College student learning behavior analysis system based on big data

Abstract

The invention belongs to application of a big data analysis technology in the field of education, discloses a college student learning behavior analysis system based on big data, and aims to provide personalized learning suggestions and education decision support by analyzing online learning data of students. The system specifically comprises a data acquisition module, a data processing module, a label module, a data storage module and a portrait generation module. According to the college student learning behavior analysis system based on the big data, an objective, real and accurate student portrait label system is constructed, so that the cognitive load brought to teachers and students by a large amount of data is reduced; the student group is deeply depicted, the generated portrait is high in accuracy, and corresponding decision support is provided for student education managers, so that student academic development is assisted, repeated and tedious work caused by manual labeling is reduced, a label system is periodically updated, and automatic updating of the portrait is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and in particular to a college student learning behavior analysis system based on big data. Background Art

[0002] With the increasing development of technologies such as artificial intelligence and big data, students' learning mode has gradually changed from the traditional classroom mode to the online mode. At present, many online learning platform achievements have been explored in the field of distance education at home and abroad, realizing fundamental changes in the way of knowledge dissemination through networking. Classroom education is no longer restricted by time and place, and ubiquitous learning is realized. However, from the data analysis of students using online learning platforms, it can be seen that there are the following shortcomings: First, a large number of learning resources cause students to be immersed in the environment of selecting resources and thus cannot accurately select course resources suitable for their learning characteristics, resulting in problems such as inability to focus on learning and poor learning results. Second, the online learning platform simply provides a large number of course resources, lacks tracking and evaluation of students' learning process, learning behavior, and learning results, and cannot analyze students' deep-dimensional learning information, so it is impossible to truly achieve closed-loop feedback to effectively assist students in learning. Therefore, on the basis of the existing online learning platform, it is an urgent task to be able to teach students in accordance with their aptitude and provide a personalized recommendation system for different types of students.

[0003] Student portrait is a user model established based on a series of actual data. It is derived from the concept of user portrait and is the use scenario of user portrait in education. By marking student labels in massive multi-source heterogeneous data, forming student portraits, evaluating students more comprehensively and accurately, using student portraits to adopt personalized teaching methods and other comprehensive understanding and guidance of students, thereby improving teaching quality, cultivating outstanding talents, enabling students to obtain personalized learning support services, and leading students to change in the direction they expect.

[0004] There are still some shortcomings in the field of student portraits: first, data mining algorithms are less used and have low accuracy, and a unified and feasible student labeling system has not yet been established, making it impossible to form a true and accurate student portrait; second, a large amount of data requires the combination of professional data processing knowledge to be deeply understood and cannot be fully used by managers; third, the current student portrait field simply copies technical research in the user portrait field, ignoring the limitations of data and forcibly portraying rich and comprehensive student portraits, which makes it impossible to provide targeted portrait descriptions, affecting the accuracy of student portraits. Summary of the invention

[0005] The purpose of the present invention is to provide a college student learning behavior analysis system based on big data, which reduces the cognitive load brought to teachers and students by a large amount of data by constructing an objective, true and accurate student portrait labeling system; deeply characterizes the student group, and the generated portrait has high accuracy, providing corresponding decision-making support for student education managers, thereby helping students' academic development, reducing the repetitive and tedious work caused by manual labeling, regularly updating the labeling system, and realizing automatic updating of portraits.

[0006] To achieve the above objectives, the present invention provides a college student learning behavior analysis system based on big data, including a data acquisition module, a data processing module, a label module, a data storage module, and a portrait generation module.

[0007] Preferably, based on the data acquisition module, learning-related data is obtained from the school's academic affairs system, online learning platform, and smart classroom system, and imported into the data processing module; wherein the learning-related data includes: student basic information, course information, online platform learning information, classroom behavior information, and academic performance information.

