Decommissioned soldier intelligent question and answer interaction management platform based on large model

Abstract

The invention discloses a decommissioned soldier intelligent question and answer interaction management platform based on a large model, which relates to the field of intelligent interaction and comprises a client, a data interaction end and a data analysis end. Obtaining decommissioned soldier identity information and interaction demand information through a client, and generating an initial user portrait according to the obtained information; obtaining decommissioned soldier service information through the data interaction terminal, obtaining corresponding decommissioned soldier service information according to the interaction demand information and the initial user portrait, and generating an initial interaction questionnaire; analyzing and processing the reply result of the initial interaction questionnaire through a data analysis terminal, constructing a feature interaction questionnaire, generating a decommissioned user portrait according to the feature interaction questionnaire and the reply result of the initial interaction questionnaire, and generating corresponding intelligent recommendation information according to the decommissioned user portrait; according to the method, the accuracy in the intelligent question answering process is improved to a certain extent.

Description

Technical Field

[0001] The present invention relates to the field of intelligent interaction technology, and in particular to a large-scale model-based intelligent question-answering interaction management platform for veterans. Background Art

[0002] With the development of society and the advancement of science and technology, people's demand for intelligence is increasing, especially among veterans, a special group that needs more attention and support. Under the traditional service model, veterans often face problems such as information silos, lack of resources, and inadequate services. These problems not only affect the quality of life of veterans, but also make their rights and interests not fully protected. By building an intelligent question-and-answer interactive management platform for veterans based on a large model, these problems can be effectively solved and the sense of achievement and happiness of veterans can be improved. After searching, the invention patent with Chinese patent number CN106570181A discloses an intelligent interaction method based on context management, which includes the following steps: S1, pre-configuring a question and answer database in the server; S2, receiving expression information input by the user end; and pre-processing the expression information; S3, sending the pre-processed expression information to the server; the server matches the question and answer database to obtain interaction information; and sends the interaction information to the user end for presentation.

[0003] Compared with the existing technology, the invention patent with Chinese patent number CN106570181A can generate corresponding interactive information according to the expression information input by the user end through the question-answering server, and present the obtained interactive information to the corresponding user.

[0004] However, in actual use, the above interaction method cannot guarantee the accuracy of the interaction process, and due to the diversity of interaction information in the server, it is impossible to accurately obtain the corresponding interaction information from the server. In addition, the error of the information expressed by the user input has not been considered, so the obtained interaction information may have a low accuracy problem. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcoming of low accuracy of interactive question and answer in the existing technology, and to propose an intelligent question and answer interactive management platform for veterans based on a large model.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: A large-scale model-based intelligent question-answering interactive management platform for veterans, including a client, a data interaction terminal, and a data analysis terminal; The client is used to obtain the identity information of veterans, verify the identity information of veterans, grant a corresponding user account based on the verification result, submit interaction demand information based on user needs through the user account, generate an initial user profile based on the interaction demand information, and obtain and send corresponding interaction information; The data interaction terminal is used to obtain veterans service information, perform feature extraction on the veterans service information, build a service interaction feature library, extract corresponding veterans service information in the service interaction feature library based on the initial user portrait and interaction demand information, generate an initial interaction questionnaire based on the veterans service information, send it to the client to obtain the corresponding initial interaction answer sheet, and generate an initial interaction data packet; The data analysis end is used to obtain an initial interaction data packet, generate a corresponding feature interaction questionnaire based on the initial interaction data packet, and generate a feature interaction data packet based on the feedback results of the feature interaction questionnaire, perform quantification processing on the initial interaction data packet and the feature interaction data packet, generate a retired user portrait based on the quantification processing results, and generate corresponding intelligent recommendation information based on the interaction demand information and the retired user portrait.

[0007] The above technical solution further includes: the process of the client generating the initial user portrait includes: Obtain veterans' identity information, including basic personal information, service-related information, and retirement placement and security information; enter the obtained veterans' identity information into the official platform for verification and analysis; if the verification passes, grant the veterans a corresponding user account; veterans submit interaction demand information through the corresponding user account, including interaction demand and interest and hobby information; Based on the semantic analysis algorithm, the identity information and interest and hobby intention information of veterans are analyzed and processed respectively, the semantic keywords in the identity information and interest and hobby intention information of veterans are obtained, the feature variables are extracted from the semantic keywords, the basic feature variables corresponding to the identity information and interest and hobby intention information of veterans are obtained, and the obtained basic feature variables are derived to obtain corresponding derived feature variables; the corresponding objective user portrait and subjective user image are respectively constructed according to the basic feature variables and derived feature variables corresponding to the identity information and interest and hobby intention information of veterans; The obtained objective user portrait and subjective user image are integrated to obtain the initial user portrait corresponding to the user account.

