Development method based on artificial intelligence education service platform

Abstract

The present invention relates to the field of data mining technology, and in particular to a development method based on an artificial intelligence education service platform. The method comprises the following steps: obtaining artificial intelligence education course data and performing topological structure analysis to establish a course model; analyzing forum access data to obtain user preferences and demand maps, thereby building a course recommendation model; extracting and analyzing user real-time interaction features to generate personalized course recommendations; using the course topological structure model to plan the best learning path for users; constructing a code simulation model based on virtual code editing environment data to divide the code practice environment; building an artificial intelligence education service platform, and performing strategy analysis based on the user's learning path, practice environment and real-time recommendation data, and finally pushing the education strategy to the platform for implementation. The present invention can improve the utilization rate of educational resources.

Description

Technical Field

[0001] The present invention relates to the field of data mining technology, and in particular to a development method based on an artificial intelligence education service platform. Background Art

[0002] With the rapid development of artificial intelligence (AI) technology, the global demand for professionals in the field of AI has increased dramatically. The application scenarios of AI technology cover multiple industries from smart healthcare, autonomous driving to natural language processing, which has promoted the urgent need for highly skilled AI talents. However, the current AI education system generally has the problem of disconnection between theoretical knowledge and practical application. Traditional AI courses mostly focus on the explanation of theoretical knowledge, ignoring the training of practical operations and the guidance of learning paths, which limits students' ability to effectively transform theoretical knowledge into practical skills. Existing AI education platforms often lack an integrated practical operation environment and personalized learning paths, resulting in insufficient practical links and skills improvement in the learning process. In addition, existing platforms are usually unable to provide personalized learning suggestions and dynamic adjustments for different students' learning progress and ability levels, which in turn affects learning effects and learning efficiency. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide a development method based on an artificial intelligence education service platform to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a development method based on an artificial intelligence education service platform includes the following steps:

[0005] Step S1: obtaining artificial intelligence education course data, and performing artificial intelligence course topology structure analysis based on the artificial intelligence education course data, thereby obtaining an artificial intelligence course topology structure model;

[0006] Step S2: Obtaining artificial intelligence forum access data, and performing user access preference analysis on the artificial intelligence forum access data, thereby obtaining user access preference data; visualizing a user demand map based on the user access preference data, thereby obtaining a user knowledge demand map, and constructing a course recommendation model based on the user knowledge demand map, thereby obtaining a course recommendation model;

[0007] Step S3: extracting user real-time interaction features from the artificial intelligence forum access data to obtain user real-time interaction data, and performing real-time user interaction behavior analysis on the user real-time interaction data to obtain user interaction behavior data; performing customized course content recommendations on the user interaction behavior data through a course recommendation model to obtain user real-time learning content recommendation data;

[0008] Step S4: Performing optimal learning path planning for the user based on the user's real-time learning content recommendation data through the artificial intelligence course topology structure model, thereby obtaining the user's optimal learning path planning data;

[0009] Step S5: obtaining virtual code editing environment data, and constructing a code running simulation model according to the virtual code editing environment; dividing the user's optimal learning path planning data into code practice operation environments through the code running simulation model, thereby obtaining the user's code practice environment data;

[0010] Step S6: constructing an artificial intelligence education service platform for the artificial intelligence course topology model, the course recommendation model, and the code operation simulation model, thereby obtaining an artificial intelligence education service platform;

[0011] Step S7: Perform user artificial intelligence education strategy analysis based on user code practice environment data, user optimal learning path planning data, and user real-time learning content recommendation data to obtain the user artificial intelligence education strategy, and upload it to the artificial intelligence education service platform to execute the education strategy push task.

[0012] The present invention can systematically understand and display the relationship between different course modules through the topological structure model, and help build a clear course system. The course topological structure model can visualize the context and knowledge context of the course, so that students can more clearly understand the learning path of each knowledge point. Based on the topological structure analysis, educational institutions can identify blanks or repetitions in the course content and optimize and adjust the content. The user demand map can accurately reflect the user's interest in certain knowledge points, making the course recommendation more targeted and personalized, and improving the user's learning interest and effect. The course recommendation model can recommend suitable courses based on the specific needs of the user to ensure that the learning content matches the actual needs of the user. Personalized course recommendation can save the user's time and energy, reduce the interference of irrelevant content, and thus improve learning efficiency. Real-time interactive data analysis can dynamically adjust the recommended content according to the user's learning progress and interaction situation, ensuring the timeliness and adaptability of learning resources. Customized content recommendation can provide targeted learning suggestions to help users master new knowledge and skills faster. Recommendations based on real-time user behavior can enhance the learning experience and make the learning process more smooth and personalized. Optimal learning path planning can help users learn knowledge points in a scientific order and improve the systematicness and coherence of learning. By planning the learning path, it can ensure that users will not miss important knowledge points, thereby improving learning effects. Optimizing the learning path can make the learning process more efficient and help users achieve their learning goals faster. The code running simulation model provides a virtual environment for code practice, which increases students' practice opportunities and helps the actual application and mastery of skills. Dividing the practical operation environment can provide a more suitable practice environment according to the learning path and user needs, and improve the practice effect. The virtual environment can reduce the technical obstacles that may be encountered in actual operation and provide students with a smoother practice experience. By integrating various models, the education service platform built can provide a comprehensive integrated learning solution and simplify the learning process. The integrated platform can provide a better user experience, making course recommendations, learning path planning and practical operations more coherent and seamless. The platform can centrally manage and analyze user data to provide more accurate learning suggestions and feedback. By analyzing user data and formulating personalized education strategies, users' learning needs can be met more accurately. Content pushed according to the strategy can help users get more effective guidance in the learning process, thereby improving the overall learning effect. The dynamic adjustment of the education strategy can be optimized in real time according to user feedback and learning progress to ensure the adaptability and flexibility of the learning process. The beneficial effects of these steps working together can greatly improve the quality and efficiency of AI education, help students better transform theoretical knowledge into practical skills, and meet the needs of personalized learning, thereby improving the utilization of teaching resources.

[0013] Optionally, step S1 specifically includes:

[0014] Step S11: Obtaining artificial intelligence education course data, and classifying the artificial intelligence education course data by course subject, thereby obtaining artificial intelligence basic course data and artificial intelligence application course data;

[0015] Step S12: Performing basic course hierarchical structure analysis on the artificial intelligence basic course data, thereby obtaining basic course hierarchical structure data;

[0016] Step S13: performing application course hierarchical structure analysis on the artificial intelligence application course data, thereby obtaining application course hierarchical structure data;

[0017] Step S14: performing course level progressive relationship analysis on the basic course level structure data and the applied course level structure data, thereby obtaining course level connection relationship data;

[0018] Step S15: Perform artificial intelligence course topology structure analysis based on course hierarchical connection relationship data to obtain an artificial intelligence course topology structure model.

[0019] The present invention systematically collects and classifies AI course data to ensure that all necessary course areas are covered, including basic knowledge and application skills. Through subject classification, courses can be divided into two categories: basic and applied, which helps to understand the course content and structure more clearly. The classified data makes subsequent analysis more accurate. For example, basic courses and applied courses may have different learning objectives and content depth. This classification helps to clarify the focus and learning path of different courses. Analyzing the hierarchical structure of basic courses can reveal how the course content is organized, such as which are elementary courses and which are advanced courses, to help students gradually build a knowledge system. Understanding the course hierarchy helps to optimize the course setting and ensure that the course content is arranged in a logical order in line with the learner's cognitive development. According to the hierarchical structure, personalized learning paths and recommended courses can be designed for learners of different levels. The hierarchical structure analysis of applied courses can help clarify the complexity of actual applications and distinguish between entry-level, intermediate and advanced application courses. Understanding the hierarchical structure of applied courses can help design courses that meet industry needs to match the needs of different career stages. Hierarchical analysis helps to develop a clear skill development path so that learners can gradually master application skills from basic to advanced. Analyzing the progressive relationship between course levels can identify the relationship between courses and ensure that learners can smoothly transition from basic courses to applied courses. Understanding the progressive relationship between courses can help reduce knowledge duplication and omissions and improve learning efficiency. It can help educational institutions design coherent teaching plans to ensure that the learning outcomes of each course can effectively support the learning of subsequent courses. The topological structure model provides an overall perspective of the course and helps identify the relationship and dependency between courses. Through the course topology structure, the course layout can be optimized to ensure the effective combination of different courses and learning modules and improve the efficiency of educational resources. Topological structure analysis can reveal gaps or deficiencies in the curriculum system, thereby providing a basis for curriculum improvement and new curriculum development.

[0020] Optionally, step S12 is specifically:

[0021] Step S121: Classifying basic courses according to the artificial intelligence basic course data, thereby obtaining mathematics basic course data and programming basic course data;

[0022] Step S122: extracting course text data from the programming basic course data and the mathematics basic course data respectively, thereby obtaining the programming basic course text data and the mathematics basic course text data;

[0023] Step S123: performing mathematical formula similarity evaluation on the mathematics basic course text data and the programming basic course text data, thereby obtaining mathematical formula similarity data;

[0024] Step S124: Associating the basic mathematics course data and the basic programming course data with high similarity courses according to the mathematical formula similarity data, thereby obtaining mathematics-programming course association data;

[0025] Step S125: construct a course hierarchy structure according to the mathematics-programming course association data, thereby obtaining basic course hierarchy structure data.

