Intelligent education support method, system and equipment based on deep learning, and medium

Abstract

The invention relates to the technical field of intelligent education, and particularly discloses an intelligent education support method, system and device based on deep learning, and a medium, and the method comprises the steps: S1, data collection; s2, constructing a model; s3, identifying and classifying service requests; s4, analyzing a resource library; s5, data interaction; s6, carrying out model iteration; according to the digital intelligence tutor APP designed by the invention, three core requirements of academic guidance (examination research, insurance research and graduation theses), scientific research competition guidance and career planning are integrated into a whole for the first time, and the application limitation of a general digital intelligence education tool is broken through; a professional exclusive digital intelligence guidance normal form taking nursing subject requirements as the core and taking undergraduate personalized growth as the guide is constructed, the theoretical and practical blank of a full-chain digital intelligence growth support system in the nursing education field is filled, and a non-universal theoretical reference is provided for digital intelligence education tool development of highly specialized subjects.

Description

Technical Field

[0001] This invention belongs to the field of intelligent education technology, specifically relating to intelligent education support methods, systems, devices, and media based on deep learning. Background Technology

[0002] With the deep integration of artificial intelligence and education, intelligent education systems have become a core supporting technology for promoting personalized teaching and optimizing the allocation of educational resources. Their applications in learning behavior analysis, learning needs prediction, and personalized resource delivery offer new pathways to overcome the homogenization dilemma of traditional large-scale teaching. The deep integration of artificial intelligence and education has become a core direction of digital transformation in education, especially in professional talent training fields such as nursing, where the demand for personalized and precise educational support is increasingly urgent. While current intelligent education support technologies have made some progress in improving the efficiency of large-scale teaching, many key shortcomings remain, making it difficult to meet the deep-seated needs of students' full-cycle growth and professional development.

[0003] Significant flaws exist in the model building and prediction stages. The "black box effect" of deep learning models is prominent, lacking interpretability to support trust from teachers and students. Existing models often focus on a single prediction task, separating short-term performance prediction from long-term career matching goals, and failing to fully integrate subject-specific professional knowledge constraints, such as the connection to the nursing curriculum and the mapping of abilities to knowledge points. This results in predictions lacking educational rationality and professional suitability. Furthermore, models often remain static after training, making it difficult to cope with performance degradation caused by the evolution of student abilities and data distribution shifts.

[0004] These shortcomings make it difficult for existing intelligent education support technologies to break through the bottleneck of integrating "data-driven" with "educational laws" and "professional characteristics," and they are unable to provide students with precise support covering the entire chain of academics, scientific research, and career. There is an urgent need for a deep learning intelligent education support method that can integrate multi-source data, strengthen professional adaptation, and achieve dynamic optimization.

[0005] In response, this application proposes a deep learning-based intelligent education support method, system, device, and medium to address the aforementioned problems. Summary of the Invention

[0006] The purpose of this invention is to provide intelligent education support methods, systems, devices, and media based on deep learning to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] Deep learning-based intelligent education support methods include:

[0009] We collected personalized needs data from undergraduate nursing students to obtain an initial needs dataset.

[0010] Based on the characteristics of the nursing profession and the initial demand dataset, a deep learning model is constructed and trained to obtain an intelligent guidance agent adapted to the nursing field.

[0011] Receive service requests initiated by users, perform type identification and classification processing on the service requests, and obtain the target requests for corresponding academic guidance, scientific research competition guidance or career planning guidance;

[0012] The intelligent guidance agent calls a preset nursing-related resource library and uses deep learning algorithms to parse and process the target request to generate personalized guidance results.

[0013] The personalized guidance results are output to the user through multiple ports, and user interaction feedback data is collected.

[0014] The deep learning model is iteratively optimized based on the interactive feedback data to update the processing capabilities of the intelligent guidance agent.

[0015] Preferably, the personalized needs data includes information on the needs of undergraduate nursing students in areas such as writing graduation theses, applying for postgraduate studies, planning for postgraduate entrance examinations, participating in scientific research competitions, choosing career paths, and optimizing learning methods, but does not include data on summaries of professional course knowledge points or systematic academic knowledge transfer.

