Self-adaptive education method based on artificial intelligence

Abstract

The invention discloses a self-adaptive education method based on artificial intelligence, and relates to the technical field of education, and the method comprises the following components: intelligent emotion monitoring, learning performance evaluation, self-adaptive teaching strategy adjustment, personalized learning path planning and a feedback and optimization mechanism. According to the method, by introducing a learning performance evaluation and prediction module and a prediction and adaptive teaching strategy adjustment module, individual differences, including learning styles, interest preferences and ability levels, of students are fully considered, so that teaching strategies are intelligently recommended and adjusted, and personalized learning paths are dynamically generated; the personalized teaching mode effectively overcomes the defects of a traditional'one-cut 'teaching mode, the teaching effect is remarkably improved, each student can obtain the most suitable learning experience, and the learning enthusiasm and the learning effect of the students are greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of educational technology, and in particular to an adaptive educational method based on artificial intelligence. Background Art

[0002] With the rapid development of artificial intelligence technology, the field of education is undergoing unprecedented changes. Traditional teaching methods often adopt a "one-size-fits-all" model, ignoring individual differences among students, such as learning styles, interest preferences, and ability levels, resulting in uneven teaching results.

[0003] Traditional teaching methods can evaluate students' learning performance to a certain extent, but they mainly rely on quantitative indicators such as test scores, homework scores and answering speed, which have many shortcomings. Traditional evaluation methods often cannot accurately predict the difficulties and interests that students may encounter in the learning process, and it is difficult to adjust teaching strategies in time to promote students' in-depth learning. Therefore, an adaptive education method based on artificial intelligence is proposed. Summary of the invention

[0004] The purpose of the present invention is to make up for the shortcomings of the prior art and to provide an adaptive education method based on artificial intelligence. The method fully considers the individual differences of students and dynamically adjusts the personalized teaching through emotional intelligence monitoring, learning performance evaluation, adaptive teaching strategy adjustment, personalized learning path planning and feedback and optimization mechanism, thereby effectively improving the teaching effect and students' learning experience.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: an adaptive education method based on artificial intelligence, the specific steps of the method are:

[0006] S100, emotional intelligence monitoring module: uses high-definition cameras and microphones to capture students' facial expressions, voice intonation and micro-expression changes, analyzes these non-verbal signals in real time, and identifies students' basic emotional state;

[0007] S200, learning performance evaluation and prediction module: Combined with academic performance and answering speed data, it comprehensively evaluates students' learning progress and ability level, and predicts students' difficulties and interests through algorithm models;

[0008] S300, Adaptive Teaching Strategy Adjustment Module: Based on the output results of the emotional intelligence monitoring module and the learning performance evaluation and prediction module, the module uses an algorithm model to predict the difficulties and interests encountered by students, intelligently recommends and adjusts teaching strategies, and stimulates interest through storytelling, gamification elements or humorous explanations when boredom is detected. When students are found to be in an excited state, challenging tasks can be added to promote in-depth learning;

[0009] S400, Personalized Learning Path Planning: Dynamically generates personalized learning paths based on students’ emotional state, learning performance, and personal preferences, allowing students to independently choose learning content within a certain range, enhancing learning autonomy and participation;

[0010] S500, Feedback and Optimization Mechanism: Establish a feedback loop for students’ learning outcomes, continuously optimize teaching strategies and learning paths, regularly collect student feedback, combine big data analysis, and continuously adjust and improve system performance.

[0011] Furthermore, the specific steps of S100 are:

[0012] S101, start the high-definition camera and microphone to ensure that they can clearly capture the student's facial expression and voice information. The camera starts to record the student's facial video stream, and the microphone simultaneously records the student's voice data. These data are captured in the form of a continuous stream and transmitted to the processing unit. The video stream captured by the camera is pre-processed, including frame extraction, face detection and tracking, and facial feature point positioning steps. Through the face recognition algorithm, it is ensured that the system can accurately track the student's facial area and extract key facial features.

