Interactive solution generator based on natural language processing

Abstract

The present invention provides an interactive solution generator based on natural language processing, which relates to the field of data processing technology, including: an interactive application receiving module; a learning intention acquisition module, which is used to perform intention recognition and obtain a first learning intention; a fitness acquisition module, which is used to match a first preliminary course version and obtain a first preliminary fitness; an instruction issuing module, which is used to issue a preliminary adjustment instruction when it does not meet a predetermined fitness limit; a cluster matching module, which is used to match a first cluster; a recommendation table acquisition module, which is used to obtain a first adjustment recommendation table; a target course acquisition module, which is used to increase or decrease the course outline and obtain a first target course; a course plan sending module, which is used to send the first course plan to a first student. The present invention solves the technical problem that course recommendations in the prior art often rely on static rules or simple keyword matching, resulting in low personalization and accuracy of recommendation results.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an interactive solution generator based on natural language processing. Background Art

[0002] Digital schools use a variety of digital tools and platforms to manage teaching, student data, course content, and teaching interactions. For example, online learning platforms, learning management systems, smart classrooms, virtual reality teaching, etc. are all key components of digital schools. In these environments, personalized education and efficient course recommendations are particularly important because they directly affect students' learning experience and educational outcomes. Traditional course recommendations are often based on simple keyword matching or static rules, which makes the recommendation results unable to fully reflect students' personal needs and learning goals. In digital schools, the lack of personalized recommendations will result in students being unable to obtain courses that suit their specific needs, affecting learning outcomes and satisfaction.

[0003] For example, application number CN202110668665.X application name: Course recommendation method and device, discloses a course recommendation method and device, which obtains user information of the user to be recommended and analyzes it to obtain learning attribute information, and outputs course recommendation results based on comprehensive portrait analysis. However, this method lacks detailed learning intention recognition and processing, personalized budget and course version matching, dynamic adjustment and optimization mechanism, and full use of historical learning data, resulting in insufficient accuracy and personalization of recommendations. Summary of the invention

[0004] This application provides an interactive solution generator based on natural language processing, aiming to solve the technical problem that course recommendations in the prior art often rely on static rules or simple keyword matching, resulting in an inability to accurately match the actual needs of students, and thus leading to low personalization and accuracy of recommendation results.

[0005] The first aspect disclosed in the present application provides an interactive solution generator based on natural language processing, and the interactive solution generator based on natural language processing includes: an interactive application receiving module, the interactive application receiving module is used to receive a first interactive application, the first interactive application includes a first course of application consultation interaction of a first student, wherein the first course has multiple course versions; a learning intention acquisition module, the learning intention acquisition module is used to perform intent recognition on a first interactive text of the first student through an interactive intention recognition model to obtain a first learning intention; a fitness acquisition module, the fitness acquisition module is used to match a first preliminary course version in the multiple course versions according to a first learning budget parameter extracted from the first learning intention based on a predetermined learning budget indicator, and obtain a first preliminary fitness of the first preliminary course version; an instruction issuing module, the The instruction issuing module is used to issue a preliminary selection adjustment instruction when the first preliminary selection fitness does not meet the predetermined fitness limit; the cluster cluster matching module is used to traverse the first cluster cluster matching the first student in the student database, the first cluster cluster includes multiple history students, and the multiple history students have a one-to-one correspondence with multiple history learning course syllabus tables; the recommendation table acquisition module is used to analyze the multiple history learning course syllabus tables based on the preliminary selection adjustment instruction to obtain a first adjustment recommendation table; the target course acquisition module is used to increase or decrease the course syllabus of the first preliminary selection course version according to the first adjustment recommendation table to obtain a first target course; the course plan sending module is used to send the first course plan obtained by analyzing the first target course to the first student.

[0006] The second aspect disclosed in the present application provides an interactive solution generation method based on natural language processing, the interactive solution generation method based on natural language processing comprising: receiving a first interactive application, the first interactive application comprising a first course of an application consultation interaction of a first student, wherein the first course has multiple course versions; performing intent recognition on the first interactive text of the first student through an interactive intention recognition model to obtain a first learning intention; matching a first preliminary course version among the multiple course versions according to a first learning budget parameter extracted from the first learning intention based on a predetermined learning budget indicator, and obtaining a first preliminary fitness of the first preliminary course version; when the first preliminary fitness does not meet a predetermined fitness limit, issuing a preliminary adjustment instruction; traversing a first clustering cluster matching the first student in a student database, the first clustering cluster comprising multiple history students, and the multiple history students having a one-to-one correspondence with multiple history learning course syllabus tables; analyzing the multiple history learning course syllabus tables based on the preliminary adjustment instruction to obtain a first adjustment recommendation table; adjusting the course syllabus of the first preliminary course version according to the first adjustment recommendation table to obtain a first target course; and sending the first course plan obtained by analyzing the first target course to the first student.

