Teaching text generation method and system based on natural language processing

Abstract

The invention provides a teaching text generation method and system based on natural language processing, and relates to the technical field of data processing, and the method comprises the steps: collecting and obtaining a target teaching knowledge base, carrying out entity recognition and relation extraction, and constructing a target teaching knowledge graph; obtaining a target teaching input text, performing text structuring processing, and determining a target teaching text structure; presetting a teaching content rule base, and performing content mapping and content filtering to obtain basic teaching text content; constructing a target student portrait based on the learning behavior historical data, performing dynamic adjustment, and determining adaptive teaching text content; and building a teaching template library, performing template matching and text injection, and generating target teaching text content. The technical problem that a teaching text generation method in the prior art generally depends on a templated question bank and a fixed rule, flexible adjustment is difficult to make according to the learning progress and learning characteristics of each student, and the teaching effect is not ideal is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for generating teaching texts based on natural language processing. Background Art

[0002] In today's education sector, with the continuous development of information technology, intelligent teaching tools have gradually become a powerful tool for teachers, students, parents, and other stakeholders. This is especially true in areas such as homework design, grading, feedback, and home-school interaction. Digital and intelligent methods are increasingly becoming the key to improving educational efficiency. Traditional teaching text generation methods typically rely on templated question banks and fixed grading rules. While efficient, this approach is less responsive to individual needs. Existing technologies often struggle to flexibly adjust homework design based on each student's learning progress and characteristics, resulting in teachers lacking sufficient flexibility and innovation in homework design. Consequently, the content of homework and feedback may not fully meet students' personalized learning needs, resulting in ineffective teaching and little significant improvement in student learning. Summary of the Invention

[0003] This application provides a teaching text generation method and system based on natural language processing, aiming to solve the technical problem that the teaching text generation methods in the prior art usually rely on templated question banks and fixed rules, which are difficult to make flexible adjustments according to the learning progress and learning characteristics of each student, resulting in a lack of sufficient flexibility and innovation in the design of teaching texts, and thus leading to unsatisfactory teaching results.

[0004] The first aspect disclosed in the present application provides a teaching text generation method based on natural language processing, the method comprising: collecting and acquiring a target teaching knowledge base, performing entity recognition and relationship extraction on the target teaching knowledge base, and constructing a target teaching knowledge graph; acquiring a target teaching input text, performing text structuring processing on the target teaching input text using a natural language processing algorithm, and determining a target teaching text structure; presetting a teaching content rule base, and using the target teaching knowledge graph to perform content mapping and content filtering on the target teaching text structure based on the teaching content rule base to obtain basic teaching text content; constructing a target student portrait based on learning behavior history data, and dynamically adjusting the basic teaching text content in combination with the target student portrait to determine the adapted teaching text content; building a teaching template library, and performing template matching and text injection on the adapted teaching text content based on the teaching template library to generate target teaching text content.

[0005] The second aspect disclosed in the present application provides a teaching text generation system based on natural language processing, which is used for the above-mentioned teaching text generation method based on natural language processing. The system includes: a relationship extraction module, which is used to collect and obtain a target teaching knowledge base, perform entity recognition and relationship extraction on the target teaching knowledge base, and construct a target teaching knowledge graph; a structured processing module, which is used to obtain a target teaching input text, perform text structured processing on the target teaching input text using a natural language processing algorithm, and determine the target teaching text structure; a content filtering module, which is used to preset a teaching content rule base, use the target teaching knowledge graph to perform content mapping and content filtering on the target teaching text structure based on the teaching content rule base, and obtain basic teaching text content; a dynamic adjustment module, which is used to construct a target student portrait based on learning behavior history data, dynamically adjust the basic teaching text content in combination with the target student portrait, and determine the adapted teaching text content; a text injection module, which is used to build a teaching template library, perform template matching and text injection on the adapted teaching text content based on the teaching template library, and generate the target teaching text content.

[0006] One or more technical solutions provided in this application have at least the following beneficial effects: By performing entity recognition and relationship extraction on the target teaching knowledge base, the key concepts in the teaching content and the relationships between them are automatically extracted, which can efficiently convert scattered teaching data into structured knowledge graphs, facilitating subsequent knowledge reasoning and content generation; by performing structured processing on the target teaching input text through natural language processing algorithms, the key information in the text can be identified and organized into a standardized format, which not only makes the original text content more standardized, but also provides a clear framework for subsequent content mapping and generation, thereby improving the comprehensibility and operability of the teaching content; through the preset teaching content rule library, the content in the target teaching text structure is automatically mapped to the corresponding concepts and terms in the teaching knowledge graph, ensuring that the generated teaching text content is consistent with the known teaching knowledge base and subject knowledge. Combined with content filtering, it helps to generate more refined basic teaching text content that meets teaching needs; based on learning Building a target student portrait based on historical learning behavior data can reflect personalized information such as students' knowledge mastery, learning style, and interest preferences, providing data support for subsequent adjustments to teaching texts. Combined with the target student portrait, the basic teaching text content is automatically adjusted to make it more in line with the needs of specific students, significantly improving the degree of personalization of teaching and helping to improve students' learning outcomes. By building a teaching template library, appropriate templates are automatically selected for filling and injection based on the adapted teaching text content, ensuring that the generation of teaching texts not only complies with the teaching syllabus, but also can efficiently generate structured and standardized teaching materials. After matching the appropriate template, the teaching content is automatically injected into the template to complete field replacement and dynamic generation. This automated generation process not only reduces the workload of manual editing, but also ensures that the generated teaching texts meet the expected teaching objectives and subject requirements, greatly improving the efficiency and accuracy of teaching design.