[0013] S102, denoising, enhancing, and framing the voice data captured by the microphone for subsequent emotion analysis. Framing is to divide the continuous voice signal into a series of short time frames, and calculate the expression-related geometric features and texture features based on the facial feature points extracted in the preprocessing stage. These features can reflect different facial expressions, and extract acoustic features from the preprocessed voice frames, as well as more advanced voice emotion features.

[0014] S103, input the extracted facial expression features and speech features into a pre-trained model, and output the probability distribution of one or more emotion categories by analyzing the input features. These emotion categories usually include basic emotional states. According to the output of the model, the student's current emotional state is determined, and the identified emotional state is passed as input to the adaptive teaching strategy adjustment module, so as to dynamically adjust the teaching strategy according to the emotional feedback, collect the students' emotion recognition results and actual teaching effect feedback during use, so as to continuously optimize the accuracy and robustness of the model, and regularly evaluate the performance of the emotional intelligence monitoring module.

[0015] Furthermore, the specific steps of S200 also include:

[0016] S201, clarify which data sources will be used for the evaluation, including but not limited to students' test scores, homework scores, answering speed, class participation records, after-class review time, and online learning behavior logs, remove duplicate data, process missing values, and correct any obvious errors, organize the data into a format suitable for analysis, calculate the proportion of completed courses and learning rate based on students' learning history and current completion status, calculate the average score, standard deviation, and ranking percentage quantitative indicators based on test scores and homework scores, analyze students' answering time data, calculate the average answering time, and the answering time distribution of each type of question or knowledge point, select algorithm models based on the complexity of the problem and the characteristics of the data, predict difficult points and points of interest, divide the data set into training set, validation set, and test set, with a ratio of 70%, 15%, and 15%, use the training set data to train the model, adjust the model parameters to optimize performance, use the validation set to evaluate the performance of the model, and adjust the model according to the evaluation results;

[0017] S202, based on the results of the validation set, perform feature selection and model parameter optimization operations to improve the generalization ability of the model, use the trained model to predict the test set or newly collected data, evaluate the difficulties and points of interest that students may encounter, analyze the prediction results, explain which factors lead to specific prediction results, and which features have the greatest impact on the prediction results. Feedback the prediction results to teachers or students, collect feedback, and adjust or optimize the model as needed. Deploy the learning performance evaluation and prediction module to the actual learning management system to ensure that it can evaluate students' learning performance in real time or regularly, continuously monitor the performance of the model, regularly check the accuracy of the prediction results, and update the model or data source as needed to collect user feedback, and continuously optimize the functions of the evaluation module and user experience.

[0018] Furthermore, the formula for predicting difficulty points and points of interest in S201 is:

[0019] Prediction difficulty formula:

[0020] Where: D represents the predicted difficulty level, A represents the average speed of answering questions, T i represents the speed of answering the i-th question, represents the average speed of all answers, α is the weight of the speed of answering questions on the prediction of difficulty, S represents the standard deviation of learning performance, C j represents the grade of the jth test or assignment, represents the average score, β is the weight of the influence of learning performance on the prediction of difficulty;

[0021] Formula for predicting points of interest:

[0022] Where: I represents the predicted interest point degree, A, T i , α has the same meaning as above, S, C j , The meaning of β is the same as above. F is an additional factor, which indicates the proportion of time students spend on independent learning on a specific subject or topic. γ is the weight of the impact of answering speed on the prediction of interest points. δ is the weight of the impact of academic performance on the prediction of interest points. ∈ is the weight of the impact of independent learning time on the prediction of interest points.