[0026] The present invention can classify course data into two categories: basic mathematics courses and basic programming courses. This classification helps subsequent processing and analysis, ensuring that the organization and management of data are more efficient. Classification allows course data in different fields to be processed separately, ensuring that the respective characteristics and needs can be paid attention to during the analysis process, thereby improving the accuracy of the analysis. Text information can be extracted from course data to extract key content and core points. This helps to understand the specific content and structure of the course. The extracted text data provides basic data for further analysis, especially in steps such as mathematical formula similarity evaluation, where accurate text data is essential. Evaluating the similarity of mathematical formulas can identify similar mathematical concepts and formulas in different courses. This helps to discover potential connections between courses. By understanding the similarity of formulas, learning paths and course arrangements can be optimized, so that learners can better understand mathematical applications in programming when mastering the basics of mathematics. Establishing associations between courses based on similarity data can help educators and learners understand the relationship between different courses. For example, certain mathematical concepts may also appear in programming courses. Through course association data, the interdisciplinary integration of mathematics and programming courses can be promoted, and guidance can be provided for course design, thereby improving teaching quality and effectiveness. By building a hierarchical structure, the course content can be organized into clearer levels to help learners systematically master knowledge from the basics to the advanced level. The construction of a course hierarchy helps course designers understand the logical structure of the course content, improve the course setting, and make the course system more reasonable. A structured course system can improve learners' learning experience, make the learning path clearer, and help learners accumulate and apply knowledge effectively.

[0027] Optionally, step S13 is specifically:

[0028] Step S131: performing application course clustering calculation according to the artificial intelligence application course data, thereby obtaining application course clustering data;

[0029] Step S132: extracting programming code structure features from the application course clustering data, thereby obtaining programming code structure data;

[0030] Step S133: performing programming code composition analysis according to the programming code structure data, thereby obtaining programming code composition structure data;

[0031] Step S134: dividing the programming code structure according to the programming code composition structure data, thereby obtaining basic programming code structure data and specific programming code structure data;

[0032] Step S135: performing programming code level integration according to the basic programming code structure data and the specific programming code structure data, thereby obtaining programming code level data;

[0033] Step S136: constructing an application course hierarchical structure for the application course clustering data according to the programming code hierarchical data, thereby obtaining application course hierarchical structure data.

[0034] The present invention can group courses according to similarity by clustering the applied course data, so as to help educational institutions better understand the relationship and similarity between courses. This classification can reveal the potential structure and hierarchy between courses, making the course design more systematic. The clustering results can be used to optimize the course arrangement and combination to ensure the logic and coherence of the course content. For example, similar courses are arranged together to form a complete learning path. Suitable course combinations can be recommended according to students' interests and learning progress to improve learning effects and satisfaction. By extracting the structural features of the programming code, a deep understanding of the code logic and organization can be obtained. This helps to identify the key parts and complexity of the code and improve the readability and maintainability of the code. The extraction of structural features can help discover potential problems in the code, such as duplicate code, complex logical structure, etc., so that optimization and reconstruction can be performed to improve the quality of the code. Structural feature data can be used as the basis for developing automated tools (such as code review tools and refactoring tools) to improve programming efficiency and accuracy. Parsing the components of the programming code (such as functions, variables, classes, etc.) can more clearly understand the structure and function of the code. This helps to identify the design patterns and functional modules of the code. By analyzing the composition of the code, a more comprehensive code audit can be performed to ensure the correctness and security of the code. The analysis results can be used to create teaching materials to help students and beginners better understand programming languages ​​and programming practices. Dividing the code structure into basic and specific programming code structures can help better organize and manage the code and improve the modularity of the code. The division of basic and specific structures can promote code reuse, reduce duplication of work, and improve development efficiency. Through clear structural division, code maintenance and expansion can be easier, reducing the complexity of the system. Integrating the hierarchical data of the programming code can form a clear code hierarchy structure, make the code more modular, and enhance the maintainability and scalability of the code. Through hierarchical integration, the system architecture can be better optimized so that the various levels of the system have a good collaborative relationship. For the development of complex systems, hierarchical integration can help developers better manage and control the complexity of the system. Constructing the hierarchical structure of the application course based on the hierarchical data of the programming code can form a systematic course system, making the course content more systematic and hierarchical. The establishment of a hierarchical structure can help better organize the teaching content, formulate a reasonable teaching plan and schedule, and improve the teaching effect. A clear course hierarchy structure helps to evaluate the effect of the course, find problems in the course, and make targeted improvements.

[0035] Optionally, step S14 is specifically:

[0036] Step S141: extracting course description texts from the artificial intelligence basic course data and the artificial intelligence application course data, respectively, to obtain basic course description text data and application course description text data;

[0037] Step S142: performing course relevance analysis based on the basic course description text data and the applied course description text data, thereby obtaining basic-applied course relevance data;

[0038] Step S143: integrating the application course pre-basic structure according to the foundation-application course association data, thereby obtaining the application course pre-basic structure data;

[0039] Step S144: formulating progressive relationship rules according to the course pre-basic structure data, thereby obtaining course progressive relationship data;

[0040] Step S145: Performing course-level progressive connection on the basic course level structure data and the applied course level structure data according to the course progressive relationship data, thereby obtaining course level connection relationship data.

[0041] The present invention extracts key information from the description text of the basic course of artificial intelligence, and extracts the course objectives, main contents, teaching methods, etc. Similar information is extracted from the description text of the applied course, including course objectives, application fields, practical contents, etc. Extracting and organizing the course description text can standardize the course information, making it easy to compare and analyze. Ensure that the core information of each course is clearly extracted, reduce information omissions or confusion, and improve the accuracy of subsequent analysis. Through correlation analysis, the support relationship of the basic course to the applied course can be clearly identified, helping to understand the logical connection between courses. Optimize the course setting based on the analysis results to ensure that students have mastered the necessary basic knowledge before learning the applied course. Formulate clear prerequisite course requirements for each applied course to help students understand the prerequisite knowledge required to complete the applied course. By integrating the prerequisite infrastructure, ensure the coherence and systematicness of the course content and reduce knowledge omissions. Formulating progressive relationship rules can standardize the course learning process and ensure that learning is carried out in a scientific knowledge progression order. Through hierarchical connection, a complete course structure is formed, which helps students understand the overall architecture of the entire course system. It helps to design a scientific and reasonable course system to ensure that each course has its clear position and role.

[0042] Optionally, step S2 specifically includes:

[0043] Step S21: obtaining artificial intelligence forum access data, and classifying the artificial intelligence forum access data by user access type, thereby obtaining forum video user viewing data and forum document user reading data;

[0044] Step S22: Counting the viewing time of tagged videos based on the forum video user viewing data, thereby obtaining tagged data of videos with high viewing time;

[0045] Step S23: Perform tag document search statistics on forum document user search data, thereby obtaining high-frequency document search tag data;

[0046] Step S24: performing a label intersection operation on the high-watching-time video label data and the high-frequency document reference label data to obtain user preference label data, and performing user access preference integration based on the user preference label data to obtain user access preference data;

[0047] Step S25: Visualizing the user demand map according to the user access preference data, thereby obtaining the user knowledge demand map;

[0048] Step S26: Construct a course recommendation model based on the user knowledge demand map to obtain a course recommendation model.

[0049] By acquiring forum access data, the present invention can fully understand the user's access behavior, including video viewing and document reading. Classifying user access data by type helps to understand the specific behavior patterns of different users. This provides a structured data basis for subsequent analysis. By dividing the data into video and document access, it helps to conduct targeted analysis and optimization for different types of content. Counting video viewing time can help identify users' interest in specific video content. Video tags with high viewing time show users' preferences and points of interest. Determining which video tags are highly concerned can help forum administrators or content creators optimize content strategies and improve user satisfaction. Video tags with high viewing time can reveal popular trends and hot topics, providing direction for content creation. Counting document reading frequency can reveal users' attention to specific documents or topics, thereby gaining insight into users' actual needs. Frequently consulted document tags can serve as the basis of a recommendation system to help provide more accurate content recommendations. Identifying frequently consulted document tags helps to improve and update these documents to better meet user needs. Through label intersection operations, users' preferences for videos and documents can be comprehensively analyzed to obtain a more comprehensive user interest portrait. After integrating user preference data, more personalized content recommendations can be provided to users to improve user experience. Understanding users' comprehensive preferences can optimize content strategies, promotional activities, and user interaction methods. Visualizing user needs in the form of a map can intuitively display users' knowledge needs and areas of interest, making it easier to analyze and make decisions. Knowledge demand maps help manage and organize related knowledge resources, making knowledge acquisition more efficient. Visualized maps can reveal the changing trends of user needs and help predict the future direction of demand development. Building a course recommendation model based on the user's knowledge demand map can provide users with more accurate and relevant course recommendations and improve learning outcomes. Personalized course recommendations can help users find courses that best suit their interests and needs, thereby improving learning satisfaction and efficiency. An effective course recommendation model can better utilize and allocate educational resources, so that resource investment is more in line with user needs.

[0050] Optionally, step S25 is specifically:

[0051] Step S251: extracting user access path features from artificial intelligence forum access data, thereby obtaining user access path data;

[0052] Step S252: performing access path statistics on the user access path data to obtain high-frequency access path data, and performing user access behavior analysis based on the high-frequency access path data to obtain user forum access behavior data;

[0053] Step S253: Associating the user forum access behavior data and the user access preference data with the user access behavior pattern, thereby obtaining the user access behavior pattern data;

[0054] Step S254: constructing a user demand map according to the user access behavior pattern data, thereby obtaining a user demand map set;

[0055] Step S255: performing group demand integration on the user demand map set, thereby obtaining a group demand map set;

[0056] Step S256: extracting the user knowledge demand graph from the group demand graph set according to the user demand graph set, thereby obtaining the user knowledge demand graph.