[0016] Preferably, the training data of the deep learning model includes rules for nursing-related scientific research competitions (including "Innovation and Entrepreneurship Training" and "Challenge Cup"), further education policies, career development trends, peer experience cases, and industry policy documents. The intelligent guidance agent is a digital mentor-specific intelligent agent adapted to the nursing profession.

[0017] Preferably, the type identification uses a natural language processing algorithm, wherein academic guidance requests correspond to guidance on graduation thesis, postgraduate recommendation, postgraduate entrance examination and learning methods consultation; scientific research competition guidance requests correspond to consultation on topic selection, preparation and application for competitions such as "Innovation and Entrepreneurship Training" and "Challenge Cup"; and career planning guidance requests correspond to consultation on career direction selection, employment prospect analysis and career development path planning related to nursing.

[0018] Preferably, the preset resource library includes anonymized examples of excellent graduation theses, experience materials on postgraduate recommendation and entrance examination, award-winning cases of scientific research competitions, employment data in the nursing industry, and career planning theory materials.

[0019] Preferably, the multi-port includes a student terminal, a tutor terminal, and an administrator terminal. The student terminal is used to initiate requests and receive guidance results, the tutor terminal is used to supplement professional guidance opinions and review guidance results, and the administrator terminal is used for resource updates and permission management.

[0020] The intelligent guidance agent's processing includes intelligent question answering, personalized resource recommendation, experience matching, and path planning suggestions.

[0021] Another aspect of the present invention is to provide a deep learning-based intelligent education support system, comprising:

[0022] The data acquisition module is used to collect personalized needs data from undergraduate nursing students to obtain an initial needs dataset.

[0023] The model building module is used to build a deep learning model based on nursing professional characteristics and the initial demand dataset, and train it to obtain an intelligent guidance agent adapted to the nursing field.

[0024] The service request identification and classification module is used to receive service requests initiated by users, perform type identification and classification processing on the service requests, and obtain the target requests for corresponding academic guidance, scientific research competition guidance or career planning guidance.

[0025] The resource library analysis module is used to call a preset nursing professional-related resource library using the intelligent guidance agent, and combine deep learning algorithms to parse and process the target request to generate personalized guidance results.

[0026] The data interaction module is used to output the personalized guidance results to the user through multiple ports and to collect user interaction feedback data.

[0027] The model iteration module is used to iteratively optimize the deep learning model based on the interactive feedback data and update the processing capabilities of the intelligent guidance agent.

[0028] An electronic device, the electronic device comprising:

[0029] At least one processor; and a memory communicatively connected to said at least one processor; wherein:

[0030] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described above.

[0031] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described above.

[0032] Compared with the prior art, the beneficial effects of the present invention are:

[0033] (1) The “Digital Mentor” APP designed in this invention integrates the three core needs of “academic guidance (graduate entrance examination / guarantee for postgraduate studies, graduation thesis), scientific research competition guidance (Challenge Cup, innovation and entrepreneurship), and career planning” into one, breaking through the application limitations of general digital education tools, and constructing a professional digital guidance paradigm “centered on the needs of nursing disciplines and guided by the personalized growth of undergraduates”. It fills the theoretical and practical gap of “full-chain digital growth support system” in the field of nursing education and provides a “non-generalized” theoretical reference for the development of digital education tools for highly specialized disciplines.

[0034] (2) This invention takes the “digital intelligent mentor” AI agent as the core and combines a multi-port collaborative design of “student end-mentor end-administrator end” to reconstruct the service logic of mentor support from a theoretical perspective: On the one hand, through the 24 / 7 response capability of the AI ​​agent, the traditional “passive appointment-based guidance” is transformed into a service model of “proactive personalized push + instant Q&A”, breaking through the limitations of time and space; on the other hand, it integrates “peer experience library (shared by senior students), expert resource library (mentor and industry expert guidance), dynamic information library (competition / graduate entrance examination / career policy)” to form a collaborative support system of “initial matching of intelligent agent - in-depth supplementation by human mentor - mutual verification by peer community”, which theoretically solves the core problems of “dispersed resources and insufficient accuracy” in the traditional mentor system and provides a reusable service logic framework for the digital transformation of the professional mentor system in universities. Attached Figure Description