[0007] According to a third aspect disclosed in the present application, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is implemented based on the first aspect disclosed in the present application when executing the computer program.

[0008] The fourth aspect disclosed in the present application provides a computer-readable storage medium having a computer program stored thereon, and the computer program is implemented based on the first aspect disclosed in the present application when executed by a processor.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] By using the intention recognition model to identify the students' interactive texts, we can accurately capture the students' learning needs and goals, and make course recommendations based on the students' specific needs, rather than relying on simple keyword matching or general recommendation rules, thereby providing a more personalized learning experience; according to the students' learning budget parameters, including budget duration and budget effectiveness, the scheduled duration and scheduled effectiveness of the initial course version are matched, ensuring that the recommended course version not only meets the students' needs, but is also within the budget, thereby improving the applicability of the recommendation; when the adaptability of the initial course version does not meet the predetermined adaptability limit, an adjustment instruction is issued. This dynamic adjustment mechanism ensures that it can respond to different situations and changes in demand, and provide continuously optimized course recommendations; through By traversing and matching historical students in the student database and analyzing multiple historical learning course syllabi, course adjustments can be made based on historical data. This data-driven method utilizes the learning situation and course feedback of historical students to improve the accuracy and effectiveness of course recommendations. Based on the analysis of the historical learning course syllabus, the preliminary course version is adjusted to obtain the target course. This optimization makes course recommendations not only based on the current needs of students, but also takes into account the actual effects and room for improvement of the course, thereby improving the overall quality and adaptability of the course; the adjusted course plan is sent to students to provide them with personalized learning plans. In this way, specific course content, structure and predicted satisfaction can be provided, thereby helping students make better learning decisions.

[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A schematic diagram of the structure of an interactive solution generator based on natural language processing provided in an embodiment of the present application;

[0013] Figure 2 A flowchart of a method for generating an interactive solution based on natural language processing provided in an embodiment of the present application;

[0014] Figure 3 This is a diagram of the internal structure of a computer device provided in an embodiment of the present application.

[0015] Explanation of the accompanying drawings: interactive application receiving module 10, learning intention acquisition module 20, adaptability acquisition module 30, instruction issuing module 40, clustering cluster matching module 50, recommendation table acquisition module 60, target course acquisition module 70, course planning sending module 80. DETAILED DESCRIPTION

[0016] The embodiment of the present application solves the technical problem that course recommendations in the prior art often rely on static rules or simple keyword matching, which makes it impossible to accurately match the actual needs of students, and thus leads to low personalization and accuracy of recommendation results, by providing an interactive solution generator based on natural language processing.

[0017] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the accompanying drawings of the specification. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0018] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides an interactive solution generator based on natural language processing, and the interactive solution generator based on natural language processing includes:

[0019] The interactive application receiving module 10 is used to receive a first interactive application, wherein the first interactive application includes a first course for which a first student applies for consultation interaction, wherein the first course has multiple course versions.

[0020] A first interactive application is received from a first student through a user interface. The first interactive application includes the student's personal information and requested course information, specifically including a first course. The first course includes multiple course versions, such as a simple version, an in-depth version, an ultimate version, etc. Each version of the course has a predetermined learning time range and a predetermined learning effectiveness. Exemplarily, the simple version is suitable for beginners or students who do not have high requirements for the course content, with a shorter learning time and a lower course difficulty; the in-depth version is suitable for students with a certain foundation, with a moderate learning time and more detailed content; the ultimate version is suitable for students who need to have an in-depth understanding of the course content, with the longest learning time and the most detailed and in-depth content.

[0021] The learning intention acquisition module 20 is used to perform intention recognition on the first interactive text of the first student through an interactive intention recognition model to obtain a first learning intention.

[0022] A first interaction text is received from a first learner. The first interaction text is a consulting question, a request or a description of a course submitted by the learner in the application, for example, "I hope to learn more about the in-depth version of the course." The first interaction text is preprocessed, including noise removal, word segmentation, vocabulary standardization, etc. The preprocessed text is input into an interaction intent recognition model for analysis. The model recognizes the learner's intent by analyzing the vocabulary, phrases and context in the text, and outputs a first learning intention. The first learning intention is the main need or expectation of the first learner.

