Intelligent service system for calibration cooperation

Abstract

The invention relates to the technical field of calibration cooperation, in particular to an intelligent service system for calibration cooperation, which comprises a judicial file analysis module, a decision optimization module, a document logic verification module and a logic defect correction module. According to the method, efficient extraction of key information in judicial documents and accurate evaluation of case complexity are realized by combining text structured processing and fuzzy logic, and the processing strategy can quickly adjust a decision support direction based on detailed data analysis, so that the decision process better meets the actual requirements of cases, and the accuracy of the decision process is improved. The application of dynamic update data optimizes decision suggestions, improves the real-time performance and pertinence of decisions, significantly reduces judicial errors, accelerates the judgment process, guarantees the accuracy of judicial documents through automatic logic verification and error correction, effectively avoids errors caused by manual omission, and improves the accuracy of the judicial documents. And the judicial processing speed and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of inspection and proofreading cooperation, and in particular to an intelligent service system for inspection and proofreading cooperation. Background Art

[0002] In the context of legal case trials, legal mediation and prosecution agencies, prosecution-procuratorial cooperation focuses on improving information accuracy and process efficiency in legal proceedings. This technical field combines legal information systems, automated trial technologies and data correction to ensure that data and information in the case handling process are correct, while optimizing processing steps to improve efficiency. In judicial practice, it includes automated evidence detection, proofreading of legal documents and real-time monitoring of procedural compliance to ensure the fairness and accuracy of the trial and mediation process.

[0003] Among them, the intelligent service system of cooperation between procuratorates and schools aims to improve the level of automation in trial and mediation of legal cases, legal mediation and procuratorial organs through integrated intelligent technology, and to enhance the efficiency and accuracy of legal procedures by introducing artificial intelligence to analyze judicial documents, automatically identify key evidence and simulate potential legal consequences. Intelligent procuratorial and school services can help judges quickly verify facts and apply laws in complex cases, reduce manual trial time, and improve the accuracy and rationality of judgments and mediation. It is of great significance to speed up legal procedures, reduce the burden on judicial personnel and enhance judicial credibility.

[0004] Traditional systems lack support for real-time data processing and are insufficient to cope with dynamic changes in cases, resulting in the inability to adjust the decision-making process in a timely manner, affecting the timeliness and accuracy of the trial. In addition, existing automated proofreading technology fails to penetrate into the logical level and is limited to superficial grammatical or format error correction. It fails to deeply identify and correct logical inconsistencies in documents, which to a certain extent limits the legal applicability of documents and the accuracy of judicial documents, and increases uncertainty and risks in the judicial process. These shortcomings highlight the limitations of traditional technology in terms of automation depth and matching. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose an intelligent service system for inspection and proofreading cooperation.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: an intelligent service system for inspection and proofreading cooperation, the system comprising:

[0007] The judicial document analysis module performs text structured decomposition based on the input judicial documents, extracts the chain of evidence and relevant legal reference information in the case, analyzes the weight of evidence and the correlation between legal provisions, and evaluates the complexity of the case through fuzzy logic to obtain the case complexity index;

[0008] The decision optimization module adjusts the decision support direction for judicial workers based on the case complexity index, matches the complexity and change trend of the case, and optimizes the decision process and decision suggestions according to the real-time updated data of the case to obtain the decision support configuration;

[0009] The document logic verification module analyzes the judicial document based on the decision support configuration, identifies and corrects grammatical errors in the document, evaluates sentence coherence, and then reviews the logical structure in the document to determine whether it meets the legal logic standards and obtains logic verification feedback results;

[0010] The logic defect correction module mines potential logic defects in the document based on the logic verification feedback results, including contradictory conditions or unclear responsibilities, and performs logical analysis on the judicial document to correct the logical errors in the document and obtain a judicial document correction record.

[0011] The present invention is improved in that the steps of text structure decomposition are specifically as follows:

[0012] Based on the input judicial documents, the text content is extracted, including the original text and the metadata of the document, to obtain a preliminary text dataset;

[0013] The preliminary text data set is analyzed to identify key text elements, including names of people, places, dates, and legal terms, using the formula:

[0014]

[0015] Evaluate the score Gs of entity recognition and obtain structured text with entity tags, where wg i represents the critical weight of the i-th entity, tg i represents the number of times the i-th entity is mentioned in the document, and n is the number of entities;

[0016] Based on the structured text with entity tags, the chain of evidence and relevant legal reference information in the case are extracted to establish a case information dataset.

