Big data-based legal consultation service information management system and method

Abstract

The invention relates to the technical field of legal consultation, in particular to a legal consultation service information management system and method based on big data, and the system comprises a consultation partition module which is used for obtaining consultation information during consultation of a user, and judging a consultation partition of the consultation information relative to legal services according to the intention of the user during consultation; the partition mapping module is used for acquiring legal partitions of the consultation information and constructing a mapping relation between consultation partitions and the legal partitions; the law article combination module is used for extracting a key partition from the consultation partition and the legal partition, and judging the combination matching degree among a plurality of law articles in the legal partition according to the key partition to obtain a law article deviation combination; the case updating module is used for carrying out information checking on the legal partitions on the basis of the legal provision deviation combination, identifying the updating conditions of cases in the legal partitions and updating the mapping relation between the consulting partition and the legal partitions according to the updating conditions of the cases; and the accuracy and comprehensiveness of the legal consultation service are improved.

Description

Technical Field

[0001] The present invention relates to the field of legal consulting technology, and in particular to a legal consulting service information management system and method based on big data. Background Art

[0002] Legal consultation refers to the activities of people engaged in legal services to explain, illustrate, propose suggestions and solutions to legal affairs. It is not limited to lawyers answering legal knowledge questions for legal seekers, but involves a wider range of legal workers to explain legal questions. However, with the rapid development of big data technology, the processing and utilization of massive legal data and case information has become possible. In order to better meet users' needs for legal consultation and improve service efficiency and accuracy, it is necessary to combine multiple legal provisions, cases and corresponding user consultation content to complete a comprehensive answer to user consultation.

[0003] For example, Chinese patent publication number CN117473074A discloses an artificial intelligence-based judicial case intelligent information matching system and method, which belongs to the field of judicial information matching. The judicial case intelligent information matching system includes a case monitoring module, a database, a case analysis module and an intelligent matching module. The case monitoring module is used to collect basic data information of historical judicial cases and retrieval information of legal service personnel. The database is used to encrypt and store the collected data information and analysis results. The case analysis module is used to analyze the key content areas and predicted standpoints of legal service personnel and analyze the cited legal provisions. The intelligent matching module is used to adjust the retrieved historical judicial case content according to the analysis results and display the cited legal provisions extracted by analysis.

[0004] For example, Chinese patent publication number CN118113853A discloses a legal consulting service management method and system for auxiliary large language models, which involves the technical field of intelligent language text processing. The method includes: obtaining legal consulting sample data, including user question and answer data, user question and answer audio data and legal knowledge data, performing data conversion, and generating a structured text data set; constructing a seed instruction database based on the user question and answer data, and dividing the structured text data set into multiple basic legal knowledge bases and multiple stylized knowledge bases; training an auxiliary large language model based on the seed instruction database, multiple basic legal knowledge bases and multiple stylized knowledge bases, performing legal knowledge matching, and outputting knowledge question and answer information.

[0005] The prior art describes a method for selecting legal provisions from a legal service personnel plan, and describes the use of a language model to set knowledge question and answer information. These methods represent the processing of legal consultation from two different angles. However, when a legal provision changes, or the direction of user consultation changes, it is difficult to identify the corresponding legal provision combinations in these cases using only the above two methods. It is necessary to promptly discover the legal provision information required by the current user, and use these legal provisions to update the database set up during user consultation to improve the efficiency and accuracy of user consultation. Summary of the invention

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a legal consulting service information management system based on big data, including: a consulting partition module, which is used to obtain consulting information during user consultation, and judge the consulting partition of the consulting information relative to the legal service according to the user's intention during consultation.

[0007] The partition mapping module is used to obtain the legal partition of the consulting information and build a mapping relationship between the consulting partition and the legal partition.

[0008] The legal provisions combination module is used to extract key partitions from the consulting partition and the legal partition, and judge the combination matching degree between multiple legal provisions in the legal partition according to the key partitions to obtain the biased combination of legal provisions.

[0009] The case update module is used to verify the information of the legal partition based on the combination of legal provisions, identify the update status of the cases in the legal partition, and update the mapping relationship between the consulting partition and the legal partition according to the update status of the cases.

[0010] The consultation management module is used to identify the consultation objectives and consultation sequence of the consultation information according to the mapping relationship between the updated consultation partitions and the legal partitions, and to generate a consultation management library.

[0011] The method for managing legal consulting service information based on big data includes: S1. obtaining consulting information submitted by users, classifying the consulting information using an intent classification model, and identifying intent categories; analyzing synonyms, polysemous words, and related words under the intent categories to form consulting partitions.

[0012] S2. Call the laws and cases related to the consulting information, generate unlabeled law recognition results, filter the law recognition results, and obtain the target law recognition results as the legal partition; according to the information representation in the consulting partition and the descriptive information and category information of the legal partition, construct the mapping relationship between the consulting partition and the legal partition.

[0013] S3. Perform cluster analysis on the consulting partition and the legal partition to determine the key partition; calculate the cosine similarity between the keywords in the key partition, extract the legal provisions, and set the legal provision bias combination.

[0014] S4. Extract the time distribution probability according to the biased combination of legal provisions, set the target path, and fit the combination occurrence probability; compare the combination occurrence probability with the citation of cases in the legal partition, identify and update the cases, and update the mapping relationship between the consultation partition and the legal partition.