[0007] 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

[0008] Figure 1 A flow chart of a method for generating teaching text based on natural language processing provided in an embodiment of the present application.

[0009] Figure 2 A schematic diagram of the structure of a teaching text generation system based on natural language processing provided in an embodiment of the present application.

[0010] Explanation of the accompanying drawings: relationship extraction module 10, structured processing module 20, content filtering module 30, dynamic adjustment module 40, text injection module 50. DETAILED DESCRIPTION

[0011] The embodiments of the present application provide a method and system for generating teaching texts based on natural language processing, which solves the technical problem that the teaching text generation methods in the prior art generally rely on templated question banks and fixed rules, making it difficult to make flexible adjustments based on the learning progress and learning characteristics of each student, resulting in a lack of sufficient flexibility and innovation in the design of teaching texts, and thus leading to unsatisfactory teaching results.

[0012] After introducing the basic principles of this application, various non-limiting embodiments of this application will be specifically described below in conjunction with the accompanying drawings. It should be understood that the specific embodiments described here are only used to explain this application and are not used to limit this application.

[0013] Example 1

[0014] like Figure 1 As shown, the embodiment of the present application provides a method for generating teaching text based on natural language processing, the method comprising: Collect and obtain the target teaching knowledge base, perform entity recognition and relationship extraction on the target teaching knowledge base, and construct a target teaching knowledge graph.

[0015] Acquire a large amount of teaching-related knowledge, including theoretical knowledge, practical cases, and teaching methods, from multiple sources such as textbooks, research papers, teaching literature, and online education platforms. This collected teaching knowledge is then integrated to form a target teaching knowledge base. Natural language processing techniques are used to perform entity recognition on the text in the target teaching knowledge base, extracting important subject-related entities from the teaching content. These entities constitute the basic units of teaching knowledge. In addition to entity recognition, relationships between entities, such as causal relationships, inclusion relationships, or similarity relationships, are extracted. Relationship extraction can be achieved through techniques such as dependency parsing and semantic role labeling. The identified entities and extracted relationships are organized into a graph structure, where nodes represent entities and edges represent relationships, to obtain the target teaching knowledge graph.

[0016] A target teaching input text is obtained, and a natural language processing algorithm is used to perform text structuring processing on the target teaching input text to determine the target teaching text structure.

[0017] Target teaching input text refers to text data from students or teachers, and can come from various forms of input such as textbooks, homework, and discussions. The target teaching input text is subjected to text structuring. Text structuring refers to converting unstructured text into information with a clear structure. This includes: parsing the text using syntactic analysis algorithms to identify basic components of a sentence, such as the subject, predicate, and object; performing contextual semantic analysis on the text to understand the meaning of each sentence; and segmenting the text based on its semantic information, breaking long sentences or paragraphs into smaller semantic units. Through structuring, the target teaching text structure is obtained, allowing the input text to match multiple dimensions, such as teaching objectives and student needs, ensuring that the generated teaching content is highly personalized and accurate.

[0018] A teaching content rule library is preset, and the target teaching knowledge graph is used to perform content mapping and content filtering on the target teaching text structure based on the teaching content rule library to obtain basic teaching text content.

[0019] A pre-set teaching content rule library is a collection of content rules, specifications, and logical frameworks for a specific teaching domain. Its content includes rules for organizing knowledge points, teaching objectives and competency requirements, teaching methods, and content updates and specifications. Based on the target teaching text structure, its content is mapped to the corresponding content modules in the teaching content rule library. For example, if the text describes "Newton's Second Law in Physics," it is mapped to the "Mechanics Foundations" module in the "Physics" domain and further refined into knowledge points related to "Force and Motion." Based on the specifications of the teaching content rule library, the target teaching text structure is filtered to remove irrelevant information or information that does not meet teaching requirements. For example, if the text content is too complex and exceeds the student's cognitive level, it is simplified. Based on the mapping and filtering results, structured basic teaching text content is generated, which strictly adheres to teaching objectives and specification requirements.

[0020] A target student portrait is constructed based on historical learning behavior data, and the basic teaching text content is dynamically adjusted in combination with the target student portrait to determine the adapted teaching text content.

[0021] Historical learning behavior data includes students' interactive behaviors, homework grades, learning progress, study time, test results, participation in class discussions, etc. It can also include students' performance inside and outside the classroom, such as the frequency of accessing learning resources, the length of time spent watching teaching videos, and interactions with other learners. The target student portrait is a personalized learning description generated based on historical learning behavior data. The portrait analysis dimensions include knowledge mastery, learning style, and interest preferences. Based on the portrait analysis dimensions, a multi-dimensional analysis of historical learning behavior data is performed to obtain student feature sets in different dimensions. Data clustering techniques, such as K-means or hierarchical clustering, can be used to classify student groups, divide different learning groups, and construct target student portraits. Through cluster analysis, we can better understand students' learning patterns and provide a basis for subsequent dynamic adjustments. Based on the target student portrait, the basic teaching text content is dynamically adjusted according to the characteristics of each student group, and the adapted teaching text content is determined to achieve the generation of personalized teaching content.

[0022] A teaching template library is built, and template matching and text injection are performed on the adapted teaching text content based on the teaching template library to generate target teaching text content.

[0023] Build a teaching template library. The teaching template library is a database that stores various teaching templates. Each teaching template defines a format for presenting teaching content. The teaching template includes modules such as teaching objectives, knowledge point introduction, examples, exercises, and summaries. Among them, teaching templates can be classified according to subjects, teaching methods, learning objectives, and student levels, such as basic teaching templates, interactive teaching templates, and advanced teaching templates.

