English personalized learning recommendation method and system based on big data

Abstract

The invention relates to the technical field of personalized learning, in particular to an English personalized learning recommendation method and system based on big data, and the method comprises the following steps: collecting interactive data of a learner on an English learning platform, sorting the number of login times, testing submission conditions and task completion data, carrying out the statistical analysis of the data according to a sorting result, and obtaining a recommendation result. And recording a relationship between task completion degree and time, and constructing an English learning behavior mode index. According to the invention, the emotion of the learner is captured in real time through the integrated facial recognition and voice analysis technology, the teaching content is dynamically adjusted according to the emotion data set, unprecedented adaptability is provided, the teaching strategy is allowed to respond to the emotion change of the learner in time, and the teaching difficulty and content are accurately adjusted in combination with refined behavior pattern analysis. The method is advantaged in that alignment with the practical ability and demand of the learner is guaranteed, through instant analysis of behavior and emotion feedback, correlation and effectiveness of teaching contents are improved, and efficiency and effect of the learning process are greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of personalized learning technology, and in particular to a method and system for recommending personalized English learning based on big data. Background Art

[0002] The field of personalized learning technology primarily involves customizing educational content and learning paths based on learners' individual interests, learning speed, knowledge background, and behavioral habits through algorithms and big data analysis. Machine learning models are used to analyze learners' behaviors and performance to provide resources and activities that best suit their learning needs. Personalized learning platforms can track learning progress, adjust course difficulty, and even recommend appropriate learning materials, thereby optimizing learning outcomes and improving learning efficiency.

[0003] Among them, the personalized English learning recommendation method refers to the use of personalized learning technology to recommend the most suitable English learning materials and activities for learners. The purpose is to analyze the learners' learning habits and progress and automatically recommend specific content that can improve English proficiency, such as vocabulary, grammar, listening and speaking exercises, etc., which can help learners effectively improve their English level, while saving learning time and making the learning process more efficient and targeted.

[0004] While existing personalized learning technologies provide customized learning paths based on learners' historical data, they often overlook the importance of learners' real-time emotions and immediate feedback, resulting in unresponsive adjustments to teaching content and an inability to adapt to learners' dynamic needs. The lack of the ability to instantly analyze emotions and behaviors limits the immediate updating of teaching strategies and the timely optimization of teaching content. Furthermore, existing technologies rely on preset learning models and paths, lacking flexibility and difficulty responding to rapid changes in learners' behavior and emotions. This often leads to poor teaching results and low learning efficiency, preventing learners from receiving the learning support that best meets their current needs, which in turn affects the depth and breadth of personalized learning. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the existing technology and propose an English personalized learning recommendation method based on big data.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for personalized English learning recommendation based on big data, comprising the following steps: S1: Collect learners' interaction data on the English learning platform, organize login times, test submission status, and task completion data, conduct statistical analysis based on the organized results, record the relationship between task completion and time, and construct English learning behavior pattern indicators; S2: Collect facial images and voice information through cameras and microphones, perform facial recognition and voice analysis, monitor learners' facial expressions and voice tones when watching English teaching videos and doing oral practice, extract key emotional indicators, and generate an English learning emotion dataset; S3: Based on the English learning behavior pattern indicators, cluster analysis is performed to analyze learners' test frequency and course interaction patterns. Based on the analysis results and combined with the English learning emotion dataset, targeted adjustment points for teaching content are determined, and the course difficulty is adjusted to match the learners' actual needs and emotional state, and personalized English learning recommendation content is formulated; S4: Apply the personalized English learning recommendation content to monitor learners' reactions to the adjusted English teaching content in real time, collect new behavioral and emotional data, evaluate the improvement effect of teaching strategies based on the collected results and compare them with the original data, continuously optimize the English teaching content, and form an optimized English teaching plan.

[0007] As a further solution of the present invention, the steps of organizing the interactive data are as follows: S111: Call the English learning platform database through the API to collect learners' login, test submission, and task completion events within the target time range, extract the timestamp, user ID, and event type, and obtain the original interaction dataset; S112: Based on the original interactive data set, perform data cleaning to identify and remove abnormal data, including duplicate login information and unmarked test submissions, to generate a cleaned interactive data set; S113: Performing database query and aggregation analysis based on the cleaned interactive data set to obtain a sorted learner behavior data set.

[0008] As a further solution of the present invention, the steps for obtaining the English learning behavior pattern indicator are: S121: performing statistical analysis based on the organized learner behavior data set to reveal basic distribution characteristics of the data and obtain a preliminary data overview; S122: Based on the preliminary data profile, perform regression analysis using the formula: ; Calculating task completion , generate regression analysis results, where Is the basic task completion rate, represents the nonlinear influence coefficient of time, is the error term, Represents the number of days; S123: Using the regression analysis results, perform factor analysis to extract key influencing factors from the login frequency, test submission frequency, and task completion data to construct an English learning behavior pattern indicator.