[0015] S5. According to the mapping relationship between the updated consultation partitions and the legal partitions, the consultation objectives are extracted and sorted according to the probability of occurrence to obtain the consultation order; the consultation objectives and the consultation order are combined to generate a consultation management library.

[0016] The beneficial effects of the present invention are as follows: 1. The present invention constructs consultation information by classifying and mapping, so that after identifying the user's intention, it can set data partitions related to the user's intention based on the synonyms, polysemous words and related words consulted by the user, and then quickly match relevant legal provisions and cases for these partitions to improve the accuracy of the service.

[0017] 2. The present invention establishes a mapping relationship between consultation partitions and legal partitions, which can automatically call up laws and cases closely related to the consultation information, generate law recognition results, and determine the laws that are highly matched with the target consultation information as the basis for legal partitions, thereby realizing intelligent mapping of consultation partitions and legal partitions; then, a mapping relationship is established based on the information representation in the consultation partition and the detailed description and category information of the legal partition to ensure the accuracy and pertinence of the legal consultation service; finally, by establishing a mapping relationship, relevant legal knowledge and cases can be quickly located to provide users with timely and efficient legal consultation services.

[0018] 3. The present invention conducts in-depth cluster analysis on consulting partitions and legal partitions, identifies key partitions, calculates the similarities between keywords in key partitions, and then extracts legal provisions corresponding to the keywords to set legal provision bias combinations, which can meet user consulting needs; sets the target path according to the time distribution probability of the legal provision bias combination, and obtains the probability of combination occurrence through fitting analysis, which is compared with the citation of cases in the legal partition, accurately identifies and updates case information, and ensures the timeliness of legal consulting services. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0020] Figure 1 It is a system framework diagram of the legal consulting service information management system based on big data.

[0021] Figure 2 It is a flow chart of the consultation partition module of the legal consultation service information management system based on big data.

[0022] Figure 3It is a flow chart of the partition mapping module of the legal consulting service information management system based on big data.

[0023] Figure 4 It is a flow chart of the legal clause combination module of the legal consulting service information management system based on big data.

[0024] Figure 5 It is a flowchart of the legal consulting service information management method based on big data. DETAILED DESCRIPTION

[0025] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. If no specific techniques or conditions are specified in the embodiments, the techniques or conditions described in the literature in the art or the product specifications are used.

[0026] See also Figure 1 The legal consulting service information management system based on big data includes: a consulting partition module, a partition mapping module, a legal article combination module, a case update module and a consulting management module; wherein the output end of the consulting partition module is connected to the partition mapping module, the output end of the partition mapping module is connected to the legal article combination module, the output end of the legal article combination module is connected to the case update module, and the output end of the case update module is connected to the consulting management module.

[0027] The consultation partition module is used to obtain the consultation information of the user during consultation, and determine the consultation partition of the consultation information relative to the legal service according to the user's intention during consultation.

[0028] The partition mapping module is used to obtain the legal partition of the consulting information and build a mapping relationship between the consulting partition and the legal partition.

[0029] The legal provisions combination module is used to extract key partitions from the consulting partition and the legal partition, and judge the combination matching degree between multiple legal provisions in the legal partition according to the key partitions to obtain the biased combination of legal provisions.

[0030] The case update module is used to verify the information of the legal partition based on the combination of legal provisions, identify the update status of the cases in the legal partition, and update the mapping relationship between the consulting partition and the legal partition according to the update status of the cases.

[0031] The consultation management module is used to identify the consultation objectives and consultation sequence of the consultation information according to the mapping relationship between the updated consultation partitions and the legal partitions, and to generate a consultation management library.

[0032] In one embodiment of the present invention, the consultation partitioning set in the consultation partitioning module will set up multiple partitions according to the user's intention during consultation, and set up multiple related partitions according to the synonyms, polysemous words and related words existing in the consultation information. These partitions are combined to obtain the partition of the current consultation information for the user, thereby completing the integration of the corresponding data.

[0033] When users obtain corresponding legal services, due to the different educational levels and environments of the users themselves, the users will have different forms of descriptions for the laws and legal issues they want to understand. At this time, it is necessary to first identify the intent from the consultation information described by the users, and find out the multiple different situations that the users want to identify; to make preliminary settings and complete the preliminary settings of consultation partitions; at this time, when obtaining partitions based on user intentions, it is easy for a word or the meaning of the intention itself to be close, resulting in multiple similar but different descriptions and answers for the current legal issues and terms that want to be consulted. At this time, it is necessary to initially obtain multiple consultation partitions. Afterwards, multiple synonyms and polysemous words with multiple meanings are likely to appear in the user's description vocabulary, and the user's own intentions can be further understood. At the same time, some users may have disordered word order and incorrect vocabulary use. These users do not want to explain directly because their own problems are more prominent. At this time, an additional related word must be obtained to complete the setting of consultation partitions.