[0008] Preferably, the basic information of the student includes, but is not limited to: student ID, student name, student gender, date of birth, registration year, student grade, student major, and student type;

[0009] Course information includes but is not limited to: course code, course name, course year, course grade, course type, and course credits;

[0010] Online platform learning information includes but is not limited to: login time, platform online time, video content watched, resource content downloaded, number of resource downloads, video viewing time, number of questions answered, and question answering accuracy;

[0011] Classroom behavior information includes but is not limited to: attendance, class participation, class performance, class discipline, quality of interaction, completion of class tasks, and teacher evaluation;

[0012] Academic performance information includes but is not limited to: student GPA, course grades, credit accumulation, and performance trends;

[0013] In addition, learning-related data includes:

[0014] Type and frequency of learning resource access: the frequency and frequency of students accessing different types of learning resources;

[0015] Learning resource completion rate: the rate at which students complete various types of learning resources;

[0016] Homework submission: timeliness and quality of homework submitted by students;

[0017] Interactive area activities: students’ behavior of posting, replying to posts, and publishing study notes in the interactive area of ​​the online learning platform.

[0018] Preferably, the big data component Sqoop is used for data collection and HUE is used for task scheduling. Sqoop tasks are executed regularly every day. Various types of student-related learning data are collected from various business systems and imported into the data processing platform by system docking, data entry, and batch import. The specific process is as follows:

[0019] S1. Statistical analysis is performed on the data structure and field contents of the source data to generate the number of null values, the number of statistical record rows, and the field rule analysis results;

[0020] S2, incremental data capture based on timestamps, triggers, and logs;

[0021] S3. Perform preliminary data processing, verify, clean, decode and rename the data, verify the validity of the data, check whether there are logical errors in the data, and find and correct identifiable errors in the data;

[0022] S4. Transfer the collected student learning data to the data processing module.

[0023] Preferably, after receiving the learning-related data acquired by the data acquisition module, the data processing module pre-processes the learning-related data to form a student feature data set, and stores it in the data storage module; wherein, based on the data processing module, the learning-related data is pre-processed based on the principles of integrity, comprehensiveness, legality, and uniqueness, including data cleaning, data integration, data conversion, and data reduction.

[0024] Preferably, the data processing module preprocesses the learning related data, and the specific process is as follows:

[0025] (1) For missing data, it is necessary to determine the range of missing values, calculate the proportion of missing values ​​for each field, and select different cleaning methods according to the missing proportion and the importance of the field, as follows:

[0026] A. For fields with high importance and low missing rate, fill them by using mean, median, automatic filling of empty values ​​0 or neighbor value filling;

[0027] B. For fields with high importance and high missing rate, use other fields to obtain them through calculation;

[0028] C. Simply fill in or do not process fields with low importance and low missing rate;

[0029] D. For fields with low importance and high missing rate, remove the field;

[0030] (2) Data integration for entity recognition with homonyms, synonyms, inconsistent units, inconsistent naming, and attribute redundancy. When a mismatch is encountered, it is converted, refined, and integrated at the bottom level;

[0031] (3) Normalize each type of data to eliminate the impact of different data dimensions and unify the format of learning data. That is, change the learning data into a value between [0, 1] and adopt the maximum and minimum value normalization linear transformation method, as shown below:

[0032]

[0033] Among them, X i ' j Indicates the normalized value of the attribute, X ij represents the unnormalized value, X max Indicates the maximum value of all samples of this attribute, X min Indicates the minimum value of all samples of this attribute; stores the processed student feature data set into the data storage module.

[0034] Preferably, the label module is used to store the established label system and label rules, and analyze the characteristic factors to calculate the weights to generate student portrait labels; wherein the label module includes a label system construction module and a label generation module.

[0035] Preferably, the label system construction module constructs a label system representing the student's learning situation according to the acquired student feature data types and association relationships, and realizes automatic update of the label system through a clustering algorithm;

[0036] Based on the label system construction module, the following dimensions are classified according to the student feature dataset, as shown below:

[0037] (1) Learning engagement dimension:

[0038] Learning activity: measured by students’ learning platform login frequency, course resource access times, and video viewing time;

[0039] Course participation: assessed by the number of times students speak in the course discussion forum, the frequency of homework submission, and the participation in quizzes;

[0040] Resource utilization: measured by the number of times learning materials are downloaded and viewed, and the number of course-related exercises completed;

[0041] (2) Learning effectiveness dimension:

[0042] Performance: Evaluated through students’ grade point average, course grade distribution, and progress trend;

[0043] Homework and tests: measured by the quality of homework completion, test scores, self-assessment and feedback;

[0044] Learning goal achievement: Evaluate by comparing the learning goals set by students with their actual achievement:

[0045] (3) Learning style dimension:

[0046] Learning time preference: Analyze students’ learning time preference through their online learning time distribution;

[0047] Learning material preference: Analyze students’ preference for different learning materials such as reading, watching videos, participating in discussions, etc.