[0009] As a further solution of the present invention, the steps for obtaining the English learning emotion dataset are: S211: Deploy cameras and microphones to collect learners' facial images and voice signals in real time while they are learning English, generating real-time audio and video data streams; S212: Processing the real-time audio and video data stream, extracting facial feature data from the video, extracting pitch frequency data from the audio, and generating facial feature data and pitch frequency data; S213: Analyze the facial feature data and the tone frequency data, identify and classify the learner's emotional state, and generate an English learning emotion dataset.

[0010] As a further embodiment of the present invention, the steps for analyzing the learner test frequency and course interaction pattern are as follows: S311: Based on the English learning behavior pattern indicators, collect the learner's login frequency, learning time, and course completion data to establish a behavior pattern data set; S312: performing cluster analysis on the behavior pattern data set, identifying different learner groups based on learner behavior characteristics, and generating cluster analysis results; S313: Analyze the cluster analysis results, extract and compare the test frequencies and course interaction patterns of different learner groups, analyze the behavioral characteristics of the different groups, and generate analysis records of learner test frequencies and course interaction patterns.

[0011] As a further solution of the present invention, the steps for obtaining the personalized English learning recommendation content are: S321: Based on the learner test frequency and course interaction pattern analysis records and the English learning emotion dataset, perform data analysis on the emotion logs and feelings survey results of each learner group to identify the emotional states of different learners in English learning activities and generate emotion feature analysis results; S322: Analyze the emotion feature analysis results, compare the emotional needs of the different groups with the current teaching settings, evaluate the degree of match between the existing teaching content and the learners' emotional states, mark the teaching content that needs to be adjusted, and generate teaching content adjustment points; S323: Based on the teaching content adjustment points, a customized English learning course is designed for learner groups with different emotional and behavioral characteristics, a corresponding teaching rhythm and course depth are selected, and personalized English learning recommendation content is generated.

[0012] As a further solution of the present invention, the steps for collecting the new behavior and emotion data are: S411: Applying the personalized English learning recommendation content, collecting learner feedback data on the recommended content in real time, and obtaining learner behavior logs; S412: Extracting interaction data from the learner behavior log, including click rate, number of swipes, and dwell time, and processing the data in combination with learner feedback to obtain comprehensive behavior data; S413: Based on the comprehensive behavioral data, analyze the learner's video or audio response when exposed to the recommended content, extract emotional features, and output a behavioral emotion feedback data set.

[0013] As a further solution of the present invention, the steps for obtaining the optimized English teaching plan are: S421: Integrate the behavioral and emotional feedback dataset with the original learning data, including initial test scores and pre-course questionnaire results, compare the data differences before and after the teaching content adjustment, and obtain learner performance difference analysis results; S422: Based on the learner performance difference analysis results, use the formula: ; Calculate the average improvement in learners' grades , identify the successful elements and content that need improvement in the teaching strategy, and generate teaching strategy optimization records, including, Represents the total number of participating learners, and represent the test scores before and after the adjustment of teaching strategies; S423: Conduct a series of strategy discussions based on the teaching strategy optimization record, iteratively update the English teaching content and methods in combination with practical feedback, and form an optimized English teaching plan.

