boundaries, and paragraph numbers. Duplicate fields, missing fields, misaligned fields, and format noise are cleaned. Chapter titles are associated with the standard number and standard name. The clause number and clause text position are located according to the chapter title. Then, the corresponding positions in the referenced document status information, implementation evidence documents, testing record documents, and rectification record documents are associated with the clause number. The positions of the referenced documents, implementation evidence documents, testing record documents, and rectification record documents are marked to generate group standard clause positioning data.

[0030] S1.3. The clause number, clause text position, reference document position, implementation evidence position, inspection record position and rectification record position in the group standard clause location data are grouped together by the same clause number, and the clause text content, reference document content, implementation evidence content, inspection record content and rectification record content are extracted according to each position to generate the group standard clause evidence base dataset.

[0031] Specifically, a collection index is established according to the clause number in the group standard clause location data. The clause text position, reference document position, implementation evidence position, test record position, and rectification record position corresponding to the same clause number are written into the same clause record. The clause text content is extracted according to the clause text position, the reference document content is extracted according to the reference document position, the implementation evidence content is extracted according to the implementation evidence position, the test record content is extracted according to the test record position, and the rectification record content is extracted according to the rectification record position. Then, the clause text content, reference document content, implementation evidence content, test record content, and rectification record content are bound with the corresponding clause number to generate the group standard clause evidence basic dataset.

[0032] S2. Extract the clause normative strength and clause evidence requirements from the group standard clause evidence base dataset, and differentiate the clause evidence requirements according to the clause normative strength to generate a clause normative strength evidence requirement table.

[0033] S2.1 Extract the text content, source of evidence, time of evidence, and object of evidence fields of the same clause number from the group standard clause evidence base dataset, and perform normative word marking and obligation action extraction on the text content of the clause to obtain the normative strength of the clause.

[0034] Specifically, the main text of clauses with the same clause number under the same clause number is extracted from the evidence base dataset of group standard clauses, as well as the evidence source information from the content of cited documents, implementation evidence, inspection records, and rectification records. The main text of the clauses is divided into sentences and segments while retaining the same clause number. Normative words such as "shall," "must," "must not," "prohibited," "appropriate," "recommended," "may," "can," and "reference" in the main text of the clauses are marked with their positions to form normative word marking results. Obligations such as performance, inspection, proof, record, rectification, review, prohibition, restriction, and explanation are extracted from the corresponding sentences and segments of the normative word marking results to form obligation action extraction results. The normative words in the normative word marking results are divided into strongly binding words, restrictive binding words, recommended binding words, and explanatory words. The obligation actions in the obligation action extraction results are divided into inspection actions, proof actions, rectification actions, record actions, prohibited actions, restricted actions, and explanatory actions. Combinations of strongly binding terms with detection, proof, or rectification actions, as well as combinations of restrictive terms with prohibited or restricted actions, are classified as high-intensity clauses. Combinations of restrictive terms with detection, proof, or recording actions are classified as medium-to-high-intensity clauses. Combinations of recommended terms with recording, explanatory, or detection actions are classified as medium-to-low-intensity clauses. Combinations of explanatory terms with explanatory actions are classified as low-intensity clauses. When the combination of normative terms and obligatory actions does not fall into the above categories, the basic intensity is determined according to the priority of the normative terms, and a first-level adjustment is made based on the category of obligatory actions to obtain the normative intensity of the clause.

[0035] Among them, strong constraint words include "should" and "must", restrictive constraint words include "must not" and "prohibited", recommended constraint words include "appropriate" and "recommended", and explanatory words include "may", "can", and "for reference".

[0036] S2.2 Based on the strength of the clause, the content of the clause text, and the source of evidence under the same clause number, the testing requirements, supporting material requirements, implementation record requirements, and review requirements are correspondingly collected to obtain the clause evidence requirements.

[0037] Specifically, based on the strength of the clause specifications, the content of the clause text, and the evidence source information in the cited documents, implementation evidence, testing records, and rectification records under the same clause number, the testing objects, testing indicators, judgment limits, and testing frequencies are extracted from the clause text and compiled into testing requirements; the document types, proof objects, and proof periods are extracted from the clause text and evidence source information and compiled into proof material requirements; the record objects, record times, and record sources corresponding to implementation process records, testing records, and rectification records are extracted from the evidence source information and compiled into implementation record requirements; the review objects, review evidence scope, review frequency, and review triggering conditions are extracted from the clause text, evidence source field, evidence time field, and evidence object field and compiled into review requirements; and the testing requirements, proof material requirements, implementation record requirements, and review requirements are grouped under the same clause number to obtain the clause evidence requirements.