[0048] Depth of learning: measured by students’ ability to handle complex problems and willingness to conduct in-depth research;

[0049] (4) Social interaction dimension:

[0050] Collaborative learning: measured by students’ performance in team projects, number of posts and replies in the interactive area, and participation in team assignments;

[0051] Knowledge sharing: assessed by how often students share their study notes, answer other people’s questions, and participate in discussions on the platform;

[0052] Social network influence: measured by the student’s influence in the social learning network and the number of citations;

[0053] (5) Self-regulated learning dimension:

[0054] Study plan and goal setting: Evaluate the rationality of the study plan set by the students, the clarity of the goals and the degree of achievement;

[0055] Self-monitoring and reflection: measured by students’ ability to self-monitor their learning progress, reflect regularly, and adjust their learning strategies;

[0056] Learning strategy use: Assessed by the frequency and effectiveness of students' use of different learning strategies.

[0057] Preferably, the label generation module in the label module divides students into different types through the feature storage module in the data storage module and the label system construction module in the label module, generates corresponding student labels, and stores them in the data storage module; by identifying the background information of multiple learning platforms registered by students and the comprehensive consideration of the evaluation indicators of students by the teaching system, the construction of the learning student portrait defines the classification indicators from the following dimensions:

[0058] 1) Learning engagement dimension:

[0059] High participation: students frequently log in to the platform, actively participate in course discussions and resource interactions, and have a high homework submission rate;

[0060] Moderate participation: students are moderately engaged, occasionally participate in discussions and interact with resources, and the assignment submission rate is stable but not high;

[0061] Low participation: students rarely log into the platform, do not often participate in course discussions and resource interactions, and have low homework submission rates;

[0062] 2) Learning effectiveness dimension:

[0063] Highly effective: students have high GPAs, a good distribution of grades, and show a clear trend of improvement;

[0064] Moderate results: students have a moderate GPA, average grade distribution, and no obvious improvement trend;

[0065] Low performance: students have low GPAs, poor grade distribution, and negative or stagnant progress trends;

[0066] 3) Learning style dimension:

[0067] Exploratory learners: students prefer a variety of learning materials, enjoy in-depth learning tasks, and show strong curiosity and desire to explore;

[0068] Traditional learners: students who prefer traditional learning materials and have a moderate interest in deep learning tasks;

[0069] Passive learners: students have low interest in learning materials and deep learning tasks, and have a more passive learning style;

[0070] 4) Social interaction dimension:

[0071] Socially active: Students are active in team projects, frequently share knowledge and answer questions, and have high social influence;

[0072] Moderately social: Students perform moderately in team projects, occasionally share knowledge and answer questions, and have average social influence;

[0073] Socially passive: Students are passive in team projects, rarely share knowledge or answer questions, and have low social influence;

[0074] 5) Self-regulated learning dimension:

[0075] Expert in self-regulation: Students can effectively set learning plans and goals, have strong self-monitoring abilities, and flexibly use a variety of learning strategies;

[0076] Average self-regulation: Students are average in setting study plans and goals, their self-monitoring ability needs to be improved, and their learning strategies are not flexible enough;

[0077] Novice Self-Regulatory Students have difficulty setting learning plans and goals, have weak self-monitoring abilities, and use single learning strategies.

[0078] Preferably, the data storage module includes a feature storage module and a tag storage module;

[0079] Among them, the feature storage module is used to store the student feature data set formed by the data processing module. According to the characteristics of the student feature data set, the database design and data set storage are carried out according to the principles of data distributed storage and separate storage of data and indexes; the label storage module is used to store the student labels generated by the label generation module in the label module, and write the label data into the database in a directional manner;

[0080] The data storage design of the data storage module uses events, student users and event attributes to record student behavior events; each event data corresponds to an event of a student user, a user can generate multiple events, and an event can contain multiple event attributes;

[0081] A complete event includes: event participants, using User_id to set the user's unique ID, event occurrence time, using the time field to record the event occurrence time, event location, event occurrence method, and event specific content accurate to milliseconds;

[0082] Each User entity corresponds to a real user and is uniquely identified by User_id. The User entity is associated with the user's behaviors and events through User_id.