[0057] The present invention can identify the behavior patterns and points of interest of users in a forum by extracting the characteristics of the user's access path. For example, data such as the order of pages visited by users and the duration of stay can reveal the user's focus. A structured data basis is provided for subsequent data analysis and statistics. More personalized content recommendations can be provided to users to improve user experience. By counting high-frequency access paths, the content or forum parts most frequently visited by users can be identified, so as to understand the main interests and needs of users. Through statistical analysis of access paths, the user's behavior patterns, such as common access processes and high-frequency pages visited, can be understood, so as to improve the structure and content layout of the forum. Based on the analysis results, the user interface can be optimized to reduce the time users spend looking for information and improve the overall usability of the forum. Through association analysis, the relationship between the user's access behavior and his preferences can be deeply understood, so as to more accurately grasp the user's needs. The typical behavior patterns of users, such as common browsing orders or behavior habits, are revealed, and these patterns can be used to optimize content recommendations and user interaction design. More personalized services and content can be provided to users, so as to improve user satisfaction and user stickiness of the forum. By constructing a demand map, the specific needs and points of interest of users can be clarified, which helps to provide targeted content and services. Provide data support for content creators so that they can create content that meets user needs. Help forum managers optimize resource allocation and invest key resources in areas with high user demand. Integrating demand data from different users can identify broader group needs and help understand overall user preferences and demand trends. Provide data support for forums or platforms to formulate strategic plans, such as deciding which features need to be developed or improved first. Identifying users' needs for specific knowledge can guide the optimization and supplementation of knowledge content and ensure that the knowledge that users really need is provided. Improve users' learning experience and satisfaction by meeting their knowledge needs.

[0058] Optionally, step S3 specifically includes:

[0059] Step S31: extracting user real-time interaction features and user historical search features from the artificial intelligence forum access data, thereby obtaining user real-time interaction data and user historical search data;

[0060] Step S32: extracting page dwell time features from the user real-time interaction data to obtain page dwell time data, and performing high-dwell time page tag statistics based on the page dwell time data to obtain high-dwell time page tag data;

[0061] Step S33: performing tag search frequency statistics on the user's historical search data, thereby obtaining high-frequency search tag data and low-frequency search tag data;

[0062] Step S34: performing an intersection operation on the high-value dwell time page tag data and the high-frequency search tag data to obtain obvious preference tag data; performing an intersection operation on the high-value dwell time page tag data and the low-frequency search tag data to obtain potential preference tag data;

[0063] Step S35: The page dwell time data integrates the real-time interaction time behavior of the obvious preference tag data and the potential preference tag data, thereby obtaining user interaction behavior data;

[0064] Step S36: Customized course content recommendation is performed on the user interaction behavior data through the course recommendation model, thereby obtaining user real-time learning content recommendation data.

[0065] The present invention extracts user real-time interaction and historical search features, which helps to establish a comprehensive view of user behavior and make subsequent data analysis more accurate. Page dwell time feature extraction can reveal the user's interest in a specific page, help identify and classify pages with high dwell time, and thus provide more targeted content. Through tag search frequency statistics, it is possible to identify topics and points of interest that users frequently pay attention to, which helps to understand users' long-term preferences. Intersection operations can integrate different data sources and identify users' explicit and potential preference tags, thereby refining user interest models. Integrating data on explicit preference and potential preference tags helps to fully understand users' interactive behaviors and improve the accuracy of personalized recommendations. Course content recommendations based on user interactive behavior data can provide tailored learning resources and enhance users' learning experience and satisfaction.

[0066] Optionally, step S5 specifically includes:

[0067] Step S51: obtaining virtual code editing environment data, and constructing a code running simulation model according to the virtual code editing environment;

[0068] Step S52: extracting the planned course connection structure features from the artificial intelligence course topology structure model according to the user's optimal learning path planning data, thereby obtaining the planned course structure data;

[0069] Step S53: performing course structure node connection statistics on the planned course structure data, thereby obtaining densely connected course structure node data and sparsely connected course structure node data;

[0070] Step S54: dividing the densely connected course structure node data into a course code complex practice environment through a code running simulation model, thereby obtaining dense course node code practice environment data; dividing the sparsely connected course structure node data into a course code simplified practice environment through a code running simulation model, thereby obtaining sparse course node code practice environment data;

[0071] Step S55: Arrange the dense course node code practice environment data and the sparse course node code practice environment data in a learning path practice environment, thereby obtaining user code practice environment data.

[0072] The present invention can provide an accurate simulation environment by acquiring virtual code editing environment data and constructing a code running simulation model, so as to help students test and practice codes in a virtual environment, thereby improving code writing and debugging capabilities. According to the user's optimal learning path planning data, the course topology structure is extracted, and the connection relationship between courses can be effectively identified and planned, so that the learning path is more in line with the user's learning needs and difficulty level, and the learning efficiency is improved. Statistics on the connection of course structure nodes are helpful to distinguish between dense and sparse connection parts in the course. This makes it possible to design corresponding learning strategies for different learning nodes and optimize the allocation of learning resources. By dividing the environment of dense and sparse connection course structures into different complexities, a more targeted practice environment can be provided. Dense nodes will get a more complex environment to challenge students' abilities; sparse nodes will provide a simplified environment to help students gradually master basic concepts. By arranging these data into a learning path practice environment, personalized code practice environment data can be provided to users in the end, improving the adaptability and effect of learning, and making the learning process more targeted and practical.

[0073] Optionally, step S6 specifically includes:

[0074] Step S61: Obtain cloud platform resource data;

[0075] Step S62: Designing an integrated platform architecture for the artificial intelligence course topology model, the course recommendation model, and the code operation simulation model, thereby obtaining integrated platform architecture data;

[0076] Step S63: Obtain artificial intelligence forum port data, and perform docking interface design on the integrated platform architecture data according to the artificial intelligence port data, so as to obtain optimized integrated platform architecture data;

[0077] Step S64: perform platform architecture resource coupling on the integrated platform architecture data according to the cloud platform resource data, thereby obtaining an artificial intelligence education service platform, and upload it to the cloud platform to execute the platform deployment task.

[0078] The present invention obtains cloud platform resource data to help understand available computing power, storage capacity, network bandwidth and other information, ensuring that the platform design can make full use of these resources. Optimize resource allocation and use according to actual resource conditions to avoid waste or shortage of resources, thereby improving the operating efficiency and stability of the platform. Integrating different models (course topology, recommendation, code simulation) can create a comprehensive education platform that integrates various functions and services. When designing the integrated platform architecture, it is possible to ensure that the data flow and functional collaboration between these models are consistent, thereby improving the compatibility and overall performance of the platform. Through the docking interface design, the seamless connection between the integrated platform and external systems such as the artificial intelligence forum can be achieved, and the function and data exchange capabilities of the platform can be enhanced. The optimized interface design can promote data sharing and interaction between the platform and external systems, and improve the interoperability of the system. Ensure that the platform architecture can make full use of the resources provided by the cloud platform to avoid performance problems caused by resource mismatch. Uploading the optimized integrated platform architecture to the cloud platform and deploying it can quickly build and run the artificial intelligence education service platform to support the actual operation and expansion of educational applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0080] Figure 1 A schematic diagram of the steps of the development method of the artificial intelligence education service platform of the present invention;

[0081] Figure 2 Detailed step flow diagram of step S1 in the present invention;

[0082] Figure 3 Detailed step flow diagram of step S2 in the present invention;

[0083] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0084] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0085] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0086] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0087] To achieve this, please refer to Figures 1 to 3 The present invention provides a development method based on an artificial intelligence education service platform, the method comprising the following steps:

[0088] Step S1: obtaining artificial intelligence education course data, and performing artificial intelligence course topology structure analysis based on the artificial intelligence education course data, thereby obtaining an artificial intelligence course topology structure model;

[0089] In this embodiment, course data about artificial intelligence (AI) are collected from major online education platforms (such as Coursera, edX, Udacity) and educational institutions. These data include course name, course content, course objectives, lecturers, course difficulty level, learning path, etc. Use data cleaning technology to process the acquired data to ensure the accuracy and consistency of the data. Remove duplicate data and fill in missing values. Use graph theory methods (such as graph models of nodes and edges) to perform topological analysis on courses. Nodes represent courses, and edges represent prerequisite relationships or knowledge point associations between courses. For example, deep learning courses depend on machine learning courses. Construct a course topology structure model through algorithms (such as Kruskal algorithm or Prim algorithm). The model shows the dependencies and learning paths between courses. For example, the course "Basics of Neural Networks" requires the completion of "Basics of Linear Algebra" first.

[0090] Step S2: Obtaining artificial intelligence forum access data, and performing user access preference analysis on the artificial intelligence forum access data, thereby obtaining user access preference data; visualizing a user demand map based on the user access preference data, thereby obtaining a user knowledge demand map, and constructing a course recommendation model based on the user knowledge demand map, thereby obtaining a course recommendation model;

[0091] In this embodiment, user access data is collected from artificial intelligence forums (such as Reddit's AI section and Stack Overflow's AI tag). This includes visit frequency, click-through rate, comments, likes, etc. Data mining techniques (such as cluster analysis and association rule mining) are used to analyze users' access preferences. Identify users' interests in specific topics (such as natural language processing or computer vision). Based on user preference data, a user knowledge demand map is constructed. The map shows the user's knowledge needs and interests in the field of artificial intelligence, such as the need for more courses on "reinforcement learning". A course recommendation model is constructed using machine learning algorithms (such as collaborative filtering and content recommendation algorithms). The model recommends relevant courses based on the user's interests and needs. For example, when the user is interested in "deep learning", the "convolutional neural network" course is recommended.