[0014] Big data-based personalized English learning recommendation system, including: The data cleaning and analysis module calls the English learning platform database, collects the original interactive data set, identifies and removes abnormal data, performs aggregate analysis, and obtains the organized learner behavior data set; The behavior pattern building module performs statistical analysis based on the organized learner behavior data set to reveal the basic distribution characteristics of the data, performs regression analysis to calculate task completion, extracts key influencing factors, and constructs English learning behavior pattern indicators; The emotion data collection module deploys cameras and microphones to collect learners' facial images and voice signals in real time while they are learning English. It extracts facial feature data from the video and pitch frequency data from the audio, identifies and classifies the learners' emotional states, and generates an English learning emotion dataset. The behavior characteristic analysis module establishes learner behavior pattern data based on the English learning behavior pattern indicators, identifies different learner groups based on the learner behavior characteristics, extracts and compares the test frequencies and course interaction patterns of the different learner groups, analyzes the behavior characteristics of the different groups, and generates analysis records of learner test frequencies and course interaction patterns; The course customization module analyzes the learners' test frequency and course interaction patterns and the English learning emotion dataset to identify the emotional states of different learners during English learning activities, evaluate the degree of match between existing teaching content and learners' emotional states, design customized English learning courses, and generate personalized English learning recommendation content; The teaching content iteration module applies the personalized English learning recommendation content, collects learners' feedback data on the recommended content in real time, extracts interaction data and combines it with learners' feedback, extracts emotional characteristics, compares the data differences before and after the teaching content adjustment, calculates the learners' average score improvement, and iteratively updates the English teaching content and methods based on practical feedback to form an optimized English teaching plan.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are: In this invention, learners' emotions are captured in real time through integrated facial recognition and voice analysis technology, and teaching content is dynamically adjusted according to the emotional data set, providing unprecedented adaptability, allowing teaching strategies to respond to learners' emotional changes in a timely manner, combined with refined behavioral pattern analysis, to accurately adjust teaching difficulty and content to ensure alignment with learners' actual abilities and needs. Through instant analysis of behavioral and emotional feedback, the relevance and effectiveness of teaching content are improved, greatly enhancing the efficiency and effectiveness of the learning process. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of the main steps of the present invention; Figure 2 A flowchart for organizing interactive data of the present invention; Figure 3 This is a flow chart for obtaining English learning behavior pattern indicators of the present invention; Figure 4 This is a flowchart for obtaining an English learning emotion dataset of the present invention; Figure 5 A flowchart for analyzing learner test frequency and course interaction patterns of the present invention; Figure 6 This is a flowchart for obtaining personalized English learning recommendation content of the present invention; Figure 7 This is a flowchart for collecting new behavior and emotion data of the present invention; Figure 8This is a flow chart for obtaining the optimized English teaching plan of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0019] See also Figure 1 ,The English personalized learning recommendation method based on big data includes the following steps: S1: Collect learners' interaction data on the English learning platform, organize login times, test submission status, and task completion data, conduct statistical analysis based on the organized results, record the relationship between task completion and time, and construct English learning behavior pattern indicators; S2: Collect facial images and voice information through cameras and microphones, perform facial recognition and voice analysis, monitor learners' facial expressions and voice tones when watching English teaching videos and doing oral practice, extract key emotional indicators, and generate an English learning emotion dataset; S3: Based on English learning behavior pattern indicators, we conduct cluster analysis to analyze learners' test frequency and course interaction patterns. Based on the analysis results and combined with the English learning emotion dataset, we determine targeted adjustments to the teaching content, adjust the course difficulty to match learners' actual needs and emotional state, and develop personalized English learning recommendations. S4: Apply personalized English learning recommendation content, monitor learners' reactions to the adjusted English teaching content in real time, collect new behavioral and emotional data, compare the collected results with the original data to evaluate the improvement effect of teaching strategies, continuously optimize English teaching content, and form an optimized English teaching plan.

[0020] English learning behavior pattern indicators include interaction frequency, completion speed and accuracy rate; English learning emotion datasets include expression recognition parameters, voice tonality indicators and emotion intensity scores; personalized English learning recommendation content includes interactive task types, teaching material difficulty levels and emotional feedback adaptation analysis results; the optimized English teaching plan includes teaching content adjustment measures, learning effect monitoring tools and continuous feedback mechanisms.

[0021] See also Figure 2 ,The steps for organizing the interactive data are: S111: Call the English learning platform database through the API to collect learners' login, test submission, and task completion events within the target time range, extract the timestamp, user ID, and event type, and obtain the original interaction dataset; The English learning platform's database is called through the API interface. The system administrator first ensures that all API calls follow the strictest data security standards, including the use of OAuth for authentication and authorization to protect data from unauthorized access. The collection process involves obtaining learners' login information, test submission records, and task completion details within a specific time range from the database. Data extraction uses efficient SQL query statements, which are specially designed to process large amounts of data to ensure performance and response speed in high-concurrency environments. The original data set contains key information such as timestamps, user IDs, and event types, which will provide a basis for subsequent data analysis, thereby supporting more complex data operations such as data merging, segmentation, and reorganization. The purpose is to generate a complete and accurate user interaction data set, laying the foundation for the next step of data cleaning and analysis.

[0022] S112: Based on the original interactive data set, perform data cleaning to identify and remove abnormal data, including duplicate login information and unmarked test submissions, to generate a cleaned interactive data set; After obtaining the original interactive data set, data cleaning technology is used for in-depth analysis to identify and remove abnormal records in the data set, such as duplicate login information or incomplete test submission records. The data cleaning process includes multiple stages. The first is to use automated scripts to scan the obvious errors in the data set, such as date and time stamp errors. Then, outliers in the data are identified through algorithms, such as using standard deviation and interquartile range (IQR) methods to determine and exclude outliers. For missing values, multiple interpolation methods are used to estimate missing values based on adjacent data to maintain data integrity. The data formatting stage involves standardizing all data into a unified format to facilitate subsequent data processing and analysis. The final result is a cleaned, high-quality data set.