[0024] According to the target student portrait and the difficulty and objectives of the teaching content, the most suitable teaching template is matched. For example, if the target student portrait shows that the group is a basic learner, a concise and clear basic teaching template is selected. According to the matched teaching template, the adapted teaching text content is injected into the teaching template. This process replaces fields and dynamically generates content according to the structure of the teaching template. For example, the specific teaching content and cases in the "Basics of Mechanics" module are inserted into the corresponding positions in the template, and the depth and complexity of the content are dynamically adjusted according to the students' learning behavior. The target teaching text content finally generated includes structured content, standardized teaching modules and personalized content adapted to students' needs. These contents can provide knowledge point explanations, case analysis, exercise training, etc. that are in line with students' abilities and interests, and can effectively improve teaching effectiveness and students' learning experience.

[0025] Furthermore, the construction of the target teaching knowledge graph includes: According to the text data application standard, a multi-level data cleaning procedure is determined; the target teaching knowledge base is cleaned according to the multi-level data cleaning procedure to obtain a standard teaching knowledge base; entity labeling and relationship extraction are performed on the standard teaching knowledge base to obtain a teaching knowledge entity set and a knowledge entity relationship set; graph construction and verification optimization are performed based on the teaching knowledge entity set and the knowledge entity relationship set to construct the target teaching knowledge graph.

[0026] The quality and format of the data in the target teaching knowledge base may be quite heterogeneous, so data cleaning is necessary to ensure the standardization of text data. A series of text data application standards should be defined to unify and standardize data processing, such as data integrity standards, data consistency standards, and data accuracy standards. This will determine a multi-level data cleaning process. For example, redundant content, noise information, and irrelevant text should be removed, while spelling checking and error correction should be performed. The text format should be standardized, such as standardizing punctuation, abbreviations, and capitalization. The semantic content in the text should be annotated and cleaned, identifying and correcting polysemy and ambiguity. Common words that have no substantive impact on semantic understanding should be removed.

[0027] The target teaching knowledge base is cleaned according to the multi-level data cleaning procedure. The standard teaching knowledge base after cleaning conforms to the unified format, has accurate content and contains no redundant or irrelevant information.

[0028] The purpose of entity labeling is to identify important entities in the text and assign corresponding categories to them. Rule-based labeling methods and deep learning models are used for entity recognition. Based on the results of entity recognition, a set of teaching knowledge entities is obtained. The purpose of relationship extraction is to identify the relationship between different entities in the text, such as the relationship between "Newton's law" and "mechanics". Pattern matching-based methods or deep learning models can be used for relationship extraction. The goal of relationship extraction is to determine which entities have specific types of relationships. Through relationship extraction, a set of knowledge entity relationships is obtained.

[0029] Based on the acquired set of teaching knowledge entities and knowledge entity relationships, a knowledge graph is constructed. A knowledge graph represents the structured relationships between knowledge using nodes and edges, where each entity is considered a node, and the edges between each pair of related entities are their relationships. Furthermore, constructing a knowledge graph involves more than just generating nodes and edges; the key lies in its optimization process, which includes removing redundant information, optimizing the representation of nodes and relationships, and cleaning up inaccurate entities and relationships. This optimization results in a target teaching knowledge graph with high accuracy, which can provide strong support for the generation of subsequent teaching content.

[0030] Furthermore, obtaining the teaching knowledge entity set and the knowledge entity relationship set includes: Build an entity standardization system, perform entity recognition and annotation on the standard teaching knowledge base according to the entity standardization system to obtain an entity annotation sample set; use the BERT network structure to proportionally divide and annotate the entity annotation sample set to obtain an entity annotation model; based on the entity annotation model, perform entity type annotation on the standard teaching knowledge base to obtain a teaching knowledge entity set; obtain an entity relationship rule template, use the entity relationship rule template to perform relationship extraction and template matching on the teaching knowledge entity set to obtain a knowledge entity relationship set.

[0031] In a teaching knowledge base, entities can be divided into different categories, such as subject knowledge points, mathematical theorems, and physical formulas. The entity standardization system's task is to clarify these entity categories and assign standardized labels to each category to avoid confusion between different categories. Following this entity standardization system, existing semi-automated annotation tools are used to perform entity recognition and annotation on the standard teaching knowledge base. After annotation, a sample entity annotation set is generated, containing the annotated entity information and its classification. This sample entity annotation set is used to train and evaluate subsequent entity annotation models.

[0032] The BERT network structure is a pre-trained language model based on the Transformer architecture. Its main advantage lies in its ability to capture contextual information in text and understand the context of vocabulary through bidirectional training. The entity annotation sample set is divided into a training dataset and a validation dataset. The common ratio is 80% for training and 20% for validation. This division ensures model validation during training while preventing overfitting. During the annotation training process, the BERT network structure is trained using the training dataset, and the cross-entropy loss function is used to calculate the gap between the predicted label output by the model and the actual label. The model goes through multiple rounds of training, continuously adjusting parameters to minimize the loss function, and ultimately reaches a state that can effectively perform entity annotation. After training is complete, the validation dataset is used to evaluate the model's performance, and the model is further optimized based on the evaluation results, ultimately resulting in an entity annotation model that can automatically identify various entities in the teaching knowledge base.

[0033] Use the trained entity annotation model to process each text in the standard teaching knowledge base and mark the entity types therein. The annotated text contains multiple entities and their corresponding category labels. These entities constitute the teaching knowledge entity set.