[0035] Figure 1 This is a flowchart of the deep learning-based intelligent education support method of the present invention;

[0036] Figure 2 This is a block diagram of the intelligent education support system based on deep learning according to the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0038] Example 1:

[0039] Please see Figure 1 As shown, a deep learning-based intelligent education support method is used:

[0040] Taking the application of intelligent guidance for undergraduate nursing students in a certain university during a regular semester as an example:

[0041] Taking the 2023-2024 academic year as a typical semester for the School of Nursing and Health at a university, this paper implements intelligent educational support methods corresponding to the "Digital Mentor" APP, targeting the daily academic, research competition, and career planning needs of undergraduate nursing students from freshman to senior year. The implementation covers all steps of claims 1-6, and the specific implementation process is as follows:

[0042] Step 1: Collect personalized needs data from undergraduate nursing students to obtain an initial needs dataset.

[0043] Data acquisition method: The combination of "semi-structured interview + online questionnaire" was adopted, combined with the "stratified purpose sampling" method mentioned in the literature

[12] , and 50 nursing undergraduates from freshman to senior year were selected (a total of 200 people, 32 males and 168 females). Among them, 100 people were in the experimental group (using this method) and 100 people were in the control group (using the traditional mentor system).

[0044] Semi-structured interviews: Based on the "interview outline" in the document (including 6 types of questions such as academics, scientific research, and career), each group of interviews lasted 40-60 minutes. The recordings were transcribed into text, and a total of 192 valid interview texts were obtained.

[0045] Online questionnaire: A needs assessment questionnaire with 30 items was designed (Cronbach's α = 0.89, reliability qualified), and 188 valid questionnaires were collected.

[0046] Data Processing and Results: Through Nvivo software coding analysis, 820 valid needs were extracted from interview texts and questionnaires, forming an initial needs dataset. The needs were categorized as follows: academic guidance needs (35%, including 12% for postgraduate entrance examination consultation, 15% for graduation thesis guidance, and 8% for learning methods), scientific research and competition guidance needs (25%, including 10% for "Innovation and Entrepreneurship Training" consultation, 8% for "Challenge Cup" preparation, and 7% for competition information), and career planning needs (40%, including 18% for employment direction, 12% for civil service examination consultation, and 10% for career assessment).

[0047] Step 2: Construct a deep learning model based on the characteristics of the nursing profession, and train it to obtain an intelligent guidance agent adapted to the nursing field.

[0048] Model building and training data:

[0049] Model architecture: It adopts a hybrid deep learning architecture of "BERT pre-trained model + CNN convolutional layer" (adapted to natural language understanding and demand matching scenarios). The input layer is nursing professional text data, and the output layer is the "demand-resource" matching result.

[0050] Training dataset: Integrates nursing-specific data, including: postgraduate entrance examination / recommendation policy documents (admission brochures of 50 nursing colleges in 2023-2024), scientific research competition cases (120 provincial and above award-winning projects of "Innovation and Entrepreneurship Training Program" and 80 nursing-related works of "Challenge Cup"), career planning materials (30 nursing industry employment reports and 25 civil service examination / entrepreneurship policies), and peer experience posts (150 anonymized experiences of outstanding senior students), with a total data volume of 1.2GB;

[0051] Training parameters: Adam optimizer, learning rate 1e-4, batch size = 32, 50 rounds of iterative training, training set to validation set ratio 8:2.

[0052] Intelligent agent generation results: The trained "digital intelligence tutor" intelligent agent achieved an accuracy rate of 88% in demand recognition on the validation set and a demand-resource matching F1 value of 0.86, which meets the basic accuracy requirements for personalized guidance in nursing.

[0053] Step 3: Receive service requests initiated by users, perform type identification and classification processing on the service requests, and obtain the target request.