[0083] Finally, based on the portrait generation module, labels are integrated to generate student portraits of different dimensions, and corresponding learning plans are formulated according to different categories of students to improve learning efficiency.

[0084] Therefore, the present invention adopts the above-mentioned college student learning behavior analysis system based on big data, and reduces the cognitive load brought to teachers and students by a large amount of data by constructing an objective, true and accurate student portrait label system; deeply characterizes the student group, and the generated portrait has high accuracy, providing corresponding decision-making support for student education managers, thereby helping students' academic development, reducing the repetitive and tedious work caused by manual labeling, and regularly updating the label system to realize automatic updating of portraits.

[0085] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 It is a system block diagram of a college student learning behavior analysis system based on big data of the present invention. DETAILED DESCRIPTION

[0087] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0088] like Figure 1 As shown, the present invention provides a college student learning behavior analysis system based on big data, including a data acquisition module, a data processing module, a label module, a data storage module, and a portrait generation module.

[0089] 1. Based on the data collection module, learning-related data is obtained from the school's academic affairs system, online learning platform, and smart classroom system, and imported into the data processing module. Among them, learning-related data includes: basic information of students, course information, online platform learning information, classroom behavior information, and academic performance information.

[0090] Among them, student basic information includes but is not limited to: student number, student name, student gender, date of birth, registration year, student grade, student major, and student type.

[0091] Course information includes but is not limited to: course code, course name, course year, course grade, course type, and course credits.

[0092] Online platform learning information includes but is not limited to: login time, platform online time, video content watched, resource content downloaded, number of resource downloads, video viewing time, number of questions answered, and question accuracy rate.

[0093] Classroom behavior information includes but is not limited to: attendance rate, class participation, class performance, class discipline, interaction quality, class task completion, and teacher evaluation.

[0094] Academic performance information includes but is not limited to: student average grade point average, course grades, accumulated credits, and performance trends.

[0095] In addition, learning-related data includes:

[0096] Type and frequency of learning resource access: The frequency and frequency of students accessing different types of learning resources.

[0097] Learning resource completion rate: the ratio of students completing various types of learning resources.

[0098] Homework submission status: the timeliness and completion quality of students’ homework submission.

[0099] Interactive area activities: students’ behavior of posting, replying to posts, and publishing study notes in the interactive area of ​​the online learning platform.

[0100] Use the big data component Sqoop for data collection and HUE for task scheduling. Execute Sqoop tasks regularly every day. Collect various types of student-related learning data from various business systems through system docking, data entry, batch import, etc., and import them into the data processing platform. The specific process is as follows:

[0101] S1. Perform statistical analysis on the data structure and field content of the source data to generate analysis results such as the number of null values, the number of statistical record rows, and field rules;

[0102] S2, incremental data capture based on timestamps, triggers, and logs;

[0103] S3. Perform preliminary data processing, check, clean, decode and rename the data, check the validity of the data, check whether there are logical errors in the data, find and correct identifiable errors in the data, such as data consistency check, invalid value and default value processing, etc.;

[0104] S4. Transfer the collected student learning data to the data processing module.

[0105] 2. After the data processing module receives the learning-related data acquired by the data acquisition module, it pre-processes the learning-related data to form a student feature data set and stores it in the data storage module.

[0106] Among them, based on the data processing module, the learning-related data is pre-processed according to the principles of integrity, comprehensiveness, legality and uniqueness, including data cleaning, data integration, data conversion and data reduction.

[0107] (1) For missing data, it is necessary to determine the range of missing values, calculate the proportion of missing values ​​for each field, and select different cleaning methods according to the missing proportion and the importance of the field, as follows:

[0108] A. For fields with high importance and low missing rate, fill them by using mean, median, automatic filling of empty values ​​0 or neighbor value filling;

[0109] B. For fields with high importance and high missing rate, use other fields to obtain them through calculation;

[0110] C. Simply fill in or do not process fields with low importance and low missing rate;

[0111] D. For fields with low importance and high missing rate, remove the field;

[0112] For example, for the grade data in the academic affairs system, some students are missing due to uncertain factors such as suspension of study and absence from exams. In order to ensure the integrity of the data, these data are cleaned. The data of students who are on suspension are eliminated based on the student status information, and the grades of students who are absent from exams are recorded as 0.