[0023] The adaptability acquisition module 30 is used to match the first preliminary course version among the multiple course versions according to the first learning budget parameter extracted from the first learning intention based on a predetermined learning budget indicator, and obtain the first preliminary adaptability of the first preliminary course version.

[0024] The planned learning budget indicators include the planned learning budget duration and the planned learning budget effectiveness, that is, the acceptable learning duration and expected learning effect of the students during the learning process. The first learning intention is analyzed, and the parameters of the planned learning budget duration and the planned learning budget effectiveness are extracted from them. The first learning budget duration parameters and the first learning budget effectiveness parameters are obtained as the first learning budget parameters. For example, if the first learning intention is to understand the basic knowledge in a shorter period of time, the learning time budget may be 5-10 hours.

[0025] A first preliminary selected course version is matched among multiple course versions. The first preliminary selected course version may be any one of the multiple course versions. The first preliminary selected course version has a first predetermined learning duration and a first predetermined learning completion effect.

[0026] Calculate the adaptability of the first preliminary course version to the first learning budget parameter. Specifically, compare the first learning budget duration parameter with the first scheduled learning duration to obtain the duration deviation; compare the first learning budget effectiveness parameter with the first scheduled learning completion effectiveness to obtain the effectiveness deviation. Perform weighted summation of the duration deviation and the effectiveness deviation to obtain the first preliminary adaptability of the first preliminary course version. The adaptability indicates whether the course version meets the student's learning time budget and expected results.

[0027] The instruction issuing module 40 is used to issue a preliminary selection adjustment instruction when the first preliminary selection fitness does not meet the predetermined fitness limit.

[0028] The predetermined suitability limit is a pre-set threshold used to determine whether the pre-selected course version meets the student's learning budget requirements. This limit can be set based on historical data, course characteristics and user needs. For example, if it is set to 80%, it means that the course's suitability must reach or exceed 80% to be considered suitable. The first pre-selected suitability is compared with the predetermined suitability limit. If the first pre-selected suitability is lower than the predetermined suitability limit, it means that the course version does not meet the student's learning budget requirements. In this case, a pre-selected adjustment instruction is generated, which instructs further adjustments to find a more suitable course version.

[0029] The cluster matching module 50 is used to traverse the first cluster matching the first student in the student database, the first cluster includes multiple history students, and the multiple history students have a one-to-one correspondence with multiple history learning course syllabus tables.

[0030] The student database is a pre-built database that stores history students and corresponding history learning course outlines. Taking the courses that students have learned and the learning intentions of students as cluster evaluation indicators, the first learned course and the first learning intention of the first student are extracted, and the first learned course and the first learning intention are matched in the student database to find a group of history students with similar characteristics to the first student to form a first cluster. The first cluster is a group containing similar history students. Multiple history student information in the first cluster is extracted from the student database. Each history student information corresponds to a history learning course outline. These history learning course outlines describe the course content, learning structure, etc. of the history students.

[0031] The recommendation table acquisition module 60 is used to analyze the multiple history learning course syllabus tables based on the preliminary adjustment instruction to obtain a first adjustment recommendation table.

[0032] Extract the first preliminary course syllabus of the first preliminary course version, compare multiple history learning course syllabus tables with the first preliminary course syllabus, find out the content sets that need to be added or reduced, and sort the content sets according to the frequency of occurrence of the added or reduced content, and generate a first adjustment recommendation table according to the sorting results. The first adjustment recommendation table includes priority recommendations for those contents that best meet the needs of students, thereby improving the accuracy and effectiveness of course recommendations.

[0033] The target course acquisition module 70 is used to adjust the course outline of the first preliminary course version by adding or reducing the course outline according to the first adjustment recommendation table to obtain a first target course.

[0034] According to the first adjustment recommendation table, the curriculum of the first preliminary course version is adjusted by adding or reducing content. Specifically, the new content in the first adjustment recommendation table is integrated into the curriculum of the first preliminary course version, including adding new chapters, modules or learning activities; unnecessary or redundant parts in the recommendation table are removed from the curriculum of the first preliminary course version. According to the adjusted curriculum, the course content is rearranged to form a new course structure to obtain the first target course.

[0035] The course planning sending module 80 is used to send the first course planning obtained by analyzing the first target course to the first student.

[0036] The first target course is analyzed to obtain a first course plan. Specifically, the first post-study satisfaction of a first history student among multiple history students is obtained, the mode value of the first post-study satisfaction is taken as the first predicted satisfaction, the first predicted satisfaction is marked to the first target course, and a first course plan is generated. The first course plan is sent to the first student so that the student can clearly understand the course schedule and be ready to participate in the course learning.