[0038] It should also be noted that high-intensity clauses trigger review due to updated evidence, abnormal indicators, and incomplete rectification; medium-to-high-intensity clauses trigger review due to expired evidence and abnormal indicators; medium-to-low-intensity clauses trigger review due to missing records; and low-intensity clauses trigger review due to changes in the source of evidence.

[0039] S2.3. Configure the evidence requirements of the clauses according to the strength of the clauses, and generate a table of evidence requirements based on the strength of the clauses.

[0040] Specifically, the evidentiary requirements for clauses under the same clause number are read according to the strength of the clause specifications. The evidentiary requirements for clauses corresponding to high-strength clauses are configured to meet the testing requirements, supporting document requirements, implementation record requirements, and review requirements simultaneously. The evidentiary requirements for clauses corresponding to medium-high-strength clauses are configured to meet the testing requirements, supporting document requirements, and implementation record requirements, and to meet the review requirements when the evidence expires or the indicators are abnormal. The evidentiary requirements for clauses corresponding to medium-low-strength clauses are configured to meet the implementation record requirements and review requirements. The evidentiary requirements for clauses corresponding to low-strength clauses are configured to meet the supporting document requirements, implementation record requirements, and review requirements for changes in the source of evidence. The configuration content under different clause specifications is bound to the same clause number to generate a clause specification strength evidentiary requirement table.

[0041] S3. Based on the evidence requirement table for the strength of the clause, extract the corresponding evidence from the basic dataset of evidence for the clauses in the group standard, and verify the coverage status, validity status and object consistency status between the corresponding evidence and the evidence requirement of the clause, and generate a clause evidence coverage matching set.

[0042] S3.1 Extract the evidence requirements for the same clause number from the clause strength evidence requirement table, and organize the evidence extraction conditions according to evidence type, evidence field, evidence time and evidence object to generate clause evidence extraction rules.

[0043] Specifically, the evidence requirements for each clause are extracted from the table of evidence requirements for the strength of the clauses, according to the same clause number. The type of evidence is determined based on the testing requirements, supporting document requirements, implementation record requirements, and review requirements in the clause evidence requirements. The object of evidence is determined based on the testing object, the object of proof, the object of record, and the object of review. The field of evidence is determined based on the testing indicators, judgment limits, supporting document type, record source, evidence conclusion field, rectification status field, and review triggering conditions. The time of evidence is determined based on the testing frequency, proof period, record time, and review frequency. The validity period of evidence is calculated according to the shortest period among the proof period, testing frequency, and review frequency. The evidence type, field of evidence, time of evidence, validity period of evidence, and object of evidence are grouped into the same clause number to generate the clause evidence extraction rules.

[0044] S3.2. According to the rules for extracting evidence in the clauses, locate the corresponding implementation evidence content, test record content, referenced document content, supporting material content, rectification record content, and review record content in the basic dataset of evidence for the clauses of the group standard, and group the implementation evidence content, test record content, referenced document content, supporting material content, rectification record content, and review record content into the same clause number to generate evidence corresponding to the clause.

[0045] Specifically, according to the evidence type, evidence field, evidence time, and evidence object in the evidence extraction rules of the clauses, the implementation evidence content, inspection record content, referenced document content, supporting material content, rectification record content, and review record content under the same clause number are detected in the basic dataset of evidence for the clauses of the group standard. Then, the detected implementation evidence content, inspection record content, referenced document content, supporting material content, rectification record content, and review record content are matched by field, time, and object. The corresponding implementation evidence content, inspection record content, referenced document content, supporting material content, rectification record content, and review record content are then grouped into the same clause number to generate clause-corresponding evidence.

[0046] S3.3 Perform coverage status verification, evidence validity status verification, and object consistency status verification on the evidence corresponding to the clause and the evidence requirements in the clause specification strength evidence requirement table, and generate a clause evidence coverage matching set.

[0047] Specifically, the evidence requirements for each clause are retrieved from the corresponding evidence and the evidence requirements table for the same clause number. The implementation evidence, testing record, cited document, and rectification record content in the corresponding clause evidence are matched item by item with the clause evidence requirements. The required evidence types and fields are verified to ensure they are covered by the corresponding clause evidence, yielding a coverage status verification result. The evidence time and validity period fields in the corresponding clause evidence are verified to meet the proof period, recording time, and review frequency requirements, yielding an evidence validity status verification result. The evidence objects in the corresponding clause evidence are verified to correspond to the testing objects, proof objects, recording objects, and review objects in the clause evidence requirements, yielding an object consistency status verification result. The coverage status verification result, evidence validity status verification result, and object consistency status verification result are then grouped under the same clause number to generate a clause evidence coverage matching set.