[0092] Step S3: extracting user real-time interaction features from the artificial intelligence forum access data to obtain user real-time interaction data, and performing real-time user interaction behavior analysis on the user real-time interaction data to obtain user interaction behavior data; performing customized course content recommendations on the user interaction behavior data through a course recommendation model to obtain user real-time learning content recommendation data;

[0093] In this embodiment, the real-time interaction data of users is obtained from the forum and learning platform, including browsing time, click path, interaction frequency, etc. The interactive behavior of users is analyzed using time series analysis technology (such as HMM model, LSTM network). The learning habits and interest changes of users are identified. The course recommendation model is used to analyze the real-time interaction data of users and generate customized learning content recommendations. For example, if the user frequently consults "deep learning", the system recommends the related "advanced deep learning application" course.

[0094] Step S4: Performing optimal learning path planning for the user based on the user's real-time learning content recommendation data through the artificial intelligence course topology structure model, thereby obtaining the user's optimal learning path planning data;

[0095] In this embodiment, the artificial intelligence course topology model is used to analyze the user's learning progress and needs. The shortest path algorithm (such as the Dijkstra algorithm) is used to plan the user's learning path to ensure the optimal learning effect. For example, the optimal learning route from "basic machine learning" to "advanced algorithms" is provided for a learner. The user's optimal learning path planning data is generated to provide personalized learning suggestions and path maps.

[0096] Step S5: obtaining virtual code editing environment data, and constructing a code running simulation model according to the virtual code editing environment; dividing the user's optimal learning path planning data into code practice operation environments through the code running simulation model, thereby obtaining the user's code practice environment data;

[0097] In this embodiment, the user's programming operation data is collected from a virtual code editing environment (such as Jupyter Notebook, Google Colab), including code submission records, running results, etc. A code running simulation model is constructed using simulation technology to simulate the impact of different programming environments on code execution. For example, the code running effect under different computing resources is simulated. According to the user's optimal learning path planning data, the code practice environment is divided into different areas (such as basic practice area and advanced challenge area). For example, a "Python Basics" practice environment is provided for beginners, and a "Deep Learning Model Optimization" practice environment is provided for advanced users.

[0098] Step S6: constructing an artificial intelligence education service platform for the artificial intelligence course topology model, the course recommendation model, and the code operation simulation model, thereby obtaining an artificial intelligence education service platform;

[0099] In this embodiment, the artificial intelligence education service platform architecture is designed to integrate the course topology model, course recommendation model and code simulation model. Web development technologies (such as React and Django) are used to develop the front-end and back-end of the platform. The front-end displays course recommendations, learning paths and code practice environments, and the back-end handles data storage and model calculations. The various models and modules are integrated into the platform to ensure smooth data flow and functions. For example, when users select courses on the platform, corresponding suggestions are provided based on the course recommendation model.

[0100] Step S7: Perform user artificial intelligence education strategy analysis based on user code practice environment data, user optimal learning path planning data, and user real-time learning content recommendation data to obtain the user artificial intelligence education strategy, and upload it to the artificial intelligence education service platform to execute the education strategy push task.

[0101] In this embodiment, artificial intelligence education strategy analysis is performed based on user code practice environment data, learning path planning data, and learning content recommendation data. Determine the user's learning strategy and goals. The user's code practice records (such as completed programming tasks, solved problems, and programming language usage) and practice environment, learning path planning data (such as learned modules, learning time allocation, etc.) and real-time learning content recommendation data (such as currently recommended learning resources and user interaction data) can be integrated into a unified data platform. Use Python's Pandas and Scikit-learn libraries for data cleaning and feature extraction. Use K-means clustering to analyze the user's learning pattern, and use collaborative filtering algorithms (such as ALS algorithms based on matrix decomposition) for personalized recommendations. Predict the user's learning needs and interests by building a deep learning model (such as a sequence model based on LSTM). According to the analysis results, the recommendation system generates learning content suitable for the user, and the practice tasks for each link are formulated through the user's code practice environment data. For example, if the user performs well in the field of data analysis, the recommendation system will push more advanced data science courses and related programming challenges. Use reinforcement learning technology to optimize the user's learning path. For example, the Q-learning algorithm is used to dynamically adjust the recommended learning modules based on the user's learning feedback to ensure that the learning difficulty and time allocation of each module are reasonable, thereby improving learning efficiency. Finally, a comprehensive user artificial intelligence education strategy is generated and pushed to the user through the platform. For example, based on the strategy, the user is recommended to participate in the "Machine Learning Project Practice" course. The strategy push task is automatically executed on the education service platform to update the user's learning suggestions and resource recommendations in real time.

[0102] Optionally, step S1 specifically includes:

[0103] Step S11: Obtaining artificial intelligence education course data, and classifying the artificial intelligence education course data by course subject, thereby obtaining artificial intelligence basic course data and artificial intelligence application course data;

[0104] In this embodiment, detailed information of artificial intelligence courses is obtained from educational platforms (such as Coursera, Udacity, edX) and academic databases (such as IEEEXplore, Google Scholar). These data include course titles, descriptions, chapter contents, target skills, etc. Text analysis of course descriptions is performed using natural language processing (NLP) technology. Topic modeling is performed using spaCy or NLTK, and courses are subject-classified using the Latent Dirichlet Allocation (LDA) algorithm to identify whether the courses belong to the basic (such as machine learning basics, linear algebra) or application (such as deep learning in image processing, reinforcement learning) fields. By extracting keywords from the course description, the classification results will include "artificial intelligence basic course data" (such as "Introduction to Machine Learning", "Basics of Data Science") and "artificial intelligence application course data" (such as "Computer Vision Applications", "Natural Language Processing Applications").

[0105] Step S12: Performing basic course hierarchical structure analysis on the artificial intelligence basic course data, thereby obtaining basic course hierarchical structure data;

[0106] In this embodiment, the courses in the basic course data are analyzed in a hierarchical structure to determine the order of the course modules. For example, the course "Basics of Data Science" includes sub-modules such as "Basics of Statistics", "Data Preprocessing" and "Data Visualization". The course hierarchy graph is constructed using the networkx library in Python. Each node represents a course module, and the edge represents the dependency relationship between modules. In the "Basics of Machine Learning" course, the hierarchical structure data includes the following relationships: "Linear Regression" is the basic module, "Logistic Regression" depends on "Linear Regression", and "Support Vector Machine" is learned after "Logistic Regression".

[0107] Step S13: performing application course hierarchical structure analysis on the artificial intelligence application course data, thereby obtaining application course hierarchical structure data;

[0108] In this embodiment, the hierarchical structure analysis of the artificial intelligence application course data is performed to determine the application level of the course content. For example, in the "Deep Learning Application" course, it can be divided into "Convolutional Neural Network Application", "Recurrent Neural Network Application" and "Generative Adversarial Network Application". The course hierarchy diagram is generated using the Graphviz tool. Each node represents a specific application course or application field, and the edges between nodes represent dependencies. In "Computer Vision Application", the hierarchical structure data includes "Object Detection" as the basic level, "Image Segmentation" depends on "Object Detection", and "Face Recognition" as an advanced application.

[0109] Step S14: performing course level progressive relationship analysis on the basic course level structure data and the applied course level structure data, thereby obtaining course level connection relationship data;

[0110] In this embodiment, the hierarchical structure data of basic courses and applied courses are integrated, and algorithms (such as the Shortest Path algorithm in Graph Theory) are used to analyze the progressive relationship between courses. For example, how to advance from "Basic Machine Learning" to "Deep Learning Applications". In the networkx library, a mixed graph is created to connect the hierarchical structure data of basic courses and applied courses to analyze the dependencies and learning paths between different courses. For example, "Linear Regression" is connected to the learning path of the "Convolutional Neural Network" application course as a basic course, and "Data Preprocessing" is used as the connection basis to form a complete course progressive path.

[0111] Step S15: Perform artificial intelligence course topology structure analysis based on course hierarchical connection relationship data to obtain an artificial intelligence course topology structure model.

[0112] In this embodiment, the topological_sort function in the networkx library is used to topologically sort the course hierarchical connection relationship to generate a topological structure model of the artificial intelligence course. The topological structure will show the learning path from basic to advanced courses. The topological structure model is visualized using Matplotlib or Plotly. Each node represents a course module, and the connection between the nodes shows the order of courses. The generated topological structure model may display, such as "Basic Machine Learning" → "Deep Learning Basics" → "Deep Learning Applications", to help students clarify the learning path and systematically master the knowledge points.

[0113] Optionally, step S12 is specifically:

[0114] Step S121: Classifying basic courses according to the artificial intelligence basic course data, thereby obtaining mathematics basic course data and programming basic course data;

[0115] In this embodiment, all course information, including course name, course description, course content, etc., is extracted from the artificial intelligence basic course data. Natural language processing (NLP) technology is used to segment and tag the course description, and keywords related to mathematics and programming are extracted from it. A machine learning classification algorithm (such as support vector machine SVM or naive Bayes classifier) ​​is applied to train the extracted features. The labeled mathematics and programming courses are used as training data. The courses are classified into the categories of basic mathematics courses or basic programming courses, and two data sets are output: basic mathematics course data and basic programming course data.

[0116] Step S122: extracting course text data from the programming basic course data and the mathematics basic course data respectively, thereby obtaining the programming basic course text data and the mathematics basic course text data;

[0117] In this embodiment, detailed text data of each course, including course content and syllabus, is extracted from the classification results. The extracted text data is preprocessed, including removing stop words, punctuation cleaning, and stem extraction. Entity recognition and topic modeling (such as Latent Dirichlet Allocation, LDA) in NLP technology are used to extract key content in the course text. Two text datasets are generated: programming basic course text data and mathematics basic course text data.