[0034] like Figure 2 As shown, the implementation method of the consultation partitioning module includes: for any consultation information during user consultation, obtaining the intent classification model corresponding to the consultation information; the intent classification model adopts logistic regression, deep learning model and the like. For example, logistic regression extracts the current consultation information using word vectors, inputs the word vectors into the logistic regression, and judges the intent probability, loss rate and attention score of all texts existing in the model for different words during training to obtain the intent categories that may be contained in the consultation information; the deep learning model uses multiple convolutional layers and neurons to extract the intent categories in the consultation information to obtain the general direction that the user wants to consult.

[0035] Use the intent classification model to classify the consulting information and obtain at least one intent category; the intent category can be a query, request for help, complaint, or other types of needs. At the same time, this category will also include classifications for a certain situation, such as criminal, civil, and other specific classification labels to determine the current user's general intention.

[0036] Identify the synonyms, polysemous words and related words in the consulting information under the corresponding intent category to form a text sample set; the text sample set will contain the synonyms, polysemous words and related words corresponding to the consulting information under the corresponding intent category. These words are related words of certain words in the consulting information that can directly express the user's intention. At this time, it is necessary to analyze the text sample set, such as synonym analysis, polysemous word analysis and related word extraction. For example, use the context of synonyms to analyze synonyms, use the meaning of polysemous words under context understanding, and related words and co-occurring words to complete the analysis of the text sample set. After that, verify the relationship between the relative entities of these three words to cover the content involved in the current user consultation. These contents will be represented as multiple data sets of synonym area, polysemous word analysis area, and related word association area, that is, the described consultation partition. This represents the distribution of different types of data sets. The purpose is to clearly show the role of each word in its specific context. Finally, these data are converted into a structured format to facilitate the subsequent setting of mapping relationships.

[0037] The synonyms, polysemous words and related words in the text sample set are analyzed respectively, and the synonym area, polysemous word analysis area and related word association area corresponding to the text sample set are obtained in turn, and used as the consulting partition of the consulting information relative to the legal service.

[0038] The synonym area adopts the form of a synonym dictionary, extracts the synonyms of the corresponding vocabulary under the currently recognized intent category, and obtains a synonym area. This synonym area lists the common synonym pairs contained in the current consultation information, and explains the synonym pairs that appear at this time, including word meaning, usage, difference, etc.; then provides example sentences to help users understand the use of synonyms in different contexts.

[0039] The polysemous word analysis area lists the polysemous words in current legal consultations, and provides a detailed explanation of each meaning of each polysemous word, including word meaning, usage, applicable objects, etc., while revealing the internal connections between the meanings, such as the core meaning, extended meaning, etc.; then provides example sentences to show the usage of polysemous words in different contexts.

[0040] The related word association area establishes an association network between words based on the semantic relationship of legal consultation words, that is, the related words that appear in the current consultation are marked to facilitate users to view the analysis of related words; then the common collocation and usage of related words in legal consultation are displayed. In this way, the consultation partition currently required is formed. The form of the consultation partition is to represent the corresponding words that may exist under the main intention of the user, and the content of the explanation and description of these words forms a data set similar to the area, so that users can get enough content description from these corresponding partition data after completing the consultation. These partitions are also the part of the subsequent processing to accurately filter out the text content required by the user.

[0041] When obtaining the synonym area, polysemous word analysis area and related word association area corresponding to the text sample set, in order to distinguish the impact of different intent categories on these partitions, the implementation method also includes: merging the synonyms, polysemous words and related words in the text sample set according to the intent category to obtain multiple intent merging results. The intent merging result indicates that the corresponding contents of the synonyms, polysemous words and related words are merged according to the intent category to obtain all the data for a consulting information, and delete the repeated and missing contents therein.

[0042] Extract synonym pairs from the intention merging results, compare the synonym pairs with the synonym dictionary, and obtain the synonym area.

[0043] The meaning change rate of polysemous words and the association strength of related words in the intention merging results are extracted, and the intention merging results are divided according to the meaning change rate of polysemous words and the association strength of related words to obtain the polysemous word analysis area and the related word association area.

[0044] The interpretation change rate of a polysemous word indicates the degree to which the interpretation of a polysemous word will change after the user's consultation information is merged according to the intent category. At this time, the semantic distance between interpretations, such as cosine similarity, is used to quantify the degree of interpretation change at this time, and the ratio of this degree to the degree of interpretation change of all current polysemous words is regarded as the interpretation change rate of the polysemous word, which reflects the degree to which the meaning of the polysemous words in the corresponding context changes after the polysemous words are merged according to the intent category in a paragraph. When the interpretation change rate of polysemous words is large, it generally means that the meaning of such words depends largely on the specific context. For example, the word "execution" may have different meanings in different legal contexts; this part of words needs to be determined by considering the context; when the interpretation change rate of polysemous words is small, it generally indicates words with fixed meanings, such as wills. Whether discussing inheritance distribution, inheritance law or family disputes, the basic definition of a will is a formal written statement of an individual's arrangements for his or her affairs after his or her death. Then, according to the interpretation change rate of polysemous words, the polysemous words at this time are divided to obtain the polysemous word analysis area corresponding to the polysemous words.