[0048] S4. Perform reverse evidence retrieval on the evidence coverage matching set of the clauses, extract evidence content that has abnormal indicators, time conflicts, inconsistent objects, and unresolved rectification with the evidence corresponding to the clauses, and determine the conflict between the evidence corresponding to the clauses and the reverse evidence to generate a positive and negative evidence conflict verification chain.

[0049] S4.1 Extract the corresponding evidence and verification results of the clauses in the clause evidence coverage matching set, and convert the content that does not meet the verification conditions into reverse retrieval trigger items to generate reverse evidence retrieval conditions.

[0050] Specifically, based on the same clause number, the corresponding evidence, coverage status verification, evidence validity status verification, and object consistency status verification are extracted from the clause evidence coverage matching set. Contents with missing evidence types or fields in the coverage status verification are marked as coverage reverse retrieval triggers. Contents with evidence time exceeding the proof period, record time, and review frequency requirements in the evidence validity status verification are marked as time reverse retrieval triggers. Contents with evidence objects not corresponding to the detection object, proof object, record object, and review object in the object consistency status verification are marked as object reverse retrieval triggers. Coverage reverse retrieval triggers, time reverse retrieval triggers, and object reverse retrieval triggers are grouped under the same clause number to generate reverse evidence retrieval conditions.

[0051] S4.2 Extract the corresponding evidence for the same clause number from the clause evidence coverage matching set, as well as the verification content of coverage status verification, evidence validity status verification and object consistency status verification, and generate clause verification content collection data.

[0052] Specifically, according to the same clause number, the corresponding evidence is extracted from the clause evidence coverage matching set, and the evidence type verification content and evidence field verification content in the coverage status verification, the evidence time validity content in the evidence validity status verification, and the evidence object consistency content in the object consistency status verification are extracted simultaneously. The clause corresponding evidence, evidence type coverage content, evidence field coverage content, evidence time validity content, and evidence object consistency content are grouped into the same clause number to generate clause verification content collection data.

[0053] S4.3. Compare the corresponding evidence in the clause verification content collection data with the clause evidence requirements in the clause specification strength evidence requirement table item by item, and mark the verification content with insufficient evidence coverage, unsatisfactory evidence validity status, and inconsistent evidence objects, and generate abnormal verification mark data.

[0054] Specifically, the corresponding evidence in the clause verification content collection data is read according to the same clause number, and the clause evidence requirements in the clause specification strength evidence requirement table are read. The evidence type verification content and evidence field verification content in the clause verification content collection data are matched with the evidence type and evidence field in the clause evidence requirements item by item. Verification content that lacks corresponding evidence type and evidence field is marked as insufficient evidence coverage. The valid evidence time content in the corresponding evidence is compared with the evidence time in the clause evidence requirements item by item. Verification content that does not meet the evidence time requirement is marked as evidence validity status not met. The consistency of evidence objects in the corresponding evidence is compared with the evidence objects in the clause evidence requirements item by item. Verification content that does not correspond to evidence objects is marked as evidence object inconsistency, generating abnormal verification mark data.

[0055] S4.4. Collect the anomaly verification marker data according to the clause number, the corresponding evidence of the clause, the anomaly type, and the location of the anomaly source, and generate the content that does not meet the verification conditions.

[0056] Specifically, an aggregation index is established according to the clause number in the anomaly verification mark data. Verification contents marked with insufficient evidence coverage, unsatisfactory evidence validity status, and inconsistent evidence objects under the same clause number are associated with the corresponding clause-related evidence. The anomaly source position of each verification content in the clause-related evidence is recorded. The clause number, clause-related evidence, anomaly type, and anomaly source position are written into the same anomaly record to generate content that does not meet the verification conditions.

[0057] S4.5. Based on the reverse evidence retrieval conditions, trace back the evidence content under the same clause number from the group standard clause evidence base dataset, and extract the associated records of the same evidence object under different clause numbers to generate candidate reverse evidence data.

[0058] Specifically, based on the clause number, evidence object, coverage reverse retrieval trigger, time reverse retrieval trigger, and object reverse retrieval trigger in the reverse evidence retrieval conditions, the implementation evidence content, inspection record content, cited document content, supporting material content, rectification record content, and review record content under the same clause number are traced back from the group standard clause evidence basic dataset. Then, according to the evidence object, the implementation evidence content, inspection record content, cited document content, supporting material content, rectification record content, and review record content with the same evidence object under different clause numbers are detected. The traced evidence content and the associated records under different clause numbers are grouped into the same evidence object to generate candidate reverse evidence data.

[0059] S4.6. Conduct anomaly checks, time conflict checks, object consistency checks, and rectification closure checks on candidate reverse evidence data, and classify the evidence content with anomalies into the same clause number to generate reverse evidence.