[0017] The present invention is improved in that the steps of obtaining the case complexity index are specifically as follows:

[0018] Analyze the chain of evidence and the complexity of legal provisions in the case, including the quantity and type of evidence and its relevance to the legal provisions, and obtain the case analysis log;

[0019] Fuzzy logic is used to process the data in the case analysis log to evaluate the interaction and correlation between each piece of evidence and legal provisions, using the formula:

[0020]

[0021] Calculate the case complexity index CH, where vh j represents the validity weight of the association between the jth piece of evidence and the legal provision, rh j is the correlation score between the jth piece of evidence and the legal provision, and m represents the total number of associations between evidence and legal provisions.

[0022] The present invention is improved in that the steps of matching the complexity and change trend of the case are specifically as follows:

[0023] Based on the case complexity index, analyze key complexity indicators, including the number of legal provisions, the type and amount of evidence, and the speed of case progress, to obtain a real-time overview of the case complexity;

[0024] Based on the real-time complexity overview of the case, the decision support direction of judicial workers is adjusted using the formula:

[0025]

[0026] The complexity and change trend of matching cases, where TQ represents the trend score of case complexity, αq is the adjustment coefficient, and cq i is the weight of the i-th complexity index, dq i is the rate of change of the ith complexity indicator during the monitoring period, and k is the total number of complexity indicators.

[0027] The present invention is improved in that the step of obtaining the decision support configuration is specifically as follows:

[0028] Perform data analysis, identify decision points, and obtain real-time case analysis information based on real-time updated data of the case, including new evidence, legal changes, and case update status;

[0029] Based on the real-time analysis information of the case, the compatibility and efficiency of the existing decision suggestions are evaluated, optimization is performed according to the evaluation results, and the effects during case handling are monitored to obtain decision support configuration.

[0030] The present invention is improved in that the step of evaluating the coherence of a sentence is specifically as follows:

[0031] Based on the case complexity index, extract sentences from judicial documents, analyze sentence structure using NLP technology, identify subject-verb-object structure, conjunction usage and logical connection between sentences, and obtain sentence structure analysis results;

[0032] The sentence structure analysis results are evaluated, the logical flow between sentences is compared, and whether there is a logical break or semantic repetition is checked using the formula:

[0033]

[0034] Calculate the coherence score CL′ of each paragraph of text to get the sentence coherence evaluation result, where rl i is the modified coherence contribution of the i-th sentence, pl i is the position parameter of the i-th sentence in the paragraph, sl i Indicates the number of sentences associated with the i-th sentence, n l is the total number of sentences in the document;

[0035] Based on the sentence coherence evaluation results, the sentence structure is adjusted, conjunctions are added or incoherent parts are rewritten to optimize the fluency and logic of the judicial document and obtain a corrected document feedback result.

[0036] The present invention is improved in that the steps of obtaining the logic check feedback result are specifically as follows:

[0037] Review the logical structure of judicial documents, identify each logical relationship and reasoning chain in the document, test whether the logical connection is reasonable, determine whether there is non-causal reasoning and circular reasoning, and obtain preliminary logical error analysis results;

[0038] Based on the preliminary logical error analysis results, each identified logical problem is evaluated to determine whether the logical structure in the document meets the legal logic standards, and obtain a logical verification feedback result.

[0039] The present invention is improved in that the steps of obtaining the judicial document amendment record are specifically as follows:

[0040] Based on the feedback results of the logic check, potential logic defects in the document are identified, including inconsistent conditions or unclear responsibilities, and the document is analyzed to record each defect and its context to obtain preliminary defect analysis information;

[0041] Based on the preliminary defect analysis information, perform a logical analysis to analyze the impact and priority of the defects using the formula:

[0042]

[0043] Get the correction suggestion result, where DB represents the severity score of the document defect, qb i is the impact weight of the i-th defect, fb i is the frequency of the ith defect, p b is the total number of logic defects;

[0044] Based on the correction suggestion results, the logical errors in the document are corrected, the sentence structure and argument logic are adjusted, and the judicial document correction record is obtained.