[0113] (2) Data integration is performed to identify entities with the same name but different meanings, synonyms with different names, inconsistent units, inconsistent naming, and redundant attributes. When mismatches are encountered, they are converted, refined, and integrated at the bottom level.

[0114] (3) Normalize each type of data to eliminate the impact of different data dimensions and unify the format of learning data. That is, change the learning data into a value between [0, 1] and adopt the maximum and minimum value normalization linear transformation method, as shown below:

[0115]

[0116] Among them, X i ' j Indicates the normalized value of the attribute, X ij represents the unnormalized value, X max Indicates the maximum value of all samples of this attribute, X min Indicates the minimum value of all samples of this attribute; stores the processed student feature data set into the data storage module.

[0117] 3. The label module is used to store the established label system and label rules, analyze the characteristic factors and calculate the weights to generate student portrait labels. The label module includes the label system construction module and the label generation module.

[0118] The label system construction module constructs a label system that represents the student's learning situation based on the acquired student feature data type and association relationship, and realizes the automatic update of the label system through the clustering algorithm. Based on the label system construction module, according to the dimension classification of the student feature data set, under each dimension, students can be divided into different types or levels according to the classification indicators.

[0119] (1) Learning engagement dimension:

[0120] Learning activity: measured by indicators such as the frequency of students logging into the learning platform, the number of times they access course resources, and the length of time they watch videos;

[0121] Course participation: evaluated by the number of times students speak in the course discussion area, the frequency of homework submission, and the degree of participation in tests;

[0122] Resource utilization: measured by the number of times learning materials are downloaded and viewed, the number of course-related exercises completed, etc.

[0123] (2) Learning effectiveness dimension:

[0124] Performance: Evaluated through students’ average grade point average, course grade distribution, progress trend, etc.

[0125] Homework and tests: measured by the quality of homework completion, test scores, self-assessment and feedback;

[0126] Learning goal achievement: assessed by comparing the learning goals set by students with their actual achievement.

[0127] (3) Learning style dimension:

[0128] Study time preference: Analyze students’ study time preference through their online study time distribution (e.g. morning, afternoon, evening);

[0129] Learning material preference: Analyze students’ preference for different learning materials such as reading, watching videos, participating in discussions, etc.

[0130] Depth of learning: measured by students’ ability to handle complex problems, willingness to conduct in-depth research, etc.

[0131] (4) Social interaction dimension:

[0132] Collaborative learning: measured by students’ performance in team projects, number of posts and replies in the interactive area, and participation in team assignments;

[0133] Knowledge sharing: assessed by the frequency with which students share their study notes, answer other people’s questions, and participate in discussions on the platform;

[0134] Social network influence: measured by the student’s influence in the social learning network and the number of citations.

[0135] (5) Self-regulated learning dimension:

[0136] Study plan and goal setting: Evaluate the rationality of the study plan set by the students, the clarity of the goals and the degree of achievement;

[0137] Self-monitoring and reflection: measured by students’ ability to self-monitor their learning progress, reflect regularly, and adjust their learning strategies;

[0138] Learning strategy use: Assessed by the frequency and effectiveness of students' use of different learning strategies (e.g., memory techniques, problem-solving strategies).

[0139] The label generation module in the label module divides students into different types through the feature storage module in the data storage module and the label system construction module in the label module, generates corresponding student labels, and stores them in the data storage module.

[0140] The present invention provides a college student learning behavior analysis system based on big data, which comprehensively considers the evaluation indicators of students by identifying the background information of multiple learning platforms registered by students and the teaching system. Therefore, the construction of the learning student portrait defines the classification indicators from the following dimensions:

[0141] 1) Learning engagement dimension:

[0142] High participation: students frequently log in to the platform, actively participate in course discussions and resource interactions, and have a high homework submission rate;

[0143] Moderate participation: students are moderately engaged, occasionally participate in discussions and interact with resources, and the assignment submission rate is stable but not high;

[0144] Low participation: Students rarely log into the platform, do not often participate in course discussions and resource interactions, and have low homework submission rates.

[0145] 2) Learning effectiveness dimension:

[0146] Highly effective: students have high GPAs, a good distribution of grades, and show a clear trend of improvement;

[0147] Moderate results: students have a moderate GPA, average grade distribution, and no obvious improvement trend;

[0148] Low performance: Students have low grade point averages, poor grade distribution, and negative or stagnant progress trends.