[0023] S113: Based on the cleaned interactive data set, perform database query and aggregate analysis to obtain a collated learner behavior data set; Based on the cleaned interactive data set, detailed behavioral analysis is conducted. Data analysts use complex SQL query language to deeply explore the data. During the query process, not only the number of logins of each user is calculated, but also the frequency of each test submission and the accuracy of task completion are analyzed in detail. Each data is calculated through aggregate functions such as SUM and COUNT to ensure that the participation and learning efficiency of each user can be measured. In addition, conditional filtering is applied to distinguish different types of user behaviors, such as distinguishing users who log in frequently but submit few tests, and users who complete many tasks but log in few times. This helps educators understand the effects of different learning paths and provides data support for optimizing course design and learning resource allocation. The generated aggregated data not only reflects the user's activity, but also provides key indicators for further educational research and the establishment of behavioral patterns.

[0024] See also Figure 3 , the steps to obtain English learning behavior pattern indicators are: S121: Based on the organized learner behavior data set, statistical analysis is performed to reveal the basic distribution characteristics of the data and obtain a preliminary data overview; Connect to the data warehouse of the English learning platform and call the organized learner behavior dataset, which covers the number of logins, test submission records and task completion data of each user; use data analysis software to perform descriptive statistics, which involves calculating the mean, median, variance and standard deviation of various types of data, aiming to reveal the basic distribution characteristics and statistical summary of learner activities. The statistical results not only lay the foundation for subsequent in-depth analysis, but also help identify outliers and trends in the dataset to adjust the educational intervention strategy and content design of the learning platform to make it more in line with the actual needs and learning habits of learners. In this way, researchers can more accurately predict and improve learners' learning outcomes and the overall educational effect of the platform.

[0025] S122: Based on the preliminary data profile, perform regression analysis using the formula, ; Calculating task completion , generate regression analysis results, where Is the basic task completion rate, represents the nonlinear influence coefficient of time, is the error term, Represents the number of days; set up , , and assuming that the task completion degree at 10 days is to be calculated, substitute the specific values: ; Through numerical calculation, if we assume is approximately 0, then: ; The results show that at 10 days, the task completion rate is expected to be 59.5%, showing the actual impact of time on task completion.

[0026] S123: Using the regression analysis results, perform factor analysis to extract key influencing factors from the login frequency, test submission frequency, and task completion data to construct an indicator of English learning behavior patterns; Using the results of regression analysis, factor analysis was performed to extract the main influencing factors from the login frequency, test submission frequency and task completion data. The core goal is to integrate these factors to construct a comprehensive English learning behavior pattern indicator. The indicator will describe and predict learners' behavior patterns by comprehensively considering the mutual influence of different learning activities. The obtained behavior pattern indicators can help educators and platform developers better understand the specific behavior and learning motivation of each learner; the process uses statistical methods and machine learning techniques for data processing and model construction to ensure that effective information is extracted from large amounts of complex data, and the information is converted into practical operational guidelines that can be directly applied to teaching improvement and curriculum design. Through insights, the platform can provide learners with more personalized learning resources and support, thereby improving learning efficiency and effectiveness.

[0027] See also Figure 4 ,The steps to obtain the English learning emotion dataset are: S211: Deploy cameras and microphones to collect learners' facial images and voice signals in real time while they are learning English, generating real-time audio and video data streams; The process of capturing learners' faces and voices begins with the use of high-resolution cameras and high-sensitivity microphones to collect learners' facial expressions and voice data in real time during the English learning process. The data is first pre-processed by a digital signal processor, including noise reduction and enhancement, to ensure the clarity of the image and sound. Real-time video analysis technology is used to identify changes in the position of learners' facial features such as eyes, nose and mouth. At the same time, voice processing software analyzes the voice waveform and identifies key voice features such as pitch, volume and rhythm. Facial and voice data are analyzed synchronously to accurately capture learners' emotional reactions, such as happiness, sadness or tension, and generate a detailed real-time audio and video data stream, providing the necessary raw data for subsequent emotional analysis.

[0028] S212: Processing the real-time audio and video data stream, extracting facial feature data from the video, extracting pitch frequency data from the audio, and generating facial feature data and pitch frequency data; After obtaining the real-time audio and video data stream, the data stream is processed, including facial recognition and voice analysis. Frame analysis is performed on the video data, and machine learning technology is used to identify and track changes in the learner's facial expressions. Each recognized expression is associated with a specific emotional label. For example, a frown may indicate confusion or dissatisfaction. At the same time, spectral analysis is performed on the voice data to extract changes in pitch, volume, and speaking speed. Audio features are used to analyze the learner's emotional state. For example, a rise in pitch may indicate excitement or anxiety. By integrating the data, a comprehensive analysis report containing facial and voice features can be generated, thereby generating facial feature data and pitch frequency data.