[0118] Step S123: performing mathematical formula similarity evaluation on the mathematics basic course text data and the programming basic course text data, thereby obtaining mathematical formula similarity data;

[0119] In this embodiment, mathematical formulas are extracted from text data using optical character recognition (OCR) technology and formula parsing tools (such as MathJax). The extracted formulas are converted into standardized representations of mathematical expressions (such as LaTeX format). A formula similarity algorithm (such as cosine similarity, Jaccard similarity, or edit distance) is applied to calculate the similarity between formulas. A similarity threshold is used to determine which formulas have a high similarity, and mathematical formula similarity data is output, including a similarity score for each pair of formulas.

[0120] Step S124: Associating the basic mathematics course data and the basic programming course data with high similarity courses according to the mathematical formula similarity data, thereby obtaining mathematics-programming course association data;

[0121] In this embodiment, the similarity information in the basic mathematics course data and the basic programming course data is combined to construct a course similarity matrix. An association rule learning algorithm (such as the Apriori algorithm) is used to identify pairs of courses with high similarity in the basic mathematics course and the basic programming course. Mathematics-programming course association data is generated, which lists high-similarity course pairs and their similarity scores. For example, if the formula of a mathematics course has a high similarity with the formula of a programming course, these courses are marked as associated courses.

[0122] Step S125: construct a course hierarchy structure according to the mathematics-programming course association data, thereby obtaining basic course hierarchy structure data.

[0123] In this embodiment, a course hierarchy is designed based on the mathematics-programming course association data. A hierarchical clustering algorithm (such as K-means or a hierarchical clustering algorithm) is used to group related courses. A course hierarchy diagram is constructed to identify the relationship between courses at different levels. For example, basic mathematics courses and introductory programming courses are used as the basic layer, and advanced mathematics and programming courses are used as advanced layers. Generate basic course hierarchical structure data, including course grouping, hierarchical relationships, and detailed information on each layer of courses. For example, a tree structure of basic mathematics courses and basic programming courses is constructed according to similarity and degree of association.

[0124] Optionally, step S13 is specifically:

[0125] Step S131: performing application course clustering calculation according to the artificial intelligence application course data, thereby obtaining application course clustering data;

[0126] In this embodiment, the artificial intelligence application course data includes features such as course title, course content, difficulty level, and study hours. First, the course data is converted into a feature vector through a preprocessing step. The features include the text representation of the course name, the topic distribution of the content (extracted by natural language processing technology), etc. Next, the K-means clustering algorithm is used to perform clustering calculations on these feature vectors to divide the courses into several categories. For example, K=5 can be selected to divide the courses into five categories: basic programming, data structure, algorithm design, web development, and machine learning. The clustering results will be used as application course clustering data for subsequent steps.

[0127] Step S132: extracting programming code structure features from the application course clustering data, thereby obtaining programming code structure data;

[0128] In this embodiment, the code of a programming language (such as Python) in the application course clustering data is taken as an example to extract the code structure features. Assuming there is a set of Python function codes, the code is converted into a tree structure using the abstract syntax tree (AST) library. By parsing the AST, the structural features of the code can be extracted, including function definitions, class definitions, statement types, control structures (such as if statements, loops), etc. The results of feature extraction include the frequency of occurrence of each structure, the nesting depth, the complexity of the code block, etc. Ultimately, these feature data form programming code structure data.

[0129] Step S133: performing programming code composition analysis according to the programming code structure data, thereby obtaining programming code composition structure data;

[0130] In this embodiment, the obtained programming code structure data is analyzed in detail. Taking the extracted Python function as an example, the parsing process includes decomposing the code block into components such as variable declaration, function call, control flow statement, etc. The code is further disassembled using a code analysis tool (such as Python's ast module) to identify the function and logical structure of each code segment. For example, the definition part, parameter list, local variables, and each statement block of a Python function are extracted separately, and their types and functions are marked. The result is detailed programming code composition structure data, which describes the various components of the code and their interrelationships.

[0131] Step S134: dividing the programming code structure according to the programming code composition structure data, thereby obtaining basic programming code structure data and specific programming code structure data;

[0132] In this embodiment, based on the obtained programming code composition structure data, the code is divided into basic and specific structures. First, the definition of "basic programming code structure" includes basic grammatical elements such as variable declaration, basic control structure (such as if-else, for loop), etc. Specific structures include complex design patterns, API calls of specific libraries, etc. Taking the code of a Python class as an example, the basic parts of the class (such as attribute definition, method definition) are classified as basic structures, and complex design patterns (such as factory mode, singleton mode) are classified as specific structures. Finally, basic programming code structure data and specific programming code structure data are formed.

[0133] Step S135: performing programming code level integration according to the basic programming code structure data and the specific programming code structure data, thereby obtaining programming code level data;

[0134] In this embodiment, the obtained basic and specific programming code structure data are hierarchically integrated. Assume that there is a set of Python project codes, and each code file is divided into basic and specific structures. During integration, the code blocks are first organized hierarchically according to the file structure, and then the basic and specific structures are integrated according to the functional modules (such as data processing modules, user interface modules). A hierarchical data structure (such as a tree structure) is used to represent the hierarchical relationship of each module. For example, a project that includes data processing and network request functions can use the data processing function as a level of the basic structure and the network request function as a level of the specific structure. The integration result is complete programming code hierarchical data.

[0135] Step S136: constructing an application course hierarchical structure for the application course clustering data according to the programming code hierarchical data, thereby obtaining application course hierarchical structure data.

[0136] In this embodiment, a hierarchical structure is constructed for the application course clustering data based on the programming code hierarchical data. Ensure that the organization of the programming code hierarchical data has been completed and that the application course data has been clustered. Combine these hierarchical data with the course clustering data, and construct a hierarchical structure based on the complexity of the course category and the target application field. For example, the "Basic Programming" course is matched with the basic programming code structure, and the "Advanced Algorithm" course is matched with a specific programming code structure. The constructed hierarchical structure can be a hierarchical relationship diagram of the course, showing the position and relevance of different courses on the learning path. Ultimately, the application course hierarchical structure data is formed, reflecting the association and hierarchical relationship between different courses and the programming code structure.

[0137] Optionally, step S14 is specifically:

[0138] Step S141: extracting course description texts from the artificial intelligence basic course data and the artificial intelligence application course data, respectively, to obtain basic course description text data and application course description text data;

[0139] In this embodiment, the description text of the basic course and the applied course of artificial intelligence is extracted from the syllabus, course introduction and teaching materials. The text extraction algorithm (e.g., Named Entity Recognition, NER) in natural language processing technology (NLP) is used to extract relevant course content from the document. Specifically, Python's spaCy library is used for text processing, and the key terms and concepts in the course description are identified and extracted through the trained NER model. These extracted texts will be used for subsequent analysis. After extraction, the description text of the basic course can be stored as "basic course description data", and the description text of the applied course can be stored as "application course description data".

[0140] Step S142: performing course relevance analysis based on the basic course description text data and the applied course description text data, thereby obtaining basic-applied course relevance data;

[0141] In this embodiment, based on the extracted description text data, text similarity calculation technology (such as cosine similarity, Jaccard similarity) is used to analyze the correlation between basic courses and applied courses. TfidfVectorizer in the sklearn library is used to convert the course description text into a vector representation, and then the similarity between these vectors is calculated. According to the calculated similarity, a correlation data table between basic courses and applied courses is generated, and the correlation scores between each pair of courses are recorded and sorted. This data table is used to determine which applied courses depend on which basic courses, thereby forming "basic-applied course correlation data".

[0142] Step S143: integrating the application course pre-basic structure according to the foundation-application course association data, thereby obtaining the application course pre-basic structure data;

[0143] In this embodiment, the basic course prerequisites required for each applied course are determined based on the "basic-applied course association data". A graph database (such as Neo4j) is used to model the relationship between basic courses and applied courses as a graph structure, where nodes represent courses and edges represent prerequisite relationships. The combination of basic courses required for each applied course is determined through a graph algorithm (such as the Dijkstra algorithm). The integration results form the "application course prerequisite infrastructure data", which includes the basic courses corresponding to each applied course and their weight information.

[0144] Step S144: formulating progressive relationship rules according to the course pre-basic structure data, thereby obtaining course progressive relationship data;

[0145] In this embodiment, based on the "application course prerequisite infrastructure data", progressive relationship rules are formulated. The criteria for progressive relationships are defined, such as how the degree of completion of basic courses (through tests, homework, etc.) affects the learning of application courses. A rule engine (such as Drools) is used to establish progressive rules. For example, it is defined that "if basic courses A and B are completed, application course X can be carried out". "Course progressive relationship data" is generated through these rules, which describes the progressive relationship and learning path between different courses.

[0146] Step S145: Performing course-level progressive connection on the basic course level structure data and the applied course level structure data according to the course progressive relationship data, thereby obtaining course level connection relationship data.

[0147] In this embodiment, according to the "course progressive relationship data", the basic courses and applied courses are integrated into the course hierarchy structure. Use data visualization tools (such as Graphviz) to draw a course hierarchy diagram to represent the progressive connection relationship between courses. Design course hierarchy data, including connection information of the basic course hierarchy and the applied course hierarchy. Through a graphical hierarchical structure diagram, show how each basic course leads to the relevant applied course, thereby forming "course hierarchy connection relationship data". This visual structure helps teaching staff understand the course schedule and its progressive relationship.