[0060] Specifically, the following checks are performed on the detection records and implementation evidence in the candidate reverse evidence data: Anomaly checks are conducted on the indicators; the evidence requirements for the same clause number are retrieved from the clause strength evidence requirement table; evidence with indicator values ​​exceeding the judgment limits in the clause evidence requirements is marked as anomaly; time conflict checks are conducted on the evidence time in the candidate reverse evidence data; evidence with a time after the validity period of the corresponding clause evidence and whose conclusion negates the corresponding clause evidence, or evidence with a time outside the time limit of the clause evidence requirement, is marked as time conflict; object consistency checks are conducted on the evidence objects in the candidate reverse evidence data; evidence with objects that do not correspond to the evidence objects in the clause evidence requirements is marked as object inconsistency; rectification record checks are conducted on the rectification closure loop in the candidate reverse evidence data; evidence without a review approval record is marked as rectification not closed loop; evidence with anomalies, time conflicts, object inconsistencies, and rectification not closed loop are grouped under the same clause number to generate reverse evidence.

[0061] Among them, the judgment limit value refers to the qualified boundary value of the indicator extracted from the testing requirements in the evidence requirements of the clause.

[0062] S4.7 Using the clause number and the evidence object as the association key, pair the evidence corresponding to the clause with the reverse evidence within the same clause and across clauses for the same evidence object to generate positive and negative evidence comparison data.

[0063] Specifically, a positive association key is established according to the clause number and evidence object in the corresponding evidence, and a negative association key is established according to the clause number and evidence object in the negative evidence. The corresponding evidence and negative evidence with the same positive association key and negative association key are paired within the same clause, and the corresponding evidence and negative evidence with the same evidence object but different clause numbers are paired across clauses. The paired content within the same clause and the paired content across clauses are collected according to the clause number, evidence object and evidence source location to generate positive and negative evidence comparison data.

[0064] S4.8 Conduct conflict verification on the indicator values, evidence time, evidence objects, and rectification closure status in the comparison data of positive and negative evidence, and divide the conflict impact level based on the evidence requirement table of the strength of the clauses and generate the degree of evidence conflict.

[0065] Specifically, the corresponding evidence and reverse evidence in the comparison data of positive and negative evidence are read according to the same clause number. Deviation verification of indicator values ​​in the corresponding evidence and reverse evidence is performed to obtain indicator conflict markers. The chronological order and validity period of the evidence in the corresponding evidence and reverse evidence are verified to obtain time conflict markers. The consistency of the evidence objects in the corresponding evidence and reverse evidence is verified to obtain object conflict markers. The completion status of the rectification closure loop in the corresponding evidence and reverse evidence is verified to obtain rectification conflict markers. Then, the clause normative strength weight is determined according to the clause normative strength evidence requirement table. The indicator conflict markers, time conflict markers, object conflict markers, rectification conflict markers and clause normative strength weights are input into the evidence conflict degree calculation formula to generate the evidence conflict degree.

[0066] The formula for calculating the degree of conflict of evidence is as follows: ; in, Indicates the first The degree of conflict of evidence corresponding to each clause number Indicates the first The clause strength weight corresponding to each clause number Indicates the first The indicator conflict flag corresponding to each clause number Indicates the first The time conflict marker corresponding to each clause number Indicates the first The object conflict flag corresponding to each clause number Indicates the first The rectification conflict markers corresponding to each clause number This indicates the weighting coefficient corresponding to the indicator conflict marker. This represents the weighting coefficient corresponding to the time conflict marker. This represents the weight coefficient corresponding to the object conflict marker. This indicates the weighting coefficient corresponding to the rectification conflict marker. An index indicating the clause number; In this embodiment, the indicator conflict marker, time conflict marker, object conflict marker, and rectification conflict marker are all set to 1 when they exist, and 0 when they do not exist. The clause specification strength weight is determined according to the clause specification strength: high-strength clauses are set to 1, medium-high-strength clauses are set to 0.8, medium-low-strength clauses are set to 0.5, and low-strength clauses are set to 0.3. The weight coefficients satisfy the following conditions: And can be set to , , , ;when When the value is not less than 0.75, the degree of conflict of evidence is classified as Level 1 conflict of evidence; when... When the value is not less than 0.4 and less than 0.75, the degree of conflict of evidence is level two; when When the value is less than 0.4, the level of conflict of evidence is classified as Level III.

[0067] S4.9. Chain-collect the degree of evidence conflict with the corresponding clause number, the corresponding evidence for the clause, and the reverse evidence to generate a chain of verification for the conflict between positive and negative evidence.

[0068] Specifically, the degree of evidence conflict, the corresponding evidence and the reverse evidence are read according to the same clause number. The corresponding evidence is used as the supporting evidence node, the reverse evidence is used as the reverse evidence node, and the degree of evidence conflict is used as the conflict determination node. The clause evidence conflict chain is established according to the sequential relationship between the corresponding evidence, the reverse evidence and the degree of evidence conflict. The clause evidence conflict chain is bound to the corresponding clause number to generate a positive and negative evidence conflict verification chain.