[0029] S213: Analyze facial feature data and tone frequency data to identify and classify learners’ emotional states and generate an English learning emotion dataset; By analyzing the integrated facial feature data and tone frequency data, the model is based on artificial intelligence algorithms and can interpret and classify complex emotional expressions. For example, the model identifies emotions such as happiness or sadness by analyzing the smile intensity and eye opening degree in facial feature data, combined with the tone changes in the voice analysis results. In addition, the rhythm and pauses of the voice are used to further confirm the tension and stability of the emotion. The multi-dimensional recognition of emotions allows the model to accurately predict and classify the learner's emotional state and generate detailed emotion analysis results. The results integrate all data on facial expressions and voice tones to form a comprehensive English learning emotion dataset, providing educators with a powerful tool to adjust teaching methods and improve learners' learning experience.

[0030] See also Figure 5 ,The analysis steps of learners’ testing frequency and course interaction patterns are: S311: Based on the English learning behavior pattern indicators, collect learners' login frequency, learning time and course completion data to establish a behavior pattern dataset; When compiling indicators of English learning behavior patterns, we first aggregate data from multiple educational platforms, including learners' login times, learning duration, and course completion records. Through this data, we can capture learners' learning activities in different time periods. Next, we clean and preprocess the data, such as removing outliers and filling in missing data, to ensure data accuracy and availability. This ensures that the basic data for subsequent analysis is not only comprehensive but also accurate. The comprehensive data will support subsequent cluster analysis, providing reliable basic data for analyzing learners' test frequency and course interaction patterns, and generating a behavior pattern dataset.

[0031] S312: performing cluster analysis on the behavior pattern data set, identifying different learner groups based on learner behavior characteristics, and generating cluster analysis results; In the cluster analysis process, the K-means clustering algorithm is selected, which is suitable for processing large-scale data sets and can effectively discover patterns and trends in the data. Before clustering begins, the number of clusters, that is, the K value, is first determined, which is usually determined based on the characteristics of the data and business needs. Then, K data points are randomly selected as the initial cluster centers. Next, the distance from each data point to each cluster center is calculated, and the data points are assigned to the nearest cluster center. After that, the center point of each cluster is updated by calculating the mean of all points belonging to the cluster. This process is repeated, and each iteration optimizes the position of the cluster center until the change in the cluster center is very small or the preset number of iterations is reached, thereby ensuring that we obtain the best clustering results and generate cluster analysis results.

[0032] S313: Analyze the cluster analysis results, extract and compare the test frequencies and course interaction patterns of different learner groups, analyze the behavioral characteristics of different groups, and generate analysis records of learner test frequencies and course interaction patterns; After obtaining the clustering results, we conducted an in-depth analysis of the behavioral characteristics of each cluster group, focusing in particular on the test frequency and course interaction patterns of learners. By recording the test participation frequency and course activity of learners in each group in detail, and using statistical methods to compare and analyze the data, we found significant differences between groups. These differences help us understand the behavioral patterns of different learners. Based on the analysis results, we compiled an analysis record of learners' test frequency and course interaction patterns. The record not only summarizes the behavioral patterns of each group, but also puts forward targeted teaching adjustment suggestions, aiming to optimize the test frequency and enhance the interactivity of the course design, so as to better meet the needs of different learners, and ultimately serve as the basis for adjusting teaching content and strategies.

[0033] See also Figure 6 , the steps to obtain personalized English learning recommendation content are: S321: Based on the learners' test frequency and course interaction pattern analysis records and English learning emotion dataset, we analyze the emotion logs and feelings survey results of each learner group to identify the emotional states of different learners in English learning activities and generate emotion feature analysis results; In the process of constructing the English learning behavior and emotion dataset, we first collected data from multiple educational technology platforms, including but not limited to learners' login frequency, learning time, and emotional state records. The data underwent rigorous preprocessing, including removing inconsistent data points, filling missing values, and smoothing outliers to enhance the quality and usability of the dataset. Then, we used advanced statistical analysis techniques, such as factor analysis and cluster analysis, to identify potential learning patterns and emotional trends. This process not only helped to deeply understand the learners' behavior and emotional dynamics, but also provided data support for the formulation of more effective teaching strategies. Through these detailed data analyses, we successfully constructed a comprehensive and accurate analysis result of behavior and emotion characteristics.

[0034] S322: Analyze the results of the emotional feature analysis, compare the emotional needs of the different groups with the current teaching settings, evaluate the degree of match between the existing teaching content and the learners' emotional state, mark the teaching content that needs to be adjusted, and generate teaching content adjustment points; When evaluating the adaptability of teaching content and course design, we conducted an in-depth analysis of the results of emotional feature analysis, paying special attention to the impact of learners' emotional fluctuations on learning outcomes. By comprehensively considering learners' emotional data and behavioral manifestations, we identified multiple key teaching adjustment points. For example, for those learners who show anxiety or stress when learning, it is recommended to reduce the course difficulty and add more interactive elements to reduce their learning burden; for more active and engaged groups, the challenge can be appropriately increased, and the depth and breadth of the course content can be enhanced. Each adjustment is based on detailed data analysis to ensure that teaching more accurately corresponds to the actual needs of each learner, and ultimately form a comprehensive teaching content adjustment plan.