[0045] Compared with the prior art, the advantages and positive effects of the present invention are:

[0046] In the present invention, by combining text structured processing and fuzzy logic, efficient extraction of key information in judicial documents and accurate assessment of case complexity are achieved. This processing strategy can quickly adjust the direction of decision support based on detailed data analysis, so that the decision-making process is more in line with the actual needs of the case. The application of dynamically updated data optimizes decision-making suggestions, improves the real-time and pertinence of decisions, significantly reduces judicial errors and accelerates the judgment process. Through automated logic verification and error correction, the accuracy of judicial documents is guaranteed, effectively avoiding errors caused by human omissions, and improving the speed and accuracy of judicial processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a system flow chart of the present invention;

[0048] Figure 2 It is a flowchart of text structure decomposition in the present invention;

[0049] Figure 3 This is a flowchart for obtaining the case complexity index in the present invention;

[0050] Figure 4 A matching flow chart of the complexity and change trend of the cases in the present invention;

[0051] Figure 5 A flowchart for obtaining a decision support configuration in the present invention;

[0052] Figure 6 A flowchart for evaluating sentence coherence in the present invention;

[0053] Figure 7 A flowchart for obtaining the feedback result of the logic check in the present invention;

[0054] Figure 8 A flowchart for obtaining judicial document amendment records in the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0056] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0057] Example

[0058] See also Figure 1 The present invention provides a technical solution: an intelligent service system for cooperation between inspection and proofreading, comprising:

[0059] The judicial document analysis module performs text structured decomposition based on the input judicial documents, extracts the chain of evidence and relevant legal reference information in the case, analyzes the weight of evidence and the correlation between legal provisions, and evaluates the complexity of the case through fuzzy logic to obtain the case complexity index;

[0060] The decision optimization module adjusts the decision support direction for judicial workers based on the case complexity index, matches the complexity and change trend of the case, and optimizes the decision process and decision suggestions based on the real-time updated data of the case to obtain the decision support configuration;

[0061] The document logic verification module is based on the decision support configuration, analyzes judicial documents, identifies and corrects grammatical errors in the text, evaluates sentence coherence, and then reviews the logical structure in the document to determine whether it meets the legal logic standards and obtains logic verification feedback results;

[0062] The logic defect correction module mines the potential logic defects in the document based on the logic verification feedback results, including contradictory conditions or unclear responsibilities, and performs logical analysis on the judicial document to correct the logical errors in the document and obtain the judicial document correction record.

[0063] Case complexity indicators include the complexity of the chain of evidence, applicability of laws, and the number of disputed points. The decision support configuration specifically includes decision adjustment direction, trend matching, and data response sensitivity. The logic verification feedback results include logical consistency, grammatical accuracy, and structural compliance. The judicial document correction records specifically refer to the error type analysis results, correction details, and correction effect evaluation results.

[0064] See also Figure 2 , the steps of text structure decomposition are as follows:

[0065] Based on the input judicial documents, the text content is extracted, including the original text and the metadata of the document (title, page number), to obtain a preliminary text dataset;

[0066] The document scanning tool is used to parse the text and formatting information in the document page by page. When extracting title information, it is necessary to match the fixed title format and font features. Regular expressions are used to match and verify the accuracy of the title text content. At the same time, the specific page number position of each page content is obtained, and the information is corresponding to the title. The text content is extracted by obtaining the main body of the document through line-by-line analysis, including processing text structures such as line breaks and paragraph symbols. At the same time, the language features of the text are parsed to mark keywords. Metadata such as page numbers and paragraph positions also need to be attached to each paragraph of text through positioning algorithms. The results of all operations are integrated into a preliminary text data set to ensure that the complete original text and metadata information are included.

[0067] The preliminary text data set was analyzed to identify key text elements, including names of people, places, dates, and legal terms, using the formula:

[0068]

[0069] Evaluate the score Gs of entity recognition and obtain structured text with entity tags, where wg i represents the critical weight of the ith entity, indicating the importance of different entities in the legal text. For example, the name of a witness or a specific legal term will be given a higher weight to reflect its key role in case analysis. i represents the number of times the i-th entity is mentioned in the document, reflecting the frequency of the entity appearing in the document. The more entities appear, the more critical they are to case analysis. n is the number of entities.

