Legal document generation system based on intelligent speech recognition

Abstract

The present invention discloses a legal document generation system based on intelligent voice recognition, and relates to the technical field of legal document generation. The legal document generation system based on intelligent voice recognition calculates the legal document keyword compliance index, constructs a legal text generation model, searches the legal database, generates legal documents, improves the efficiency and quality of legal services, reduces the cost of legal document generation, optimizes the legal service process, calculates the legal document compliance index through user feedback and expert advice, proofreads and corrects the generated legal documents, improves the accuracy of legal documents, optimizes the legal document generation model, continuously optimizes the legal document generation system, and improves the document quality and user satisfaction.

Description

Technical Field

[0001] The present invention relates to the field of legal document generation, and specifically to a legal document generation system based on intelligent speech recognition. Background Art

[0002] With the continuous advancement of artificial intelligence and natural language processing technologies, technologies in fields such as speech recognition, text analysis and automatic generation have developed rapidly. At the same time, with the development of society and the increase in legal affairs, the legal industry has an increasing demand for informatization. Traditional document generation methods usually take a lot of time and manpower.

[0003] For example, the publication number: CN1912902A discloses a legal document generation system and method, including a query module, providing multiple detailed searches to assist users in querying legal documents and their status; an information entry module, providing a guide for the creation of legal documents, instructing users to enter relevant information; a template import module, used to import the template of the legal document corresponding to the above-mentioned information; a clause editing module, used to edit the clauses listed in the template of the above-mentioned imported legal document; a consultation module, used to submit the newly created legal document to other users for review, and other users can express their opinions and conduct consultations on the clauses in the newly created legal document; a storage module, used to generate formal legal documents and store data related to the legal document. Utilizing the present invention, the generation and signing of legal documents can be realized in an electronic way, saving the time of writing and reviewing legal documents.

[0004] For example, Publication No.: CN111274779A discloses a legal document generation system and method based on user experience prediction, which uses an expression form conversion model to query the same professional vocabulary and popular vocabulary in a professional vocabulary library and a popular vocabulary library respectively according to word information, and determines the expression form of the user's legal document according to the proportion between the professional vocabulary and the popular vocabulary obtained by the query and the preset rules, and then processes and converts the legal document according to the expression form, and feeds back the converted legal document to the interactive terminal. The present invention can identify the consulting information provided by the user and provide the user with the required legal document according to the user's understanding ability.

[0005] Based on the above solutions, it is found that there are still some shortcomings in the generation of legal documents, which are specifically reflected in the following aspects: the legal document generation system still has some shortcomings in legal service efficiency and quality, legal document generation cost, legal service process, and user interface and interaction design. Summary of the invention

[0006] In view of the deficiencies of the prior art, the present invention provides a legal document generation system based on intelligent voice recognition, which solves the problems of the above-mentioned background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a legal document generation system based on intelligent voice recognition, including the following modules: a voice recognition module, a model building module, a legal text generation module, a legal text correction module, and a legal database module; the voice recognition module is used to collect the user's initial voice data and obtain the user's voice data; the model building module is used to convert the user's voice data into text form, obtain the user's voice text data, and perform preprocessing to build a legal text generation model; the legal text generation module searches the legal database based on the legal text generation model, generates legal documents, and displays the generated legal documents to users and experts through a visual interface; the legal text correction module is used to obtain user feedback data and expert suggestion data through user feedback on legal documents and expert suggestions on legal documents, calculate the legal document compliance index, which is used to measure the degree of compliance of legal documents, and proofread and correct the generated legal documents; the legal database module is used to update and store legal-related information in the legal database.

[0008] Furthermore, the specific process of converting the user voice data into text form, obtaining the user voice text data, and performing preprocessing is as follows: converting the user voice data into text form: using voice recognition technology to call the voice recognition API to convert the user's voice data into text form; preprocessing the user voice data: cleaning the user voice text data by using regular expressions and text processing libraries to remove duplications and improve incomplete sentences; performing word segmentation on the cleaned user voice text data, dividing the continuous text into single words or phrases, and performing feature extraction, the features include the word frequency, part of speech, and syntactic structure of the user voice text, and serving as the input of the model.