[0069] S5. Determine the risk status and credibility status of the clauses based on the chain of conflict between positive and negative evidence, and classify the rectification level and review requirements according to the degree of conflict of evidence in the chain of conflict between positive and negative evidence, and generate a dynamic assessment report on the compliance of the group standard.

[0070] S5.1 Extract the clause number, clause normative strength, degree of evidence conflict, corresponding evidence and reverse evidence from the positive and negative evidence conflict verification chain, and merge and organize them according to the same clause number to generate clause conflict merging data.

[0071] Specifically, according to the clause evidence conflict chain in the positive and negative evidence conflict verification chain, the clause number, the degree of evidence conflict, the corresponding evidence of the clause, and the reverse evidence are extracted. The clause normative strength corresponding to the same clause number is retrieved from the clause normative strength evidence requirement table. The clause normative strength, degree of evidence conflict, corresponding evidence of the clause, and reverse evidence under the same clause number are merged and organized to generate clause conflict merged data.

[0072] S5.2. Based on the degree of evidence conflict and the strength of clause norms in the clause conflict merging data, classify and determine the indicator deviation status, time conflict status, object conflict status and rectification closure status, and generate clause risk status.

[0073] Specifically, the degree of evidence conflict and the strength of clause norms in the clause conflict merging data are read according to the same clause number. The indicator conflict mark in the degree of evidence conflict is converted into the indicator deviation state, the time conflict mark in the degree of evidence conflict is converted into the time conflict state, the object conflict mark in the degree of evidence conflict is converted into the object conflict state, and the rectification conflict mark in the degree of evidence conflict is converted into the rectification closed loop state. According to the strength of clause norms, the indicator deviation state, time conflict state, object conflict state and rectification closed loop state are classified as high risk, medium risk and low risk to generate the clause risk status.

[0074] It should also be noted that high risk refers to clauses with high regulatory strength and multiple key conflicts of evidence; medium risk refers to clauses with medium-to-high or medium-to-low regulatory strength and a single key conflict of evidence; and low risk refers to clauses with low regulatory strength and only minor conflicts of evidence. Key conflicts include indicator conflict markers, object conflict markers, and rectification conflict markers; minor conflicts include time conflict markers and abnormalities in the validity of evidence that do not affect the consistency of the evidence object.

[0075] S5.3. Using the risk status of the clause as the verification entry point, verify the source credibility, time validity, object consistency and conflict admissibility of the clause-related evidence and reverse evidence in the clause conflict merging data, and generate the credibility status of the evidence.

[0076] Specifically, the corresponding evidence and reverse evidence in the clause risk status and clause conflict merging data are read according to the same clause number. The clause risk status is used as the verification entry point to trigger the verification scope of the corresponding evidence. The source credibility of the evidence source of the corresponding evidence and reverse evidence is verified. The time validity of the evidence time of the corresponding evidence and reverse evidence is verified. The object consistency of the evidence object of the corresponding evidence and reverse evidence is verified. The admissibility of the conflict content of the corresponding evidence and reverse evidence is verified. Then, the verification content of source credibility verification, time validity verification, object consistency verification and conflict admissibility verification is classified into the same clause number to generate the evidence credibility status.

[0077] It should also be noted that the scope of evidence verification is configured according to the risk level of the clauses: high risk corresponds to verifying all evidence corresponding to the same clause number, all reverse evidence, records of different clause numbers related to the same evidence object within the past 24 months, and all records of rectification not yet closed; medium risk corresponds to verifying all evidence corresponding to the same clause number, all reverse evidence, and records of different clause numbers related to the same evidence object within the past 12 months; low risk corresponds to verifying the latest evidence corresponding to the same clause number, the latest reverse evidence, and the valid content of evidence within the past 6 months. The configuration of the scope of evidence verification is based on the clause specification strength, the degree of evidence conflict, the validity status of the evidence, and the rectification closure status. The past 24 months, the past 12 months, and the past 6 months are determined according to the review cycle multiples corresponding to the clause specification strength, with high risk corresponding to four 6-month review cycles, medium risk corresponding to two 6-month review cycles, and low risk corresponding to one 6-month review cycle. The credibility of the source is determined according to the order of third-party testing records, public platform records, records stamped by the implementing unit, and records uploaded manually; the admissibility of conflict is determined by the credibility of the evidence source, the validity of the evidence time, and the consistency of the evidence object.

[0078] S5.4. Based on the degree of evidence conflict, the risk status of the clause, and the credibility status of the evidence, merge and sort the clauses to be rectified under the same clause number to generate clause rectification ranking data.