[0035] S323: Based on the adjustment points of teaching content, we design customized English learning courses for learner groups with different emotional and behavioral characteristics, select the corresponding teaching pace and course depth, and generate personalized English learning recommendation content; Based on the adjustment points of teaching content, exclusive courses are designed for learners with different emotional and learning behavior characteristics. For example, more inclusive and supportive course activities are designed for learners with large emotional fluctuations, and more independent and exploratory learning tasks are provided for learners with strong learning motivation. In addition, the teaching rhythm and depth are adjusted to ensure that every learner can learn in the most suitable environment. By implementing these personalized recommendations, not only the learning effect is improved, but also the learner's emotional state is optimized, making teaching more humane and effective. Ultimately, through continuous evaluation and optimization, a set of dynamically updated personalized teaching recommendation content is formed.

[0036] See also Figure 7 , the steps for collecting new behavior and emotion data are: S411: Apply personalized English learning recommendation content, collect learner feedback data on the recommended content in real time, and obtain learner behavior logs; Personalized English learning recommendation content is implemented in the teaching platform, and the implementation status is recorded in real time. The teaching materials selected and interacted with by learners after each login are captured by the logging system. The log includes the timestamps of when learners start and complete the course, the frequency of participating in online quizzes and their results, and the details of feedback submission. The data is initially screened to remove the time records of login and logout, and retain the data directly related to learning activities, thus forming a detailed record of learners' learning behaviors and preferences. This dataset provides basic information for subsequent behavioral analysis and learning recommendation adjustments.

[0037] S412: Extract interaction data from the learner behavior log, including click rate, number of swipes, and dwell time, and combine it with learner feedback to process the data and obtain comprehensive behavior data; In the teaching platform, learners' behavioral data are collected through front-end technology, including detailed page visit records, specific content interaction operations, and input feedback information. These data are first formatted to convert unstructured log information into structured data. Then, invalid and duplicate records are removed through a data cleaning program. Finally, data mining technology is used to analyze user behavior patterns, identify learners' preferences and learning efficiency for different teaching contents, and establish a more accurate learning portrait for learners. The comprehensive behavioral data obtained can more accurately reflect the learners' actual learning status and provide a basis for further personalization of teaching content.

[0038] S413: Based on the comprehensive behavioral data, analyze the learner's video or audio reactions when they are exposed to the recommended content, extract emotional features, and output a behavioral emotion feedback dataset; The platform monitors learners' facial expressions and voice changes in real time while they watch teaching videos. It collects data by connecting cameras and microphones, and analyzes learners' emotional responses through algorithms. From smiles and frowns to changes in speech speed, every subtle change in expression and voice is recorded and analyzed. The emotional analysis results are then integrated with behavioral data to form a comprehensive behavioral emotion feedback dataset. This data includes not only learners' behavioral patterns but also their emotional states, providing educators with a deep understanding framework, enabling them to better adjust teaching strategies to suit the emotions and learning needs of different learners.

[0039] See also Figure 8 , the steps to obtain the optimized English teaching plan are: S421: Integrate the behavioral and emotional feedback dataset with the original learning data, including initial test scores and pre-course questionnaire results, and compare the data differences before and after the teaching content adjustment to obtain the results of the learner performance difference analysis; The comprehensive behavioral and emotional data set is integrated with the original learning data on the teaching platform, mainly including the learners' initial test scores and the questionnaire survey results before the start of teaching. The data are compared to identify changes in learners' performance after receiving personalized teaching content, including changes in emotional responses. This comparison takes into account the baseline differences between different learners and ensures the consistency and comparability of the data through standardization. This process not only helps to identify which teaching strategies are effective, but also points out which ones may need further adjustment or improvement, thereby providing data support and decision-making basis for the optimization of teaching content. This comprehensive analysis result is the first step in optimizing teaching strategies and can clearly point to specific teaching links that need to be focused on for improvement in the future.

[0040] S422: Based on the results of the learner performance difference analysis, the formula is used. ; Calculate the average improvement in learners' grades , identify the successful elements and content that need improvement in the teaching strategy, and generate teaching strategy optimization records, including, Represents the total number of participating learners, and represent the test scores before and after the adjustment of teaching strategies; There are 10 students, and their average scores before and after adjustment are 75 and 85 respectively. Then: ; This indicates that the average score increased by 1 point, and the results can help verify the effectiveness of teaching strategies.