[0009] Furthermore, the specific process of constructing a legal text generation model is as follows: obtaining legal text data from the legal database including legal clauses, case documents, and legal documents, and preprocessing the collected legal text data, including cleaning, word segmentation, and removal of stop words; using a recurrent neural network model as a baseline model for the legal text generation model, using the legal text data as a training set, training the baseline model of the legal text generation model, learning the associations and rules between texts, performing multiple training and fitting, and constructing a legal text generation model.

[0010] Furthermore, based on the legal text generation model, the legal database is searched, and the specific process of generating legal documents is as follows: based on the legal text generation model, the legal text data in the legal database is connected, and the user voice text data is input; the user voice text data is matched and retrieved with the legal text in the database, and the legal clauses, case documents, and legal documents related to the legal text in the database and the user voice text data are queried through text similarity algorithms and information retrieval technology; based on the matched legal clauses, case documents, and legal documents, the legal text generation model is used to generate legal documents.

[0011] Furthermore, the specific process of displaying the generated legal documents to users and experts through a visual interface is as follows: using front-end languages ​​HTML, CSS and JavaScript to create a visual interface for generating legal documents, creating a user voice text input box, and uploading user voice text data; loading the user voice text data into the legal text generation model, producing legal documents, organizing legal documents using titles, paragraphs and list elements, reasonably setting line spacing, character spacing and paragraph spacing, using citation format to annotate relevant legal text data, and displaying legal documents through visual coding.

[0012] Furthermore, the specific calculation process of obtaining user feedback data and expert suggestion data through user feedback on legal documents and expert suggestions on legal documents is as follows: the user feedback data includes the user's score for the completeness of the legal document, the user's score for the accuracy of the legal document, and the user's score for the readability of the legal document; the expert suggestion data includes the expert's score for the standardization of the legal document, the expert's score for the professionalism of the legal document, and the expert's score for the practicality of the legal document.

[0013] Furthermore, the calculation process of the compliance index of the legal document is as follows: based on user feedback data, evaluate the user's satisfaction with the legal document and calculate the user's feedback index for the legal document; based on expert advice data, evaluate the quality of the legal document and calculate the expert's advice index for the legal document; based on the user feedback index and the expert advice index, assign weights to the user feedback index and the expert advice index to calculate the compliance index of the legal document.

[0014] Furthermore, the calculation formula of the legal instrument compliance index is as follows:

[0015]

[0016] In the formula, λ represents the compliance index of legal documents, δ represents the user feedback index on legal documents, γ represents the expert suggestion index on legal documents, β1 represents the weight factor of the user feedback index on legal documents corresponding to the compliance index of legal documents, and β2 represents the weight factor of the expert suggestion index on legal documents to the compliance index of legal documents.

[0017] Furthermore, the specific process of proofreading and correcting the generated legal document is as follows: based on the legal document compliance index, compare it with the standard value of the legal document compliance index, and calculate the difference between the legal document compliance index and the standard value of the legal document compliance index; compare the difference between the legal document compliance index and the standard value of the legal document compliance index with the allowable deviation value of the legal document compliance index to calculate the legal document generation compliance value; based on the legal document generation compliance value, compare it with the set standard value of the legal document generation compliance value. If the legal document generation compliance value is less than the set standard value of the legal document generation compliance value, send the generated legal document to the expert to proofread and correct the generated legal document; the calculation formula for calculating the legal document generation compliance value is as follows:

[0018]

[0019] In the formula, ψ represents the compliance value generated by the legal instrument, which is used to evaluate the compliance index of the legal instrument, λ represents the compliance index of the legal instrument, λ1 represents the standard value of the compliance index of the legal instrument, and Δλ represents the allowable deviation value of the compliance index of the legal instrument.

[0020] Furthermore, the specific process of updating and storing legal-related information in the legal database is as follows: the legal-related information in the legal database includes legal text data, expert advice data, user feedback index on legal documents, expert advice index on legal documents, legal document compliance index, standard value of legal document compliance index and allowable deviation value of legal document compliance index.