[0034] The goal of relationship extraction is to extract the logical relationships between entities from the set of teaching knowledge entities. Common relationship types include definition relationships, causal relationships, belonging relationships, and dependency relationships. Entity relationship rule templates are pre-defined patterns used to extract relationships between different entities. Entity relationship rule templates can be static or dynamically learned through machine learning methods. For example, dependency parsing in natural language processing can be used to understand the grammatical relationships between different entities. In the annotated set of teaching knowledge entities, the defined entity relationship rule templates are applied to identify and extract the relationships between entities. That is, the corresponding relationships are extracted through template matching. After relationship extraction, the generated knowledge entity relationship set contains entity pairs and the relationship types between them.

[0035] Furthermore, determining the target teaching text structure includes: A natural language processing algorithm is used to perform denoising and format conversion on the target teaching input text to obtain a standard teaching input text; PCFG is used to perform syntactic analysis and semantic recognition on the standard teaching input text to obtain semantic information of the teaching input text; a preset text segmentation rule is used to segment the standard teaching input text based on the semantic information of the teaching input text to obtain a teaching text segmentation set; entity extraction and relationship classification are performed on the teaching text segmentation set based on NER, and the target teaching text structure is structured and output.

[0036] De-noising involves removing irrelevant content, correcting spelling and grammar, and removing stop words. Specifically, the target teaching input text may contain unnecessary or irrelevant content, such as advertisements and copyright notices. This often affects the quality of text processing and needs to be removed, leaving only the main text. The target teaching input text may also contain spelling errors or grammatical irregularities, which can affect subsequent processing. Spelling errors are corrected through spell checking. Stop words are common words that do not contribute to semantic understanding and often interfere with text analysis results, so they need to be removed. Format conversion is performed, converting the raw text into a standardized structure. For example, free text can be converted into JSON, XML, or other structured formats suitable for computer processing. After denoising and format conversion, the resulting standard teaching input text has a uniform format, is free of redundant content, and is suitable for further processing, providing high-quality input data for subsequent semantic recognition and syntactic analysis.

[0037] PCFG is a probability-based context-free grammar that assigns probabilities to grammar rules to represent the likelihood of each rule generating text in a certain context. PCFG is widely used in syntactic analysis, especially when processing the grammatical structure of complex text, as it can provide better predictive capabilities.

[0038] The goal of syntactic analysis is to convert the input natural language text into a syntactic tree that shows the structural relationship between words. PCFG infers the most likely syntactic structure based on the probability value. Specifically, the standard teaching input text is processed into lexical units through word segmentation, and PCFG is applied to perform syntactic analysis on these lexical units to generate multiple possible syntactic trees. The nodes of the syntactic tree represent words or phrases, and the edges represent the relationship between words. Finally, the one with the highest probability is selected as the optimal structure.

[0039] Semantic recognition is performed based on syntactic analysis. Specifically, by analyzing the syntactic tree, semantic roles such as subject, predicate, object, etc. are identified, and the relationship between each component is understood. For example, in the sentence "Newton's second law is the basis of physics", PCFG is used to identify "Newton's second law" as the subject, "is" as the predicate, and "the basis of physics" as the object, which helps to understand the semantic structure of the sentence and ultimately obtain the semantic information of the teaching input text, providing basic data for subsequent teaching content generation and processing.

[0040] Text segmentation rules are used to divide the text into multiple units with independent meanings based on the semantic information and structure of the teaching text. These units can be based on themes, concepts, chapters, paragraphs, sentences or grammatical structures. The design of the rules should be formulated according to the characteristics of the teaching content. For example, for textbooks or lesson plan texts, they can be segmented according to the following rules: divide according to different teaching topics or chapters, each topic or chapter contains relevant knowledge points and concepts; divide the text by sentences, each sentence represents a basic semantic unit; divide the text into several parts based on keywords or concepts, each part revolves around a core concept; based on the results of semantic analysis, identify the interrelated parts of the text and process them as a segment.

[0041] The above-mentioned text segmentation rules are used to segment the standard teaching input text into blocks. During the segmentation process, combined with the semantic information of the teaching input text, it is possible to identify which parts should belong to the same block. The segmented text becomes multiple smaller semantic units, forming a teaching text segment set, which serves as input data for subsequent processing.

[0042] NER is a natural language processing task that aims to identify meaningful entities from text, such as names of people, places, organizations, times, dates, subject names, formulas, and so on. NER is used to extract entities from a collection of instructional text blocks. Following entity extraction, relationship classification is performed, identifying the relationships between entities. Relationship classification can be based on predefined rules or using machine learning models. For example, a BERT-based relationship classification model is used to identify and classify relationships between two entities. After entity extraction and relationship classification, structural processing is performed to obtain the target instructional text structure, which includes clear entities and their relationships. This provides a foundation for subsequent instructional content generation, knowledge query, and recommendation systems.

[0043] Furthermore, obtaining the basic teaching text content includes: The target teaching knowledge graph is used to perform content mapping on the target teaching text structure to obtain matching teaching text content; content specification rules, teaching logic rules and capability adaptation rules are determined according to the teaching content rule library; based on the content specification rules, teaching logic rules and capability adaptation rules, the matching teaching text content is content filtered and integrated to obtain the basic teaching text content.