[0070] There are three entities in the sample data, wg1 = 0.5, tg1 = 10 (representing that the witness name appears 10 times, with a weight of 0.5), wg2 = 0.8, tg2 = 5 (representing that the key legal term appears 5 times, with a weight of 0.8), wg3 = 0.3, tg3 = 15 (representing that the date appears 15 times, with a weight of 0.3), then the calculation formula is:

[0071] Gs=(0.5×10)+(0.8×5)+(0.3×15)=5+4+4.5=13.5;

[0072] The result indicates that the total score of the key entities in the document is 13.5, reflecting the overall impact of the importance and frequency of the entities in the document, which is helpful for further legal analysis and evidence extraction.

[0073] Based on structured text with entity tags, extract the chain of evidence and relevant legal reference information in the case to establish a case information dataset;

[0074] By analyzing the legal terms and key evidence fields in structured text, the evidence fields are defined as text fragments that can reflect the relevance of the case. Matching rules such as "witness statements" and "monitoring records" are used to identify the chain of evidence. The matching fields are classified and processed through logical judgment. When extracting legal references, it is necessary to cross-check the entries in the legal terminology library, select legal clauses that are consistent with the case context, and accurately match them by combining the legal entry number and applicable conditions. All extracted evidence chains and legal reference information are organized into a case information data set to ensure that the key evidence and legal basis of the case are recorded completely and accurately.

[0075] See also Figure 3 , the specific steps for obtaining the case complexity index are:

[0076] Analyze the chain of evidence and the complexity of legal provisions in the case, including the quantity and type of evidence and its relevance to the legal provisions, and obtain the case analysis log;

[0077] A relationship diagram between evidence and laws is established for the input case documents, and preliminary information on all evidence chains and corresponding legal references is extracted. The quantity and type of evidence are identified through classification extraction, including written evidence, electronic evidence, physical evidence, etc. A unique identifier is assigned to each piece of evidence using a specific marking method, and legal references are identified through one-by-one analysis of the document structure and keyword matching. The reference characteristics of each piece of evidence are determined in combination with the logical associations in the evidence chain. The quantity record of evidence can be completed directly through classification statistics, while the type classification depends on the key descriptors in the evidence content, such as mentions of time, place, people or specific events. For the processing of legal references, the corresponding relationship between each piece of evidence and the legal article is established by extracting the legal article number or specific legal article expression marked in the document. Finally, all extracted information is summarized as a case analysis log.

[0078] Fuzzy logic is used to process the data in the case analysis log to evaluate the interaction and correlation between each piece of evidence and legal provisions, using the formula:

[0079]

[0080] Calculate the case complexity index CH, where vh j represents the validity weight of the association between the jth piece of evidence and the legal provision. The weight reflects the importance of the association between a piece of evidence and a specific legal provision. jis the relevance score between the jth piece of evidence and the legal provision. The score is based on the degree of direct relevance between the evidence and the content of the legal provision. For example, a high score will be given if the evidence directly supports or refutes a legal provision. m represents the total number of relevances between the evidence and the legal provision.

[0081] There is a case with 3 pieces of evidence. The validity weight and relevance score of each piece of evidence are as follows: vh1=0.5, rh1=8 (the evidence supports important legal provisions), vh2=0.3, rh2=5 (the evidence indirectly relates to legal provisions), vh3=0.2, rh3=3 (the evidence has a low relevance to legal provisions). The calculation process is:

[0082] CH=(0.5×8)+(0.3×5)+(0.2×3)=4+1.5+0.6=6.1;

[0083] The result shows that the complexity index of the case is 6.1, which reflects the comprehensive correlation between evidence and laws in the case and the difficulty of handling. The higher the index, the more complex the case is and more attention and resources are required to handle it.

[0084] See also Figure 4 ,The specific steps for matching the complexity and change trends of the cases are:

[0085] Based on the case complexity index, analyze key complexity indicators, including the number of legal provisions, the type and amount of evidence, and the speed of case progress, to obtain a real-time overview of the case complexity;

[0086] All update records and basic information are extracted from the case data. The main laws involved in the current case are determined by comparing the records of the number of legal clauses, and classified according to the frequency of citation of the laws and the position of the cited paragraphs in the case. Then, the type and quantity of evidence are screened, and the evidence contained in the case is divided into documentary evidence, physical evidence, and electronic evidence, and their quantity and specific type are recorded. The time series analysis method is used to analyze the speed of case progress. The interval time of each case update and the current case stage are compared with the historical average processing time, and the advancement rate of the current case compared with historical cases is calculated. In this process, all data points of the case need to be normalized to ensure that all data are counted with a unified time window and indicator standard. Finally, the three indicators of the number of legal clauses, the number and type of evidence, and the speed of case progress are classified and recorded to generate a real-time complexity overview of the case.