[0023] Furthermore, the specific steps of S300 are:

[0024] S301, obtaining students' real-time emotional data and learning performance data from the emotional intelligence monitoring and prediction module and the learning performance evaluation and prediction module, cleaning the collected data, removing noise and outliers, ensuring the accuracy and reliability of the data, integrating the emotional data and learning performance data into a unified database, building an algorithm model, predicting the difficulties and points of interest that students may encounter in the future learning process, extracting key features from the integrated data, such as learning progress, error rate, and emotional changes, as inputs to the prediction model, running the prediction model, and generating a prediction report on the difficulties and points of interest in students' future learning;

[0025] S302, match the corresponding teaching strategies according to the difficulties and points of interest in the prediction report. For example, for the predicted difficulties, more detailed explanations, additional exercises or one-on-one tutoring can be recommended. For the points of interest, related extended reading or practical activities can be recommended. Combined with the results of emotional intelligence monitoring, the teaching strategies are adjusted to be emotionally adaptive. For example, when students show boredom, storytelling, gamification elements or humorous explanations are recommended to stimulate interest. When students are in an excited state, challenging tasks are added to promote deep learning. The recommended teaching strategies are deployed to the teaching platform or learning environment to ensure that teachers and students can easily access and use them. During the teaching process, students' learning status and emotional changes are continuously monitored to ensure the effective implementation of teaching strategies. An instant feedback mechanism is established to allow students and teachers to provide feedback on the effectiveness of teaching strategies. These feedbacks can be used as a basis for subsequent adjustments to teaching strategies.

[0026] Furthermore, the specific steps of S300 also include:

[0027] S303, after the teaching strategy has been implemented for a period of time, evaluate its impact on students' learning outcomes and emotional states. This can be achieved by comparing the learning performance data and emotional data before and after implementation. According to the effect evaluation results, the teaching strategy is iteratively optimized. If the effect is significant, it can be maintained or further promoted. If the effect is not good, it is necessary to re-analyze the students' difficulties, interests and emotional states and adjust the teaching strategy.

[0028] Furthermore, the specific steps of S400 are:

[0029] S401, collect students' emotional state data in the emotional intelligence monitoring module, including basic emotional state and its changing trend, obtain students' learning progress, ability level, answering speed, and predicted difficulty and interest points from the learning performance evaluation and prediction module, and collect students' personal preference information through questionnaires, interviews or system interaction records, including but limited to learning style, interest areas, and learning speed preference;

[0030] S402, comprehensively analyzing the collected data, using machine learning to identify the correlation and pattern between students' emotional states, learning performance and personal preferences, and constructing a generative model of personalized learning paths, which can predict the learning path based on the students' current states and personal preferences;

[0031] S403, dynamically generate personalized learning paths based on the results of data analysis and modeling, which includes selecting learning content that suits students’ current emotional state and learning progress, as well as incorporating learning activities and resources that suit students’ personal preferences. The learning path should have a certain degree of flexibility and be able to be dynamically adjusted based on students’ real-time feedback and learning progress.

[0032] Furthermore, the specific steps of S400 also include:

[0033] S404, presenting the generated personalized learning path to students in an intuitive manner, allowing students to independently select learning content and learning sequence within a certain range to enhance learning autonomy and sense of participation. At the same time, providing necessary guidance and suggestions to help students make reasonable choices;

[0034] S405, students study according to the personalized learning path. The system continuously monitors the students' emotional state and learning performance during the learning process, and collects students' feedback data during the learning process for subsequent optimization and adjustment of the learning path;

[0035] S406, based on students' feedback data and new learning data, regularly evaluate the difficulties and points of interest encountered in the learning path, and optimize and iterate the learning path according to the evaluation results.

[0036] Furthermore, the specific steps of S500 are:

[0037] S501, real-time monitoring and collection of students’ learning performance data, including academic performance, learning progress, and task completion, and simultaneous collection of students’ emotional state data, derived from the output of the emotional intelligence monitoring module, and regular collection of students’ subjective feedback through questionnaires, interviews, or system-built-in feedback tools, including learning experience and satisfaction with teaching strategies;

[0038] S502, conduct big data analysis on the collected learning achievement data and emotional state data, identify the correlation pattern between learning achievement and emotional state, analyze student feedback, summarize common problems and personalized needs, use machine learning algorithms to extract key indicators from the data, and evaluate and predict the effectiveness of teaching strategies and learning paths;

[0039] S503, based on the data analysis results, establish a feedback loop of student learning outcomes and emotional state changes. When it is found that the learning outcomes decline or the emotional state is not good, trigger the feedback mechanism, analyze the bottlenecks and deficiencies in the feedback loop, and identify the teaching elements or learning paths that need to be adjusted;

[0040] S504, dynamically adjust the teaching strategy based on the results of the feedback loop. When it is detected that students have a poor grasp of a certain knowledge point, adjust the depth of the teaching content or the teaching method, optimize the learning path, dynamically adjust the difficulty and sequence of learning tasks based on the difficulties and interests encountered by students, and consider introducing new teaching resources or tools.