[0021] The present invention has the following beneficial effects:

[0022] (1) The legal document generation system based on intelligent voice recognition calculates the legal document keyword matching index based on user voice text data, constructs a legal text generation model, searches the legal database, generates legal documents, improves the efficiency and quality of legal services, reduces the cost of legal document generation, and optimizes the legal service process. It has important application value and promotion prospects.

[0023] (2) The legal document generation system based on intelligent voice recognition obtains user feedback data and expert advice data through user feedback and expert advice, calculates the legal document compliance index, proofreads and corrects the generated legal documents, improves the accuracy of legal documents, optimizes the legal document generation model, continuously optimizes the legal document generation system, and improves the document quality and user satisfaction.

[0024] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a flow chart of the legal document generation system based on intelligent voice recognition of the present invention.

[0026] Figure 2 Flowchart for calculating the legal instrument compliance index for the present invention. DETAILED DESCRIPTION

[0027] The embodiment of the present application solves the problem of manual input and editing in the process of writing legal documents, which consumes a lot of time and human resources, by using a legal document generation system based on intelligent voice recognition.

[0028] The overall idea of ​​the problem in the embodiment of this application is as follows:

[0029] Collect the user's initial voice data, obtain the user's voice data, convert the user's voice data into text form, obtain the user's voice text data, and perform preprocessing to build a legal text generation model.

[0030] Based on the legal text generation model, the legal database is searched to generate legal documents, which are then presented to users and experts through a visual interface.

[0031] Based on legal documents, through user feedback on legal documents and expert suggestions on legal documents, we obtain user feedback data and expert suggestion data, calculate the legal document compliance index, and proofread and revise the generated legal documents.

[0032] See also Figure 1 The embodiment of the present invention provides a technical solution: a legal document generation system based on intelligent voice recognition, comprising the following modules: a voice recognition module, a model building module, a legal text generation module, a legal text correction module, and a legal database module; the voice recognition module is used to collect the user's initial voice data and obtain the user's voice data; the model building module is used to convert the user's voice data into text form, obtain the user's voice text data, and perform preprocessing to build a legal text generation model; the legal text generation module searches the legal database based on the legal text generation model, generates legal documents, and displays the generated legal documents to users and experts through a visual interface; the legal text correction module is used to obtain user feedback data and expert suggestion data through user feedback on legal documents and expert suggestions on legal documents, calculate the legal document compliance index, which is used to measure the degree of compliance of legal documents, and proofread and correct the generated legal documents; the legal database module is used to update and store legal-related information in the legal database.

[0033] Specifically, the specific process of converting user voice data into text form, obtaining user voice text data, and preprocessing is as follows: Convert user voice data into text form: Use voice recognition technology to call the voice recognition API to convert the user's voice data into text form; Preprocess user voice data: Clean the user voice text data by using regular expressions and text processing libraries to remove duplications and improve incomplete sentences; Perform word segmentation on the cleaned user voice text data, divide the continuous text into single words or phrases, and perform feature extraction. The features include word frequency, part of speech, and syntactic structure of the user voice text, and serve as input to the model.

[0034] In this implementation scheme, by performing feature extraction on user voice text data, the performance of the model can be further optimized and the quality of documents generated by the system can be improved. The purpose of using regular expressions and text processing libraries for cleaning is to remove useless content in user voice text data, including repeated sentences and missing words, to ensure clean input data; the purpose of word segmentation is to divide continuous text into individual words or phrases, and extract the features of these words and phrases, including word frequency, part of speech and syntactic structure. These features can be used as input to the model to help the model better understand the meaning and structure of user voice text data, thereby improving the accuracy and quality of text generation.

[0035] Specifically, the specific process of constructing a legal text generation model is as follows: obtain legal text data from the legal database, including legal clauses, case documents, and legal documents, and preprocess the collected legal text data, including cleaning, word segmentation, and removal of stop words; use a recurrent neural network model as the benchmark model of the legal text generation model, use the legal text data as a training set, train the benchmark model of the legal text generation model, learn the associations and rules between texts, perform multiple training and fitting, and construct a legal text generation model.