[0054] Request Reception and Processing Process: 100 students in the experimental group initiated service requests through the "Digital Mentor" APP student terminal. The request format was natural language input (such as "How to prepare for the 2024 Nursing Postgraduate Entrance Examination Re-examination?" "What should be paid attention to when choosing a topic for the Innovation and Entrepreneurship Project?").

[0055] Classification Processing Techniques and Results: The deep learning model constructed in step 2 was used for type recognition. Through the entity recognition and intent classification modules of natural language processing (NLP), the requests were classified into three target requests: "academic guidance", "scientific research competition guidance" and "career planning guidance".

[0056] During the implementation period (16 weeks), the experimental group initiated a total of 420 valid requests, averaging 0.26 requests per person per week. The requests were categorized as follows: 189 requests for academic guidance (45%, including 85 for postgraduate entrance examination consultation, 72 for graduation thesis guidance, and 32 for learning methods); 84 requests for scientific research and competition guidance (20%, including 38 for "Innovation and Entrepreneurship Training" consultation, 30 for "Challenge Cup" preparation, and 16 for competition information); and 147 requests for career planning guidance (35%, including 68 for employment direction, 45 for civil service examination consultation, and 34 for career assessment).

[0057] Classification accuracy verification: 80 random manual review requests were made, and the classification was correct 73 times, with an accuracy rate of 91.25% and an F1 score of 0.90, indicating stable classification performance.

[0058] Step 4: The intelligent guidance agent calls upon a pre-set nursing-related resource library, combines deep learning algorithms to parse and process the target request, and generates personalized guidance results.

[0059] Pre-defined resource library structure: Based on the document "Functional Module Design", the resource library only contains non-course resources for the nursing profession, specifically:

[0060] Academic guidance resources: 80 examples of desensitized excellent graduation theses, 50 sets of postgraduate entrance examination university matching tables (including application-to-admission ratio and re-examination subjects), and 20 study method manuals (such as "Nursing Professional Literature Reading Skills").

[0061] Scientific research competition resources: 30 sets of "Innovation and Entrepreneurship Training Program" application templates, 50 PPT / reports (de-sensitized) of award-winning works in the "Challenge Cup" nursing category, and a competition time reminder database (15 provincial and above nursing competitions in 2024);

[0062] Career planning resources: 15 nursing industry employment prospect analysis reports, 20 civil service exam job postings (nursing related), and 5 sets of career assessment scales (such as the Holland Occupational Themes test for nursing).

[0063] Specific exclusions include: course content resources such as professional course knowledge point videos, knowledge graphs, and AI question banks.

[0064] Analysis and guidance result generation: The agent extracts key information (such as "2024 postgraduate entrance examination" and "innovation topic selection") from the request through deep learning algorithms (such as attention mechanism), matches the corresponding resources in the resource library, and generates structured guidance results (such as postgraduate entrance examination re-examination suggestions including "review focus of re-examination subjects", "tutor contact skills" and "common interview questions", and competition guidance including "recommendation of topic direction", "application process" and "material writing points").

[0065] Results quality verification: Five nursing faculty members (associate professors and above) were invited to rate 100 guidance results (1-5 points). The average score was 4.1 points, with 4.2 points for "relevance" and 4.0 points for "practicality", indicating that the guidance results meet the needs of the nursing profession.

[0066] Step 5: Output personalized guidance results to users through multiple platforms and collect user interaction feedback data:

[0067] Multi-port output implementation: Based on the "three ports" design in the document, collaborative output is achieved from the student end, the tutor end, and the administrator end.

[0068] For students: Guidance results are pushed through the "Messages" interface of the APP, with an average response time of 1.8 seconds (from initiating the request to receiving the result). It supports online viewing and downloading of resources (such as paper templates and competition application forms).

[0069] On the tutor side: the tutor conducts a second review of the guidance results involving complex needs (such as "thesis topic argumentation" and "revision of research project plan"), with an average review time of 2.0 hours (the average review time of the traditional tutor system is 24 hours), and pushes it to the student after the review is approved;

[0070] Administrator side: Update the resource library once a week (such as adding new competition notices and updating postgraduate entrance examination policies) to ensure the timeliness of resources.