[0087] Based on the real-time overview of case complexity, adjust the decision support direction of judicial workers, using the formula:

[0088]

[0089] Match the complexity and changing trends of cases, where TQ represents the trend score of case complexity, αq is the adjustment coefficient, which is obtained based on the analysis of historical case data and is used to adjust the sensitivity of the trend score to match different types of cases and changing environments, and cq i is the weight of the i-th complexity index, which determines the influence of each complexity index in the total score. i is the rate of change of the ith complexity indicator during the monitoring period, calculated as the ratio of the difference between the current value and the historical average, reflecting the speed and magnitude of change of each complexity indicator over time, k is the total number of complexity indicators;

[0090] The weights and change rates of the three complexity indicators are as follows: cq1 = 0.4, dq1 = 0.05 (change in the amount of evidence), cq2 = 0.3, dq2 = 0.07 (update of legal provisions), cq3 = 0.3, dq3 = 0.02 (case progress speed), and the adjustment coefficient αq = 1.5, which is used to enhance the response to changes, is substituted into the formula for calculation:

[0091] TQ=1.5·((0.4×0.05)+(0.3×0.07)+(0.3×0.02));

[0092] TQ=1.5·(0.02+0.021+0.006);

[0093] TQ = 1.5 0.047 = 0.0705;

[0094] The result shows that the trend score of case complexity is 0.0705, which reflects the comprehensive changes in the complexity of current cases, which helps legal experts to adjust their strategies and decision-making processes in a timely manner.

[0095] See also Figure 5 , the specific steps for obtaining decision support configuration are:

[0096] Perform data analysis to identify decision points and obtain real-time case analysis information based on real-time updated case data, including new evidence, legal changes, and case update status;

[0097] Carry out real-time monitoring of case data sources, extract each change information, such as newly added evidence types, updated legal clauses, and current case status records, classify the extracted information, and classify new evidence into relevant categories (such as written evidence, physical evidence, etc.), update legal clauses and mark specific changes, such as new clauses, revised content, or abolished clauses, and record the progress of case updates, such as entering trial, waiting for judgment, or other related stages. For all classified and marked information, use structured analysis to form a basic information set for real-time case analysis. By analyzing the correlation of data in the basic information set, further explore the trend of case changes and store it in the form of trend judgment data. Finally, complete the recording and organization of real-time case analysis information.

[0098] Based on real-time case analysis information, evaluate the compatibility and efficiency of existing decision suggestions, optimize according to the evaluation results, and monitor the effects of case handling to obtain decision support configuration;

[0099] Classify and analyze the decision suggestions involved in the information, determine whether the existing suggestions match the current status and trends of the case, for example, determine whether the existing suggestions are applicable to the current newly added types of evidence and legal changes, or whether they provide effective solutions for the current progress of the case, compare the analysis results with the case trend records, identify the inapplicable parts of the decision suggestions, and optimize and update them. During the optimization process, introduce weight parameters according to the complexity and urgency of the case, recalculate the priority ranking of each suggestion, verify the optimized suggestions with the current status of the case, and evaluate the specific effects of the optimized suggestions in case handling through dynamic monitoring, adjust the insufficient parts and form the final optimized decision suggestions, and finally complete the construction of the decision support configuration.

[0100] See also Figure 6 , the steps to evaluate sentence coherence are as follows:

[0101] Based on the case complexity index, extract sentences from judicial documents, use NLP technology to analyze sentence structure, identify subject-verb-object structure, conjunction usage and logical connection between sentences, and obtain sentence structure analysis results;

[0102] The document is split into independent sentences, and each sentence is marked as a structural unit. The subject, predicate and object structures are extracted by decomposing the sentences. The frequency of conjunctions is counted and the sentences before and after each conjunction are recorded. The logical connections between sentences are matched through association analysis methods to generate a sentence connection relationship table. For each incorrectly connected or unconnected part in the connection relationship table, the document context is used to infer the logic between the supplementary sentences. By analyzing the position of the sentence in the document and the degree of logical dependence on other sentences, detailed sentence structure analysis results are generated, including the quantitative distribution of subject, predicate and object, the use of conjunctions and the completeness of logical connections, and finally the sentence structure analysis results are obtained.