[0008] Furthermore, the process of the client obtaining and sending corresponding interaction information includes: Set up a user interaction module, obtain interaction demand information through the user interaction module, generate a corresponding interaction demand application based on the interaction demand information and send it to the data interaction terminal; The user interaction module obtains and sends the corresponding interaction information in the user account.

[0009] Furthermore, the process of building a service interaction feature library on the data interaction end includes: Establish a service retrieval platform set, conduct real-time monitoring of the service retrieval platform set based on big data retrieval technology, and obtain corresponding veterans service information; Obtain keyword information corresponding to veterans service information based on a semantic analysis algorithm, perform feature extraction on the obtained keyword information, obtain keyword feature data, perform classification processing on the keyword feature data, obtain classification results of the corresponding veterans service information, the classification results including a single classification result and multiple classification results, and mark each classification result in turn; A service interaction feature library is set up, and the marked veterans service information is stored in the service interaction feature library. An interactive search sub-window is set up, and the interactive search sub-window is used to retrieve the corresponding veterans service information in the service interaction feature library according to the marked classification results.

[0010] Furthermore, the process of generating the initial interaction data packet at the data interaction terminal includes: Obtaining an interactive demand application submitted by a corresponding user account, performing feature extraction on the interactive demand information in the interactive demand application, obtaining keyword feature data corresponding to the interactive demand information, obtaining a type of the interactive demand information based on the keyword feature data, performing information retrieval through a corresponding interactive search subwindow based on the type corresponding to the interactive demand information, and obtaining corresponding veterans service information; Obtain an initial user portrait, filter the obtained veterans service information based on the objective user portrait within the initial user portrait, obtain a matching level between the objective user portrait corresponding to the interactive demand information and the veterans service information based on the filtering result, and set a corresponding objective set of veterans service information based on the matching level; Matching the obtained objective set of veterans service information with the subjective user profile in the initial user profile to obtain the corresponding matching level, and generating the corresponding subjective subset of veterans service information based on the matching level; Setting an initial serial number, setting a quiz question based on the initial serial number, setting option serial numbers based on the quiz question, setting corresponding feature labels for different option serial numbers, selecting veterans service information from each subjective subset of veterans service information within each objective set of veterans service information according to a preset ratio based on the initial serial number, matching the obtained veterans service information with each feature label, generating corresponding quiz questions, and generating a corresponding initial interactive questionnaire based on the quiz questions; The obtained initial interactive questionnaire is sent to the client, and the veteran corresponding to the user account selects and answers the questionnaire, obtains the corresponding initial interactive answer sheet and answer process, and generates an initial interactive data packet.

[0011] Furthermore, the process of generating the feature interaction questionnaire on the data analysis side includes: Obtain the corresponding initial serial number in the initial interactive questionnaire, set mutually related feature serial numbers according to the initial serial number, and use the feature label of the option serial number corresponding to the question and answer corresponding to the initial serial number as the initial input label; According to the initial input tags, the corresponding question and answer questions and feature tags corresponding to the option numbers in the mutually associated feature numbers are analyzed and processed to obtain the feature tags corresponding to each option number, and the corresponding veterans service information is obtained according to the feature tags. According to the question and answer questions corresponding to the feature numbers, the corresponding feature interaction questionnaire is set, and the obtained feature interaction questionnaire is sent to the user account corresponding to the client to obtain the corresponding feature interaction answer sheet and reply process, and the obtained feature interaction questionnaire, feature interaction answer sheet and reply process are generated into a feature interaction data packet.

[0012] Furthermore, the process of generating retired user portraits on the data analysis side includes: Respectively obtain the corresponding initial interaction answer sheet and feature interaction answer sheet in the initial interaction data packet and the feature interaction data packet; analyze and process the answer results corresponding to the mutually associated initial serial number and feature serial number, obtain the feature label corresponding to the answer result option of the feature serial number, and mark the obtained feature label as the feature input label; Quantize the initial input labels and feature input labels to obtain initial quantized features. Based on the initial quantized features, determine the matching results between the corresponding feature numbers and the question-answer responses corresponding to the initial numbers. Based on the matching results, determine whether the corresponding responses in the initial interactive answer sheet and the feature interactive answer sheet are valid. The service information of veterans corresponding to the reply results corresponding to the valid initial serial numbers and characteristic serial numbers marked in the initial interactive answer sheet and the characteristic interactive answer sheet is analyzed and processed to obtain the corresponding characteristic tags, and a profile of retired users is constructed based on the corresponding characteristic tags.