[0071] Feedback data collection: User feedback was collected through the "My - Feedback" module of the APP. A total of 86 valid feedback items were collected during the implementation period, of which 70 were "Satisfied / Fairly Satisfied" (81.4%) and 16 were "Needs Improvement" (18.6%, mainly focusing on "Personalized Recommendation Accuracy" and "Resource Diversity").

[0072] Step 6: Iteratively optimize the deep learning model based on interactive feedback data, and update the processing capabilities of the intelligent guidance agent:

[0073] Optimization Direction and Methods: Two optimization measures were implemented for the 16 feedback items that indicated "needs improvement":

[0074] Supplementing training data: Added 30 sets of past exam questions for postgraduate entrance examination in nursing and 15 sets of mid-term report templates for "Innovation and Entrepreneurship" projects, expanding the training dataset to 1.5GB;

[0075] Adjusting model parameters: The attention mechanism weights were adjusted to "professional matching degree 0.6 + demand urgency degree 0.4" to improve the accuracy of personalized recommendations.

[0076] Optimization result verification: After optimization and retraining the model (20 iterations), the agent's demand recognition accuracy increased to 92%, and the demand-resource matching F1 score increased to 0.91; user feedback was collected again (40 comments), and the percentage of "satisfied / relatively satisfied" increased to 89.5%, demonstrating significant optimization effect.

[0077] As can be seen from the above, the designed "Digital Mentor" APP integrates the three core needs of "academic guidance (graduate entrance examination / guarantee for postgraduate studies, graduation thesis), scientific research competition guidance (Challenge Cup, Innovation and Entrepreneurship Training Program), and career planning" into one, breaking through the application limitations of general digital education tools. It constructs a professional and exclusive digital guidance paradigm "centered on the needs of nursing disciplines and guided by the personalized growth of undergraduates", filling the theoretical and practical gap of "full-chain digital growth support system" in the field of nursing education, and providing "non-generalized" theoretical reference for the development of digital education tools for highly specialized disciplines.

[0078] Example 2:

[0079] A smart guidance application specifically designed for the graduation season of nursing undergraduates at a certain university:

[0080] This example uses the 2024 graduation season (March-June) of a university's School of Nursing and Health as its background, focusing on the core issues of "postgraduate entrance examination interview + graduation thesis + employment choice" for senior undergraduate nursing students, and implements intelligent education support methods. The specific implementation process is as follows:

[0081] Step 1: Collect personalized needs data from graduating nursing undergraduates to obtain the initial needs dataset:

[0082] Data acquisition method: 150 senior nursing undergraduates were selected (75 in the experimental group and 75 in the control group), and the following methods were used: in-depth interviews and a needs priority ranking questionnaire.

[0083] In-depth interviews: Targeting three core issues, namely “preparation for postgraduate entrance examination re-examination”, “thesis defense”, and “choice of employment offer”, each group of interviews lasted 60 minutes, and 72 valid interview transcripts were obtained;

[0084] Priority questionnaire: A Likert 5-point scale questionnaire was designed (Cronbach's α = 0.91), and 145 valid questionnaires were collected.

[0085] Data processing and results: 580 valid requirements were extracted, and the requirements were categorized as follows: academic guidance (90%, including 60% for postgraduate entrance examination interviews and 30% for graduation theses), scientific research competition guidance (5%, only guidance for the completion of "Innovation and Entrepreneurship Training" projects), career planning (5%, job offer comparison), and no professional course knowledge points.

[0086] Step 2: Build a deep learning model based on graduation season needs, and train it to obtain a smart guidance agent specifically for nursing graduates:

[0087] Model building and training data: Based on the model in Example 1, supplementary data specifically for graduation season is provided:

[0088] The training dataset has been updated with the following additions: 30 sets of nursing postgraduate entrance examination re-examination questions, 20 graduation thesis defense PPT templates, and 15 nursing job offer comparison indicators (such as salary and promotion path), with a total data volume of 1.8GB.