[0044] The target teaching knowledge graph can provide information such as the relationship between knowledge points, conceptual structure, and contextual associations; the target teaching text structure includes annotated entities and the grammatical structure between them. Using the target teaching knowledge graph, each teaching module in the target teaching text structure is matched with the corresponding node in the graph. The goal of this process is to associate the specific content in the teaching text with the abstract knowledge points, theorems, formulas, definitions, etc. in the knowledge graph. Specifically, the entities in the target teaching text structure are directly matched with the entities in the target teaching knowledge graph. If the entities in the knowledge graph are not directly mentioned in the text, they can be mapped through semantic similarity matching or synonym matching. Through the above mapping process, each part of the target teaching text structure is associated with the information in the target teaching knowledge graph, and finally a matching teaching text content consistent with the target teaching knowledge graph is obtained.

[0045] The teaching content rule base is a database containing various teaching norms and rules. It covers how to organize teaching content, how to design teaching structure, and how to adjust teaching materials to meet the needs of different student groups. According to the teaching content rule base, content norms, teaching logic rules and ability adaptation rules are determined. Among them, content norms are used to determine the standardization and consistency of teaching content, such as the writing standards of formulas, unified definitions of terms, standardization of paragraph structures, etc.; teaching logic rules are used to ensure that teaching content is presented in a reasonable logical order, for example, basic knowledge is introduced first, and then complex concepts are gradually introduced to ensure that students can understand the content step by step; ability adaptation rules are used to adjust the difficulty and presentation of teaching content according to students' learning ability, interests and knowledge background. For example, for beginners, more concise and clear explanations are needed, while for advanced learners, more in-depth and challenging content is required.

[0046] Content filtering is performed according to the aforementioned content specification rules, teaching logic rules, and ability adaptation rules. The purpose is to remove non-compliant or unnecessary parts, so that the final teaching text is more concise and efficient. The filtered content includes terms or symbols that do not conform to content specifications, paragraphs with unclear logic or unreasonable order, and content that is too complex or too simple that is not suitable for specific student groups. After content filtering, the remaining content is integrated together according to the specifications and logical order to form a structured and easy-to-understand basic teaching text content, which will meet all content specification rules, teaching logic rules, and ability adaptation rules.

[0047] Furthermore, the target student profile constructed based on historical learning behavior data includes: Define a portrait dimension set, which includes knowledge mastery, learning style, and interest preference; perform multidimensional analysis on the learning behavior history data according to the portrait dimension set to obtain a multidimensional portrait parameter set; perform cluster analysis on the multidimensional portrait parameter set to obtain student group clustering results, and construct the target student portrait based on the student group clustering results.

[0048] The portrait dimension set is a collection of key factors that need to be considered when constructing a student portrait. These dimensions are used to describe the multidimensional characteristics of students and provide detailed information required for personalized teaching. Among them, the degree of knowledge mastery indicates the degree to which students have mastered the course content and subject knowledge points, which can be measured by students' homework scores, test scores, exam results, etc.; learning style indicates the way and method preferred by students in the learning process, such as visual learners (obtaining information through images, charts, etc.), auditory learners (learning through lectures, discussions, etc.) or hands-on learners (learning through experiments and actual operations); interest preferences indicate students' interest in specific subjects or knowledge areas, and students' interest areas can be evaluated through the courses they choose, the activities they participate in, the reading materials, etc.

[0049] Historical learning behavior data refers to the various records and data generated by students during their learning process. Using the aforementioned profile dimension set, we conduct multidimensional analysis of this data. For example, we can analyze students' mastery of different subjects and knowledge points based on their academic performance data. We can also infer their learning styles through their learning activity records and participation patterns. For example, some students frequently watch instructional videos and take notes, indicating a visual learning style. We can also infer their interests and preferences based on the courses they choose, the reading materials they read, and the learning activities they participate in. This multidimensional analysis yields a multidimensional profile parameter set encompassing each dimension, reflecting each student's characteristics and performance across these dimensions.

[0050] Cluster analysis is an unsupervised learning method that aims to divide a dataset into clusters such that members within each cluster are as similar as possible in certain characteristics, while clusters are as diverse as possible. In this step, cluster analysis methods, such as K-means and hierarchical clustering, are used to group students based on a multidimensional profile parameter set, identifying groups of similar learners. The cluster analysis results reveal the characteristics of different student groups. For example, in one group, most students have mastered basic knowledge and prefer to learn through videos; in another group, students have a higher level of knowledge but are very interested in physics and prefer to learn through experiments. Based on the results of the cluster analysis, a target student profile is constructed for each student group, reflecting detailed characteristics of the students along different dimensions.

[0051] Furthermore, the determination of the adapted teaching text content includes: An adjustment strategy is performed on the dimensional information of each dimension in the portrait dimension set to obtain a multi-dimensional teaching adjustment strategy set; a strategy is matched on the target student portrait based on the multi-dimensional teaching adjustment strategy set to determine the target teaching adjustment strategy; the basic teaching text content is dynamically adjusted using the target teaching adjustment strategy to determine the adapted teaching text content.

[0052] Adjustment strategies aim to tailor instructional content and methods to the diverse learning characteristics and needs of students, making them more tailored to the personalized learning needs of specific students. First, a detailed analysis of each dimension in the profile dimension set is conducted to identify how each dimension influences the adjustment of instructional strategies. For example, if a student has a poor grasp of a particular knowledge point, more detailed explanations, simpler examples, or additional practice and guidance may be needed. If a student has a good grasp of a particular knowledge point, more challenging content or direct access to deeper learning may be provided. If a student prefers visual learning, adjustment strategies may include more diagrams, video explanations, and animations. If an auditory learner is more inclined to use more explanations or audio resources. If a student has a strong interest in a particular area, instructional content can be personalized to that student's interests, incorporating relevant content and examples to stimulate their learning motivation. Combining information from these dimensions, a multidimensional set of instructional adjustment strategies tailored to different student groups is developed, aiming to precisely adjust the instructional process to maximize its adaptation to students' individual needs.