[0041] Furthermore, the specific steps of S500 also include:

[0042] S505, regularly evaluate the overall performance of the system, including data processing capabilities, model prediction accuracy, and user feedback response speed. Based on the evaluation results, make necessary optimization and upgrades to the system, including algorithm improvement, system architecture adjustment, and user interface optimization. Continuously iterate and improve the system to adapt to the ever-changing learning needs of students and improve the efficiency and quality of education and teaching.

[0043] S506, maintain continuous monitoring of system operation, ensure the effective operation of the feedback and optimization mechanism, continuously collect new student data for subsequent analysis and optimization work, regularly iterate the system, introduce new technologies and new concepts, and improve the intelligence level of the system and user satisfaction.

[0044] Compared with the existing technology, an adaptive education method based on artificial intelligence has the following beneficial effects:

[0045] 1. This method introduces a learning performance evaluation and prediction module and a prediction and adaptive teaching strategy adjustment module. This method fully considers the individual differences of students, including learning styles, interest preferences and ability levels, so as to intelligently recommend and adjust teaching strategies and dynamically generate personalized learning paths. This personalized teaching method effectively solves the shortcomings of the traditional "one-size-fits-all" teaching model, significantly improves the teaching effect, enables each student to obtain the learning experience that suits him or her best, and greatly improves students' learning enthusiasm and learning effect.

[0046] 2. This method dynamically generates personalized learning paths, allowing students to independently choose learning content and learning order within a certain range, which greatly enhances students' learning autonomy and sense of participation. At the same time, according to students' emotional state and learning performance, the system can adjust teaching strategies in a timely manner, such as introducing storytelling or gamification elements to stimulate interest when students feel bored, and adding challenging tasks to promote in-depth learning when they are excited. This flexible and changeable teaching method helps to stimulate students' interest in learning, cultivate their autonomous learning ability, and thus promote students' all-round development. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 Flowchart of an AI-based adaptive education approach. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] Embodiment 1

[0051] Install high-definition cameras and microphones in classrooms or learning areas to ensure that students' facial expressions and voice intonations can be clearly captured, and the data captured by the cameras and microphones can be analyzed in real time to identify students' basic emotional states. The system can accurately capture students' micro-expression changes and provide precise emotional feedback for teaching strategy adjustments. The identified emotional states are fed back to the teaching strategy adjustment module in real time so that the system can dynamically adjust teaching content and methods according to emotional changes.

[0052] The system automatically collects students' test scores, homework completion, and answering speed learning data to fully reflect students' learning progress and ability level. It cleans, denoises, and standardizes the collected data to ensure data quality. It uses algorithm models to evaluate students' learning performance and predict the difficulties and interests that students may encounter in mathematics learning. The formula for predicting difficulties is:

[0053] Where: D represents the predicted difficulty level, A represents the average speed of answering questions, T i represents the speed of answering the i-th question, represents the average speed of all answers, α is the weight of the speed of answering questions on the prediction of difficulty, S represents the standard deviation of learning performance, C j represents the grade of the jth test or assignment, represents the average score, β is the weight of the influence of learning performance on the prediction of difficulty;

[0054] Formula for predicting points of interest:

[0055] Where: I represents the predicted interest point degree, A, T i , α has the same meaning as above, S, C j , The meaning of β is the same as above. F is an additional factor, which indicates the proportion of time students spend on independent learning on a specific subject or topic. γ is the weight of the impact of answering speed on the prediction of interest points. δ is the weight of the impact of academic performance on the prediction of interest points. ∈ is the weight of the impact of independent learning time on the prediction of interest points. The model can learn the complex relationship between student performance and various factors based on historical data, thereby improving the accuracy of prediction.