[0089] Training parameters: learning rate adjusted to 8e-5, 30 rounds of iterative training, training set to validation set ratio 9:1.

[0090] Intelligent agent generation results: The demand recognition accuracy of the "digital intelligence mentor" intelligent agent, which is dedicated to graduation season, reached 90%, and the resource matching F1 value for postgraduate entrance examination re-examination / thesis guidance was 0.89.

[0091] Step 3: Receive graduation season service requests, perform type identification and classification processing, and obtain the target request:

[0092] Request reception and processing: 75 students in the experimental group initiated graduation season requests through the student APP. During the implementation period (12 weeks), a total of 315 valid requests were initiated, with an average of 0.35 requests per person per week.

[0093] Classification results: 252 instances of academic guidance (80%, including 189 instances of postgraduate entrance examination re-examination and 63 instances of graduation thesis), 16 instances of scientific research competition guidance (5%), and 47 instances of career planning guidance (15%); the classification accuracy rate reached 93.3% after manual review (60 times).

[0094] Step 4: The intelligent guidance agent accesses the graduation season-specific resource library to generate personalized guidance results:

[0095] Graduation Season Exclusive Resource Library: New resources include "Postgraduate Entrance Examination Interview Simulation Question Bank (Nursing Major)", "Common Questions and Answering Strategies for Graduation Thesis Defense", and "Nursing Employment Offer Evaluation Form";

[0096] Guidance Result Generation: For the request of "Answering Professional Questions in Postgraduate Entrance Examination Interview", the agent generates a structured result of "Problem Breakdown + Core Knowledge Points + Answer Framework"; for the request of "Graduation Thesis Defense", it generates "Defense Process + Judges' Focus Points + Emergency Handling Plan". The average score of the teacher's evaluation (50 results) is 4.3 points.

[0097] Step 5: Multi-port output and feedback data acquisition:

[0098] Multi-port output:

[0099] For students: the response time for postgraduate entrance examination interview guidance results is 1.5 seconds, and the download rate of graduation thesis templates is 92%.

[0100] From the supervisor's perspective: The review of graduation theses takes 1.5 hours, which is 25% shorter than a regular semester.

[0101] Administrator side: Update "Nursing Spring Recruitment Job Information" every 3 days, improving resource timeliness by 60%.

[0102] Feedback data: 68 valid feedback items were collected, with 86.8% of respondents expressing satisfaction or satisfaction.

[0103] Step 6: Based on the feedback optimization model, update the processing capabilities of the intelligent guidance agent.

[0104] Optimization measures: Supplement "Nursing Postgraduate Entrance Examination Interview Supervisor Preference Data" (20 sets) and "Graduation Thesis Blind Review Comments Response Templates" (15 sets), and adjust the "Demand Urgency" weight of the model to 0.5;

[0105] Optimization results: The accuracy of intelligent agent demand recognition has been improved to 94%, and the percentage of users who are "satisfied / relatively satisfied" has increased to 92%.

[0106] As can be seen from the above, this invention takes the "digital intelligent mentor" AI agent as its core and combines a multi-port collaborative design of "student end-mentor end-administrator end" to theoretically reconstruct the service logic of mentor support: On the one hand, through the 24 / 7 response capability of the AI ​​agent, the traditional "passive appointment-based guidance" is transformed into a service model of "proactive personalized push + instant Q&A", breaking through the limitations of time and space; on the other hand, it integrates "peer experience database (shared by senior students), expert resource database (mentor and industry expert guidance), and dynamic information database (competition / postgraduate entrance examination / career policy)" to form a collaborative support system of "initial matching of intelligent agents - in-depth supplementation by human mentors - peer community mutual assistance verification", which theoretically solves the core problems of "dispersed resources and insufficient accuracy" in the traditional mentor system and provides a reusable service logic framework for the digital transformation of the professional mentor system in universities.