[0079] Specifically, the degree of evidence conflict, the risk status of the clause, and the credibility status of the evidence are read according to the same clause number. The conflict indicators, time conflict indicators, object conflict indicators, and rectification conflict indicators in the degree of evidence conflict are used as sorting factors. The high risk, medium risk, and low risk in the risk status of the clause are used as sorting levels. The source credibility, time validity, object consistency, and conflict admissibility in the credibility status of the evidence are used as correction factors. Clauses with insufficient evidence coverage, unsatisfactory evidence validity status, inconsistent evidence objects, abnormal indicators, time conflicts, inconsistent objects, and unclosed rectification are marked as clauses to be rectified. The clauses to be rectified are merged and sorted to generate clause rectification sorting data.

[0080] S5.5 Divide the rectification sorting data of the clauses into rectification levels, and configure the review frequency, review objects, review evidence scope and review pass conditions according to the rectification level to generate rectification review configuration data.

[0081] Specifically, clauses requiring rectification that are high-risk and whose evidence is partially credible or unreliable are classified as Level 1 rectification; clauses requiring rectification that are medium-risk and those that are high-risk and whose evidence is credible are classified as Level 2 rectification; and clauses requiring rectification that are low-risk and whose evidence is credible are classified as Level 3 rectification. Level 1 rectification is configured to be reviewed every 15 days, with the review covering all evidence corresponding to the same clause number and all reverse evidence. The scope of the reviewed evidence includes implementation evidence, testing records, rectification records, and review records from the past 24 months. A successful review requires that the indicator conflict marker, object conflict marker, and rectification conflict marker all be 0, and that the evidence is credible. Level 2 rectification is configured to be reviewed every 30 days. The review targets are the corresponding evidence and reverse evidence under the same clause number that have conflict markers. The scope of the review evidence includes the contents of the inspection records, rectification records, and review records within the past twelve months. The review pass condition is that there is a review pass record in the rectification record content and the rectification conflict marker value is 0. The third-level rectification configuration is reviewed once every sixty days. The review targets are the latest corresponding evidence selected from the corresponding evidence of the clause according to the evidence time. The scope of the review evidence includes the latest corresponding evidence of the clause and the latest rectification record content under the same clause number. The review pass condition is that the evidence validity status is met and the evidence credibility status is credible. The rectification level, review frequency, review targets, review evidence scope, and pass conditions are grouped into the same clause number to generate rectification review configuration data.

[0082] S5.6 Organize the rectification level and review requirements in the rectification and review configuration data into a report format, and write the corresponding clause number, clause risk status, evidence credibility status and evidence conflict degree to generate a dynamic assessment report on the compliance of group standards.

[0083] Specifically, the clause-level assessment records are sorted according to the rectification level and included in the dynamic assessment report for compliance of group standards. This includes the clause number, clause risk status, evidence credibility status, degree of evidence conflict, location of the corresponding evidence source, location of the reverse evidence source, location of the anomaly source, rectification level, review frequency, review object, scope of review evidence, and review approval conditions. A report index is established according to the clause number, enabling the dynamic assessment report for compliance of group standards to trace back to the corresponding evidence, reverse evidence, anomaly verification mark data, and rectification review configuration data to generate the dynamic assessment report for compliance of group standards.

[0084] This embodiment also provides a data mining system for assessing the compliance of group standards, including: The module comprises four parts: a clause database construction module, a clause library construction module, and a dynamic evaluation module. The first part collects relevant data from group standards, performs format governance and clause location, and generates a basic dataset of group standard clause evidence. The second part extracts the clause's normative strength and evidentiary requirements from the basic dataset and configures these requirements differently according to their strength, generating a clause normative strength evidentiary requirement table. The third part verifies the evidence, extracting corresponding evidence from the dataset based on the evidence requirement table and verifying the coverage, validity, and object consistency between the evidence and the required evidence, generating a clause evidence coverage matching set. The fourth part searches the evidence coverage matching set for reverse evidence, extracting evidence that exhibits abnormal indicators, time conflicts, object inconsistencies, or incomplete rectification of the corresponding evidence. It then compares the corresponding evidence with the reverse evidence to generate a positive and negative evidence conflict verification chain. The fifth part assesses the clause's risk status and evidence credibility based on the positive and negative evidence conflict verification chain, and classifies rectification levels and review requirements according to the degree of evidence conflict in the chain, generating a dynamic evaluation report on the group standard's compliance.