[0045] The association strength of related words indicates the probability of related words and words corresponding to the intent category appearing, and measures whether the association strength value of related words will change after merging, that is, when two words corresponding to the intent category correspond to one related word, the association strength of related words will increase, or when two related words correspond to one word corresponding to the intent category, the association strength of related words will also increase. Then, the probability of co-occurrence of these merged words is calculated to change the original association strength value. Finally, the related words in the intent merging result will be divided according to the association strength of related words to obtain related word association areas under multiple different values. If the association strength of related words is large, it generally means that these related words will indicate more specific legal case details in the co-occurring legal clauses or specific legal issues; when the association strength of related words is small, it generally means some loose information, which may be from a specific issue to a broad background or overview. These will indicate that the focus of the content expressed by the user is relatively scattered, or the relevant issues are not clearly presented. Based on the correlation strength of these related words and the rate of change in the meaning of polysemous words, it is possible to improve the adjustment of the user's answer type when the user's description and emphasis change or differ, thereby improving the efficiency of communication with the user.

[0046] In order to make the content displayed in the consultation partition more comprehensive, when judging the consultation partition of consultation information relative to legal services, it can also include: judging the consultation partition, analyzing the number of consultations and consultation time of users in the consultation partition. The number of consultations and consultation time obtained here are the synonym areas, polysemous word analysis areas and related word association areas in the consultation partition. According to the number of times the data in this partition is consulted when the user consults, and the corresponding time of the consultation, the consultation partition is fitted according to the number of consultations and consultation time, so that the corresponding vocabulary in the consultation partition gradually approaches the actual intention of the user; constructing a mapping relationship from the consultation partition to the actual intention of the user to form a mapping situation between the current consultation partitions, and completing the judgment of the consultation partition of consultation information relative to legal services.

[0047] When fitting consultation partitions according to the number of consultations and consultation time, it is necessary to generate data points for the consultation partitions based on the distribution of consultation time, and fit the data of the number of consultations changing over time to obtain the trend of the number of consultations changing over time. Then, the slope value of the trend of the number of consultations changing over time is used as the mapping between the current consultation partitions. This method is to show the changing trend of users' behavior when consulting on legal issues, so as to facilitate the prediction of the information that users want to consult later.

[0048] In one embodiment of the present invention, the partition mapping module mainly sets up partitions for the French and cases that may correspond to the current consultation information, and forms a mapping relationship between the relevant content in this partition and the currently divided consultation partition, so as to facilitate subsequent verification and adjustment, find the accurate French and cases in the user consultation, and complete the comprehensive classification management of the user's actual answers and related information.

[0049] When mapping the zones, what is actually happening is finding cases and laws that may have the same intent as the current consultation zone, in order to determine the key issues that need to be described and addressed in the current consultation zone.

[0050] When constructing legal partitions, the partition mapping module will find corresponding laws and cases from the consulting information, and set this corresponding content as legal partitions. Then, it will use a structured formal language to set a mapping relationship between the legal partitions and the consulting partitions. At the same time, in order to improve the specific content of the division, the consulting information will generate multiple unlabeled legal article recognition results, and when the number of legal article recognition results is greater than a certain threshold, the corresponding content will be combined into legal partitions. Then, the similarity value of these unlabeled data will be calculated, and the legal article recognition results will be adjusted according to the similarity value, and finally a relatively comprehensive legal partition will be obtained.

[0051] like Figure 3 As shown, the implementation method of the partition mapping module includes: calling the laws and cases corresponding to the consulting information, generating multiple unlabeled law recognition results, and the unlabeled law recognition results represent the laws and cases that are not associated with the vocabulary in the consulting partition.

[0052] It is determined whether multiple unlabeled legal article recognition results are target legal article recognition results. If they are target legal article recognition results, the target legal article recognition results are regarded as legal partitions of the consulting information.

[0053] The implementation method of judging whether multiple unlabeled legal article recognition results are target legal article recognition results includes: judging in turn whether the number of current legal article recognition results is greater than a first number threshold, and if it is greater than the first number threshold and the similarity value of the current legal article recognition result is greater than the first similarity threshold, then the current legal article recognition result is used as the target legal article recognition result. At this time, when the number of legal article recognition results is large, it means that there are enough independent results to form one or more meaningful legal partitions; when the number is small, it means that the recognized results are not enough to form a comprehensive or meaningful legal partition alone; at the same time, the similarity value will represent the similarity of the legal article recognition results in content, and the similarity value will calculate the corresponding legal articles and cases in the legal article recognition results, and then judge the relationship between these data one by one as a data set corresponding to a required target legal article recognition result. The similarity value of this legal article recognition result will use the average value of the similarity value of the data in the entire result to represent the overall similarity. If the similarity value of the legal article recognition results is greater than the first similarity threshold, it means that these results point to the same legal problem or solution. If it is less than the first similarity threshold, it means that these results involve different legal clauses, cases or interpretations, or they are related but have different focuses. At this time, these results must be screened to form the legal partitions currently required. After these legal partitions are processed accordingly, their mapping relationships will be used as a reference for future legal consultation processing to improve the overall processing effect. At the same time, the above-mentioned first quantity threshold and first similarity threshold will use historical data to generate or manage legal information for consultation for users, and the average quantity and similarity of the data sent to users will be used as the first quantity threshold and first similarity threshold used at this time.