[0148] Optionally, step S2 specifically includes:

[0149] Step S21: obtaining artificial intelligence forum access data, and classifying the artificial intelligence forum access data by user access type, thereby obtaining forum video user viewing data and forum document user reading data;

[0150] In this embodiment, the API interface is used to extract user access log data from the artificial intelligence forum. It includes user ID, access time, access content type (video or document), video viewing time, document review time and other information. By analyzing the content type field in the access data, the data is divided into two categories: video user viewing data records the video ID, viewing time, user ID, etc. watched by the user; document user review data records the document ID, review time, user ID, etc. reviewed by the user. Python's Pandas library can be used to process the data, and regular expressions can be used to extract relevant fields and store them in the database for further analysis.

[0151] Step S22: Counting the viewing time of tagged videos based on the forum video user viewing data, thereby obtaining tagged data of videos with high viewing time;

[0152] In this embodiment, the video ID and viewing time are extracted from the stored user viewing data, and the average viewing time of each video is calculated. A threshold is set, for example, a video with a viewing time of more than 10 minutes is marked as a high viewing time video. These video IDs and their labels are stored as high viewing time video label data. SQL query statistics viewing time data can be used, combined with Python's data processing library to calculate the average value, and the result is output to a CSV file or database.

[0153] Step S23: Perform tag document search statistics on forum document user search data, thereby obtaining high-frequency document search tag data;

[0154] In this embodiment, the document ID and the number of times the document is viewed are extracted from the user document viewing data, and the frequency of each document is calculated. A threshold is set, for example, a document with more than 50 views is marked as a high-frequency document. These document IDs and their labels are stored as high-frequency document viewing label data. The aggregation function of the SQL database can be used for statistics, and the Python script can be used to process the data, filter out high-frequency documents, and generate a report.

[0155] Step S24: performing a label intersection operation on the high-watching-time video label data and the high-frequency document reference label data to obtain user preference label data, and performing user access preference integration based on the user preference label data to obtain user access preference data;

[0156] In this embodiment, a set operation is used to calculate the intersection of high-watching-time video tags and high-frequency document reference tags to obtain a common tag. According to the common tag, user access data is aggregated to analyze user preferences. For example, users who follow both high-watching-time videos and high-frequency documents are found. The intersection of tags can be calculated using Python's set operation, data aggregation can be performed using the Pandas library, and a visual chart can be generated using Matplotlib to display user preference data.

[0157] Step S25: Visualizing the user demand map according to the user access preference data, thereby obtaining the user knowledge demand map;

[0158] In this embodiment, based on user preference data, a graph construction tool such as Gephi or Graphviz is used to draw a user demand graph. Nodes represent tags or topics, and edges represent the user's attention relationship to these tags. The knowledge demand graph is exported as an image file or an interactive web page to display the user's knowledge demand pattern. The graph structure can be defined using the dot language of Graphviz, and Gephi can be used to visualize and analyze the graph to generate an easy-to-understand knowledge demand graph.

[0159] Step S26: Construct a course recommendation model based on the user knowledge demand map to obtain a course recommendation model.

[0160] In this embodiment, the user knowledge demand map is used to analyze the user's knowledge needs, and a course recommendation model is built based on these needs. The model can be based on collaborative filtering, content recommendation, or a hybrid recommendation algorithm. Select a suitable recommendation algorithm, such as a content-based recommendation algorithm that analyzes the degree of match between the course content and user needs. A machine learning library such as Scikit-learn or TensorFlow can be used to build and train a recommendation model, which is combined with user knowledge demand map data for training and evaluation, and finally a course recommendation system is generated.

[0161] Optionally, step S25 is specifically:

[0162] Step S251: extracting user access path features from artificial intelligence forum access data, thereby obtaining user access path data;

[0163] In this embodiment, user access records are extracted from the server log of the artificial intelligence forum. These records contain user ID, access timestamp, access page URL, and click sequence. The collected data is cleaned to remove duplicate access records and incomplete data. The page URL is standardized and the click sequence is converted into an ordered path list. Feature engineering technology is used to convert each access path into a feature vector. The URL is encoded using TF-IDF (term frequency-inverse document frequency), and the time series features in the path are extracted using sliding window technology. Finally, principal component analysis (PCA) is used for dimensionality reduction to simplify path features. For example, a user visited the "Artificial Intelligence Basics" page of the forum, then visited the "Deep Learning" page, and then the "Neural Network" page. The extracted features include access time interval, page type (basic, advanced), and jump frequency between pages.

[0164] Step S252: performing access path statistics on the user access path data to obtain high-frequency access path data, and performing user access behavior analysis based on the high-frequency access path data to obtain user forum access behavior data;

[0165] In this embodiment, a frequent pattern mining algorithm (such as FP-Growth) is used to count the extracted access paths and identify the access paths with higher frequency. For example, the number of times the path "Artificial Intelligence Basics->Deep Learning->Neural Network" appears is counted. Combined with the statistical results, cluster analysis (such as K-means algorithm) is applied to group user access behaviors. By analyzing high-frequency access paths, a user access behavior model is constructed to identify common behavior patterns and access habits. For example, it is found that 70% of users usually continue to visit the "Deep Learning" page after visiting the "Artificial Intelligence Basics" page. The analysis results show that most users stay on the "Deep Learning" page for an average of more than 5 minutes when viewing the page.

[0166] Step S253: Associating the user forum access behavior data and the user access preference data with the user access behavior pattern, thereby obtaining the user access behavior pattern data;

[0167] In this embodiment, an association rule mining algorithm (such as the Apriori algorithm) is used to extract association patterns from user access behavior data. For example, a pattern is found that "after visiting the 'machine learning' page, users often visit the 'data processing' page." Combined with the user's access preference data (such as user interest tags), the identified behavior patterns are associated with user characteristics to form more accurate user behavior pattern data. Through association rule mining, it can be found that if a user visits the "machine learning" page, there is a 60% probability that they will continue to visit the "data processing" page. These patterns are associated with user interest tags (such as interest in data science) to verify the validity of the pattern.

[0168] Step S254: constructing a user demand map according to the user access behavior pattern data, thereby obtaining a user demand map set;

[0169] In this embodiment, a user demand graph is constructed based on user access behavior pattern data. The nodes in the graph represent user needs, and the edges represent the relationship between needs. For example, there is a direct correlation between the "deep learning" demand and the "neural network" demand. Integrate user behavior patterns, interest tags, and access path data into the demand graph to ensure that the graph accurately reflects user needs and preferences. For example, create a graph in which nodes include "Deep Learning Tutorial", "AI Application Cases", and "Programming Practice". These nodes are connected by edges, indicating the user's jump relationship and degree of association between these contents.

[0170] Step S255: performing group demand integration on the user demand map set, thereby obtaining a group demand map set;

[0171] In this embodiment, data aggregation technology (such as cluster analysis) is used to integrate group needs on the user demand map. Similar needs are classified together to identify the main group demand patterns. According to the identified group demand patterns, the demand map is optimized so that it can better reflect the overall demand trend of the group. For example, through cluster analysis, it is found that there are three main demand groups: one is users who have a strong interest in "deep learning", the second is users who are interested in "machine learning applications", and the third is users who have a demand for "AI basic knowledge". These needs are integrated into the group demand map to form three main demand branches.

[0172] Step S256: extracting the user knowledge demand graph from the group demand graph set according to the user demand graph set, thereby obtaining the user knowledge demand graph.

[0173] In this embodiment, a user knowledge demand map is extracted based on a group demand map set. The nodes in the knowledge demand map represent specific knowledge points, and the edges represent the relationships between knowledge points. For example, there is a direct relationship between the "convolutional neural network" knowledge point and the "image recognition" knowledge point. The extracted knowledge demand map is analyzed to identify the association between the knowledge points, and the map is adjusted and optimized according to the actual needs of the user. For example, a knowledge demand map is created that includes nodes such as "convolutional neural network", "image recognition technology" and "deep learning algorithm". The edges between the nodes represent the association between these knowledge points, for example, "convolutional neural network" is a key component of "image recognition technology".

[0174] Optionally, step S3 specifically includes:

[0175] Step S31: extracting user real-time interaction features and user historical search features from the artificial intelligence forum access data, thereby obtaining user real-time interaction data and user historical search data;

[0176] In this embodiment, in the artificial intelligence forum, it is first necessary to collect users' real-time interaction data and historical search data. Users' real-time interaction data includes their activity records in the forum, such as post browsing, commenting, liking and other behaviors. Historical search data records users' search keywords and their frequency. Technically, real-time interaction data can be captured in real time on each page through embedded analysis scripts (such as JavaScript) and stored in server logs; while historical search data can be extracted through the search history in the user account record system. Data processing frameworks such as ApacheKafka can be used for data stream processing, and these data can be stored and managed through MySQL databases.

[0177] Step S32: extracting page dwell time features from the user real-time interaction data to obtain page dwell time data, and performing high-dwell time page tag statistics based on the page dwell time data to obtain high-dwell time page tag data;

[0178] In this embodiment, in each page visited by the user, the user's page dwell time, that is, the length of time the user stays on each page, is captured by the front-end JavaScript code. The dwell time data can be recorded as a timestamp and sent to the server through the API interface. In terms of data storage and processing, a Redis cache system can be used to record and count the dwell time data in real time. Subsequently, based on the statistical analysis of the dwell time data, page tags with high dwell time are identified. For example, a page with a dwell time of more than 60 seconds is marked as a high-amount dwell time page, and data processing and statistics are performed through Python's data analysis library such as Pandas to obtain high-amount dwell time page tag data.