[0037] Furthermore, the performing intention recognition on the first interactive text of the first learner by using the interactive intention recognition model to obtain the first learning intention includes:

[0038] The first interactive text is preprocessed to obtain a first preprocessing result, which includes a plurality of words with interaction frequency identifiers; words with a predetermined ranking threshold in an interactive vocabulary sequence obtained by analyzing the plurality of words with interaction frequency identifiers are extracted to form an intention recognition vocabulary group; the intention recognition vocabulary group is input into the interactive intention recognition model to obtain the first learning intention.

[0039] The first interactive text of the first student is obtained. The first interactive text is the consulting questions, requests or course descriptions submitted by the student in the application. The first interactive text is preprocessed, including removing noise, such as removing punctuation marks, special characters and unnecessary spaces; segmenting the text, segmenting the text by vocabulary according to language characteristics and application requirements; standardizing the vocabulary, and removing stop words. All unique words are extracted from the preprocessed text, a vocabulary is constructed, and the frequency of each word in the vocabulary in the text is calculated to generate a word frequency matrix. Each row of the matrix represents a word, and each column represents a word frequency, showing the number of times each word appears in the text. The generated word frequency matrix is ​​integrated into the first preprocessing result.

[0040] For multiple words with interaction frequency identifiers, sort the words according to their frequencies in the text, and arrange them from high to low in frequency to generate an interaction word sequence. Set a predetermined ranking threshold to select words with higher rankings in the vocabulary. This threshold can be an absolute value of word frequency, such as the number of occurrences exceeding a certain number, or a relative ranking, such as the top N words. Extract words with a ranking higher than the predetermined ranking threshold from the interaction word sequence to form an intent recognition word group, which includes keywords that have a significant impact on intent recognition.

[0041] The words in the intention recognition vocabulary group are converted into standard words, and the interactive intention recognition model is input. The standard words are retrieved and matched using the intention database embedded in the interactive intention recognition model. According to the matching results, the first historical standard intention recognition vocabulary group that meets the similarity limit is obtained, and the first historical learning intention corresponding to the first historical standard intention recognition vocabulary group is obtained as the first learning intention. The first learning intention is output through the interactive intention recognition model to represent the learner's learning goals or needs.

[0042] Furthermore, the step of inputting the intention recognition vocabulary group into the interactive intention recognition model to obtain the first learning intention includes:

[0043] Randomly extract the first word in the intention recognition vocabulary group; read a predetermined standard vocabulary replacement strategy, and replace and adjust the first word based on the predetermined standard vocabulary replacement strategy to obtain a first standard vocabulary; form a standard intention recognition vocabulary group based on the first standard vocabulary; extract the first intention data group from the intention database, wherein the intention database is embedded in the interactive intention recognition model, and the first intention data group includes a first historical standard intention recognition vocabulary group and a first historical learning intention; when the similarity between the first word of the standard intention recognition vocabulary group and the first historical standard intention recognition vocabulary group reaches a similarity limit, use the first historical learning intention as the first learning intention; output the first learning intention through the interactive intention recognition model.

[0044] A word is randomly selected from the intent recognition word group using a random sampling method as the first word.

[0045] The predetermined standard vocabulary replacement strategy is a set of rules for replacing vocabulary into a standardized form, including rules such as synonym replacement, terminology unification, and spelling correction. The first vocabulary is adjusted according to the predetermined standard vocabulary replacement strategy. For example, expressions such as "car", "automobile", and "car" are replaced with the standardized "vehicle". After replacement and adjustment, the first standard vocabulary is obtained.

[0046] The same replacement adjustment method as the first standard vocabulary is adopted to traverse all the words in the intent recognition vocabulary group to obtain the standard intent recognition vocabulary group.

[0047] The intent database is a database containing historical intent data, which is embedded in the interactive intent recognition model and is used to store and retrieve intent-related data. A historical record is randomly extracted from the intent database as the first intent data group, which includes a first historical standard intent recognition vocabulary group and a first historical learning intent, wherein the first historical standard intent recognition vocabulary group is a standard vocabulary group used for intent recognition in history, which represents the standardized vocabulary used to recognize learning intent in the past; the first historical learning intent is the actual learning intent data associated with the first historical standard intent recognition vocabulary group.

[0048] The similarity between the standard intent recognition vocabulary group and the first historical standard intent recognition vocabulary group is compared, for example, by calculating the cosine angle between the two vocabulary group vectors to measure their similarity; or calculating the ratio of the intersection and the union of two sets to evaluate their similarity. The first vocabulary similarity is obtained by calculation.