[0054] If it is greater than the first quantity threshold and the similarity value of the current law recognition result is less than the first similarity threshold, the intersection of the target law recognition result generated last time and the current law recognition result is regarded as the target law recognition result. The intersection is used here to indicate finding the laws or cases that are commonly recognized in the two results, and these laws or cases are considered to be related to the consultation information in both recognitions. The existence of the intersection indicates that these laws or cases have a high relevance and importance because they are confirmed in different batches of recognition results.

[0055] If it is less than the first quantity threshold and the similarity value of the current legal article recognition result is greater than the first similarity threshold, the current legal article recognition result is superimposed with the adjacent legal article recognition result as the target legal article recognition result. The situation here is that when the number of legal articles or data to be recognized is small, the data is supplemented from the adjacent legal article recognition results to form a more complete and comprehensive legal article set.

[0056] If it is less than the first quantity threshold and the similarity value of the current legal article recognition result is less than the first similarity threshold, it is determined that the current legal article recognition result is not the target legal article recognition result.

[0057] After the target legal article identification result is completed here, the required legal partition is obtained, and then the consulting partition and the legal partition are mapped according to the legal logic. The legal logic described here is represented by the types of legal articles involved, such as substantive law, procedural law, administrative regulations, etc. According to these contents, the legal relationships involved in the consulting partition, such as contractual relationships, tort relationships, labor relationships, etc., are gradually mapped to the legal partition to complete the mapping of consulting partitions and legal partitions.

[0058] Therefore, the implementation method of constructing the mapping relationship between the consultation partition and the legal partition includes: using the information representation of the synonym area, polysemous word analysis area and related word association area existing in the consultation partition, the description information of the legal partition and the category of the legal partition to construct a mapping relationship between the consultation partition and the legal partition.

[0059] The information representation of the synonym area, polysemous word analysis area and related word association area in the consultation partition refers to the information corresponding to these synonyms, polysemous words and related words in the consultation partition. After mapping this information with the description of a certain law or case in the legal partition, it is mapped with the category of the legal partition, thereby completing the mapping relationship between the consultation partition and the legal partition.

[0060] In one embodiment of the present invention, the legal article combination module is mainly used to extract key partitions from the consulting partition and the legal partition, and then judge the corresponding legal articles in these key partitions to identify multiple groups of legal articles that can be combined. The combination matching degree is to judge whether the legal articles in the legal partition can be combined together when the key partition combines the legal articles at this time, and whether they can be used as the content of the answer to the same consulting information, and then find out the consulting goals under a certain matching degree after these legal article combinations to complete the recommendation and combination of legal articles.

[0061] In the law article combination module, the key partition is the main content extracted from the consulting partition and the legal partition. This part of the content will represent the subject that needs to be identified later, and the form of multiple law article combinations will be judged according to this subject to complete the law article information needed later. Then, based on the information represented by these laws, a relatively comprehensive combination of integrated information is obtained as embedded information for subsequent information verification, and the adjustment and update of the overall data is completed.

[0062] When calculating the combination matching degree, the length of the longest common subsequence between the legal provision combinations is used for judgment, so as to obtain the longest common subsequence under different data in the key partition, and obtain the biased combination of legal provisions at this time; at the same time, according to the keywords extracted under the key partition, the cosine similarity method is used for multiple groups of data corresponding to the keywords to obtain the common sequence, and then the corresponding common subsequence is obtained to complete the setting of the combination matching degree.

[0063] like Figure 4 As shown, the implementation method of the legal article combination module includes: clustering the consultation partition and the legal partition according to the service type, legal type, and legal function, and setting the largest cluster center after the cluster analysis as the key partition. The service type here mainly refers to the specific way and content of providing legal consulting services, for example, it may include face-to-face legal consultation, telephone consultation, online legal consultation, etc.; the legal type refers to legal consultation involving different legal fields, such as civil legal consultation, criminal legal consultation, administrative legal consultation, etc.; the legal function mainly refers to the role of legal consultation in solving legal problems, safeguarding legal rights and interests, and preventing legal risks, such as providing legal advice, drafting and reviewing contracts, and acting as an agent in litigation.

[0064] Extract keywords from key partitions, calculate the cosine similarity between keywords, and set a common sequence related to the cosine similarity between keywords; the common sequence related to the cosine similarity between keywords refers to the data in the consulting partition and the legal partition corresponding to each keyword, find the intersection of these data, and use the intersection to describe the common sequence at this time; the acquisition of keywords will use word frequency analysis technology, use the documents stored in the database to extract the text in the current key partition, and mark multiple keywords according to the extracted word frequency. The keyword frequency will use the average value of the keyword frequency marked as the keyword in the document stored in the database as the basis for extracting keywords from the key partition at this time, and then calculate the cosine similarity between the extracted keywords; this step can be represented by vectorization , use cosine similarity to complete the setting of cosine similarity between keywords, then select keywords according to cosine similarity, find the intersection of the data corresponding to these keywords, and obtain the common sequence related to the cosine similarity between each keyword. At the same time, the cosine similarity between each keyword represents the similarity between two keywords. This similarity needs to be greater than 0.6, and the data of each keyword with cosine similarity greater than 0.6 are combined to extract multiple common sequences. It should be noted that the common sequence extracted at this time can be a common sequence of the data combination of two keywords with cosine similarity greater than 0.6, or a common sequence of multiple groups of keywords with cosine similarity greater than 0.6. Then, these combined common sequences are sorted according to the cosine similarity values ​​of the keywords.