[0036] In this implementation scheme, the legal text data in the legal database is obtained to provide a rich data foundation for the legal document generation system; preprocessing the collected legal text data is an important step in building a legal text generation model. Preprocessing operations such as cleaning, word segmentation, and removal of stop words can extract useful information and reduce noise, thereby improving the effect and performance of the model; using a recurrent neural network (RNN) as the benchmark model of the legal text generation model can help the system understand the associations and rules between texts. Through multiple training and fitting, the model can be gradually optimized to improve the accuracy and fluency of the generated text; based on the constructed legal text generation model, the legal document generation system with intelligent voice recognition can generate document content that meets legal requirements. The model can generate accurate legal basis and references based on user needs and inputs, standardize the format and content of documents, and improve the quality of document generation.

[0037] Specifically, based on the legal text generation model, the legal database is searched and the specific process of generating legal documents is as follows: based on the legal text generation model, the legal text data in the legal database is connected and the user voice text data is input; the user voice text data is matched and retrieved with the legal text in the database, and the legal clauses, case documents, and legal documents related to the legal text in the database and the user voice text data are queried through text similarity algorithms and information retrieval technology; based on the matched legal clauses, case documents, and legal documents, the legal text generation model is used to generate legal documents.

[0038] In this implementation scheme, by connecting to the legal database and using text similarity algorithms and information retrieval technology to match and retrieve the user's voice text data with the legal text in the database, an automated legal document retrieval process can be realized to improve the accuracy and efficiency of the retrieval; by querying the legal clauses, case documents and legal documents related to the user's voice text data in the database, relevant legal basis is provided for the generation of legal documents, ensuring that the generated legal documents meet legal requirements and provide accurate legal references and basis; according to the matched legal clauses, case documents and legal documents, the constructed legal text generation model is used to generate corresponding legal documents. The model can generate legal documents that meet the requirements based on the input legal basis and content requirements, and improve the quality and efficiency of document generation; through the process of automating the retrieval and generation of legal documents, the legal document generation system based on intelligent voice recognition can improve the user's experience and work efficiency. The user only needs to input relevant requirements through voice, and the system will automatically retrieve relevant legal texts and generate corresponding legal documents, eliminating the tedious process of manual retrieval and writing; the text similarity algorithm is an algorithm used to compare the similarity between two texts. Commonly used text similarity algorithms include cosine similarity and Jaccard similarity. These algorithms can calculate the similarity between two texts, so as to perform text classification, clustering, matching, deduplication and other operations; information retrieval technology is a technology used to process large-scale text data, including document preprocessing, index construction, query processing and evaluation strategies. The information retrieval system can convert text content into structured information that can be quickly processed by computers by establishing an index structure, so as to facilitate rapid retrieval and matching of large amounts of text data.

[0039] Specifically, the specific process of displaying the generated legal documents to users and experts through a visual interface is as follows: use the front-end languages ​​HTML, CSS and JavaScript to create a visual interface for generating legal documents, create a user voice text input box, and upload user voice text data; load the user voice text data into the legal text generation model, produce legal documents, organize legal documents using titles, paragraphs and list elements, reasonably set line spacing, character spacing and paragraph spacing, use citation format to annotate relevant legal text data, and display legal documents through visual coding.

[0040] In this implementation, by creating a user voice text input box and uploading user voice text data, users can easily input or upload their needs, improving user convenience and experience; loading user voice text data into the legal text generation model can automatically generate legal documents that meet user needs. By rationally organizing the document structure, setting styles, and annotating related content, legal documents can be made more standardized, easy to read, and easy to understand; by displaying legal documents through visual coding, the structure and content of legal documents can be displayed in a graphical way, making it easier for users and experts to understand and evaluate the quality and compliance of legal documents; by displaying the generated legal documents through a visual interface, users and experts can view, evaluate, and make suggestions or modifications to the documents, thereby promoting the effectiveness of communication and feedback, and further optimizing and improving the quality of legal documents.

[0041] Specifically, through user feedback on legal documents and expert suggestions on legal documents, the specific calculation process of obtaining user feedback data and expert suggestion data is as follows: user feedback data includes user's score for the completeness of legal documents, user's score for the accuracy of legal documents, and user's score for the readability of legal documents; expert suggestion data includes expert's score for the standardization of legal documents, expert's score for the professionalism of legal documents, and expert's score for the practicality of legal documents.