[0103] Evaluate the results of sentence structure analysis, compare the logical flow between sentences, and check whether there are logical breaks or semantic repetitions, using the formula:

[0104]

[0105] Calculate the coherence score CL′ of each paragraph of text to get the sentence coherence evaluation result, where rl i is the modified coherence contribution of the i-th sentence, which is dynamically calculated based on the strength of the association between the sentence and the context. i is the position parameter of the ith sentence in the paragraph. The closer to the beginning of the paragraph, the higher its value, reflecting the importance of the sentence in the text structure. i Indicates the number of sentences associated with the i-th sentence, reflecting the sentence dependency depth, n l is the total number of sentences in the document;

[0106] There are three sentences in the document l =3, the contribution of each sentence to corrected coherence is rl1=0.8, rl2=0.6, rl3=0.9, the position parameters are pl1=1, pl2=2, pl3=3, the number of related sentences is sl1=2, sl2=1, sl3=0, substitute into the formula for calculation:

[0107]

[0108] CL′=0.8×0.577+0.6×1.414+0.9×3;

[0109] CL′=0.462+0.849+2.7;

[0110] CL′≈4;

[0111] The result shows that the total coherence score is 4, reflecting the overall coherence of the document in terms of logical flow and structure. The higher the value, the stronger the coherence and logic of the document, which provides a basis for further document optimization.

[0112] Based on the results of sentence coherence assessment, adjust sentence structure, add conjunctions or rewrite incoherent parts to optimize the fluency and logic of judicial documents and obtain feedback results of the corrected documents;

[0113] Analyze each problematic sentence one by one, check the starting and ending points of logical breaks, rearrange the sentence structure of incoherent parts, add appropriate conjunctions to supplement the continuity of logical flow, delete redundant repeated sentences, verify the content of the rewritten part according to the subject-predicate-object structure, and cross-verify with the sentence coherence results to ensure that all adjusted content is consistent with the overall logical framework of the document. In this way, the overall content of the document is fully logically optimized, all adjusted parts are reintegrated and saved as revised versions, and finally the corrected document feedback results are obtained.

[0114] See also Figure 7 , the specific steps for obtaining the logic check feedback result are:

[0115] Review the logical structure of judicial documents, identify each logical relationship and reasoning chain in the document, test whether the logical connection is reasonable, determine whether there is non-causal reasoning and circular reasoning, and obtain preliminary logical error analysis results;

[0116] The document content is extracted in segments, and the logical relationships in each paragraph of text are identified, including causal relationships, progressive relationships, and contrast relationships. Through a combination of manual annotation and automatic analysis tools, the logical connectives in the sentences are detected, and the starting and ending points of the logical connections are recorded. Then, a check is made to see if there are any logical errors, such as non-causal reasoning and circular reasoning. Specifically, the time sequence and correlation of the previous and subsequent events in the causal reasoning are analyzed to ensure that the causal relationship is logical. At the same time, in the detection of circular reasoning, the repeated starting points and conclusions of reasoning are marked, and the rationality of the logical structure is annotated and classified to finally complete the sorting of the logical structure.

[0117] Based on the preliminary logical error analysis results, each identified logical problem is evaluated to determine whether the logical structure in the document meets the legal logic standards and obtain the logical verification feedback results;

[0118] Select each identified logical problem and classify it into types such as argumentation error, causal relationship error or insufficient supporting evidence, and assign the problem to different logic verification standard templates for further analysis. By comparing the specific characteristics of the error with the requirements of the legal logic standard, such as whether the sufficiency and necessity conditions in the argument are met, and the integrity requirements of the logical chain, the severity of each logical problem is gradually quantified and scored. Combined with the verification scores of each part, the overall logical structure of the document is reviewed, and the errors involved in the logic verification process are classified and summarized into a report. Finally, the parts of the document that meet the legal logic standards are marked.