[0179] Step S33: performing tag search frequency statistics on the user's historical search data, thereby obtaining high-frequency search tag data and low-frequency search tag data;

[0180] In this embodiment, the searched tags are extracted from the user's historical search data, and the search frequency of each tag is calculated. Text analysis technology (such as NLP toolkit) can be used to parse and classify the search keywords. The number of occurrences of each tag is counted through SQL query or MapReduce operation in NoSQL database to determine high-frequency and low-frequency search tags. For example, the search data is grouped and counted through the aggregation framework of MongoDB to obtain high-frequency and low-frequency search tag data.

[0181] Step S34: performing an intersection operation on the high-value dwell time page tag data and the high-frequency search tag data to obtain obvious preference tag data; performing an intersection operation on the high-value dwell time page tag data and the low-frequency search tag data to obtain potential preference tag data;

[0182] In this embodiment, the obtained high-dwelling time page tag data and the obtained high-frequency search tag data are subjected to an intersection operation. This intersection calculation can be implemented using the set operation function in Python. Similarly, the high-dwelling time page tag data and the low-frequency search tag data are subjected to an intersection operation. During implementation, a set intersection algorithm, such as the set.intersection() method, is used to obtain explicit preference tag data and potential preference tag data. For example, an explicit preference tag appears in both a high-dwelling time page and a high-frequency search, while a potential preference tag appears in a high-dwelling time page but in a low-frequency search.

[0183] Step S35: The page dwell time data integrates the real-time interaction time behavior of the obvious preference tag data and the potential preference tag data, thereby obtaining user interaction behavior data;

[0184] In this embodiment, the user's page dwell time data is integrated with the obtained explicit preference and potential preference tag data. By analyzing the user's interactive time behavior on the high-dwelling time page, a time series analysis method (such as sliding window technology) can be used to obtain the user's real-time interactive behavior pattern. Data integration can be performed on a big data platform such as Apache Spark to achieve real-time processing and analysis and generate user interactive behavior data. These data can be used to further understand the user's preference strength for different tags.

[0185] Step S36: Customized course content recommendation is performed on the user interaction behavior data through the course recommendation model, thereby obtaining user real-time learning content recommendation data.

[0186] In this embodiment, a course recommendation model is constructed based on the integrated user interaction behavior data using machine learning recommendation algorithms (such as collaborative filtering, matrix decomposition, etc.). The recommendation model can be implemented through the Scikit-learn or TensorFlow library in Python to generate customized course recommendations based on the user's explicit preferences and potential preference tags. In specific implementation, the model can input the user's interaction data and output a list of recommended courses. For example, if the user has a high dwell time and frequent searches on the "deep learning" tag, the system will give priority to recommending related deep learning courses.

[0187] Optionally, step S5 specifically includes:

[0188] Step S51: obtaining virtual code editing environment data, and constructing a code running simulation model according to the virtual code editing environment;

[0189] In this embodiment, the API interface of the virtual code editing environment (such as Visual Studio Code, PyCharm or Jupyter Notebook) is used to collect data such as configuration information of the code editor, plug-in installation status, project file structure, code library content, etc. These data are used to build a code running simulation model, which can simulate the execution of the code in the actual development environment. The source code files, dependency package information and editor configuration information of the project can be extracted through the API and converted into an operational data model. This model should include environment variables during code execution, performance indicators at runtime (such as CPU and memory usage), and possible error logs.

[0190] Step S52: extracting the planned course connection structure features from the artificial intelligence course topology structure model according to the user's optimal learning path planning data, thereby obtaining the planned course structure data;

[0191] In this embodiment, the topological structure of the artificial intelligence course is analyzed using user learning path data (such as the user's learning progress and feedback data on the online education platform). The course topological structure model includes course nodes (each course) and connections between courses (learning order). The characteristics of the course connection structure are extracted through graph theory algorithms (such as the shortest path algorithm or the network flow algorithm). These characteristics include dependencies between courses, prerequisite course requirements, logical flow of course content, etc. For example, using graph structure extraction technology, analyze user learning records to determine which courses are key nodes in the learning path, thereby forming an optimal course structure data set.

[0192] Step S53: performing course structure node connection statistics on the planned course structure data, thereby obtaining densely connected course structure node data and sparsely connected course structure node data;

[0193] In this embodiment, the extracted course structure data is statistically analyzed to calculate the connection density between the course structure nodes. Specifically, the number of connections of each course node (i.e., the number of direct connections between it and other course nodes) is counted, and the nodes are classified as "densely connected nodes" or "sparsely connected nodes" according to the number of connections. Densely connected nodes refer to nodes that are connected to multiple courses (for example, a certain course is associated with ten courses), while sparsely connected nodes are nodes that are connected to a small number of courses (for example, a certain course is only related to two courses). Through this statistics, the core courses and edge courses in the course network can be identified, thereby generating densely and sparsely connected course structure data.

[0194] Step S54: dividing the densely connected course structure node data into a course code complex practice environment through a code running simulation model, thereby obtaining dense course node code practice environment data; dividing the sparsely connected course structure node data into a course code simplified practice environment through a code running simulation model, thereby obtaining sparse course node code practice environment data;

[0195] In this embodiment, the simulation model is run using the constructed code to divide the practice environment for the densely connected and sparsely connected course structure node data. For densely connected course nodes, the simulation model generates a complex practice environment, which contains multiple programming tasks and code snippets with high dependencies (such as comprehensive programming tasks that require multiple libraries and tools). For sparsely connected course nodes, the simulation model generates a simplified practice environment, which contains relatively simple programming tasks and a small amount of dependencies (such as programming tasks that only require a single library). Through this environment division, practical environment data suitable for dense and sparse course nodes can be obtained respectively to support learning activities of different difficulty levels.

[0196] Step S55: Arrange the dense course node code practice environment data and the sparse course node code practice environment data in a learning path practice environment, thereby obtaining user code practice environment data.

[0197] In this embodiment, the code practice environment data of dense and sparse course nodes are integrated and arranged according to the user's learning path data. Specifically, according to the user's completed learning progress and ability level, the practice environment data is sorted according to the learning difficulty and course dependency. For example, complex code practice tasks are arranged after the user has mastered the basic knowledge, while simple code practice tasks are arranged before the basic knowledge learning. Through this arrangement, personalized user code practice environment data can be generated, which provides users with a progressive learning experience, gradually improving from basic to complex, to meet the learning needs of different users.

[0198] Optionally, step S6 specifically includes:

[0199] Step S61: Obtain cloud platform resource data;

[0200] In this embodiment, when obtaining cloud platform resource data, you first need to access the API provided by the cloud platform (such as AWS's CloudWatch API or Azure Monitor API). Through these APIs, you can regularly query and obtain the following resource data: computing resources (such as CPU and memory usage of virtual machines and container instances), storage resources (such as disk usage and storage space utilization), network resources (such as traffic statistics and bandwidth usage), and database resources (such as query response time and number of transactions). During the data acquisition process, you can use API keys for authentication and configure appropriate permissions to ensure data security. The acquired data will be returned in JSON format, and then the data will be extracted and organized into a unified format through data parsing tools for use in subsequent steps.

[0201] Step S62: Designing an integrated platform architecture for the artificial intelligence course topology model, the course recommendation model, and the code operation simulation model, thereby obtaining integrated platform architecture data;

[0202] In this embodiment, when designing the integrated platform architecture, it is first necessary to define the main components of the artificial intelligence course topology model (such as course content module, teaching resource management module, learning progress tracking module), and set the corresponding algorithm for the course recommendation model (such as collaborative filtering algorithm or content recommendation algorithm). The code running simulation model requires an integrated simulation execution environment (such as a Docker container or a virtual environment). In order to design an efficient integrated platform architecture, UML (Unified Modeling Language) diagrams can be used to describe the interaction relationship and data flow of each model component. During specific implementation, a platform architecture tool (such as Kubernetes for container management and Spring Boot for service construction) can be selected, and an interface protocol (such as RESTful API) can be defined to achieve seamless integration of each model component.

[0203] Step S63: Obtain artificial intelligence forum port data, and perform docking interface design on the integrated platform architecture data according to the artificial intelligence port data, so as to obtain optimized integrated platform architecture data;

[0204] In this embodiment, port data is obtained from the API document of the artificial intelligence forum. Assume that the forum provides a RESTful API interface, which includes port data such as user message interface, post data interface, comment interface, etc. Use API access tools (such as Postman) to test these ports to ensure that the data format and return results meet expectations. Next, design the docking interface of the integration platform based on these interface data. This may include defining the data transmission format (such as JSON or XML), the structure of interface requests and responses, and error handling mechanisms. For example, in order to connect the forum data with the course recommendation system, a data receiving interface can be designed so that new posts in the forum can update the input data of the recommendation model in real time, thereby providing more timely course recommendations.

[0205] Step S64: perform platform architecture resource coupling on the integrated platform architecture data according to the cloud platform resource data, thereby obtaining an artificial intelligence education service platform, and upload it to the cloud platform to execute the platform deployment task.

[0206] In this embodiment, based on the obtained cloud platform resource data, evaluate whether the existing resource configuration (such as CPU, memory, storage) meets the requirements of the integrated platform architecture. Assume that the integrated platform architecture requires 20 virtual machine instances, each instance is configured with 4GB of memory and 2CPU cores. According to the cloud platform resource data, check whether the current cloud environment has enough available resources or whether it needs to be expanded. Then, deploy the integrated platform architecture to the cloud platform. You can choose automation tools (such as Terraform) to configure resources and define the resource requirements (such as instance type, storage type). During the platform deployment process, you also need to configure a load balancer and monitoring tools to ensure the stability and high availability of the platform. After the deployment is completed, continuously monitor the platform operation status and adjust the resource configuration in time to cope with load changes.