[0085] In summary, this invention, by performing reverse evidence retrieval on the matching set of clause evidence coverage and by determining conflicts between the clause-corresponding evidence and the reverse evidence, can incorporate evidence such as abnormal indicators, time conflicts, inconsistent objects, and incomplete rectification into the verification chain under the same clause number. This ensures that the compliance judgment of each clause has both supporting and negative verification evidence. The generated positive and negative evidence conflict verification chain can continuously carry the clause-corresponding evidence, reverse evidence, and the degree of evidence conflict, providing a unified data foundation for clause risk status determination, evidence credibility verification, rectification level division, and review requirement configuration. This improves the traceability, stability, and review relevance of the dynamic compliance assessment report for group standards.

[0086] It should be noted that 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A data mining method for assessing the compliance of group standards, characterized in that: include, Collect data related to group standards, perform format governance and clause location, and generate a basic dataset of evidence for group standard clauses; Extract the clause normative strength and clause evidence requirements from the group standard clause evidence base dataset, and differentiate the clause evidence requirements according to the clause normative strength to generate a clause normative strength evidence requirement table. Based on the evidence requirement table for the strength of the clauses, the corresponding evidence is extracted from the basic dataset of evidence for the clauses in the group standard, and the coverage status, validity status and object consistency status between the corresponding evidence and the evidence requirement of the clauses are verified to generate a clause evidence coverage matching set. Reverse evidence retrieval is performed on the evidence coverage matching set of the clauses to extract evidence content that has abnormal indicators, time conflicts, inconsistent objects, and unclosed rectification loops with the evidence corresponding to the clauses. The evidence corresponding to the clauses is then used to determine the conflict between the reverse evidence and generate a positive and negative evidence conflict verification chain. Based on the chain of verification of conflicting positive and negative evidence, the risk status and credibility status of the clauses are determined, and the rectification level and review requirements are divided according to the degree of conflict of evidence in the chain of verification of conflicting positive and negative evidence, and a dynamic assessment report on the compliance of the group standard is generated.

2. The data mining method for assessing the compliance of group standards as described in claim 1, characterized in that: The dataset that forms the basis for generating evidence of group standard clauses specifically includes... Collect group standard text files, compilation instructions, referenced document status information, implementation evidence files, testing record files, and rectification record files, and uniformly mark them to generate group standard related data; The data associated with group standards is converted in format, restored in layout, cleaned in fields, and located in clauses. The standard number, standard name, chapter title, clause number, clause text position, reference document position, implementation evidence position, test record position, and rectification record position are marked to generate group standard clause location data. The clause numbers, clause text locations, reference document locations, implementation evidence locations, inspection record locations, and rectification record locations in the group standard clause location data are grouped together using the same clause number. Based on each location, the clause text content, reference document content, implementation evidence content, inspection record content, and rectification record content are extracted to generate the group standard clause evidence base dataset.

3. The data mining method for assessing the compliance of group standards as described in claim 2, characterized in that: The generated clause strength evidence requirement table specifically includes, Extract the text content, source of evidence, time of evidence, and object of evidence fields of the same clause number from the group standard clause evidence base dataset, and perform normative word marking and obligation action extraction on the text content of the clause to obtain the normative strength of the clause; Based on the strength of the clauses, the content of the clause text, and the source of evidence under the same clause number, the testing requirements, supporting material requirements, implementation record requirements, and review requirements are correspondingly categorized to obtain the clause evidence requirements. Based on the strength of the clause's normative provisions, the evidentiary requirements for each clause are configured differently, generating a table of evidentiary requirements for the strength of the clause's normative provisions.

4. The data mining method for assessing the compliance of group standards as described in claim 3, characterized in that: The generated terms evidence covers a matching set, specifically including: Extract the evidence requirements for the same clause number from the table of evidence requirements for the strength of the clause, and organize the evidence extraction conditions according to the evidence type, evidence field, evidence time and evidence object to generate the clause evidence extraction rules. According to the rules for extracting evidence in the clauses, locate the corresponding implementation evidence content, test record content, referenced document content, supporting material content, rectification record content, and review record content in the basic dataset of evidence for the clauses of the group standard, and group the implementation evidence content, test record content, referenced document content, supporting material content, rectification record content, and review record content into the same clause number to generate evidence corresponding to the clause; The evidence requirements for the corresponding clauses are verified against the clause evidence requirements in the clause strength evidence requirement table. The verification includes coverage status, evidence validity status, and object consistency status, generating a clause evidence coverage matching set.

5. The data mining method for assessing the compliance of group standards as described in claim 4, characterized in that: The extracted evidence, which contains evidence of abnormal indicators, time conflicts, inconsistent objects, and incomplete rectification, specifically includes: Extract the corresponding evidence and verification results of the clauses in the clause evidence coverage matching set, and transform the content that does not meet the verification conditions into reverse retrieval trigger items to generate reverse evidence retrieval conditions. Based on the reverse evidence retrieval criteria, the evidence content under the same clause number is traced back from the group standard clause evidence base dataset, and the associated records of the same evidence object under different clause numbers are extracted to generate candidate reverse evidence data. The candidate reverse evidence data is checked for anomalies in indicators, time conflicts, consistency of objects, and rectification closure loop. The evidence content with anomalies is classified into the same clause number to generate reverse evidence.