[0041] S423: Conduct a series of strategy discussions based on the teaching strategy optimization records. Combined with practical feedback, iteratively update the English teaching content and methods to form an optimized English teaching plan. Based on the teaching strategy optimization report, educational experts and the content development team organized a series of strategy discussions. The discussions and workshops were based on the latest educational theories and feedback collected from actual teaching. They analyzed which teaching methods were most effective in actual application and which content needed to be updated or modified. Through the collision of these collective wisdom, new teaching strategies and content were gradually formed and initially tested through small-scale pilot projects. The test results were confirmed through data analysis to show their actual impact on improving learning efficiency and student satisfaction. The process included collecting and analyzing students' academic performance, participation and feedback during the pilot period. All activities ensured the effectiveness and feasibility of the new strategies before large-scale implementation, and ultimately formed a fully verified and updated English teaching plan.

[0042] Big data-based personalized English learning recommendation system, including: The data cleaning and analysis module calls the English learning platform database, collects the original interactive data set, identifies and removes abnormal data, performs aggregate analysis, and obtains the organized learner behavior data set; The behavior model building module conducts statistical analysis based on the organized learner behavior data set to reveal the basic distribution characteristics of the data, performs regression analysis to calculate task completion, extracts key influencing factors, and constructs English learning behavior model indicators; The emotion data collection module deploys cameras and microphones to collect learners' facial images and voice signals in real time while they are learning English. It extracts facial feature data from the video and pitch frequency data from the audio, identifies and classifies the learners' emotional states, and generates an English learning emotion dataset. The behavior characteristic analysis module establishes learner behavior pattern data based on English learning behavior pattern indicators, identifies different learner groups based on their behavior characteristics, extracts and compares the test frequencies and course interaction patterns of different learner groups, analyzes the behavior characteristics of different groups, and generates analysis records of learner test frequencies and course interaction patterns; The course customization module analyzes learners' test frequency and course interaction patterns based on English learning emotion datasets. It identifies the emotional states of different learners during English learning activities, evaluates the degree of match between existing teaching content and learners' emotional states, designs customized English learning courses, and generates personalized English learning recommendations. The teaching content iteration module applies personalized English learning recommendation content, collects learners' feedback data on recommended content in real time, extracts interaction data and combines it with learners' feedback, extracts emotional characteristics, compares the data differences before and after the teaching content adjustment, calculates the average improvement in learners' scores, and iteratively updates English teaching content and methods based on practical feedback to form an optimized English teaching plan.

[0043] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A personalized English learning recommendation method based on big data, characterized by: The following steps are involved: Collect learners' interactive data on the English learning platform, organize login times, test submission status, and task completion data, conduct statistical analysis based on the results, record the relationship between task completion and time, and construct English learning behavior pattern indicators; Facial images and voice information are collected through cameras and microphones, and facial recognition and voice analysis are performed to monitor learners' facial expressions and voice tones while they watch English teaching videos and practice speaking. Key emotional indicators are extracted to generate an English learning emotion dataset. Based on the English learning behavior pattern indicators, cluster analysis is performed to analyze learners' test frequency and course interaction patterns. Based on the analysis results and combined with the English learning emotion dataset, targeted adjustment points for teaching content are determined, and the course difficulty is adjusted to match learners' actual needs and emotional states, and personalized English learning recommendation content is formulated; Apply the personalized English learning recommendation content, monitor learners' reactions to the adjusted English teaching content in real time, collect new behavioral and emotional data, compare the collected results with the original data to evaluate the improvement effect of the teaching strategy, continuously optimize the English teaching content, and form an optimized English teaching plan.

2. The method for personalized English learning recommendation based on big data according to claim 1, characterized in that: The steps for collating the interactive data are as follows: We call the English learning platform database through the API to collect learners' login, test submission, and task completion events within the target time range, extract timestamps, user IDs, and event types, and obtain the original interaction dataset. Based on the original interactive data set, data cleaning is performed to identify and remove abnormal data, including duplicate login information and unmarked completed test submissions, to generate a cleaned interactive data set; Based on the cleaned interactive data set, a database query is performed and aggregate analysis is performed to obtain a sorted learner behavior data set.

3. The method for personalized English learning recommendation based on big data according to claim 2, characterized in that: The steps for obtaining the English learning behavior pattern indicator are as follows: Based on the organized learner behavior data set, statistical analysis is performed to reveal the basic distribution characteristics of the data and obtain a preliminary data overview; Based on the preliminary data profile, a regression analysis was performed using the formula, ； Calculating task completion , generate regression analysis results, where Is the basic task completion rate, represents the nonlinear influence coefficient of time, is the error term, Represents the number of days; Using the regression analysis results, factor analysis was performed to extract key influencing factors from the login frequency, test submission frequency, and task completion data to construct an English learning behavior pattern indicator.

4. The method for personalized English learning recommendation based on big data according to claim 3, characterized in that: The steps for obtaining the English learning emotion dataset are as follows: Deploy cameras and microphones to collect learners' facial images and voice signals in real time during English learning, generating real-time audio and video data streams; Processing the real-time audio and video data stream, extracting facial feature data from the video, extracting pitch frequency data from the audio, and generating facial feature data and pitch frequency data; The facial feature data and tone frequency data are analyzed to identify and classify the learner's emotional state and generate an English learning emotion dataset.