[0042] In this implementation plan, users score the completeness, accuracy and readability of legal documents to help with the overall satisfaction and quality assessment of the generated legal documents; experts score the standardization, professionalism and practicality of legal documents to help professionals evaluate and advise on legal documents, provide valuable professional opinions to the system, and guide the system to improve and optimize the standardization, professionalism and practicality of legal documents to improve the quality and accuracy of legal documents; the system can conduct comprehensive evaluation and analysis by integrating user feedback data and expert advice data. The system can perform weighted calculations on user feedback data and expert advice data according to the weight factors of different indicators to obtain a comprehensive legal document quality assessment. The system can optimize and improve the legal document generation algorithm and model based on the evaluation results to provide legal documents that better meet user needs and professional standards.

[0043] Specifically, the calculation process of the compliance index of the legal document is as follows: based on user feedback data, evaluate the user's satisfaction with the legal document and calculate the user's feedback index for the legal document; based on expert advice data, evaluate the quality of the legal document and calculate the expert's advice index for the legal document; based on the user feedback index and the expert advice index, assign weights to the user feedback index and the expert advice index to calculate the compliance index of the legal document.

[0044] In this implementation scheme, based on user feedback data, each user feedback is numbered 1, 2, 3..., i,... n, where n represents the total number of feedbacks. The calculation formula for the user feedback index on legal documents is as follows:

[0045]

[0046] In the formula, δ represents the user's feedback index on legal documents, PLc i represents the score of the completeness of the legal document given by the i-th user feedback, PLa i represents the accuracy score of the legal document given by the i-th user feedback, PLf i represents the readability score of the legal document in the i-th user feedback, PLc1 represents the standard value of the user's score for the completeness of the legal document, PLa1 represents the standard value of the user's score for the accuracy of the legal document, PLf1 represents the standard value of the user's score for the readability of the legal document, α1 represents the weight factor of the user's score for the completeness of the legal document, α2 represents the weight factor of the user's score for the accuracy of the legal document, and α3 represents the weight factor of the user's score for the readability of the legal document; based on the expert advice data, each expert advice is numbered 1, 2, 3..., j,... m, m represents the total number of advices, and the calculation formula for the user's feedback index on the legal document is as follows:

[0047]

[0048] In the formula, γ represents the expert's suggestion index for legal documents, ZLs j represents the jth expert recommendation on the normativeness of the legal document, ZLd j represents the professional rating of the legal document given by the expert in the jth time, ZLu j represents the j-th expert recommendation on the practicality score of the legal document, ZLs1 represents the standard value of the expert's score on the normativeness of the legal document, ZLd1 represents the standard value of the expert's score on the professionalism of the legal document, ZLu1 represents the standard value of the expert's score on the practicality of the legal document, ΔZLu represents the allowable deviation value of the expert's score on the practicality of the legal document, δ1 represents the weight factor of the expert's score on the normativeness of the legal document, δ2 represents the weight factor of the expert's score on the professionalism of the legal document, and δ3 represents the weight factor of the expert's score on the practicality of the legal document.

[0049] See also Figure 2 Specifically, the calculation formula of the legal instrument compliance index is as follows:

[0050]

[0051] In the formula, λ represents the legal document compliance index, which is used to measure the degree of compliance of legal documents, δ represents the user feedback index on legal documents, γ represents the expert suggestion index on legal documents, β1 represents the weight factor of the user feedback index on legal documents corresponding to the legal document compliance index, and β2 represents the weight factor of the expert suggestion index on legal documents corresponding to the legal document compliance index.

[0052] In this implementation scheme, through the user's feedback index on legal documents, the system can understand the user's satisfaction and needs, and then make improvements based on the user's feedback to improve the quality and compliance of legal documents. The expert's suggestion index on legal documents can adjust and improve the legal documents based on the expert's evaluation to improve their quality and accuracy: the weight factor can be used to balance the impact of user feedback and expert advice on the compliance index of legal documents. The system can set different weights according to actual needs to better reflect the opinions of users and experts.