[0056] The system intelligently recommends and adjusts teaching strategies based on students' emotional states and learning performance evaluation results. For example, when it detects that students are confused, the system not only provides detailed graphical explanations or video tutorials, but also helps students consolidate knowledge points through interactive questions and answers. When students are bored, the system introduces math competitions and problem-solving gamification elements to stimulate students' interest and participation in learning. The system monitors students' learning performance and emotional changes in real time. Once it is found that the teaching strategy does not achieve the expected results, it will immediately adjust it and try new methods. At the same time, it collects students' feedback after using different teaching strategies for continuous optimization of teaching strategies.

[0057] Collect students' emotional state data in the emotional intelligence monitoring module, learning data in the learning performance evaluation and prediction module, and personal preference information obtained through questionnaires and interviews, use machine learning algorithms to identify the correlation and pattern between students' emotional state, learning performance and personal preferences, and build a generation model for personalized learning paths. This model can dynamically generate personalized learning paths that include basic consolidation, advanced challenges and interest expansion content based on students' current status and personal preferences, and present the generated personalized learning paths to students in an intuitive way, allowing students to independently choose learning content and learning order within a certain range. At the same time, the system provides necessary guidance and suggestions to help students make reasonable choices.

[0058] The system regularly collects students' learning outcomes data and emotional state data, as well as students' subjective feedback, conducts big data analysis on the collected data, identifies the correlation pattern between learning outcomes and emotional states, analyzes common problems and personalized needs in students' feedback, and establishes a feedback loop of students' learning outcomes and emotional state changes based on the data analysis results. When it is found that learning outcomes have declined or the emotional state is not good, the feedback mechanism is triggered, the bottlenecks and deficiencies are analyzed, and the teaching strategies and learning paths are adjusted accordingly. At the same time, the system is optimized and upgraded as necessary, including algorithm improvements, system architecture adjustments, and user interface optimization, to adapt to the ever-changing student learning needs and improve the efficiency and quality of education and teaching.

[0059] Embodiment 2

[0060] A combination of high-definition cameras and microphones is deployed next to students' reading materials to ensure that students' reading expressions and voice changes can be clearly captured. The camera records the student's facial video stream, and extracts key facial features through frame extraction and face detection. The microphone records the student's voice data, and after denoising, enhancement, and framing, the acoustic features and voice emotion features are extracted. The extracted facial expression features and voice features are input into a pre-trained deep learning model, and the model outputs the student's current emotional state.

[0061] Integrate students' reading comprehension test scores, reading speed, and answer accuracy data to form a comprehensive learning performance evaluation and prediction system. Use algorithm models to predict students' comprehension difficulty and interest points for different types of articles based on historical data and learning behavior characteristics. The formula for predicting difficulty points is:

[0062] Where: D represents the predicted difficulty level, A represents the average speed of answering questions, T i represents the speed of answering the i-th question, represents the average speed of all answers, α is the weight of the speed of answering questions on the prediction of difficulty, S represents the standard deviation of learning performance, C j represents the grade of the jth test or assignment, represents the average score, β is the weight of the influence of learning performance on the prediction of difficulty;

[0063] Formula for predicting points of interest:

[0064] Where: I represents the predicted interest point degree, A, T i , α has the same meaning as above, S, C j , The meaning of β is the same as above. F is an additional factor, which indicates the proportion of time students spend on independent learning on a specific subject or topic. γ is the weight of the impact of answering speed on the prediction of interest points. δ is the weight of the impact of academic performance on the prediction of interest points. ∈ is the weight of the impact of independent learning time on the prediction of interest points. Special attention is paid to students' mastery of complex sentences, new words, and cultural background.