[0053] Based on the multi-dimensional teaching adjustment strategy set, strategy matching is carried out for the target student portrait. The goal is to find the teaching adjustment strategy that best suits each student so that the teaching content and methods can meet their learning needs. Through this process, a specific and personalized target teaching adjustment strategy can be determined for each student.

[0054] Using the determined target teaching adjustment strategy, the basic teaching text content is dynamically adjusted to make it more suitable for the personalized learning needs of each student. Finally, adaptive teaching text content is generated to suit each student. The adaptive teaching text content includes personalized learning resources, appropriate teaching methods and reasonable knowledge depth to ensure that the teaching content meets the needs of students to the greatest extent possible.

[0055] Furthermore, generating target teaching text content includes: Template matching rules are defined, and the template matching rules include teaching objective matching, subject matching, and difficulty matching; the template matching rules are used to perform template matching and text injection on the adapted teaching text content based on the teaching template library to obtain the target teaching text content.

[0056] Template matching rules are defined to ensure that teaching content can be adjusted and optimized to match different teaching objectives, subject requirements, and student abilities, thereby generating personalized and standard teaching texts. Template matching rules include teaching objective matching, subject matching, and difficulty matching. Teaching objectives are the core outcomes or abilities to be achieved in teaching activities. Teaching objective matching rules ensure that teaching content is consistent with teaching objectives. For example, if the teaching objective is "Master Newton's Second Law of Physics," the template matching rules must ensure that the selected template effectively conveys this objective, ensuring that the required theoretical explanations, examples, and exercises are centered around Newton's Second Law. Subject matching rules ensure that the selected teaching template is consistent with the subject being taught. For example, a teaching template for physics differs in structure and content from a template for mathematics or chemistry. Based on the subject category being studied, a template relevant to the student is selected. Difficulty matching rules select teaching templates of appropriate difficulty based on the student's learning level. For example, beginners require more basic content and more examples, while advanced students require more challenging tasks or in-depth analysis. These three matching rules—teaching objective, subject, and difficulty—work together to accurately select the most appropriate template and provide students with personalized learning resources.

[0057] The adapted teaching text content is used as input, and template matching is performed from a library of teaching templates based on template matching rules. For example, if the teaching objective is "Mastering Newton's Second Law" and the student is a beginner, a template containing basic definitions, diagrams, and simple application examples is selected. After template matching is complete, the adapted teaching text content is inserted into the template, ultimately generating personalized teaching content that meets the student's needs, namely the target teaching text content. This process ensures that the teaching content not only meets the requirements in terms of structure, but also meets the student's learning needs in terms of depth, difficulty, subject relevance, and teaching objectives.

[0058] Furthermore, obtaining the target teaching text content includes: The template matching rule is used to perform similarity calculation and template optimization on the adapted teaching text content based on the teaching template library to obtain the target teaching template; the adapted teaching text content is injected into the target teaching template for field replacement and dynamic generation to obtain the target teaching text content.

[0059] The text similarity calculation method in natural language processing, such as cosine similarity, can be used to evaluate the degree of match between the adapted teaching text content and the template. For example, text similarity calculation can evaluate the similarity between the template and the adapted teaching content by analyzing keywords, syntactic structure, semantic information, etc. After calculating the similarity between all available templates and the adapted teaching text content, the templates are sorted according to the similarity score. The template with the highest similarity is the template that is most suitable for the teaching content and is used as the target teaching template.

[0060] Target teaching templates typically contain multiple placeholders or dynamic fields. These fields can be dynamic parts of the template, waiting to be filled in with specific content. The adapted teaching text content is inserted into the corresponding fields in the template. In addition to simple field replacement, some fields in the template are dynamically generated based on the specific teaching content. For example, the "Example Questions" section in the template can automatically select relevant example questions for display based on the teaching content, or automatically generate exercises related to the knowledge points. All field replacements and dynamically generated content are integrated to form the complete target teaching text content. At this point, the target teaching text content already contains all the necessary information, and has a clear structure and complete content, which meets the teaching objectives, subject requirements and students' learning needs.

[0061] In summary, the teaching text generation method based on natural language processing provided by the embodiment of the present application has the following technical effects: By performing entity recognition and relationship extraction on the target teaching knowledge base, the key concepts in the teaching content and the relationships between them are automatically extracted, which can efficiently convert scattered teaching data into structured knowledge graphs, facilitating subsequent knowledge reasoning and content generation; by performing structured processing on the target teaching input text through natural language processing algorithms, the key information in the text can be identified and organized into a standardized format, which not only makes the original text content more standardized, but also provides a clear framework for subsequent content mapping and generation, thereby improving the comprehensibility and operability of the teaching content; through the preset teaching content rule library, the content in the target teaching text structure is automatically mapped to the corresponding concepts and terms in the teaching knowledge graph, ensuring that the generated teaching text content is consistent with the known teaching knowledge base and subject knowledge. Combined with content filtering, it helps to generate more refined basic teaching text content that meets teaching needs; based on learning Building a target student portrait based on historical learning behavior data can reflect personalized information such as students' knowledge mastery, learning style, and interest preferences, providing data support for subsequent adjustments to teaching texts. Combined with the target student portrait, the basic teaching text content is automatically adjusted to make it more in line with the needs of specific students, significantly improving the degree of personalization of teaching and helping to improve students' learning outcomes. By building a teaching template library, appropriate templates are automatically selected for filling and injection based on the adapted teaching text content, ensuring that the generation of teaching texts not only complies with the teaching syllabus, but also can efficiently generate structured and standardized teaching materials. After matching the appropriate template, the teaching content is automatically injected into the template to complete field replacement and dynamic generation. This automated generation process not only reduces the workload of manual editing, but also ensures that the generated teaching texts meet the expected teaching objectives and subject requirements, greatly improving the efficiency and accuracy of teaching design.