[0119] See also Figure 8 , the specific steps for obtaining judicial document amendment records are:

[0120] Based on the feedback from logic verification, identify potential logic defects in the document, including inconsistent conditions or unclear responsibilities, analyze the document, record each defect and its context, and obtain preliminary defect analysis information;

[0121] The document is analyzed sentence by sentence to identify all potential logical defects, including those with contradictory conditions or unclear responsibilities. For each logical defect, relevant contextual information is extracted, such as the relevance of the establishment of the condition and the corresponding argument, as well as the statement body and conditional variables involved in the division of responsibilities. The context of each defect is marked and classified by establishing a mapping relationship to ensure that the source and scope of influence of the logical contradiction are accurately tracked. At the same time, the structure of repeated statements or invalid arguments is analyzed to further quantify the frequency and distribution of repeated defects in the document. These analysis results are stored as standardized defect descriptions and their contextual relationship indexes, thereby obtaining preliminary defect analysis information.

[0122] Based on the preliminary defect analysis information, perform logical analysis to analyze the impact and priority of the defects using the formula:

[0123]

[0124] Get the correction suggestion result, where DB represents the severity score of the document defect, qb i is the impact weight of the ith defect, which is set according to the degree of impact of the defect on the overall logic and understanding of the document. i is the frequency of the ith defect, that is, the number of times the defect appears in the document. The calculation of the frequency reflects the prevalence of a certain logical error in the document. The higher the frequency, the more common the error is in the document. b is the total number of logic defects;

[0125] Three major logic defects were found in the analyzed documents. The impact weight and frequency of each defect are as follows: qb1=0.5, fb1=3 (critical logic break), qb2=0.3, fb2=5 (repeated statements), qb3=0.2, fb3=2 (conditional contradiction), substitute into the formula for calculation:

[0126] No. 1:

[0127] qb1·log(fb1+1)=0.5·log(3+1)=0.5·log(4)=0.5·0.602=0.301;

[0128] No. 2:

[0129] qb2·log(fb2+1)=0.3·log(5+1)=0.3·log(6)=0.3·0.778=0.2334;

[0130] No. 3:

[0131] qb3·log(fb3+1)=0.2·log(2+1)=0.2·log(3)=0.2·0.477=0.0954;

[0132] Sum calculation

[0133] DB=0.301+0.2334+0.0954≈0.6;

[0134] The results show that the document's defect severity score is 0.6, indicating that the overall impact of logical errors in the document on understanding and logical structure is moderate to high, and correcting the errors will significantly improve the clarity and professionalism of the document.

[0135] Based on the correction suggestion results, correct the logical errors in the document, adjust the sentence structure and argument logic, and obtain the judicial document correction record;

[0136] Correct the logical errors in the document step by step, analyze the type and impact of each logical error, redefine the logical expression of the constraints for the conditional contradictions, ensure that the logical expression is consistent with the context, and for issues where responsibility is unclear, adjust the reasoning process by clarifying the responsible party or reconstructing the sentence structure. Reorganize the content of repetitive logic and invalid argument paragraphs to avoid information redundancy, and add appropriate conjunctions to ensure the coherence of the corrected sentences. For each modification, record the comparison of the documents before and after the correction, including the specific adjustment content and correction logic, and organize them into a systematic document correction list, ultimately forming a judicial document correction record.

[0137] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent service system for cooperation between inspection and proofreading, characterized in that: The system comprises: The judicial document analysis module performs text structured decomposition based on the input judicial documents, extracts the chain of evidence and relevant legal reference information in the case, analyzes the weight of evidence and the correlation between legal provisions, and evaluates the complexity of the case through fuzzy logic to obtain the case complexity index; The decision optimization module adjusts the decision support direction for judicial workers based on the case complexity index, matches the complexity and change trend of the case, and optimizes the decision process and decision suggestions according to the real-time updated data of the case to obtain the decision support configuration; The document logic verification module analyzes the judicial document based on the decision support configuration, identifies and corrects grammatical errors in the document, evaluates sentence coherence, and then reviews the logical structure in the document to determine whether it meets the legal logic standards and obtains logic verification feedback results; The logic defect correction module mines potential logic defects in the document based on the logic verification feedback results, including contradictory conditions or unclear responsibilities, and performs logical analysis on the judicial document to correct the logical errors in the document and obtain a judicial document correction record.

2. The intelligent service system for cooperation between inspection and proofreading according to claim 1 is characterized in that: The steps of text structure decomposition are specifically as follows: Based on the input judicial documents, the text content is extracted, including the original text and the metadata of the document, to obtain a preliminary text dataset; The preliminary text data set is analyzed to identify key text elements, including names of people, places, dates, and legal terms, using the formula: Evaluate the score Gs of entity recognition and obtain structured text with entity tags, where wg i represents the critical weight of the i-th entity, tg i represents the number of times the i-th entity is mentioned in the document, and n is the number of entities; Based on the structured text with entity tags, the chain of evidence and relevant legal reference information in the case are extracted to establish a case information dataset.