[0149] 3) Learning style dimension:

[0150] Exploratory learners: students prefer a variety of learning materials, enjoy in-depth learning tasks, and show strong curiosity and desire to explore;

[0151] Traditional learners: students who prefer traditional learning materials, such as text and videos, and have a moderate interest in deep learning tasks;

[0152] Passive learners: Students have low interest in learning materials and deep learning tasks, and have a more passive learning style.

[0153] 4) Social interaction dimension:

[0154] Socially active: Students are active in team projects, frequently share knowledge and answer questions, and have high social influence;

[0155] Moderately social: Students perform moderately in team projects, occasionally share knowledge and answer questions, and have average social influence;

[0156] Socially passive: Students are passive in team projects, rarely share knowledge or answer questions, and have low social influence.

[0157] 5) Self-regulated learning dimension:

[0158] Expert in self-regulation: Students can effectively set learning plans and goals, have strong self-monitoring abilities, and can flexibly use a variety of learning strategies;

[0159] Average self-regulation: Students are average in setting study plans and goals, their self-monitoring ability needs to be improved, and their learning strategies are not flexible enough;

[0160] Novice Self-Regulatory Students have difficulty setting learning plans and goals, have weak self-monitoring abilities, and use single learning strategies.

[0161] 4. The data storage module includes a feature storage module and a label storage module. Among them, the feature storage module is used to store the student feature data set formed by the data processing module. According to the characteristics of the student feature data set, the database design and data set storage are carried out according to the principles of data distributed storage and separate storage of data and indexes. The label storage module is used to store the student labels generated by the label generation module in the label module, and write the label data into the database in a directional manner.

[0162] The data storage design of the data storage module uses events, student users and event attributes to record student behavior events. Each event data corresponds to an event of a student user. A user can generate multiple events, and an event can contain multiple event attributes.

[0163] A complete event includes: event participants (specific student users who participated in this event. In the system's data interface, User_id is used to set the user's unique ID), event time (the actual time when this event occurred. In the data interface, the time field is used to record the event time accurate to milliseconds. If not actively set, the current time will be automatically obtained as the value of the time field), event location, event method, and event specific content.

[0164] Each User entity corresponds to a real user and is uniquely identified by User_id. The User entity is associated with the user's behaviors and events through User_id.

[0165] 5. Finally, based on the portrait generation module, labels are integrated to generate student portraits of different dimensions, and corresponding learning plans are formulated according to different categories of students to improve learning efficiency.

[0166] The specific process of constructing student portraits is as follows:

[0167] First, by obtaining students’ basic information data from the school’s academic affairs system, online learning platform, and smart classroom system;

[0168] Secondly, further statistical analysis of the data is carried out by extracting and cleaning the acquired data;

[0169] Then, by learning the portrait label system, the data is classified and statistically displayed according to the required labels;

[0170] Finally, the cluster analysis algorithm is used to model and analyze the data generated by the students, and an online learning student portrait is constructed from the dimensions of learning participation, learning effectiveness, learning style, social interaction, and self-regulated learning.

[0171] Therefore, the present invention adopts the above-mentioned college student learning behavior analysis system based on big data, and reduces the cognitive load brought to teachers and students by a large amount of data by constructing an objective, true and accurate student portrait label system; deeply characterizes the student group, and the generated portrait has high accuracy, providing corresponding decision-making support for student education managers, thereby helping students' academic development, reducing the repetitive and tedious work caused by manual labeling, and regularly updating the label system to realize automatic updating of portraits.

[0172] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A college student learning behavior analysis system based on big data, characterized by: It includes data acquisition module, data processing module, label module, data storage module and image generation module.

2. According to the big data-based college student learning behavior analysis system of claim 1, it is characterized by: Based on the data acquisition module, learning-related data is obtained from the school's academic affairs system, online learning platform, and smart classroom system, and imported into the data processing module; among them, learning-related data includes: student basic information, course information, online platform learning information, classroom behavior information, and academic performance information.