[0207] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0208] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. 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 invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A development method based on artificial intelligence education service platform, It is characterized in that The following steps are involved: Step S1: obtaining artificial intelligence education course data, and performing artificial intelligence course topology structure analysis based on the artificial intelligence education course data, thereby obtaining an artificial intelligence course topology structure model; Step S2: Obtaining artificial intelligence forum access data, and performing user access preference analysis on the artificial intelligence forum access data, thereby obtaining user access preference data; Visualize the user demand map based on the user access preference data to obtain the user knowledge demand map, and build a course recommendation model based on the user knowledge demand map to obtain the course recommendation model; Step S3: extracting user real-time interaction features from the artificial intelligence forum access data to obtain user real-time interaction data, and performing real-time user interaction behavior analysis on the user real-time interaction data to obtain user interaction behavior data; performing customized course content recommendations on the user interaction behavior data through a course recommendation model to obtain user real-time learning content recommendation data; Step S4: Performing optimal learning path planning for the user based on the user's real-time learning content recommendation data through the artificial intelligence course topology structure model, thereby obtaining the user's optimal learning path planning data; Step S5: obtaining virtual code editing environment data, and constructing a code running simulation model according to the virtual code editing environment; dividing the user's optimal learning path planning data into code practice operation environments through the code running simulation model, thereby obtaining the user's code practice environment data; Step S6: constructing an artificial intelligence education service platform for the artificial intelligence course topology model, the course recommendation model, and the code operation simulation model, thereby obtaining an artificial intelligence education service platform; Step S7: Perform user artificial intelligence education strategy analysis based on user code practice environment data, user optimal learning path planning data, and user real-time learning content recommendation data to obtain the user artificial intelligence education strategy, and upload it to the artificial intelligence education service platform to execute the education strategy push task.

2. According to the development method based on the artificial intelligence education service platform according to claim 1, It is characterized in that Step S1 is specifically as follows: Step S11: Obtaining artificial intelligence education course data, and classifying the artificial intelligence education course data by course subject, thereby obtaining artificial intelligence basic course data and artificial intelligence application course data; Step S12: Performing basic course hierarchical structure analysis on the artificial intelligence basic course data, thereby obtaining basic course hierarchical structure data; Step S13: performing application course hierarchical structure analysis on the artificial intelligence application course data, thereby obtaining application course hierarchical structure data; Step S14: performing course level progressive relationship analysis on the basic course level structure data and the applied course level structure data, thereby obtaining course level connection relationship data; Step S15: Perform artificial intelligence course topology structure analysis based on course hierarchical connection relationship data to obtain an artificial intelligence course topology structure model.

3. According to the development method based on artificial intelligence education service platform according to claim 2, It is characterized in that Step S12 is specifically as follows: Step S121: Classifying basic courses according to the artificial intelligence basic course data, thereby obtaining mathematics basic course data and programming basic course data; Step S122: extracting course text data from the programming basic course data and the mathematics basic course data respectively, thereby obtaining the programming basic course text data and the mathematics basic course text data; Step S123: performing mathematical formula similarity evaluation on the mathematics basic course text data and the programming basic course text data, thereby obtaining mathematical formula similarity data; Step S124: Associating the basic mathematics course data and the basic programming course data with high similarity courses according to the mathematical formula similarity data, thereby obtaining mathematics-programming course association data; Step S125: construct a course hierarchy structure according to the mathematics-programming course association data, thereby obtaining basic course hierarchy structure data.

4. The development method based on the artificial intelligence education service platform according to claim 2, It is characterized in that Step S13 is specifically as follows: Step S131: performing application course clustering calculation according to the artificial intelligence application course data, thereby obtaining application course clustering data; Step S132: extracting programming code structure features from the application course clustering data, thereby obtaining programming code structure data; Step S133: performing programming code composition analysis according to the programming code structure data, thereby obtaining programming code composition structure data; Step S134: dividing the programming code structure according to the programming code composition structure data, thereby obtaining basic programming code structure data and complex programming code structure data; Step S135: performing programming code level integration according to the basic programming code structure data and the complex programming code structure data, thereby obtaining programming code level data; Step S136: constructing an application course hierarchical structure for the application course clustering data according to the programming code hierarchical data, thereby obtaining application course hierarchical structure data.

5. The development method based on the artificial intelligence education service platform according to claim 2, It is characterized in that Step S14 is specifically as follows: Step S141: extracting course description texts from the artificial intelligence basic course data and the artificial intelligence application course data, respectively, to obtain basic course description text data and application course description text data; Step S142: performing course relevance analysis based on the basic course description text data and the applied course description text data, thereby obtaining basic-applied course relevance data; Step S143: integrating the application course pre-basic structure according to the foundation-application course association data, thereby obtaining the application course pre-basic structure data; Step S144: formulating progressive relationship rules according to the course pre-basic structure data, thereby obtaining course progressive relationship data; Step S145: Performing course-level progressive connection on the basic course level structure data and the applied course level structure data according to the course progressive relationship data, thereby obtaining course level connection relationship data.

6. The development method based on the artificial intelligence education service platform according to claim 1, It is characterized in that Step S2 is specifically as follows: Step S21: obtaining artificial intelligence forum access data, and classifying the artificial intelligence forum access data by user access type, thereby obtaining forum video user viewing data and forum document user reading data; Step S22: Counting the viewing time of tagged videos based on the forum video user viewing data, thereby obtaining tagged data of videos with high viewing time; Step S23: Perform tag document search statistics on forum document user search data, thereby obtaining high-frequency document search tag data; Step S24: performing a label intersection operation on the high-watching-time video label data and the high-frequency document reference label data to obtain user preference label data, and performing user access preference integration based on the user preference label data to obtain user access preference data; Step S25: Visualizing the user demand map according to the user access preference data, thereby obtaining the user knowledge demand map; Step S26: Construct a course recommendation model based on the user knowledge demand map to obtain a course recommendation model.

7. The development method based on the artificial intelligence education service platform according to claim 6, It is characterized in that Step S25 is specifically as follows: Step S251: extracting user access path features from artificial intelligence forum access data, thereby obtaining user access path data; Step S252: performing access path statistics on the user access path data to obtain high-frequency access path data, and performing user access behavior analysis based on the high-frequency access path data to obtain user forum access behavior data; Step S253: Associating the user forum access behavior data and the user access preference data with the user access behavior pattern, thereby obtaining the user access behavior pattern data; Step S254: constructing a user demand map according to the user access behavior pattern data, thereby obtaining a user demand map set; Step S255: performing group demand integration on the user demand map set, thereby obtaining a group demand map set; Step S256: extracting the user knowledge demand graph from the group demand graph set according to the user demand graph set, thereby obtaining the user knowledge demand graph.

8. The development method based on the artificial intelligence education service platform according to claim 1, It is characterized in that Step S3 is specifically as follows: Step S31: extracting user real-time interaction features and user historical search features from the artificial intelligence forum access data, thereby obtaining user real-time interaction data and user historical search data; Step S32: extracting page dwell time features from the user real-time interaction data to obtain page dwell time data, and performing high-dwell time page tag statistics based on the page dwell time data to obtain high-dwell time page tag data; Step S33: performing tag search frequency statistics on the user's historical search data, thereby obtaining high-frequency search tag data and low-frequency search tag data; Step S34: performing an intersection operation on the high-value dwell time page tag data and the high-frequency search tag data to obtain obvious preference tag data; performing an intersection operation on the high-value dwell time page tag data and the low-frequency search tag data to obtain potential preference tag data; Step S35: The page dwell time data integrates the real-time interaction time behavior of the obvious preference tag data and the potential preference tag data, thereby obtaining user interaction behavior data; Step S36: Customized course content recommendation is performed on the user interaction behavior data through the course recommendation model, thereby obtaining user real-time learning content recommendation data.

9. The development method based on the artificial intelligence education service platform according to claim 8, It is characterized in that Step S5 is specifically as follows: Step S51: obtaining virtual code editing environment data, and constructing a code running simulation model according to the virtual code editing environment; Step S52: extracting the planned course connection structure features from the artificial intelligence course topology structure model according to the user's optimal learning path planning data, thereby obtaining the planned course structure data; Step S53: performing course structure node connection statistics on the planned course structure data, thereby obtaining densely connected course structure node data and sparsely connected course structure node data; Step S54: dividing the densely connected course structure node data into a course code complex practice environment through a code running simulation model, thereby obtaining dense course node code practice environment data; dividing the sparsely connected course structure node data into a course code simplified practice environment through a code running simulation model, thereby obtaining sparse course node code practice environment data; Step S55: Arrange the dense course node code practice environment data and the sparse course node code practice environment data in a learning path practice environment, thereby obtaining user code practice environment data.

10. The development method based on the artificial intelligence education service platform according to claim 1, It is characterized in that Step S6 is specifically as follows: Step S61: Obtain cloud platform resource data; Step S62: Designing an integrated platform architecture for the artificial intelligence course topology model, the course recommendation model, and the code operation simulation model, thereby obtaining integrated platform architecture data; Step S63: Obtain artificial intelligence forum port data, and perform docking interface design on the integrated platform architecture data according to the artificial intelligence port data, so as to obtain optimized integrated platform architecture data; Step S64: perform platform architecture resource coupling on the integrated platform architecture data according to the cloud platform resource data, thereby obtaining an artificial intelligence education service platform, and upload it to the cloud platform to execute the platform deployment task.

Images