[0013] Furthermore, the process of generating corresponding intelligent recommendation information based on the retired user profile by the data analysis end includes: Obtain the retired user portrait corresponding to the user account, store the corresponding retired user portrait, and obtain the interaction demand information in the corresponding interaction demand application; Build a corresponding intelligent question-answering model based on big data algorithms, input interaction demand information and retired user profiles into the intelligent question-answering model, output corresponding intelligent recommendation information, and feed the obtained intelligent recommendation information back to the client through the data interaction terminal; Obtain feedback from the client and complete intelligent Q&A for veterans.

[0014] The present invention has the following beneficial effects: 1. In the present invention, veterans service information is obtained, analyzed and processed, corresponding classification results are obtained, and marking is performed according to the classification results. The set classification results may include multiple types. An interactive search sub-window is set, and the corresponding classification results are searched through the interactive search sub-window. The veterans service information corresponding to the corresponding classification results is obtained according to the search results, so that the veterans service information corresponding to the multiple classification results can be retrieved, thereby improving the utilization rate of veterans service information and avoiding analysis errors caused by missing veterans service information due to insufficient retrieval.

[0015] 2. In the present invention, the corresponding initial user portrait is constructed by setting an objective user portrait and a subjective user portrait, and the corresponding veterans service information in the service interaction feature library is retrieved, screened and matched through the objective user portrait and the subjective user portrait to generate a corresponding initial interaction questionnaire. The feature interaction questionnaire is set according to the response results of the initial interaction questionnaire, and the option results corresponding to the corresponding initial serial numbers in the initial interaction questionnaire are set to mutually related feature serial numbers to generate a corresponding feature interaction questionnaire. The effectiveness of the user in the interactive response process is judged according to the response results of the feature interaction questionnaire, thereby improving the accuracy of data analysis and processing in the intelligent question-and-answer interaction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a schematic diagram of the structure of the large-scale model-based intelligent question-answering interactive management platform for veterans proposed by the present invention; Figure 2 It is an operation flow chart of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0018] Example 1 like Figure 1As shown, the intelligent question-answer interactive management platform for veterans based on a large model proposed in the present invention includes a client, a data interaction terminal and a data analysis terminal.