[0065] Based on the output results of the emotional intelligence monitoring module and the learning performance evaluation and prediction module, the system intelligently recommends and adjusts reading training strategies. The system automatically identifies and marks the key sentences and difficult vocabulary in the article, provides detailed analysis and vocabulary expansion, and helps students overcome reading difficulties. When students’ fatigue is detected, short and relaxing essays or music clips are recommended to help students relieve fatigue and refocus their attention. When students show a high degree of concentration and interest, higher-level reading materials are recommended, such as excerpts from classic literary works and current affairs commentary, to promote in-depth learning.

[0066] Through questionnaires and system interactive records, we collect students' reading preferences, learning style and personal information, and integrate their emotional reactions, learning performance and personal preferences to dynamically generate personalized reading training plans and materials. We provide articles of different difficulty levels, supporting exercises and extended reading resources, and encourage students to challenge and choose according to their actual situation.

[0067] Through regular questionnaires, interviews and built-in feedback tools in the system, we collect students’ satisfaction and suggestions on reading materials and teaching strategies. Based on student feedback and data analysis results, we continuously optimize the selection of reading materials and the adjustment of teaching strategies. At the same time, we regularly evaluate the accuracy of the emotional intelligence monitoring module to ensure the accuracy of emotional feedback, thereby continuously improving the effectiveness of reading training and student satisfaction.

[0068] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. An adaptive education method based on artificial intelligence, characterized in that: The specific steps of this method are: S100, emotional intelligence monitoring module: uses high-definition cameras and microphones to capture students' facial expressions, voice intonation and micro-expression changes, analyzes these non-verbal signals in real time, and identifies students' basic emotional state; S200, learning performance evaluation and prediction module: Combined with academic performance and answering speed data, it comprehensively evaluates students' learning progress and ability level, and predicts students' difficulties and interests through algorithm models; S300, Adaptive Teaching Strategy Adjustment Module: Based on the output results of the emotional intelligence monitoring module and the learning performance evaluation and prediction module, the module uses an algorithm model to predict the difficulties and interests encountered by students, intelligently recommends and adjusts teaching strategies, and stimulates interest through storytelling, gamification elements or humorous explanations when boredom is detected. When students are found to be in an excited state, challenging tasks can be added to promote in-depth learning; S400, Personalized Learning Path Planning: Dynamically generates personalized learning paths based on students’ emotional state, learning performance, and personal preferences, allowing students to independently choose learning content within a certain range, enhancing learning autonomy and participation; S500, Feedback and Optimization Mechanism: Establish a feedback loop for students’ learning outcomes, continuously optimize teaching strategies and learning paths, regularly collect student feedback, combine big data analysis, and continuously adjust and improve system performance.

2. The adaptive education method based on artificial intelligence according to claim 1, characterized in that: The specific steps of S100 are: S101, start the high-definition camera and microphone, the camera starts to record the student's facial video stream, and the microphone simultaneously records the student's voice data, which are captured in the form of a continuous stream and transmitted to the processing unit, and the video stream captured by the camera is pre-processed, including frame extraction, face detection and tracking, facial feature point positioning steps, and key facial features are extracted; S102, performing denoising, enhancement, and framing operations on the voice data captured by the microphone for subsequent emotion analysis. Framing is to divide the continuous voice signal into a series of short time frames, and calculate the expression-related geometric features and texture features based on the facial feature points extracted in the preprocessing stage. These features can reflect different facial expressions, and extract acoustic features and more advanced voice emotion features from the preprocessed voice frames; S103, input the extracted facial expression features and speech features into a pre-trained model, and output the probability distribution of one or more emotion categories by analyzing the input features. These emotion categories usually include basic emotional states. According to the output of the model, the student's current emotional state is determined, and the identified emotional state is passed as input to the adaptive teaching strategy adjustment module, so as to dynamically adjust the teaching strategy according to the emotional feedback, collect the students' emotion recognition results and actual teaching effect feedback during use, so as to continuously optimize the accuracy and robustness of the model, and regularly evaluate the performance of the emotional intelligence monitoring module.