[0065] Using the common sequence related to the cosine similarity between the keywords, extract the legal provisions in the common sequence, set the longest common subsequence between the legal provisions, and set the length value of the longest common subsequence as the combined matching degree between the legal provisions. After obtaining the common sequence, identify which legal provision the keyword under the current common sequence belongs to, and set the legal provision as the data contained in the legal partition, because the current key partition is the data extracted from the legal partition and the consulting partition, then the keyword set at this time represents at least the content of one legal provision, then at this time, extracting the longest common subsequence between the legal provisions is essentially to directly divide the common sequence according to the multiple common sequences calculated by the cosine similarity of the keyword, extract the subset of these common sequences, and finally obtain a longest common subsequence to know what form of common subsequence these legal provisions will exist in. At this time, the longest common subsequence between the legal provisions needs to contain the longest common subsequence of a specific legal provision, that is, when multiple specific legal provisions are included, how much data will this longest common subsequence contain, and these data will be calculated according to the total amount of relevant data.

[0066] The combination matching degree between each legal article is set according to the time distribution probability of each legal article to set the legal article bias combination. At this time, the legal article bias combination is set by dividing the time distribution probability of the legal article into multiple intervals. The time distribution probability of the legal article is the probability of occurrence of these legal articles in different time periods after the combination. If the time distribution probability is different in a certain time period, it is considered that these legal articles are commonly used in the recent time period, which can reflect the evolution trend of the legal articles over time. In this way, the form of the legal article bias combination set in different time periods can be understood. At this time, the legal article bias combination is set, according to the time distribution probability of multiple legal articles in different time periods corresponding to the combination matching degree between multiple legal articles, according to the size of the time distribution probability, these legal articles are combined into legal article bias combination; the legal article bias combination contains not only the common subsequence after the legal article extraction processing of multiple keywords, but also the corresponding legal articles and time distribution probability to reflect the trend of different legal article combinations at different times. It should be noted that the combined laws here tend to be combined, and the data is not modified according to the time distribution probability. At this time, the corresponding data is marked, and the combination of these multiple laws is determined to complete the certification of the combination form of these laws.

[0067] In one embodiment of the present invention, the case update module is mainly used to compare the corresponding reference situations of corresponding cases in the legal partition based on the biased combination of legal provisions, to verify whether the currently set legal provision combination is normally used in actual cases, and to consider the classification of these situations in normal cases when users consult, to complete the mapping relationship established for user consultation.

[0068] The module set up at this time is actually to solve the problem of legal consultation. In the process of advancement and updating, the meaning of the vocabulary set for the user will change, so that the legal provisions corresponding to some related words and synonyms will change. At this time, it is necessary to find the content that needs to be updated in the consultation partition according to the keywords, and after deleting and updating this part of the content, find the mapping of the updated cases, legal provisions and corresponding consultation content, so as to complete the processing and classification of various data. For example, due to the emergence of certain social cases, the cases of legal provisions corresponding to some related words have changed. At this time, when consulting on legal issues, the legal provisions mapped by some keywords and related words may change. At this time, the corresponding case mapping needs to be adjusted to adjust the mapping relationship between the consultation partition and the legal partition in the current database.

[0069] The current method of verifying legal partitions and updating cases based on the biased combination of legal provisions is to complete the combination and matching of relevant data through the time distribution probability and combination matching degree set in the biased combination of legal provisions. After that, the details in these data will be processed to adjust the relationship between legal provisions and user consultations. Finally, the mapping of legal provisions and cases set for user languages ​​in the current database can be adjusted in real time according to user consultations, thereby improving the effectiveness of user consultations.

[0070] The implementation method of the case update module includes: extracting the time distribution probability of each legal provision from the legal provision bias combination; setting the target path of the legal provision bias combination according to the time period corresponding to the time distribution probability of each legal provision; each path point on the target path represents the time distribution probability of the legal provision bias combination at that point, as well as the specific legal provision information existing at that point. According to the time distribution probability extracted in the previous step, the legal provisions in the legal provision bias combination are sorted or classified according to the time period they are in, and a path representing the change of the legal provision bias combination over time is set accordingly. This path can be a time series graph or a curve graph, which is used to intuitively show the change trend of the legal provision bias combination; through this step, you can intuitively see the change of legal bias in different time periods, which helps to understand the trend of the legal provision bias combination in processing user consultation situations, and to understand the change trend of the content of the biased consultation and the legal provisions that give explanations when users consult.

[0071] The target paths of each law in the law bias combination are fitted to obtain the fitted target path. The probability value of the fitted target path in each time period is set as the combined occurrence probability of the law bias combination. The target path will be interpolated to combine the target paths corresponding to each law in the law bias combination to obtain a combined target path. The target path will be connected according to the time distribution probability of different laws in the corresponding time period, and the path points representing the target path will be continuously connected. After the connection, a relatively smooth path is obtained to show a combined trend line for all the laws in the law bias combination; it is convenient to predict the consultation information that the subsequent user will tend to consult when consulting services, thereby improving the response speed and accuracy of user consultation. At the same time, the probability value of the fitted target path in each time period represents the common occurrence probability value of multiple laws fitted by the current target path in each time period. This value represents the occurrence of multiple laws in the current combination being consulted by users or provided to users in different time periods.