[0019] In this embodiment, the veterans intelligent question-and-answer interactive management platform based on a large model provides veterans with a better service experience and consulting assistance through the client, data interaction terminal, and data analysis terminal. The specific implementation process includes: like Figure 2 As shown, the process of data analysis performed by the client, data interaction terminal, and data analysis terminal of the intelligent question-answering interactive management platform for veterans based on a large model proposed by the present invention includes the following: The client is used to obtain the identity information of veterans, verify the identity information of veterans, grant corresponding user accounts based on the verification results, and generate an initial user profile. The specific implementation process includes: Setting up an information collection module, an information processing module and a user interaction module; The information collection module is used to obtain the identity information of retired military personnel, which includes basic personal information, service-related information, and retirement placement and security information, including: Basic personal information includes name, gender, age, ID number, registered place of residence, and mobile phone number for real-name authentication; Service-related information includes time of enlistment, time of discharge, military unit served, branch of service, rank, job changes, and awards and honors received; Retirement placement and security information includes retirement placement methods, retirement pension receipt status, etc. It should be further explained that, in the specific implementation process, when the information collection module obtains the identity information of veterans, the information submitted by the veterans' corresponding clients based on their personal needs is voluntary and informed. The information processing module is provided with an information verification unit and an information processing unit; The information verification unit is used to obtain the identity information of veterans, obtain the basic personal information, service-related information, and retirement placement and security information included in the veterans' identity information, and verify and analyze the corresponding information. The specific implementation process includes: The information verification unit is interconnected with the corresponding official platform, and the obtained personal basic information and service-related information are entered into the official platform for verification and analysis. If the verification is passed, the corresponding user account will be granted to the veteran; if the verification fails, the corresponding user account will not be granted to the veteran; The information processing unit is used to analyze and process the identity information of veterans who have passed identity verification, and obtain an initial user portrait of the corresponding veterans; Obtaining veterans' identity information, analyzing and processing the veterans' identity information based on a semantic analysis algorithm, obtaining semantic keywords within the veterans' identity information, extracting feature variables from the semantic keywords, and obtaining basic feature variables corresponding to the veterans' identity information, wherein the basic feature variables include age, gender, years of service, military rank, field of service, educational level, skill certificates, and reason for retirement; Perform derivative processing on the obtained basic characteristic variables, for example: Obtain the basic characteristic variables corresponding to age, perform feature derivation on the basic characteristic variables corresponding to age, obtain the age groups of veterans, and divide the age into different intervals; Obtain basic characteristic variables corresponding to years of service and military rank, perform feature derivation on basic characteristic variables corresponding to age, obtain the service experience of veterans, and evaluate the richness of service experience; Obtain the basic characteristic variables corresponding to educational level and skill certificates, perform feature derivation on the basic characteristic variables corresponding to age, obtain the skill level of veterans, and evaluate the level of skill; Obtain corresponding derived feature variables according to the feature derivation results; Perform portrait visualization processing on the obtained basic feature variables and derived feature variables, and construct an objective user portrait based on the portrait visualization processing results; Veterans submit corresponding interaction demand information through their user accounts. The interaction demand information includes interaction demand and interest and hobby intention information, including: Interactive needs include policy resource services, employment and entrepreneurship services, education and training services, life rights and interests services, and information query services; Interest and hobby information includes user interests and hobbies, user education and training experience, and user employment intentions; Obtaining a user account corresponding to the veteran's identity information, obtaining the corresponding interaction demand information based on the user account, analyzing and processing the obtained interaction demand information based on a semantic analysis algorithm, obtaining basic feature variables and derived feature variables corresponding to the user interaction demand information, and performing image visualization processing on the obtained interaction demand information. The image visualization processing is to represent the corresponding basic feature variables and derived feature variables and construct a corresponding subjective user image. Visually integrate the obtained subjective user portrait and objective user portrait to obtain the initial user portrait corresponding to the user account; The user interaction module is provided with an interaction request unit, an interaction process unit and an interaction feedback unit, wherein: The interaction request unit inputs the interaction requirements in the interaction requirement information into the interaction requirement application form, and sends the initial user portrait corresponding to the veteran identity information to the data interaction terminal; The interactive process unit is used to conduct question-and-answer interaction with the data interactive terminal according to the interactive demand application, obtain the interactive questionnaire sent by the data interactive terminal through the interactive process unit, reply to the interactive questionnaire, obtain the question-and-answer results, and send the obtained question-and-answer results to the data interactive terminal through the interactive process unit; The interactive feedback unit is used to obtain the data analysis results obtained by the data analysis terminal and feed back the obtained data analysis results to the veteran corresponding to the corresponding user account; The user interaction module obtains and sends the corresponding interaction information in the user account.

[0020] The data interaction terminal is used to obtain veterans service information, extract features of the veterans service information, build a service interaction feature library, extract corresponding veterans service information in the service interaction feature library based on the initial user portrait and interaction demand information, generate an initial interaction questionnaire based on the veterans service information, send it to the client to obtain the corresponding initial interaction answer sheet, and generate an initial interaction data packet. The specific implementation process includes: Set up service management module and interaction management module; The service management module is provided with an information retrieval unit and a database construction unit; The information retrieval unit is used to retrieve the service information for veterans corresponding to the corresponding official channels and public channels, and the service information for veterans includes employment and entrepreneurship service information, retirement placement service information, education and training service information, medical and health service information, life preferential service information, legal support service information, preferential and compassionate service information, and social integration service information; Establish a service search platform set, which includes multiple public government channels, online service platforms, and official social platforms, and set up informed consent forms with the corresponding channels or platforms; Based on big data retrieval technology, real-time monitoring of veterans service information from various channels within the service retrieval platform set is carried out, keyword retrieval information is set, and veterans service information within the corresponding service retrieval platform set is obtained based on the keyword retrieval information; Sending the obtained veterans service information to a database building unit; The database construction unit is used to obtain corresponding veterans service information and construct a service interaction feature library based on the veterans service information. The specific implementation process includes: Obtaining veterans service information, and obtaining keyword information corresponding to each veterans service information based on a semantic analysis algorithm for the obtained veterans service information; Perform feature extraction on the obtained keyword information to obtain corresponding keyword feature data; Classify the obtained keyword feature data to obtain classification results of corresponding veterans service information, wherein the classification results include single classification results and multiple classification results, and mark the corresponding veterans service information according to the classification results; It should be further explained that, in a specific implementation process, the classification result corresponds to the type of corresponding interaction requirement.