3. The intelligent service system for cooperation between inspection and proofreading according to claim 1 is characterized in that: The specific steps for obtaining the case complexity index are as follows: Analyze the chain of evidence and the complexity of legal provisions in the case, including the quantity and type of evidence and its relevance to the legal provisions, and obtain the case analysis log; Fuzzy logic is used to process the data in the case analysis log to evaluate the interaction and correlation between each piece of evidence and legal provisions, using the formula: Calculate the case complexity index CH, where vh j represents the validity weight of the association between the jth piece of evidence and the legal provision, rh j is the correlation score between the jth piece of evidence and the legal provision, and m represents the total number of associations between evidence and legal provisions.

4. The intelligent service system for cooperation between inspection and proofreading according to claim 1 is characterized in that: The steps for matching the complexity and change trend of the case are as follows: Based on the case complexity index, analyze key complexity indicators, including the number of legal provisions, the type and amount of evidence, and the speed of case progress, to obtain a real-time overview of the case complexity; Based on the real-time complexity overview of the case, the decision support direction of judicial workers is adjusted using the formula: The complexity and change trend of matching cases, where TQ represents the trend score of case complexity, αq is the adjustment coefficient, and cq i is the weight of the i-th complexity index, dq i is the rate of change of the ith complexity indicator during the monitoring period, and k is the total number of complexity indicators.

5. The intelligent service system for cooperation between inspection and proofreading according to claim 1 is characterized in that: The steps for obtaining the decision support configuration are specifically as follows: Perform data analysis, identify decision points, and obtain real-time case analysis information based on real-time updated data of the case, including new evidence, legal changes, and case update status; Based on the real-time analysis information of the case, the compatibility and efficiency of the existing decision suggestions are evaluated, optimization is performed according to the evaluation results, and the effects during case handling are monitored to obtain decision support configuration.

6. The intelligent service system for cooperation between inspection and proofreading according to claim 1 is characterized in that: The steps of evaluating sentence coherence are specifically as follows: Based on the case complexity index, extract sentences from judicial documents, analyze sentence structure using NLP technology, identify subject-verb-object structure, conjunction usage and logical connection between sentences, and obtain sentence structure analysis results; The sentence structure analysis results are evaluated, the logical flow between sentences is compared, and whether there is a logical break or semantic repetition is checked using the formula: Calculate the coherence score CL′ of each paragraph of text to get the sentence coherence evaluation result, where rl i is the modified coherence contribution of the i-th sentence, pl i is the position parameter of the i-th sentence in the paragraph, sl i Indicates the number of sentences associated with the i-th sentence, n l is the total number of sentences in the document; Based on the sentence coherence evaluation results, the sentence structure is adjusted, conjunctions are added or incoherent parts are rewritten to optimize the fluency and logic of the judicial document and obtain a corrected document feedback result.

7. The intelligent service system for cooperation between inspection and proofreading according to claim 1 is characterized in that: The steps for obtaining the logic check feedback result are specifically as follows: Review the logical structure of judicial documents, identify each logical relationship and reasoning chain in the document, test whether the logical connection is reasonable, determine whether there is non-causal reasoning and circular reasoning, and obtain preliminary logical error analysis results; Based on the preliminary logical error analysis results, each identified logical problem is evaluated to determine whether the logical structure in the document meets the legal logic standards, and obtain a logical verification feedback result.

8. The intelligent service system for cooperation between inspection and proofreading according to claim 1 is characterized in that: The specific steps for obtaining the judicial document amendment record are: Based on the feedback results of the logic check, potential logic defects in the document are identified, including inconsistent conditions or unclear responsibilities, and the document is analyzed to record each defect and its context to obtain preliminary defect analysis information; Based on the preliminary defect analysis information, perform a logical analysis to analyze the impact and priority of the defects using the formula: Get the correction suggestion result, where DB represents the severity score of the document defect, qb i is the impact weight of the i-th defect, fb i is the frequency of the ith defect, p b is the total number of logic defects; Based on the correction suggestion results, the logical errors in the document are corrected, the sentence structure and argument logic are adjusted, and the judicial document correction record is obtained.

Images