[0062] Example 2

[0063] Based on the same inventive concept as the teaching text generation method based on natural language processing in the above embodiment, Figure 2 As shown, the embodiment of the present application provides a teaching text generation system based on natural language processing, the system comprising: The relationship extraction module 10 is used to collect and obtain the target teaching knowledge base, perform entity recognition and relationship extraction on the target teaching knowledge base, and construct a target teaching knowledge graph; the structured processing module 20 is used to obtain the target teaching input text, use a natural language processing algorithm to perform text structured processing on the target teaching input text, and determine the target teaching text structure; the content filtering module 30 is used to preset a teaching content rule base, use the target teaching knowledge graph to perform content mapping and content filtering on the target teaching text structure based on the teaching content rule base, and obtain the basic teaching text content; the dynamic adjustment module 40 is used to construct a target student portrait based on learning behavior history data, dynamically adjust the basic teaching text content in combination with the target student portrait, and determine the adapted teaching text content; the text injection module 50 is used to build a teaching template library, perform template matching and text injection on the adapted teaching text content based on the teaching template library, and generate the target teaching text content.

[0064] Furthermore, the relationship extraction module 10 is configured to perform the following steps: According to the text data application standard, a multi-level data cleaning procedure is determined; the target teaching knowledge base is cleaned according to the multi-level data cleaning procedure to obtain a standard teaching knowledge base; entity labeling and relationship extraction are performed on the standard teaching knowledge base to obtain a teaching knowledge entity set and a knowledge entity relationship set; graph construction and verification optimization are performed based on the teaching knowledge entity set and the knowledge entity relationship set to construct the target teaching knowledge graph.

[0065] Furthermore, the relationship extraction module 10 is configured to perform the following steps: Build an entity standardization system, perform entity recognition and annotation on the standard teaching knowledge base according to the entity standardization system to obtain an entity annotation sample set; use the BERT network structure to proportionally divide and annotate the entity annotation sample set to obtain an entity annotation model; based on the entity annotation model, perform entity type annotation on the standard teaching knowledge base to obtain a teaching knowledge entity set; obtain an entity relationship rule template, use the entity relationship rule template to perform relationship extraction and template matching on the teaching knowledge entity set to obtain a knowledge entity relationship set.

[0066] Furthermore, the structured processing module 20 is configured to perform the following steps: A natural language processing algorithm is used to perform denoising and format conversion on the target teaching input text to obtain a standard teaching input text; PCFG is used to perform syntactic analysis and semantic recognition on the standard teaching input text to obtain semantic information of the teaching input text; a preset text segmentation rule is used to segment the standard teaching input text based on the semantic information of the teaching input text to obtain a teaching text segmentation set; entity extraction and relationship classification are performed on the teaching text segmentation set based on NER, and the target teaching text structure is structured and output.

[0067] Furthermore, the content filtering module 30 is configured to perform the following steps: The target teaching knowledge graph is used to perform content mapping on the target teaching text structure to obtain matching teaching text content; content specification rules, teaching logic rules and capability adaptation rules are determined according to the teaching content rule library; based on the content specification rules, teaching logic rules and capability adaptation rules, the matching teaching text content is content filtered and integrated to obtain the basic teaching text content.

[0068] Furthermore, the dynamic adjustment module 40 is configured to perform the following steps: Define a portrait dimension set, which includes knowledge mastery, learning style, and interest preference; perform multidimensional analysis on the learning behavior history data according to the portrait dimension set to obtain a multidimensional portrait parameter set; perform cluster analysis on the multidimensional portrait parameter set to obtain student group clustering results, and construct the target student portrait based on the student group clustering results.

[0069] Furthermore, the dynamic adjustment module 40 is configured to perform the following steps: An adjustment strategy is performed on the dimensional information of each dimension in the portrait dimension set to obtain a multi-dimensional teaching adjustment strategy set; a strategy is matched on the target student portrait based on the multi-dimensional teaching adjustment strategy set to determine the target teaching adjustment strategy; the basic teaching text content is dynamically adjusted using the target teaching adjustment strategy to determine the adapted teaching text content.

[0070] Furthermore, the text injection module 50 is configured to perform the following steps: Template matching rules are defined, and the template matching rules include teaching objective matching, subject matching, and difficulty matching; the template matching rules are used to perform template matching and text injection on the adapted teaching text content based on the teaching template library to obtain the target teaching text content.

[0071] Furthermore, the text injection module 50 is configured to perform the following steps: The template matching rule is used to perform similarity calculation and template optimization on the adapted teaching text content based on the teaching template library to obtain the target teaching template; the adapted teaching text content is injected into the target teaching template for field replacement and dynamic generation to obtain the target teaching text content.