[0107] Example 3:

[0108] Comparative Example: Traditional Nursing Undergraduate Mentorship System:

[0109] Using the 2023-2024 academic year (regular semester) and the 2024 graduation season of a university's School of Nursing and Health as a case study, this study employs a traditional offline mentorship system (one mentor guiding 15-20 students, with one fixed offline consultation per week) as a comparative example. The core indicators are compared with those of Examples 1 and 2, as detailed below:

[0110] Implementation method:

[0111] Step 1: Needs gathering relies on "offline scheduled interviews," once a month, with an average of 1.2 needs collected per person per semester;

[0112] Step 2: No deep learning model is used; guidance is provided based on the mentor's personal experience.

[0113] Step 3: Requests to receive information are made via "WeChat / phone appointment," with an average response time of 24 hours. Classification is based on the tutor's subjective judgment.

[0114] Step 4: Resources are manually shared by the supervisor (e.g., sending thesis templates via email), and the amount of resources is only 1 / 5 of that in Example 1;

[0115] Step 5: Only two-way communication between "mentor and student" is allowed, with no administrator interface, and the resource update cycle is 1 month;

[0116] Step 6: Optimization relies on the "end-of-semester tutor summary meeting," which only takes place once per semester, resulting in low optimization efficiency.

[0117] The core metrics of the examples and comparative examples are compared in Table 1 below:

[0118] Table 1

[0119]

[0120] As shown in the table above, both Embodiment 1 and Embodiment 2 focus on three major modules: "academic guidance (postgraduate entrance examination / thesis / learning methods), scientific research competitions (innovation and entrepreneurship / challenge cup), and career planning." The "digital intelligence mentor" intelligent agent, which is dedicated to nursing majors, is used to identify and guide needs, and the accuracy rate has been improved from 88% to 94%, which reflects the core role of the intelligent agent.

[0121] Furthermore, the collaborative output of the student, tutor, and administrator terminals reduces the request response time from 24 hours to 1.5-1.8 seconds and the tutor review time is reduced by 93.75%, solving the problem of "limited resources and fixed time" in the traditional tutor system.

[0122] The data was obtained through a combination of interviews, questionnaires, app logs, and teacher ratings, resulting in a 38% increase in satisfaction and an 86.7% increase in participation in research competitions, which corresponds to "personalized guidance and efficient service".

[0123] Example 4:

[0124] like Figure 2 As shown, another aspect of the present invention is to provide a deep learning-based intelligent education support system, comprising:

[0125] The data acquisition module is used to collect personalized needs data from undergraduate nursing students to obtain an initial needs dataset.

[0126] The model building module is used to build a deep learning model based on nursing professional characteristics and the initial demand dataset, and train it to obtain an intelligent guidance agent adapted to the nursing field.

[0127] The service request identification and classification module is used to receive service requests initiated by users, perform type identification and classification processing on the service requests, and obtain the target requests for corresponding academic guidance, scientific research competition guidance or career planning guidance.

[0128] The resource library analysis module is used to call a preset nursing professional-related resource library using the intelligent guidance agent, and combine deep learning algorithms to parse and process the target request to generate personalized guidance results.

[0129] The data interaction module is used to output the personalized guidance results to the user through multiple ports and to collect user interaction feedback data.

[0130] The model iteration module is used to iteratively optimize the deep learning model based on the interactive feedback data and update the processing capabilities of the intelligent guidance agent.

[0131] As shown above, this application, based on the needs analysis of undergraduate nursing students (such as competition for postgraduate entrance exams / guaranteed admission, innovation in scientific research competitions, and career path selection), deeply integrates digital technologies (AI intelligent agents, personalized recommendations, and dynamic resource aggregation) with the characteristics of the nursing discipline (such as nursing industry prospect analysis, nursing theme adaptation in scientific research competitions, and connection to professional qualification examinations) at the theoretical level. For example, the scientific research competition guidance module focuses on nursing-related competitions such as the "Challenge Cup" and "Innovation and Entrepreneurship Training Program," the career planning module integrates nursing industry policies and employment directions, and the academic guidance module closely follows the characteristics of nursing graduation thesis writing and postgraduate entrance exam preparation. This design logic of "demand-driven - discipline adaptation - technology implementation" establishes for the first time a "theoretical adaptation path" for the application of digital education tools in highly specialized disciplines, providing theoretical guidance for the practice of interdisciplinary fields such as "nursing information and artificial intelligence" to ensure that "technology serves discipline needs," thus avoiding the problem of "generalized application" of digital technologies in professional education.