[0049] Define a similarity limit, which is used to determine whether two vocabulary groups are similar. This limit can be determined based on experience. Compare the calculated first vocabulary similarity with the similarity limit. If the first vocabulary similarity reaches or exceeds the similarity limit, the first historical learning intention is used as the first learning intention, which means that the current intention matches the historical intention, and the historical learning intention can be used as a prediction of the current learning intention.

[0050] The first learning intention is output through the interaction intention recognition model.

[0051] Furthermore, the predetermined learning budget indicators include a predetermined learning budget duration and a predetermined learning budget effectiveness.

[0052] The planned learning budget indicators include the planned learning budget duration and the planned learning budget results. The planned learning budget duration refers to the total amount of time that students are expected to invest in course learning, which may specifically include the overall learning time required for the course, and the students' daily or weekly study time arrangements during the learning process; the planned learning budget results refer to the results or effects that students expect to obtain from learning, which may specifically include the specific goals that students hope to achieve through course learning, such as knowledge mastery, skill improvement, etc., as well as course results, such as test scores, skill certification, project completion, etc.

[0053] Furthermore, traversing the first cluster matching the first student in the student database includes:

[0054] Reading predetermined student characteristics, the predetermined student characteristics include the student's already learned courses and the student's learning intention; using the student's already learned courses and the student's learning intention as clustering evaluation indicators, using the first learned course of the first student and the first learning intention as clustering constraints, clustering multiple historical students in the student database to obtain the first cluster cluster.

[0055] The expected student characteristics include the courses the student has taken and the student's learning intention. Among them, the courses the student has taken refer to the courses that the student has completed or is taking before. This information helps to understand the student's learning background and the knowledge or skills he has mastered, including the course name, completion time, course version, learning performance, learning progress, etc.; the student's learning intention refers to the student's expectations or goals for future learning, including the learning goals he hopes to achieve, his interest in the course content, the expected learning outcomes, time budget, etc.

[0056] The learned course data and learning intention data of all historical students are extracted from the student database, and the learned course data and learning intention data are standardized to ensure the consistency and comparability of the data, such as converting the learned courses into category labels and converting the learning intention into keywords or numerical representations.

[0057] Use the first learned course and the first learning intention of the first student as clustering constraints, compare its features with those of other historical students, calculate the feature similarity, and use a clustering algorithm to divide the historical students into different clusters based on the feature similarity calculation results, where each cluster contains students with similar features, and extract the clustering cluster of students with similar features to the first learned course and the first learning intention of the first student as the first clustering cluster.

[0058] Furthermore, the analyzing the plurality of history learning course syllabus tables based on the preliminary selection adjustment instruction to obtain a first adjustment recommendation table includes:

[0059] Compare the first history learning course syllabus table among the multiple history learning course syllabus tables with the first preliminary course version to obtain a first increase and decrease list; compare the second history learning course syllabus table among the multiple history learning course syllabus tables with the first preliminary course version to obtain a second increase and decrease list; take the union of the first increase and decrease list and the second increase and decrease list as the increase and decrease course syllabus set; arrange the increase and decrease course syllabus set in descending order based on the first increase and decrease frequency of the first course syllabus in the increase and decrease course syllabus set to obtain the first adjustment recommendation table.

[0060] A random one is selected from a plurality of history learning course outline tables as the first history learning course outline table, and a course outline table of the first preliminary course version is selected as the first preliminary course outline table.

[0061] Compare the outlines in the first history learning course outline table with the outlines in the first preliminary course outline table, such as each chapter or content, determine that the chapters or contents that exist in the first preliminary course outline table but do not exist in the first history learning course outline table are marked as deleted; determine that the chapters or contents that do not exist in the first preliminary course outline table but exist in the first history learning course outline table are marked as added. Create a list to list all the added and deleted contents to obtain the first addition and deletion list.

[0062] A second history learning course syllabus which is different from the first history learning course syllabus is randomly extracted from multiple history learning course syllabus tables, and the same analysis method as the first history learning course syllabus is adopted to compare the second history learning course syllabus with the first preliminary course syllabus to obtain a second increase and decrease list.

[0063] All the new contents in the first increase and decrease list and the second increase and decrease list are merged, and duplicates are removed to obtain a new content set, and then all the deleted contents are merged, and duplicates are removed to obtain a deleted content set. The new content set and the deleted content set are integrated to obtain an increase and decrease course outline set.