3. According to the big data-based college student learning behavior analysis system of claim 2, it is characterized by: Basic student information includes but is not limited to: student ID, student name, student gender, date of birth, registration year, student grade, student major, and student type; Course information includes but is not limited to: course code, course name, course year, course grade, course type, and course credits; Online platform learning information includes but is not limited to: login time, platform online time, video content watched, resource content downloaded, number of resource downloads, video viewing time, number of questions answered 、 The accuracy rate of answering questions; Classroom behavior information includes but is not limited to: attendance, class participation, class performance, class discipline, quality of interaction, completion of class tasks, and teacher evaluation; Academic performance information includes but is not limited to: student GPA, course grades, credit accumulation, and performance trends; In addition, learning-related data includes: Type and frequency of learning resource access: the frequency and frequency of students accessing different types of learning resources; Learning resource completion rate: the rate at which students complete various types of learning resources; Homework submission: timeliness and quality of homework submitted by students; Interactive area activities: students’ behavior of posting, replying to posts, and publishing study notes in the interactive area of ​​the online learning platform.

4. The system for analyzing the learning behavior of college students based on big data according to claim 2 is characterized in that: Use the big data component Sqoop for data collection and HUE for task scheduling. Execute Sqoop tasks regularly every day. Collect various types of student-related learning data from various business systems through system docking, data entry, and batch import, and import them into the data processing platform. The specific process is as follows: S1. Statistical analysis is performed on the data structure and field contents of the source data to generate the number of null values, the number of statistical record rows, and the field rule analysis results; S2, incremental data capture based on timestamps, triggers, and logs; S3. Perform preliminary data processing, verify, clean, decode and rename the data, verify the validity of the data, check whether there are logical errors in the data, and find and correct identifiable errors in the data; S4. Transfer the collected student learning data to the data processing module.

5. The system for analyzing college students' learning behavior based on big data according to claim 1 is characterized by: After receiving the learning-related data acquired by the data acquisition module, the data processing module pre-processes the learning-related data to form a student feature data set and stores it in the data storage module; wherein, based on the data processing module, the learning-related data is pre-processed based on the principles of integrity, comprehensiveness, legality, and uniqueness, including data cleaning, data integration, data conversion, and data reduction.

6. A college student learning behavior analysis system based on big data according to claim 5, characterized in that: The data processing module preprocesses the learning-related data. The specific process is as follows: (1) For missing data, it is necessary to determine the range of missing values, calculate the proportion of missing values ​​for each field, and select different cleaning methods according to the missing proportion and the importance of the field, as follows: A. For fields with high importance and low missing rate, fill them by using mean, median, automatic filling of empty values ​​0 or neighbor value filling; B. For fields with high importance and high missing rate, use other fields to obtain them through calculation; C. Simply fill in or do not process fields with low importance and low missing rate; D. For fields with low importance and high missing rate, remove the field; (2) Data integration for entity recognition with homonyms, synonyms, inconsistent units, inconsistent naming, and attribute redundancy. When a mismatch is encountered, it is converted, refined, and integrated at the bottom level; (3) Normalize each type of data to eliminate the impact of different data dimensions and unify the format of learning data. That is, change the learning data into a value between [0, 1] and adopt the maximum and minimum value normalization linear transformation method, as shown below: Among them, X i ' j Indicates the normalized value of the attribute, X ij represents the unnormalized value, X max Indicates the maximum value of all samples of this attribute, X min Indicates the minimum value of all samples of this attribute; stores the processed student feature data set into the data storage module.

7. The system for analyzing college students' learning behavior based on big data according to claim 1 is characterized by: The label module is used to store the established label system and label rules, analyze characteristic factors to calculate weights, and generate student portrait labels; among them, the label module includes a label system construction module and a label generation module.