[0132] In one embodiment, this application also provides an electronic device, the electronic device comprising:

[0133] At least one processor; and a memory communicatively connected to said at least one processor; wherein:

[0134] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described above.

[0135] In one embodiment, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described above.

[0136] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A deep learning-based intelligent education support method, characterized in that, include: We collected personalized needs data from undergraduate nursing students to obtain an initial needs dataset. Based on the characteristics of the nursing profession and the initial demand dataset, a deep learning model is constructed and trained to obtain an intelligent guidance agent adapted to the nursing field. Receive service requests initiated by users, perform type identification and classification processing on the service requests, and obtain the target requests for corresponding academic guidance, scientific research competition guidance or career planning guidance; The intelligent guidance agent invokes a preset nursing-related resource library, and combines deep learning algorithms to parse and process the target request to generate personalized guidance results. The personalized guidance results are output to the user through multiple ports, and user interaction feedback data is collected. The deep learning model is iteratively optimized based on the interactive feedback data to update the processing capabilities of the intelligent guidance agent.

2. The intelligent education support method based on deep learning according to claim 1, characterized in that, The personalized needs data includes information on the needs of undergraduate nursing students in areas such as thesis writing, postgraduate recommendation, postgraduate entrance examination planning, participation in scientific research competitions, career path selection, and optimization of learning methods. It does not include data on summaries of professional course knowledge points or systematic academic knowledge transfer.

3. The intelligent education support method based on deep learning according to claim 1, characterized in that, The training data for the deep learning model includes nursing-related scientific research competition rules, further education policies, career development trends, and industry policy documents. The intelligent guidance agent is a digital mentor-specific agent adapted to the nursing profession.

4. The intelligent education support method based on deep learning according to claim 1, characterized in that, The type recognition uses a natural language processing algorithm. Among them, academic guidance requests correspond to guidance on graduation thesis, postgraduate recommendation, postgraduate entrance examination and learning methods; scientific research competition guidance requests correspond to consultation on competition topic selection, preparation and application; and career planning guidance requests correspond to consultation on career direction selection, employment prospect analysis and career development path planning related to nursing.

5. The intelligent education support method based on deep learning according to claim 1, characterized in that, The preset resource library includes anonymized examples of excellent graduation theses, experience materials on postgraduate recommendation and entrance examination, award-winning cases of scientific research competitions, employment data in the nursing industry, and career planning theory materials.

6. The intelligent education support method based on deep learning according to claim 1, characterized in that, The multi-port includes a student terminal, a tutor terminal, and an administrator terminal. The student terminal is used to initiate requests and receive guidance results, the tutor terminal is used to supplement professional guidance opinions and review guidance results, and the administrator terminal is used for resource updates and permission management. The intelligent guidance agent's processing includes intelligent question answering, personalized resource recommendation, experience matching, and path planning suggestions.

7. The intelligent education support system based on deep learning according to claim 1, characterized in that, include: The data acquisition module is used to collect personalized needs data from undergraduate nursing students to obtain an initial needs dataset. The model building module is used to build a deep learning model based on nursing professional characteristics and the initial demand dataset, and train it to obtain an intelligent guidance agent adapted to the nursing field. The service request identification and classification module is used to receive service requests initiated by users, perform type identification and classification processing on the service requests, and obtain the target requests for corresponding academic guidance, scientific research competition guidance or career planning guidance. The resource library analysis module is used to call a preset nursing professional-related resource library using the intelligent guidance agent, and combine deep learning algorithms to parse and process the target request to generate personalized guidance results. The data interaction module is used to output the personalized guidance results to the user through multiple ports and to collect user interaction feedback data. The model iteration module is used to iteratively optimize the deep learning model based on the interactive feedback data and update the processing capabilities of the intelligent guidance agent.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; wherein: The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The medium is a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, it implements the method as described in any one of claims 1-6.

Images