[0021] Obtain the keyword feature data proportion results corresponding to the corresponding classification results of the corresponding veterans service information, and mark the corresponding classification results according to the proportion results; Setting up a service interaction feature library, storing the marked veterans' service information in the service interaction feature library, and setting up an interactive search sub-window based on the marking results; The interactive search sub-window is used to search for corresponding veterans service information based on corresponding marked results. The interactive search sub-window and marked results can be used to classify veterans service information into multiple categories, avoiding excessive classification and isolating veterans service information. The interaction management module is provided with an initial interaction unit and an interaction reply unit; The initial interaction unit is used to obtain the initial user portrait corresponding to the user account and generate a corresponding initial interaction questionnaire based on the initial user portrait. The specific implementation process includes: Obtain the interaction demand application submitted by the corresponding user account, and generate an initial interaction questionnaire based on the interaction demand application and the corresponding initial user profile; Obtain the interactive requirements in the interactive requirements application, perform feature extraction on the interactive requirements, obtain the type corresponding to the interactive requirements, perform information retrieval through the corresponding interactive search subwindow based on the type corresponding to the interactive requirements, obtain the corresponding veterans service information, and set the initial interactive answer sheet based on the obtained veterans service information: Obtain an initial user portrait, filter the obtained veterans service information based on the objective user portrait within the initial user portrait, obtain a matching level between the objective user portrait corresponding to the interactive demand information and the veterans service information based on the filtering result, and set a corresponding objective set of veterans service information based on the matching level; Obtain an objective set of veterans service information, perform matching analysis on the obtained objective set of veterans service information according to the subjective user profile in the initial user profile, and set matching levels in sequence according to the matching degree, wherein the matching levels include three levels: high match, moderate match, and no match; Integrate the corresponding veterans service information in the objective set of veterans service information according to the matching level to generate the corresponding subjective subset of veterans service information, and mark the matching level corresponding to each subjective subset of veterans service information; An initial serial number is set, and veterans service information in each subjective subset of veterans service information in each objective set of veterans service information is selected according to a preset ratio based on the initial serial number, where: The higher the matching level, the greater the preset proportion, and the preset proportion data of the initial interactive questionnaire; Select veterans service information in the corresponding subjective subset of veterans service information based on the proportion data to set question-and-answer questions in the corresponding initial sequence number; Setting the option numbers of the quiz questions, setting corresponding feature tags for different option numbers, and keyword feature data corresponding to the feature tags; Matching the veteran service information subjective subset corresponding to the veteran service information with the corresponding feature tags, and inputting the corresponding veteran service information into the corresponding option sequence number according to the matching result of the feature tags; Obtain the completed question-and-answer test questions corresponding to each initial serial number, integrate each initial serial number, obtain the initial interactive questionnaire, and send the obtained initial interactive questionnaire to the client, so that the veteran corresponding to the user account can answer, obtain the corresponding initial interactive answer sheet, and mark the answer process, and send the initial interactive questionnaire, initial interactive answer sheet and initial interactive data packet generated by the answer process to the data analysis end, which will analyze and process the initial interactive data packet.