5. The method for personalized English learning recommendation based on big data according to claim 3, characterized in that: The analysis steps of learner testing frequency and course interaction pattern are as follows: Based on the English learning behavior pattern indicators, data on learners' login frequency, learning time, and course completion are collected to establish a behavior pattern dataset; performing cluster analysis on the behavioral pattern data set, identifying different learner groups based on learner behavioral characteristics, and generating cluster analysis results; Analyze the cluster analysis results, extract and compare the test frequencies and course interaction patterns of different learner groups, analyze the behavioral characteristics of the different groups, and generate analysis records of learner test frequencies and course interaction patterns.

6. The method for personalized English learning recommendation based on big data according to claim 5, characterized in that: The steps for obtaining the personalized English learning recommendation content are as follows: Based on the learners' test frequency and course interaction pattern analysis records and the English learning emotion dataset, data analysis is performed on the emotion logs and feelings survey results of each learner group to identify the emotional states of different learners in English learning activities and generate emotion feature analysis results; Analyze the emotional characteristics analysis results, compare the emotional needs of different groups with the current teaching settings, evaluate the degree of match between the existing teaching content and the learners' emotional state, mark the teaching content that needs to be adjusted, and generate teaching content adjustment points; Based on the teaching content adjustment points, customized English learning courses are designed for learner groups with different emotional and behavioral characteristics, the corresponding teaching rhythm and course depth are selected, and personalized English learning recommendation content is generated.

7. The method for personalized English learning recommendation based on big data according to claim 6, characterized in that: The steps for collecting the new behavior and emotion data are as follows: Applying the personalized English learning recommendation content, collecting learners' feedback data on the recommended content in real time, and obtaining learners' behavior logs; Extracting interaction data from the learner behavior log, including click rate, number of swipes, and dwell time, and combining this with learner feedback to process the data and obtain comprehensive behavior data; Based on the comprehensive behavioral data, the learner's video or audio reaction when exposed to the recommended content is analyzed, emotional features are extracted, and a behavioral emotion feedback data set is output.

8. The method for personalized English learning recommendation based on big data according to claim 7, characterized in that: The steps for obtaining the optimized English teaching plan are: Integrate the behavioral and emotional feedback dataset with the original learning data, including initial test scores and pre-course questionnaire results, and compare the data differences before and after the teaching content adjustment to obtain the results of the learner performance difference analysis; According to the results of the learner performance difference analysis, the formula is used. ； Calculate the average improvement in learners' grades , identify the successful elements and content that need improvement in the teaching strategy, and generate teaching strategy optimization records, including, Represents the total number of participating learners, and represent the test scores before and after the adjustment of teaching strategies; Based on the teaching strategy optimization records, a series of strategy discussions are conducted, and the English teaching content and methods are iteratively updated in combination with practical feedback to form an optimized English teaching plan.

9. The personalized English learning recommendation system based on big data is characterized by: The system is used to execute the big data-based personalized English learning recommendation method according to any one of claims 1 to 7, comprising: The data cleaning and analysis module calls the English learning platform database, collects the original interactive data set, identifies and removes abnormal data, performs aggregate analysis, and obtains the organized learner behavior data set; The behavior pattern building module performs statistical analysis based on the organized learner behavior data set to reveal the basic distribution characteristics of the data, performs regression analysis to calculate task completion, extracts key influencing factors, and constructs English learning behavior pattern indicators; The emotion data collection module deploys cameras and microphones to collect learners' facial images and voice signals in real time while they are learning English. It extracts facial feature data from the video and pitch frequency data from the audio, identifies and classifies the learners' emotional states, and generates an English learning emotion dataset. The behavior characteristic analysis module establishes learner behavior pattern data based on the English learning behavior pattern indicators, identifies different learner groups based on the learner behavior characteristics, extracts and compares the test frequencies and course interaction patterns of the different learner groups, analyzes the behavior characteristics of the different groups, and generates analysis records of learner test frequencies and course interaction patterns; The course customization module analyzes the learners' test frequency and course interaction patterns and the English learning emotion dataset to identify the emotional states of different learners during English learning activities, evaluate the degree of match between existing teaching content and learners' emotional states, design customized English learning courses, and generate personalized English learning recommendation content; The teaching content iteration module applies the personalized English learning recommendation content, collects learners' feedback data on the recommended content in real time, extracts interaction data and combines it with learners' feedback, extracts emotional characteristics, compares the data differences before and after the teaching content adjustment, calculates the learners' average score improvement, and iteratively updates the English teaching content and methods based on practical feedback to form an optimized English teaching plan.

Images