[0072] The probability of occurrence of the combination of law-biased combinations is compared with the citation of cases in the legal partition, the difference values ​​that appear are marked, and the cases in the legal partition are classified according to the difference values ​​that appear, thereby completing the update of the cases in the legal partition.

[0073] At this time, we compare the probability of occurrence of the combination of legal provisions with the probability of occurrence of these legal provisions in the case to find out which parts need to be focused on. For example, we classify the difference between the probability of occurrence of the combination and the citation of the case in the legal partition according to the relative proportion of the difference, such as the difference value of high-attention cases is greater than +10%, the difference value of low-attention cases is less than -10%, and the difference value of regular cases is between ±10%. And the corresponding classification of this part of the classified cases is mapped to the legal partition, and the legal partition marks the keywords corresponding to these classifications in the consultation partition, and finally completes the update of the mapping relationship between the consultation partition and the legal partition.

[0074] In one embodiment of the present invention, the consulting management module is mainly used to generate a storage and classification system for managing the updated data based on the mapping relationship between the updated consulting partitions and the legal partitions.

[0075] The consulting management library is used to update and adjust the relationship between different consulting information and relevant laws and cases in real time, and to obtain a management library based on these data. This management library is used to output user demand information to complete the legal services required by users.

[0076] After completing the update of the mapping relationship between the consultation partition and the legal partition, the corresponding data stored at this time can already indicate the possible key points and areas that need attention in these consultations before and after the user consultation. Then, it is necessary to identify the key points in these areas and generate a consultation management library about these contents, so that subsequent user consultations can be directly called from the data in the currently set mapping relationship to improve data processing speed and efficiency.

[0077] At this time, the implementation method of the consultation management module is: according to the mapping relationship between the updated consultation partition and the legal partition, the consultation target in the consultation information is extracted. Here, the updated part of the data is found in the mapping relationship between the updated consultation partition and the legal partition, and the corresponding content in the consultation information is found from this part of the data, and these contents are set as consultation targets. According to the probability of occurrence of the consultation targets, the consultation targets are sorted to obtain the consultation order in the consultation information; the consultation targets and the consultation order are combined in a structured form to obtain a consultation management library.

[0078] The consultation management library generated at this time will contain multiple contents, such as consultation ID, user ID, legal division, combination of laws involved, consultation date, processing status, related case ID, classification label, etc., to represent the consultation management library set up by the user after consultation. This library will contain comprehensive data content of synonyms, polysemous words, and related words of the current user during consultation to meet the user's needs for corresponding laws.

[0079] like Figure 5 As shown, the present invention also provides a legal consulting service information management method based on big data, including: S1, obtaining consulting information submitted by users, classifying the consulting information using an intent classification model, and identifying intent categories; analyzing synonyms, polysemous words, and related words under the intent categories to form consulting partitions.

[0080] S2. Call the laws and cases related to the consulting information, generate unlabeled law recognition results, filter the law recognition results, and obtain the target law recognition results as the legal partition; according to the information representation in the consulting partition and the descriptive information and category information of the legal partition, construct the mapping relationship between the consulting partition and the legal partition.

[0081] S3. Perform cluster analysis on the consulting partition and the legal partition to determine the key partition; calculate the cosine similarity between the keywords in the key partition, extract the legal provisions, and set the legal provision bias combination.

[0082] S4. Extract the time distribution probability according to the biased combination of legal provisions, set the target path, and fit the combination occurrence probability; compare the combination occurrence probability with the citation of cases in the legal partition, identify and update the cases, and update the mapping relationship between the consultation partition and the legal partition.

[0083] S5. According to the mapping relationship between the updated consultation partitions and the legal partitions, the consultation objectives are extracted and sorted according to the probability of occurrence to obtain the consultation order; the consultation objectives and the consultation order are combined to generate a consultation management library.

[0084] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention and they are still covered by the protection scope of the present invention.

Claims

1. Legal consulting service information management system based on big data, characterized by: include: The consultation partitioning module is used to obtain the consultation information of the user during consultation and determine the consultation partition of the consultation information relative to the legal service according to the user's intention during consultation; A partition mapping module is used to obtain the legal partitions of consulting information and construct a mapping relationship between consulting partitions and legal partitions; The legal article combination module is used to extract key partitions from the consulting partition and the legal partition, and judge the combination matching degree between multiple legal articles in the legal partition according to the key partition to obtain the biased combination of legal articles; The case update module is used to verify the information of the legal partition based on the combination of legal provisions, identify the update status of the cases in the legal partition, and update the mapping relationship between the consulting partition and the legal partition according to the update status of the cases; The consultation management module is used to identify the consultation objectives and consultation sequence of the consultation information according to the mapping relationship between the updated consultation partitions and the legal partitions, and to generate a consultation management library.