[0022] The data analysis terminal is used to obtain an initial interaction data packet, generate a corresponding feature interaction questionnaire based on the initial interaction data packet, and generate a feature interaction data packet based on the feedback results of the feature interaction questionnaire, perform quantification processing on the initial interaction data packet and the feature interaction data packet, generate a retired user profile based on the quantification processing results, and generate corresponding intelligent recommendation information based on the interaction demand information and the retired user profile. The specific implementation process includes: Set up data analysis module and intelligent management module; The data analysis module is used to obtain an initial interaction data packet, analyze and process the obtained initial interaction data packet, generate a corresponding feature interaction questionnaire based on the initial interaction data packet, and generate a feature interaction data packet based on the feedback results of the feature interaction questionnaire, and send the generated feature interaction data packet to the intelligent management module. The specific implementation process includes: Obtaining the corresponding initial serial number in the initial interactive questionnaire, setting mutually associated characteristic serial numbers based on the initial serial number, wherein the characteristic serial numbers may be multiple, and each characteristic serial number is mutually associated with the corresponding initial serial number, and using the characteristic label of the option serial number corresponding to the question and answer corresponding to the initial serial number as the initial input label, wherein the characteristic label is the keyword characteristic data corresponding to the veteran service information corresponding to the corresponding option serial number; Analyze and process the corresponding question and answer questions and feature labels corresponding to the option numbers in the mutually associated feature serial numbers based on the initial input labels, obtain the feature labels corresponding to the option numbers, obtain the corresponding veterans service information based on the feature labels, and set up the corresponding feature interactive questionnaire based on the question and answer questions corresponding to the feature serial numbers; Feedback the obtained feature interaction questionnaire to the data interaction terminal, which sends the obtained feature interaction questionnaire to the client for reply, obtains the corresponding feature interaction answer sheet, sends the obtained feature interaction answer sheet to the data interaction terminal, and then the data interaction terminal generates a feature interaction data packet generated by the obtained feature interaction answer sheet and the corresponding reply process and sends it to the data analysis module; The data analysis module sends the obtained feature interaction data packet to the intelligent management module; The intelligent management module is used to obtain the initial interaction data packet and the characteristic interaction data packet, perform quantization processing on the initial interaction data packet and the characteristic interaction data packet, generate a retired user profile based on the quantization processing result, and generate corresponding intelligent interaction information based on the retired user profile. The specific implementation process includes: Obtaining an initial interaction data packet and a characteristic interaction data packet, and performing quantization processing on the obtained initial interaction data packet and characteristic interaction data packet; Get the initial interactive answer sheet and the feature interactive answer sheet, quantify the obtained question and answer answers according to the question and answer answers corresponding to the interrelated question and answer questions corresponding to the feature serial number and the initial serial number, and mark the initial input label corresponding to the initial serial number in the initial interactive answer sheet. ; Get the question and answer answers corresponding to the question and answer questions corresponding to the corresponding feature serial numbers in the feature interactive answer sheet, get the veteran service information corresponding to the option serial numbers corresponding to the question and answer answers, get the corresponding feature input labels based on the veteran service information, and mark the obtained feature input labels as , where i corresponds to the corresponding initial serial number, i=1,2,3,,,n; the initial quantitative features are marked as ,in: ; Obtain the initial quantitative feature. If the initial quantitative feature is 1, use the feature label as the feature user label corresponding to the user account, and set the retired user profile based on the feature user label. If the initial quantitative feature is 0, the feature number and the question and answer corresponding to the initial number are analyzed and processed to obtain the matching level of the feature number and the question and answer corresponding to the initial number; if the corresponding matching levels of the objective set of veterans' service information are inconsistent, the corresponding question and answer test questions are eliminated; If the matching levels corresponding to the subjective subset of veterans’ service information are inconsistent, the corresponding response process is obtained; The response process of input labels and feature labels is quantified, and their browsing data, answer data and click data are obtained respectively, which are marked as 、 and ; Perform normalization processing on browsing data, answer data and click data, obtain corresponding answer error data based on the normalization processing results, and mark the answer error data as WC; The preset browsing coefficient, answer coefficient and click coefficient are marked as 、 and ,in: ; Preset the response error threshold and compare and analyze the obtained response error data with the response error threshold: If the response error data is greater than or equal to the response error threshold, the initial input label or feature input label is valid; If the response error data is less than the response error threshold, the initial input label or feature input label is invalid; Analyze the validity of the initial input labels and feature input labels; If both the initial input label and the feature input label are invalid, mark both the initial input label and the feature input label as invalid; If both the initial input tag and the feature input tag are valid, the matching level corresponding to the corresponding veteran service information is obtained and the matching level is marked as valid; If only one of the initial input label and the feature input label is valid, the matching level corresponding to the valid label is marked as possible; Obtain the analysis results corresponding to the initial interaction answer sheet and the feature interaction answer sheet, obtain the proportion data of the matching level marked as possible, and determine the corresponding matching label based on the proportion data to determine whether to continue the interaction processing; If the proportion data is greater than the preset proportion threshold, the interactive processing continues to obtain the feature input label corresponding to the corresponding initial input label, and multiple analysis processes are performed to determine whether the feature label corresponding to the corresponding veteran service information is a feature user label; If the percentage data is less than or equal to the preset percentage threshold, the interactive processing will not continue, and the characteristic label corresponding to the corresponding veteran service information will not be marked as a characteristic user label; It should be further explained that, in the specific implementation process, each valid feature tag is obtained, and each valid feature tag obtained is marked as a feature user tag corresponding to the corresponding user account; Obtain analysis results of the initial interaction questionnaire and the characteristic interaction questionnaire, obtain veteran service information corresponding to the corresponding characteristic user tags, analyze and process the obtained veteran service information, and construct corresponding veteran user portraits; Obtain interaction demand information and retired user portraits, build a corresponding intelligent question-answering model based on big data algorithms, input the interaction demand information and retired user portraits into the intelligent question-answering model, output corresponding intelligent recommendation information, and feed the obtained intelligent recommendation information back to the client through the data interaction terminal; Obtain feedback from the client and complete intelligent Q&A for veterans.