[0053] Specifically, the specific process of proofreading and correcting the generated legal document is as follows: based on the legal document compliance index, compare it with the standard value of the legal document compliance index, calculate the difference between the legal document compliance index and the standard value of the legal document compliance index; if the difference between the legal document compliance index and the standard value of the legal document compliance index is less than the allowable deviation value of the legal document compliance index, send the generated legal document to an expert for proofreading and correction.

[0054] In this implementation scheme, based on the legal document compliance index, the difference between the legal document compliance index and the standard value of the legal document compliance index is calculated; the difference between the legal document compliance index and the standard value of the legal document compliance index is compared with the allowed deviation value of the legal document compliance index to calculate the legal document generation compliance value; based on the legal document generation compliance value, the comparison is made with the set standard value of the legal document generation compliance value. If the legal document generation compliance value is less than the set standard value of the legal document generation compliance value, the generated legal document is sent to an expert to proofread and correct the generated legal document; the calculation formula for calculating the legal document generation compliance value is as follows:

[0055]

[0056] In the formula, ψ represents the compliance value generated by the legal instrument, which is used to evaluate the compliance index of the legal instrument, λ represents the compliance index of the legal instrument, λ1 represents the standard value of the compliance index of the legal instrument, and Δλ represents the allowable deviation value of the compliance index of the legal instrument.

[0057] Specifically, the specific process of updating and storing legal related information in the legal database is as follows: the legal related information in the legal database includes legal text data, expert advice data, user feedback index on legal documents, expert advice index on legal documents, legal document compliance index, standard value of legal document compliance index and allowable deviation value of legal document compliance index.

[0058] In summary, this application has at least the following effects:

[0059] The legal document generation system based on intelligent voice recognition calculates the legal document keyword compliance index based on user voice text data, builds a legal text generation model, searches the legal database, generates legal documents, improves the efficiency and quality of legal services, reduces the cost of legal document generation, and optimizes the legal service process. Through user feedback and expert advice, user feedback data and expert advice data are obtained, and the legal document compliance index is calculated to measure the degree of compliance of legal documents. The generated legal documents are proofread and revised to improve the accuracy of legal documents, optimize the legal document generation model, and continuously optimize the legal document generation system to improve document quality and user satisfaction.

[0060] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0061] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0062] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0063] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0064] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0065] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. The legal document generation system based on intelligent speech recognition is characterized by: Includes the following modules: Speech recognition module, model building module, legal text generation module, legal text revision module, legal database module; The speech recognition module is used to collect the initial speech data of the user and obtain the user speech data; The model building module is used to convert user voice data into text form, obtain user voice text data, and perform preprocessing to build a legal text generation model; The legal text generation module searches the legal database based on the legal text generation model, generates legal documents, and displays the generated legal documents to users and experts through a visual interface; The legal text correction module is used to obtain user feedback data and expert suggestion data through user feedback on legal documents and expert suggestions on legal documents, calculate the legal document compliance index, which is used to measure the degree of compliance of legal documents, and proofread and correct the generated legal documents; The legal database module is used to update and store legal related information in the legal database; The specific process of proofreading and revising the generated legal document is as follows: Based on the legal document compliance index, the difference between the legal document compliance index and the standard value of the legal document compliance index is calculated by comparing the legal document compliance index with the standard value of the legal document compliance index; The difference between the legal document compliance index and the standard value of the legal document compliance index is compared with the allowable deviation value of the legal document compliance index to calculate the compliance value generated by the legal document; Based on the compliance value generated by the legal document, the compliance value is compared with the set standard value of the compliance value generated by the legal document. If the compliance value generated by the legal document is less than the set standard value of the compliance value generated by the legal document, the generated legal document is sent to an expert for proofreading and correction of the generated legal document; The calculation formula for calculating the compliance value generated by the legal document is as follows: In the formula, ψ represents the compliance value generated by the legal instrument, which is used to evaluate the compliance index of the legal instrument, λ represents the compliance index of the legal instrument, λ1 represents the standard value of the compliance index of the legal instrument, and Δλ represents the allowable deviation value of the compliance index of the legal instrument.