[0064] For each course outline in the set of added and reduced course outlines, count its frequency of occurrence in the first increase and decrease list and the second increase and decrease list, the frequency includes the frequency of addition and deletion, randomly extract the first course outline in the set of added and reduced course outlines, obtain the frequency of occurrence of the first course outline, and use it as the first increase and decrease frequency. Sort the set of added and reduced course outlines in descending order according to the first increase and decrease frequency, ensuring that the course outline with the highest frequency of occurrence is placed in front. Generate a first adjustment recommendation table based on the sorting results, including all course outlines and their increase and decrease frequencies, which helps determine the course content that needs to be adjusted.

[0065] Further, analyzing the first target course to obtain a first course plan includes:

[0066] Obtain a first post-learning satisfaction level of a first history student among the multiple history students; take a mode value of the first post-learning satisfaction level, recorded as a first predicted satisfaction level; mark the first predicted satisfaction level to the first target course, and generate the first course plan.

[0067] A first history student is randomly selected from multiple history students to obtain the first history student's post-course satisfaction. The post-course satisfaction refers to the student's satisfaction score after completing the course, which can be obtained through a questionnaire, feedback form or scoring system.

[0068] Summarize the satisfaction scores of all history students on similar courses, including the first history student’s post-course satisfaction, and calculate the mode value of the satisfaction scores, that is, the most frequently occurring satisfaction score, as the first predicted satisfaction score, which is used to predict the satisfaction level in similar situations.

[0069] The first predicted satisfaction level is associated with the first target course, and a tag can be added to the course in the course planning system to record the predicted satisfaction level, and a final first course plan is generated according to the first target course marked with the first predicted satisfaction level.

[0070] In summary, the interactive solution generator based on natural language processing provided in the embodiments of the present application has the following technical effects:

[0071] By using the intention recognition model to identify the students' interactive texts, we can accurately capture the students' learning needs and goals, and make course recommendations based on the students' specific needs, rather than relying on simple keyword matching or general recommendation rules, thereby providing a more personalized learning experience; according to the students' learning budget parameters, including budget duration and budget effectiveness, the scheduled duration and scheduled effectiveness of the initial course version are matched, ensuring that the recommended course version not only meets the students' needs, but is also within the budget, thereby improving the applicability of the recommendation; when the adaptability of the initial course version does not meet the predetermined adaptability limit, an adjustment instruction is issued. This dynamic adjustment mechanism ensures that it can respond to different situations and changes in demand, and provide continuously optimized course recommendations; through By traversing and matching historical students in the student database and analyzing multiple historical learning course syllabi, course adjustments can be made based on historical data. This data-driven method utilizes the learning situation and course feedback of historical students to improve the accuracy and effectiveness of course recommendations. Based on the analysis of the historical learning course syllabus, the preliminary course version is adjusted to obtain the target course. This optimization makes course recommendations not only based on the current needs of students, but also takes into account the actual effects and room for improvement of the course, thereby improving the overall quality and adaptability of the course; the adjusted course plan is sent to students to provide them with personalized learning plans. In this way, specific course content, structure and predicted satisfaction can be provided, thereby helping students make better learning decisions.

[0072] Embodiment 2 is based on the same inventive concept as the interactive solution generator based on natural language processing in the above embodiment. Figure 2 As shown, the embodiment of the present application provides an interactive solution generation method based on natural language processing, the method comprising:

[0073] A first interactive application is received, wherein the first interactive application includes a first course for application consultation interaction by a first student, wherein the first course has multiple course versions; the first interactive text of the first student is identified through an interactive intention recognition model to obtain a first learning intention; a first preliminary course version is matched among the multiple course versions according to a first learning budget parameter extracted from the first learning intention based on a predetermined learning budget indicator, and a first preliminary fitness of the first preliminary course version is obtained; when the first preliminary fitness does not meet the predetermined fitness limit, a preliminary adjustment instruction is issued; a first clustering cluster matching the first student is traversed in a student database, wherein the first clustering cluster includes multiple history students, and the multiple history students have a one-to-one correspondence with multiple history learning course syllabus tables; the multiple history learning course syllabus tables are analyzed based on the preliminary adjustment instruction to obtain a first adjustment recommendation table; the course syllabus of the first preliminary course version is adjusted to increase or decrease according to the first adjustment recommendation table to obtain a first target course; and a first course plan obtained by analyzing the first target course is sent to the first student.