[0072] Through the above detailed description of the teaching text generation method based on natural language processing in this specification, those skilled in the art can clearly understand the teaching text generation system based on natural language processing in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0073] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one 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 is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for generating teaching texts based on natural language processing, characterized in that: The method comprises: Acquire a target teaching knowledge base, perform entity recognition and relationship extraction on the target teaching knowledge base, and construct a target teaching knowledge graph; Obtaining a target teaching input text, performing text structuring processing on the target teaching input text using a natural language processing algorithm, and determining a target teaching text structure; A teaching content rule library is preset, and the target teaching knowledge graph is used to perform content mapping and content filtering on the target teaching text structure based on the teaching content rule library to obtain basic teaching text content; Constructing a target student profile based on historical learning behavior data, dynamically adjusting the content of the basic teaching text based on the target student profile, and determining the adapted teaching text content; A teaching template library is built, and template matching and text injection are performed on the adapted teaching text content based on the teaching template library to generate target teaching text content.

2. The method for generating teaching texts based on natural language processing according to claim 1, characterized in that: The construction of the target teaching knowledge graph includes: Determine the multi-level data cleaning procedures based on the text data application standards; Performing data cleaning on the target teaching knowledge base according to the multi-level data cleaning procedure to obtain a standard teaching knowledge base; Performing entity annotation and relationship extraction on the standard teaching knowledge base to obtain a teaching knowledge entity set and a knowledge entity relationship set; Based on the teaching knowledge entity set and the knowledge entity relationship set, graph construction and verification optimization are performed to construct the target teaching knowledge graph.

3. The method for generating teaching text based on natural language processing according to claim 2, characterized in that: The obtaining of the teaching knowledge entity set and the knowledge entity relationship set includes: Building an entity standardization system, performing entity recognition and annotation on the standard teaching knowledge base according to the entity standardization system, and obtaining an entity annotation sample set; Using the BERT network structure to perform proportional division and annotation training on the entity annotation sample set to obtain an entity annotation model; Perform entity type annotation on the standard teaching knowledge base based on the entity annotation model to obtain a teaching knowledge entity set; An entity relationship rule template is obtained, and the entity relationship rule template is used to perform relationship extraction and template matching on the teaching knowledge entity set to obtain a knowledge entity relationship set.

4. The method for generating teaching text based on natural language processing according to claim 1, characterized in that: Determining the target teaching text structure includes: Using a natural language processing algorithm to perform denoising and format conversion on the target teaching input text to obtain a standard teaching input text; Using PCFG to perform syntax analysis and semantic recognition on the standard teaching input text to obtain semantic information of the teaching input text; Preset text segmentation rules, and use the text segmentation rules to segment the standard teaching input text into blocks based on the semantic information of the teaching input text to obtain a teaching text segmentation set; Based on NER, entity extraction and relationship classification are performed on the teaching text block set, and the target teaching text structure is structured and output.

5. The method for generating teaching text based on natural language processing according to claim 1, characterized in that: The basic teaching text content is obtained, including: Using the target teaching knowledge graph to perform content mapping on the target teaching text structure to obtain matching teaching text content; Determining content specification rules, teaching logic rules, and ability adaptation rules based on the teaching content rule library; The matching teaching text content is filtered and integrated based on the content specification rules, teaching logic rules and ability adaptation rules to obtain the basic teaching text content.

6. The method for generating teaching text based on natural language processing according to claim 1, characterized in that: The target student profile is constructed based on historical learning behavior data, including: Defining a set of portrait dimensions, including knowledge mastery, learning style, and interest preferences; Performing a multi-dimensional analysis on the learning behavior history data according to the portrait dimension set to obtain a multi-dimensional portrait parameter set; A cluster analysis is performed on the multi-dimensional portrait parameter set to obtain a student group clustering result, and the target student portrait is constructed based on the student group clustering result.

7. The method for generating teaching text based on natural language processing according to claim 6, characterized in that: The step of determining the adapted teaching text content includes: Performing adjustment strategy analysis on each dimension information in the portrait dimension set to obtain a multi-dimensional teaching adjustment strategy set; Performing strategy matching on the target student portrait based on the multi-dimensional teaching adjustment strategy set to determine the target teaching adjustment strategy; The target teaching adjustment strategy is adopted to dynamically adjust the basic teaching text content to determine the adapted teaching text content.

8. The method for generating teaching text based on natural language processing according to claim 1, characterized in that: Generating target teaching text content includes: Defining template matching rules, wherein the template matching rules include teaching objective matching, subject matching, and difficulty matching; The template matching rule is used to perform template matching and text injection on the adapted teaching text content based on the teaching template library to obtain the target teaching text content.

9. The method for generating teaching text based on natural language processing according to claim 8, characterized in that: The obtaining of target teaching text content includes: Using the template matching rule to perform similarity calculation and template optimization on the adapted teaching text content based on the teaching template library to obtain a target teaching template; The adapted teaching text content is injected into the target teaching template for field replacement and dynamic generation to obtain the target teaching text content.

10. The teaching text generation system based on natural language processing is characterized by: For implementing the method for generating teaching text based on natural language processing according to any one of claims 1 to 9, the system comprises: A relationship extraction module is used to collect and obtain a target teaching knowledge base, perform entity recognition and relationship extraction on the target teaching knowledge base, and construct a target teaching knowledge graph; A structured processing module is used to obtain a target teaching input text, perform text structure processing on the target teaching input text using a natural language processing algorithm, and determine the target teaching text structure; A content filtering module is used to preset a teaching content rule library, use the target teaching knowledge graph to perform content mapping and content filtering on the target teaching text structure based on the teaching content rule library, and obtain basic teaching text content; A dynamic adjustment module is used to construct a target student portrait based on the historical learning behavior data, and dynamically adjust the content of the basic teaching text in combination with the target student portrait to determine the adapted teaching text content; The text injection module is used to build a teaching template library, perform template matching and text injection on the adapted teaching text content based on the teaching template library, and generate target teaching text content.

Images