[0023] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. The intelligent question-answering interactive management platform for veterans based on a large model is characterized by: Including client, data interaction and data analysis end; The client is used to obtain the identity information of veterans, verify the identity information of veterans, grant a corresponding user account based on the verification result, submit corresponding interaction demand information based on user needs through the user account, generate an initial user profile based on the interaction demand information, and obtain and send corresponding interaction information; The data interaction terminal is used to obtain veterans service information, perform feature extraction on the veterans service information, build a service interaction feature library, extract corresponding veterans service information in the service interaction feature library based on the initial user portrait and interaction demand information, generate an initial interaction questionnaire based on the veterans service information, send it to the client to obtain the corresponding initial interaction answer sheet, and generate an initial interaction data packet; The data analysis end is used to obtain an initial interaction data packet, generate a corresponding feature interaction questionnaire based on the initial interaction data packet, and generate a feature interaction data packet based on the feedback results of the feature interaction questionnaire, perform quantification processing on the initial interaction data packet and the feature interaction data packet, generate a retired user portrait based on the quantification processing results, and generate corresponding intelligent recommendation information based on the interaction demand information and the retired user portrait.

2. The large-scale model-based intelligent question-answering interactive management platform for veterans according to claim 1 is characterized in that: The process of generating the initial user portrait on the client side includes: Obtaining veterans' identity information, including basic personal information, service-related information, and retirement placement and security information; entering the obtained veterans' identity information into the official platform for verification and analysis; if the verification passes, granting the veterans a corresponding user account; and submitting interactive demand information through the corresponding user account, including interactive demand and interest and hobby information; Based on the semantic analysis algorithm, the identity information and interest and hobby intention information of veterans are analyzed and processed respectively, the semantic keywords in the identity information and interest and hobby intention information of veterans are obtained, the feature variables are extracted from the semantic keywords, the basic feature variables corresponding to the identity information and interest and hobby intention information of veterans are obtained, and the obtained basic feature variables are derived to obtain corresponding derived feature variables; the corresponding objective user portrait and subjective user image are respectively constructed according to the basic feature variables and derived feature variables corresponding to the identity information and interest and hobby intention information of veterans; The obtained objective user portrait and subjective user image are integrated to obtain the initial user portrait corresponding to the user account.

3. The large-scale model-based intelligent question-answering interactive management platform for veterans according to claim 2 is characterized in that: The process of the client obtaining and sending corresponding interaction information includes: Set up a user interaction module, obtain interaction demand information through the user interaction module, generate a corresponding interaction demand application based on the interaction demand information and send it to the data interaction terminal; The user interaction module obtains and sends the corresponding interaction information in the user account.

4. The large-scale model-based intelligent question-answering interactive management platform for veterans according to claim 3 is characterized in that: The process of building a service interaction feature library on the data interaction end includes: Establish a service retrieval platform set, conduct real-time monitoring of the service retrieval platform set based on big data retrieval technology, and obtain corresponding veterans service information; Obtain keyword information corresponding to veterans service information based on a semantic analysis algorithm, perform feature extraction on the obtained keyword information, obtain keyword feature data, perform classification processing on the keyword feature data, obtain classification results of the corresponding veterans service information, the classification results including a single classification result and multiple classification results, and mark each classification result in turn; A service interaction feature library is set up, and the marked veterans service information is stored in the service interaction feature library. An interactive search sub-window is set up, and the interactive search sub-window is used to retrieve the corresponding veterans service information in the service interaction feature library according to the marked classification results.