[0074] Further, including:

[0075] The first interactive text is preprocessed to obtain a first preprocessing result, which includes a plurality of words with interaction frequency identifiers; words with a predetermined ranking threshold in an interactive vocabulary sequence obtained by analyzing the plurality of words with interaction frequency identifiers are extracted to form an intention recognition vocabulary group; the intention recognition vocabulary group is input into the interactive intention recognition model to obtain the first learning intention.

[0076] Further, including:

[0077] Randomly extract the first word in the intention recognition vocabulary group; read a predetermined standard vocabulary replacement strategy, and replace and adjust the first word based on the predetermined standard vocabulary replacement strategy to obtain a first standard vocabulary; form a standard intention recognition vocabulary group based on the first standard vocabulary; extract the first intention data group in the intention database, wherein the intention database is embedded in the interactive intention recognition model, and the first intention database includes a first historical standard intention recognition vocabulary group and a first historical learning intention; when the similarity between the first word of the standard intention recognition vocabulary group and the first historical standard intention recognition vocabulary group reaches a similarity limit, use the first historical learning intention as the first learning intention; output the first learning intention through the interactive intention recognition model.

[0078] Furthermore, the predetermined learning budget indicators include a predetermined learning budget duration and a predetermined learning budget effectiveness.

[0079] Further, including:

[0080] Reading predetermined student characteristics, the predetermined student characteristics include the student's already learned courses and the student's learning intention; using the student's already learned courses and the student's learning intention as clustering evaluation indicators, using the first learned course of the first student and the first learning intention as clustering constraints, clustering multiple historical students in the student database to obtain the first cluster cluster.

[0081] Further, including:

[0082] Compare the first history learning course syllabus table among the multiple history learning course syllabus tables with the first preliminary course version to obtain a first increase and decrease list; compare the second history learning course syllabus table among the multiple history learning course syllabus tables with the first preliminary course version to obtain a second increase and decrease list; take the union of the first increase and decrease list and the second increase and decrease list as the increase and decrease course syllabus set; arrange the increase and decrease course syllabus set in descending order based on the first increase and decrease frequency of the first course syllabus in the increase and decrease course syllabus set to obtain the first adjustment recommendation table.

[0083] Further, including:

[0084] Obtain a first post-learning satisfaction level of a first history student among the multiple history students; take a mode value of the first post-learning satisfaction level, recorded as a first predicted satisfaction level; mark the first predicted satisfaction level to the first target course, and generate the first course plan.

[0085] Through the above detailed description of the interactive solution generator based on natural language processing in this specification, those skilled in the art can clearly understand the interactive solution generating method based on natural language processing in this embodiment. Since it corresponds to the interactive solution generator based on natural language processing disclosed in Example 1, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0086] Embodiment 3: This embodiment of the present application provides a computer device, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown. The computer device includes a processor, a memory and a network interface connected through a system bus, wherein the processor of the computer device is used to provide computing and control capabilities; the memory of the computer device includes a non-volatile storage medium and an internal memory, the non-volatile storage medium stores an operating system, a computer program and a database, and the internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium; the network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it is implemented based on the interactive solution generator based on natural language processing in the first embodiment.

[0087] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0088] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0089] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An interactive solution generator based on natural language processing, characterized by: The interactive solution generator based on natural language processing includes: An interactive application receiving module, the interactive application receiving module is used to receive a first interactive application, the first interactive application includes a first course of application consultation interaction of a first student, wherein the first course has multiple course versions; A learning intention acquisition module, wherein the learning intention acquisition module is used to perform intention recognition on the first interactive text of the first student through an interactive intention recognition model to obtain a first learning intention; A fitness acquisition module, the fitness acquisition module is used to match a first preliminary course version among the multiple course versions according to a first learning budget parameter extracted from the first learning intention based on a predetermined learning budget indicator, and obtain a first preliminary fitness of the first preliminary course version; An instruction issuing module, the instruction issuing module is used to issue a preliminary selection adjustment instruction when the first preliminary selection fitness does not meet a predetermined fitness limit; A cluster matching module, wherein the cluster matching module is used to traverse a first cluster matching the first student in a student database, wherein the first cluster includes a plurality of history students, and the plurality of history students have a one-to-one correspondence with a plurality of history learning course syllabus tables; A recommendation table acquisition module, the recommendation table acquisition module is used to analyze the plurality of history learning course syllabus tables based on the preliminary selection adjustment instruction to obtain a first adjustment recommendation table; a target course acquisition module, the target course acquisition module being used to adjust the first preliminary course version by adding or reducing course items according to the first adjustment recommendation table to obtain a first target course; a course planning sending module, the course planning sending module being used to send a first course planning obtained by analyzing the first target course to the first student; The interactive solution generator based on natural language processing also includes a first learning intention acquisition module to perform the following operation steps: Preprocessing the first interactive text to obtain a first preprocessing result, wherein the first preprocessing result includes a plurality of words with interaction frequency identifiers; Extracting and analyzing the words with a predetermined ranking threshold in the interactive word sequence obtained by the plurality of words with interaction frequency identifiers to form an intention recognition word group; Inputting the intention recognition vocabulary group into the interactive intention recognition model to obtain the first learning intention; The interactive solution generator based on natural language processing also includes a first learning intention output module to perform the following operation steps: randomly extracting a first word from the intention recognition vocabulary group; Reading a predetermined standard vocabulary replacement strategy, and performing replacement adjustment on the first vocabulary based on the predetermined standard vocabulary replacement strategy to obtain a first standard vocabulary; forming a standard intent recognition vocabulary group based on the first standard vocabulary; Extracting a first intention data group from an intention database, wherein the intention database is embedded in the interactive intention recognition model, and the first intention data group includes a first historical standard intention recognition vocabulary group and a first historical learning intention; When the similarity between the first vocabulary of the standard intent recognition vocabulary group and the first historical standard intent recognition vocabulary group reaches a similarity limit, taking the first historical learning intent as the first learning intent; The first learning intention is output through the interaction intention recognition model.