3. The adaptive education method based on artificial intelligence according to claim 1, characterized in that: The specific steps of S200 are: S201, clarify which data sources will be used for the evaluation, including but not limited to students' test scores, homework scores, answering speed, class participation records, after-class review time, and online learning behavior logs, remove duplicate data, process missing values, and correct any obvious errors, organize the data into a format suitable for analysis, calculate the proportion of completed courses and learning rate based on students' learning history and current completion status, calculate the average score, standard deviation, and ranking percentage quantitative indicators based on test scores and homework scores, analyze students' answering time data, calculate the average answering time, and the answering time distribution of each type of question or knowledge point, select algorithm models based on the complexity of the problem and the characteristics of the data, predict difficult points and points of interest, divide the data set into training set, validation set, and test set, with a ratio of 70%, 15%, and 15%, use the training set data to train the model, adjust the model parameters to optimize performance, use the validation set to evaluate the performance of the model, and adjust the model according to the evaluation results; S202, based on the results of the validation set, perform feature selection and model parameter optimization operations to improve the generalization ability of the model, use the trained model to predict the test set or newly collected data, evaluate the difficulties and points of interest that students may encounter, analyze the prediction results, explain which factors lead to specific prediction results, and which features have the greatest impact on the prediction results. Feedback the prediction results to teachers or students, collect feedback, and adjust or optimize the model as needed, deploy the learning performance evaluation and prediction module to the actual learning management system, continuously monitor the performance of the model, regularly check the accuracy of the prediction results, and update the model or data source as needed to collect user feedback, and continuously optimize the functions of the evaluation module and user experience.

4. The adaptive education method based on artificial intelligence according to claim 3, characterized in that: The formula for predicting difficulty points and interest points in S201 is: Prediction difficulty formula: Where: D represents the predicted difficulty level, A represents the average speed of answering questions, T i represents the speed of answering the i-th question, T represents the average speed of all answering questions, α is the weight of the influence of answering speed on difficulty prediction, S represents the standard deviation of learning performance, C j represents the grade of the jth test or assignment, represents the average score, β is the weight of the influence of learning performance on the prediction of difficulty; Formula for predicting points of interest: Where: I represents the predicted interest point degree, A, T i , α has the same meaning as above, S, C j , The meaning of β is the same as above. F is an additional factor, which indicates the proportion of time students spend on independent learning on a specific subject or topic. γ is the weight of the impact of answering speed on the prediction of interest points. δ is the weight of the impact of academic performance on the prediction of interest points. ∈ is the weight of the impact of independent learning time on the prediction of interest points.

5. The adaptive education method based on artificial intelligence according to claim 1, characterized in that: The specific steps of S300 are: S301, obtaining students' real-time emotional data and learning performance data from the emotional intelligence monitoring and prediction module and the learning performance evaluation and prediction module, cleaning the collected data, removing noise and outliers, integrating the emotional data and learning performance data into a unified database, building an algorithm model, predicting the difficulties and points of interest that students may encounter in the future learning process, extracting key features from the integrated data, including but not limited to learning progress, error rate, and emotional changes, as inputs to the prediction model, running the prediction model, and generating a prediction report on the difficulties and points of interest in the students' future learning; S302, match the corresponding teaching strategies according to the difficulties and points of interest in the prediction report. For the predicted difficulties, more detailed explanations, additional exercises or one-on-one tutoring can be recommended. For the points of interest, related extended reading or practical activities can be recommended. Combined with the results of emotional intelligence monitoring, the teaching strategies are adjusted to be emotionally adaptive. When students show boredom, story-based learning, gamification elements or humorous explanations are recommended to stimulate interest. When students are in an excited state, challenging tasks are added to promote deep learning. The recommended teaching strategies are deployed to the teaching platform or learning environment. During the teaching process, students' learning status and emotional changes are continuously monitored, and an instant feedback mechanism is established to allow students and teachers to provide feedback on the effectiveness of the teaching strategies. These feedbacks can be used as a basis for subsequent adjustments to the teaching strategies.