6. The data mining method for assessing the compliance of group standards as described in claim 5, characterized in that: The content that did not meet the verification conditions specifically includes, Extract the corresponding evidence for the same clause number from the clause evidence coverage matching set, as well as the verification content of coverage status verification, evidence validity status verification and object consistency status verification, and generate clause verification content collection data. The corresponding evidence in the data collection of clause verification content is compared with the evidence requirements of the clause in the table of evidence requirements for the strength of clause specifications, and the verification content with insufficient evidence coverage, unsatisfactory evidence validity status and inconsistent evidence objects is marked to generate abnormal verification mark data. The abnormal verification marker data is collected according to the clause number, the corresponding evidence for the clause, the abnormality type, and the location of the abnormality source, and the content that does not meet the verification conditions is generated.

7. The data mining method for assessing the compliance of group standards as described in claim 5, characterized in that: The generation of the conflict verification chain for both positive and negative evidence specifically includes, Using clause number and evidence object as the association key, the evidence corresponding to the clause is paired with the reverse evidence within the same clause and across clauses for the same evidence object, generating comparative data of positive and negative evidence; Conflict verification is conducted on the indicator values, evidence time, evidence objects and rectification closure status in the data comparing positive and negative evidence, and the conflict impact level is divided based on the evidence requirement table of the strength of the clauses and norms, and the degree of evidence conflict is generated. The degree of evidence conflict is chained together with the corresponding clause number, the corresponding evidence for the clause, and the reverse evidence to generate a chain for verifying the conflict between positive and negative evidence.

8. The data mining method for assessing the compliance of group standards as described in claim 7, characterized in that: The risk status and credibility status of the aforementioned judgment clauses specifically include, Extract clause number, clause normative strength, degree of evidence conflict, corresponding evidence and reverse evidence from the chain of conflict verification between positive and negative evidence, and merge and organize them according to the same clause number to generate clause conflict merging data; Based on the degree of evidence conflict and the strength of clause norms in the clause conflict consolidation data, the index deviation status, time conflict status, object conflict status and rectification closure status are classified and judged to generate clause risk status. Using the risk status of the clauses as the verification entry point, the source credibility, time validity, object consistency and conflict admissibility of the corresponding evidence and reverse evidence in the clause conflict merging data are verified to generate the credibility status of the evidence.

9. The data mining method for assessing the compliance of group standards as described in claim 8, characterized in that: The generation of the dynamic assessment report on compliance with group standards specifically includes, Based on the degree of conflict of evidence, the risk status of the clause, and the credibility status of the evidence, the clauses to be rectified under the same clause number are merged and sorted to generate clause rectification sorting data. The rectification data of the clauses is sorted into rectification levels, and the frequency of review, review objects, scope of review evidence and review pass conditions are configured according to the rectification level to generate rectification review configuration data; The rectification level and review requirements in the rectification and review configuration data are organized into reports and written with the corresponding clause number, clause risk status, evidence credibility status and evidence conflict degree, generating a dynamic assessment report on the compliance of group standards.

10. A data mining system for assessing compliance with group standards, based on the data mining method for assessing compliance with group standards according to any one of claims 1 to 9, characterized in that: include, The clause database module is used to collect data related to group standards, perform format governance and clause location, and generate a basic dataset of evidence for group standard clauses. The strength configuration module is used to extract the clause normative strength and clause evidence requirements from the group standard clause evidence basic dataset, and to differentiate the clause evidence requirements according to the clause normative strength to generate a clause normative strength evidence requirement table. The evidence verification module is used to extract the corresponding evidence from the basic dataset of group standard clause evidence according to the evidence requirement table of clause specification strength, and to verify the coverage status, validity status and object consistency status between the corresponding evidence and the clause evidence requirement, and generate a clause evidence coverage matching set. The counter-evidence chain module is used to perform reverse evidence retrieval on the clause evidence coverage matching set, extract evidence content that has abnormal indicators, time conflicts, object inconsistencies and rectification loops that are not closed with the clause corresponding evidence, and then perform conflict judgment between the clause corresponding evidence and the reverse evidence to generate a positive and negative evidence conflict verification chain. The dynamic assessment module is used to determine the risk status and credibility status of the clauses based on the verification chain of conflicting positive and negative evidence, and to divide the rectification level and review requirements according to the degree of evidence conflict in the verification chain of conflicting positive and negative evidence, and generate a dynamic assessment report on the compliance of the group standard.

Images