2. The legal document generation system based on intelligent voice recognition according to claim 1 is characterized by: The specific process of converting user voice data into text form, obtaining user voice text data, and performing preprocessing is as follows: Converting the user's voice data into text form: using voice recognition technology to call the voice recognition API to convert the user's voice data into text form; Preprocessing the user voice data: cleaning the user voice text data by using regular expressions and text processing libraries to remove repetitions and improve incomplete sentences; The cleaned user voice text data is segmented to divide the continuous text into single words or phrases, and feature extraction is performed. The features include word frequency, part of speech, and syntactic structure of the user voice text, and serve as the input of the model.

3. The legal document generation system based on intelligent voice recognition according to claim 2 is characterized by: The specific process of building a legal text generation model is as follows: Obtain legal text data from the legal database, including legal clauses, case documents, and legal documents, and pre-process the collected legal text data, including cleaning, word segmentation, and removal of stop words; The recurrent neural network model is used as the benchmark model of the legal text generation model. The legal text data is used as the training set to train the benchmark model of the legal text generation model, learn the associations and rules between texts, perform multiple training and fitting, and construct a legal text generation model.

4. The legal document generation system based on intelligent voice recognition according to claim 1 is characterized by: Based on the legal text generation model, the specific process of searching the legal database and generating legal documents is as follows: Based on the legal text generation model, connect the legal text data in the legal database and input the user's voice text data; Match and retrieve the user's voice text data with the legal text in the database, and query the legal clauses, case documents, and legal documents in the database that are related to the legal text and the user's voice text data through text similarity matching and information retrieval technology; Based on the matched legal clauses, case documents, and legal documents, the legal text generation model is used to generate legal documents.

5. The legal document generation system based on intelligent voice recognition according to claim 1 is characterized by: The specific process of generating legal documents and presenting them to users and experts through a visual interface is as follows: Use front-end languages ​​HTML, CSS and JavaScript to create a visual interface for generating legal documents, create a user voice and text input box, and upload user voice and text data; Load user voice text data into the legal text generation model to produce legal documents, organize legal documents using titles, paragraphs, and list elements, reasonably set line spacing, character spacing, and paragraph spacing, annotate relevant legal text data using citation formats, and display legal documents through visual coding.

6. The legal document generation system based on intelligent voice recognition according to claim 1 is characterized by: The specific calculation process of obtaining user feedback data and expert suggestion data through user feedback on legal documents and expert suggestion on legal documents is as follows: Based on user feedback on legal documents and expert suggestions on legal documents, user feedback data and expert suggestion data are obtained; The user feedback data includes the user's score on the completeness of the legal document, the user's score on the accuracy of the legal document, and the user's score on the readability of the legal document; The expert advice data include experts' scores on the normativeness of legal documents, experts' scores on the professionalism of legal documents, and experts' scores on the practicality of legal documents.

7. The legal document generation system based on intelligent voice recognition according to claim 1 is characterized by: The calculation process of the compliance index of the legal documents in question is as follows: Based on user feedback data, evaluate user satisfaction with legal documents and calculate the user feedback index for legal documents; Based on the expert advice data, the quality of legal documents is evaluated and the expert advice index for legal documents is calculated; Based on the user feedback index and the expert advice index, weights are assigned to the user feedback index and the expert advice index, and the legal document compliance index is calculated.

8. The legal document generation system based on intelligent voice recognition according to claim 7 is characterized by: The calculation formula of the legal instrument compliance index is as follows: In the formula, λ represents the compliance index of legal documents, δ represents the user feedback index on legal documents, γ represents the expert suggestion index on legal documents, β1 represents the weight factor of the user feedback index on legal documents corresponding to the compliance index of legal documents, and β2 represents the weight factor of the expert suggestion index on legal documents to the compliance index of legal documents.

9. The legal document generation system based on intelligent voice recognition according to claim 1 is characterized by: The specific process of updating and storing legal related information in the legal database is as follows: The legal related information in the legal database includes legal text data, expert advice data, user feedback index on legal documents, expert advice index on legal documents, legal document compliance index, standard value of legal document compliance index and allowed deviation value of legal document compliance index, legal document generation compliance value and standard value of legal document compliance index.

Images