6. The adaptive education method based on artificial intelligence according to claim 5, characterized in that: The specific steps of S300 also include: S303, after the teaching strategy has been implemented for a period of time, evaluate its impact on students' learning outcomes and emotional states. This can be achieved by comparing the learning performance data and emotional data before and after implementation. According to the effect evaluation results, iteratively optimize the teaching strategy. If the effect is significant or not, it is necessary to re-analyze the students' difficulties, interests and emotional states and adjust the teaching strategy.

7. The adaptive education method based on artificial intelligence according to claim 1, characterized in that: The specific steps of S400 are: S401, collect students' emotional state data in the emotional intelligence monitoring module, including basic emotional state and its changing trend, obtain students' learning progress, ability level, answering speed, and predicted difficulty and interest points from the learning performance evaluation and prediction module, and collect students' personal preference information through questionnaires, interviews or system interaction records, including but limited to learning style, interest areas, and learning speed preference; S402, comprehensively analyzing the collected data, using machine learning to identify the correlation and pattern between students' emotional states, learning performance and personal preferences, and constructing a generative model of personalized learning paths, which can predict the learning path based on the students' current states and personal preferences; S403, dynamically generate personalized learning paths based on the results of data analysis and modeling, which includes selecting learning content that suits students’ current emotional state and learning progress, as well as incorporating learning activities and resources that suit students’ personal preferences. The learning path should have a certain degree of flexibility and be able to be dynamically adjusted based on students’ real-time feedback and learning progress.

8. The adaptive education method based on artificial intelligence according to claim 7, characterized in that: The specific steps of S400 also include: S404, presenting the generated personalized learning path to students in an intuitive manner, allowing students to independently select learning content and learning sequence within a certain range to enhance learning autonomy and sense of participation. At the same time, providing necessary guidance and suggestions to help students make reasonable choices; S405, students study according to the personalized learning path. The system continuously monitors the students' emotional state and learning performance during the learning process, and collects students' feedback data during the learning process for subsequent optimization and adjustment of the learning path; S406, based on students' feedback data and new learning data, regularly evaluate the difficulties and points of interest encountered in the learning path, and optimize and iterate the learning path according to the evaluation results.

9. The adaptive education method based on artificial intelligence according to claim 1, characterized in that: The specific steps of S500 are: S501, real-time monitoring and collection of students’ learning performance data, including academic performance, learning progress, and task completion, and simultaneous collection of students’ emotional state data, derived from the output of the emotional intelligence monitoring module, and regular collection of students’ subjective feedback through questionnaires, interviews, or system-built-in feedback tools, including learning experience and satisfaction with teaching strategies; S502, conduct big data analysis on the collected learning achievement data and emotional state data, identify the correlation pattern between learning achievement and emotional state, analyze student feedback, summarize common problems and personalized needs, use machine learning algorithms to extract key indicators from the data, and evaluate and predict the effectiveness of teaching strategies and learning paths; S503, based on the data analysis results, establish a feedback loop of student learning outcomes and emotional state changes. When it is found that the learning outcomes decline or the emotional state is not good, trigger the feedback mechanism, analyze the bottlenecks and deficiencies in the feedback loop, and identify the teaching elements or learning paths that need to be adjusted; S504, dynamically adjust the teaching strategy based on the results of the feedback loop. When it is detected that students have a poor grasp of a certain knowledge point, adjust the depth of the teaching content or the teaching method, optimize the learning path, dynamically adjust the difficulty and sequence of learning tasks based on the difficulties and interests encountered by students, and consider introducing new teaching resources or tools.

10. The adaptive education method based on artificial intelligence according to claim 9, characterized in that: The specific steps of S500 also include: S505, regularly evaluate the overall performance of the system, including data processing capabilities, model prediction accuracy, and user feedback response speed. Based on the evaluation results, make necessary optimization and upgrades to the system, including algorithm improvement, system architecture adjustment, and user interface optimization. Continuously iterate and improve the system to adapt to the ever-changing learning needs of students and improve the efficiency and quality of education and teaching. S506, maintain continuous monitoring of system operation, continuously collect new student data for subsequent analysis and optimization, regularly iterate the system, introduce new technologies and new concepts, and improve the system's intelligence level and user satisfaction.

Images