8. The system for analyzing college students' learning behavior based on big data according to claim 7 is characterized by: The label system construction module constructs a label system that represents the student's learning situation based on the acquired student feature data types and association relationships, and realizes automatic update of the label system through clustering algorithms; Based on the label system construction module, the following dimensions are classified according to the student feature dataset, as shown below: (1) Learning engagement dimension: Learning activity: measured by students’ learning platform login frequency, course resource access times, and video viewing time; Course participation: assessed by the number of times students speak in the course discussion forum, the frequency of homework submission, and the participation in quizzes; Resource utilization: measured by the number of times learning materials are downloaded and viewed, and the number of course-related exercises completed; (2) Learning effectiveness dimension: Performance: Evaluated through students’ grade point average, course grade distribution, and progress trend; Homework and tests: measured by the quality of homework completion, test scores, self-assessment and feedback; Learning goal achievement: Evaluate by comparing the learning goals set by students with their actual achievement: (3) Learning style dimension: Learning time preference: Analyze students’ learning time preference through their online learning time distribution; Learning material preference: Analyze students’ preference for different learning materials such as reading, watching videos, participating in discussions, etc. Depth of learning: measured by students’ ability to handle complex problems and willingness to conduct in-depth research; (4) Social interaction dimension: Collaborative learning: measured by students’ performance in team projects, number of posts and replies in the interactive area, and participation in team assignments; Knowledge sharing: assessed by how often students share their study notes, answer other people’s questions, and participate in discussions on the platform; Social network influence: measured by the student’s influence in the social learning network and the number of citations; (5) Self-regulated learning dimension: Study plan and goal setting: Evaluate the rationality of the study plan set by the students, the clarity of the goals and the degree of achievement; Self-monitoring and reflection: measured by students’ ability to self-monitor their learning progress, reflect regularly, and adjust their learning strategies; Learning strategy use: Assessed by the frequency and effectiveness of students' use of different learning strategies.

9. The system for analyzing college students' learning behavior based on big data according to claim 7 is characterized by: The label generation module in the label module divides students into different types through the feature storage module in the data storage module and the label system construction module in the label module, generates corresponding student labels, and stores them in the data storage module; by identifying the background information of multiple learning platforms registered by students and comprehensively considering the evaluation indicators of students by the teaching system, the construction of learning student portraits defines classification indicators from the following dimensions: 1) Learning engagement dimension: High participation: students frequently log in to the platform, actively participate in course discussions and resource interactions, and have a high homework submission rate; Moderate participation: students are moderately engaged, occasionally participate in discussions and interact with resources, and the assignment submission rate is stable but not high; Low participation: students rarely log into the platform, do not often participate in course discussions and resource interactions, and have low homework submission rates; 2) Learning effectiveness dimension: Highly effective: students have high GPAs, a good distribution of grades, and show a clear trend of improvement; Moderate results: students have a moderate GPA, average grade distribution, and no obvious improvement trend; Low performance: students have low GPAs, poor grade distribution, and negative or stagnant progress trends; 3) Learning style dimension: Exploratory learners: students prefer a variety of learning materials, enjoy in-depth learning tasks, and show strong curiosity and desire to explore; Traditional learners: students who prefer traditional learning materials and have a moderate interest in deep learning tasks; Passive learners: students have low interest in learning materials and deep learning tasks, and have a more passive learning style; 4) Social interaction dimension: Socially active: Students are active in team projects, frequently share knowledge and answer questions, and have high social influence; Moderately social: Students perform moderately in team projects, occasionally share knowledge and answer questions, and have average social influence; Socially passive: Students are passive in team projects, rarely share knowledge or answer questions, and have low social influence; 5) Self-regulated learning dimension: Expert in self-regulation: Students can effectively set learning plans and goals, have strong self-monitoring abilities, and flexibly use a variety of learning strategies; Average self-regulation: Students are average in setting study plans and goals, their self-monitoring ability needs to be improved, and their learning strategies are not flexible enough; Novice Self-Regulatory Students have difficulty setting learning plans and goals, have weak self-monitoring abilities, and use single learning strategies.

10. The college student learning behavior analysis system based on big data according to claim 1 is characterized by: The data storage module includes a feature storage module and a tag storage module; Among them, the feature storage module is used to store the student feature data set formed by the data processing module. According to the characteristics of the student feature data set, the database design and data set storage are carried out according to the principles of data distributed storage and separate storage of data and indexes; the label storage module is used to store the student labels generated by the label generation module in the label module, and write the label data into the database in a directional manner; The data storage design of the data storage module uses events, student users and event attributes to record student behavior events; each event data corresponds to an event of a student user, a user can generate multiple events, and an event can contain multiple event attributes; A complete event includes: event participants, using User_id to set the user's unique ID, event occurrence time, using the time field to record the event occurrence time, event location, event occurrence method, and event specific content accurate to milliseconds; Each User entity corresponds to a real user and is uniquely identified by User_id. The User entity is associated with the user's behaviors and events through User_id. Finally, based on the portrait generation module, labels are integrated to generate student portraits of different dimensions, and corresponding learning plans are formulated according to different categories of students to improve learning efficiency.

Images