5. The large-scale model-based intelligent question-answering interactive management platform for veterans according to claim 4 is characterized in that: The process of generating the initial interaction data packet at the data interaction end includes: Obtaining an interactive demand application submitted by a corresponding user account, performing feature extraction on the interactive demand information in the interactive demand application, obtaining keyword feature data corresponding to the interactive demand information, obtaining a type of the interactive demand information based on the keyword feature data, performing information retrieval through a corresponding interactive search subwindow based on the type corresponding to the interactive demand information, and obtaining corresponding veterans service information; Obtain an initial user portrait, preliminarily screen the obtained veterans service information based on the objective user portrait within the initial user portrait, obtain a matching level between the objective user portrait corresponding to the interactive demand information and the veterans service information based on the screening result, and set a corresponding objective set of veterans service information based on the matching level; Matching the obtained objective set of veterans service information with the subjective user profile in the initial user profile to obtain the corresponding matching level, and generating the corresponding subjective subset of veterans service information based on the matching level; Setting an initial serial number, setting a quiz question based on the initial serial number, setting option serial numbers based on the quiz question, setting corresponding feature labels for different option serial numbers, selecting veterans service information from each subjective subset of veterans service information within each objective set of veterans service information according to a preset ratio based on the initial serial number, matching the obtained veterans service information with each feature label, generating corresponding quiz questions, and generating a corresponding initial interactive questionnaire based on the quiz questions; The obtained initial interactive questionnaire is sent to the client, and the veteran corresponding to the user account selects and answers the questionnaire, obtains the corresponding initial interactive answer sheet and answer process, and generates an initial interactive data packet.

6. The large-scale model-based intelligent question-answering interactive management platform for veterans according to claim 5 is characterized in that: The process of generating a feature interaction questionnaire on the data analysis side includes: Obtain the corresponding initial serial number in the initial interactive questionnaire, set mutually related feature serial numbers according to the initial serial number, and use the feature label of the option serial number corresponding to the question and answer corresponding to the initial serial number as the initial input label; According to the initial input tags, the feature tags corresponding to the question and answer questions and option numbers in the mutually associated feature numbers are analyzed and processed, and the feature tags corresponding to the option numbers in the question and answer questions corresponding to the feature numbers are set according to the analysis results. The corresponding veterans service information is obtained according to the feature tags, and the corresponding feature interaction questionnaire is set according to the question and answer questions corresponding to the feature numbers. The obtained feature interaction questionnaire is sent to the user account corresponding to the client, and the corresponding feature interaction answer sheet and reply process are obtained. The obtained feature interaction questionnaire, feature interaction answer sheet and reply process are generated into a feature interaction data packet.

7. The large-scale model-based intelligent question-answering interactive management platform for veterans according to claim 6 is characterized in that: The process of generating retired user profiles on the data analysis side includes: Respectively obtain the corresponding initial interaction answer sheet and feature interaction answer sheet in the initial interaction data packet and the feature interaction data packet; analyze and process the answer results corresponding to the mutually associated initial serial number and feature serial number, obtain the feature label corresponding to the answer result option of the feature serial number, and mark the obtained feature label as the feature input label; Quantize the initial input labels and feature input labels to obtain initial quantized features. Based on the initial quantized features, determine the matching results between the corresponding feature numbers and the question-answer responses corresponding to the initial numbers. Based on the matching results, determine whether the corresponding responses in the initial interactive answer sheet and the feature interactive answer sheet are valid. The service information of veterans corresponding to the reply results corresponding to the valid initial serial numbers and characteristic serial numbers marked in the initial interactive answer sheet and the characteristic interactive answer sheet is analyzed and processed to obtain the corresponding characteristic tags, and a profile of retired users is constructed based on the corresponding characteristic tags.

8. The large-scale model-based intelligent question-answering interactive management platform for veterans according to claim 7 is characterized in that: The process by which the data analysis end generates corresponding intelligent recommendation information based on the retired user profile includes: Obtain the retired user portrait corresponding to the user account, store the corresponding retired user portrait, and obtain the interaction demand information in the corresponding interaction demand application; Build a corresponding intelligent question-answering model based on big data algorithms, input interaction demand information and retired user profiles into the intelligent question-answering model, output corresponding intelligent recommendation information, and feed the obtained intelligent recommendation information back to the client through the data interaction terminal; Obtain feedback from the client and complete intelligent Q&A for veterans.

Images