2. The interactive solution generator based on natural language processing according to claim 1, characterized in that: The predetermined learning budget indicators include a predetermined learning budget duration and a predetermined learning budget effectiveness.

3. The interactive solution generator based on natural language processing according to claim 1, characterized in that: The interactive solution generator based on natural language processing further includes a first cluster acquisition module to perform the following operation steps: Reading predetermined student characteristics, wherein the predetermined student characteristics include the courses the student has learned and the student's learning intention; Taking the courses already learned by the student and the learning intention of the student as clustering evaluation indicators, taking the first learned course and the first learning intention of the first student as clustering constraints, clustering multiple historical students in the student database is performed to obtain the first cluster cluster.

4. The interactive solution generator based on natural language processing according to claim 1, characterized in that: The interactive solution generator based on natural language processing further includes a first adjustment recommendation table acquisition module to perform the following operation steps: Comparing a first history learning course outline table among the plurality of history learning course outline tables with the first preliminary course version to obtain a first addition and deletion list; Comparing a second history learning course syllabus table among the plurality of history learning course syllabus tables with the first preliminary course version to obtain a second addition and deletion list; Taking the union of the first increase and decrease list and the second increase and decrease list as the increase and decrease course syllabus set; The set of added and reduced course syllabi is arranged in descending order based on the first increase and decrease frequency of the first course syllabus in the set of added and reduced course syllabi to obtain the first adjustment recommendation table.

5. The interactive solution generator based on natural language processing according to claim 1, characterized in that: The interactive solution generator based on natural language processing also includes a first course planning acquisition module to perform the following operation steps: Obtaining a first post-study satisfaction level of a first history student among the plurality of history students; Take the mode value of the first post-study satisfaction and record it as the first predicted satisfaction; The first predicted satisfaction level is marked to the first target course to generate the first course plan.

6. An interactive solution generation method based on natural language processing, characterized in that: The method is implemented based on the interactive solution generator based on natural language processing according to any one of claims 1 to 5, and the method comprises: Receiving a first interactive application, wherein the first interactive application includes a first course of an application for consultation interaction by a first student, wherein the first course has multiple course versions; Performing intention recognition on the first interactive text of the first learner through an interactive intention recognition model to obtain a first learning intention; According to the first learning budget parameter extracted from the first learning intention based on the predetermined learning budget indicator, matching a first preliminary selected course version among the multiple course versions, and obtaining a first preliminary selected adaptability of the first preliminary selected course version; When the first preliminary selection fitness does not meet the predetermined fitness limit, issuing a preliminary selection adjustment instruction; Traversing a first cluster matching the first student in a student database, wherein the first cluster includes a plurality of history students, and the plurality of history students have a one-to-one correspondence with a plurality of history learning course syllabus tables; Analyzing the plurality of history learning course syllabus tables based on the preliminary selection adjustment instruction to obtain a first adjustment recommendation table; According to the first adjustment recommendation table, the first preliminary course version is adjusted to increase or decrease the course outline to obtain a first target course; The first course plan obtained by analyzing the first target course is sent to the first student.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the interactive solution generator based on natural language processing is implemented according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the interactive solution generator based on natural language processing according to any one of claims 1 to 5 is implemented.

Images