2. The legal consulting service information management system based on big data according to claim 1 is characterized in that: The implementation of the consultation partition module includes: For any consultation information during user consultation, obtain the intent classification model corresponding to the consultation information; Use the intent classification model to classify the consultation information and obtain at least one intent category; Identify synonyms, polysemous words and related words in the consultation information under the corresponding intent category to form a text sample set; The synonyms, polysemous words and related words in the text sample set are analyzed respectively, and the synonym area, polysemous word analysis area and related word association area corresponding to the text sample set are obtained in turn, and used as the consulting partition of the consulting information relative to the legal service.

3. The legal consulting service information management system based on big data according to claim 2 is characterized in that: The implementation method of obtaining the synonym area, the polysemous word analysis area and the related word association area corresponding to the text sample set also includes: The synonyms, polysemous words and related words in the text sample set are merged according to the intent category to obtain multiple intent merging results; Extract synonym pairs from the intention merging result, compare the synonym pairs with the synonym dictionary, and obtain the synonym area; The meaning change rate of polysemous words and the association strength of related words in the intention merging results are extracted, and the intention merging results are divided according to the meaning change rate of polysemous words and the association strength of related words to obtain the polysemous word analysis area and the related word association area.

4. The legal consulting service information management system based on big data according to claim 1 is characterized in that: The consultation partitioning implementation methods for judging consultation information relative to legal services also include: Judge the consultation partitions, analyze the number of consultations and consultation time of users in the consultation partitions, fit the consultation partitions according to the number of consultations and consultation time, and build a mapping relationship from the consultation partitions to the users' actual intentions.

5. The legal consulting service information management system based on big data according to claim 1 is characterized in that: The implementation of the partition mapping module includes: Call the legal provisions and cases corresponding to the consultation information to generate multiple unlabeled legal provision recognition results, where the unlabeled legal provision recognition results represent legal provisions and cases that are not associated with the vocabulary in the consultation partition; It is determined whether multiple unlabeled legal article recognition results are target legal article recognition results. If they are target legal article recognition results, the target legal article recognition results are regarded as legal partitions of the consulting information.

6. The legal consulting service information management system based on big data according to claim 3 is characterized in that: The implementation method of constructing the mapping relationship between the consultation partition and the legal partition includes: using the information representation of the synonym area, polysemous word resolution area and related word association area existing in the consultation partition, the description information of the legal partition and the category of the legal partition to construct a mapping relationship between the consultation partition and the legal partition.

7. The legal consulting service information management system based on big data according to claim 1 is characterized in that: The implementation methods of the legal article combination module include: Conduct cluster analysis on the consulting and legal partitions according to service type, legal type, and legal function, and set the largest cluster center after cluster analysis as the key partition; Extract keywords of key partitions, calculate cosine similarities between keywords, and set common sequences related to the cosine similarities between keywords; Using the common sequence related to the cosine similarity between the keywords, extract the legal provisions in the common sequence, set the longest common subsequence between the legal provisions, and set the length value of the longest common subsequence as the combined matching degree between the legal provisions; The combination matching degree between each legal provision is determined according to the time distribution probability of each legal provision, and the biased combination of legal provisions is set.

8. The legal consulting service information management system based on big data according to claim 7 is characterized in that: The implementation of the case update module includes: Extract the time distribution probability of each legal provision from the legal provision bias combination; set the target path of the legal provision bias combination according to the time period corresponding to the time distribution probability of each legal provision; Fitting the target path of each legal provision in the legal provision bias combination to obtain a fitted target path, and setting the probability value of the fitted target path in each time period as the combined occurrence probability of the legal provision bias combination; The probability of occurrence of the combination of law-biased combinations is compared with the citation of cases in the legal partition, the difference values ​​that appear are marked, and the cases in the legal partition are classified according to the difference values ​​that appear, thereby completing the update of the cases in the legal partition.

9. The legal consulting service information management system based on big data according to claim 1 is characterized in that: The implementation of the consultation management module is as follows: According to the updated mapping relationship between the consultation partition and the legal partition, the consultation targets in the consultation information are extracted, and the consultation targets are sorted according to the occurrence probability of the consultation targets to obtain the consultation order in the consultation information; The consulting objectives and consulting sequences are combined in a structured form to obtain a consulting management library.

10. A method for managing legal consulting service information based on big data, characterized in that: include: S1. Obtain consultation information submitted by users, classify the consultation information using the intent classification model, and identify the intent category; Analyze synonyms, polysemous words and related words under the intent category to form consultation partitions; S2. Call the laws and cases related to the consultation information, generate unlabeled law recognition results, filter the law recognition results, and obtain the target law recognition results as the legal partition; According to the information representation in the consultation partition and the description information and category information of the legal partition, a mapping relationship between the consultation partition and the legal partition is constructed; S3. Perform cluster analysis on the consulting and legal partitions to determine the key partitions; Calculate the cosine similarity between keywords in the key partition, extract legal provisions, and set the legal provision bias combination; S4. Extract the time distribution probability according to the bias combination of legal provisions, set the target path, and fit the combination occurrence probability; compare the combination occurrence probability with the citation of the case in the legal partition, identify and update the case, and update the mapping relationship between the consultation partition and the legal partition; S5. Extract the consultation targets according to the updated mapping relationship between the consultation partitions and the legal partitions, and obtain the consultation order by sorting them according to the probability of occurrence; Combine consultation objectives and consultation sequence to generate a consultation management library.

Images