A method and system for recommending higher vocational course content based on big data analysis

Abstract

The present invention relates to the field of educational content recommendation technology, specifically a method and system for recommending content for higher vocational courses based on big data analysis, comprising the following steps: obtaining course task labels and matching knowledge points, extracting task-type knowledge points, analyzing job responsibility information to define job capability levels, identifying learning behavior records to detect silent learning behavior, evaluating capability decay to adjust content push priorities, extracting operation intensity to adjust push cycles, and generating resource delivery adjustment results. In the present invention, by constructing a matching relationship between task labels and knowledge points, combining the job responsibility level division with the learning behavior response state, the adaptation accuracy between course content and capability requirements is improved, and by utilizing capability decay evaluation and learning load adjustment mechanisms, the push rhythm and content priority ranking of recommended content are optimized, making course resource distribution more in line with learning rhythm and job orientation, and enhancing the dynamic response capability of recommendation paths to individual states.

Description

Technical Field

[0001] The present invention relates to the technical field of educational content recommendation, and in particular to a method and system for recommending higher vocational course content based on big data analysis. Background Art

[0002] The field of educational content recommendation technology includes recommendation methods for personalized course content push and organization based on learners' individual characteristics and behavior data, teaching resource structure and learning path model. The core content revolves around learner data collection, feature modeling, content label extraction, course resource classification, learning path matching and recommendation mechanism. By analyzing learners' historical learning behavior, current learning status and resource usage, content distribution and resource scheduling are realized in the educational information platform. It is applied to various scenarios such as online education platforms, vocational education systems, and intelligent learning systems, supporting data-driven teaching decisions and personalized content configuration, involving data modeling, feature calculation, resource structure coding, recommendation rule formulation and the construction of intelligent push processes.

[0003] Among them, a method for recommending content for higher vocational courses based on big data analysis refers to a technical solution that aims to address the adaptability of course content in the teaching system of vocational colleges. It uses a resource labeling mechanism based on teaching tasks and student learning behavior trajectory data, combined with content structure labels and job competency standards, to generate multi-level course resource recommendation paths. The method covers technical matters such as task behavior data collection, small unit learning module labeling, course resource classification system establishment, job competency indicator extraction, course content and professional competency mapping relationship construction, course recommendation path generation and recommendation priority sorting. It is specifically designed and implemented through job task analysis models, course content structured decomposition models, learning task completion analysis indicators and content recommendation logic rules.

[0004] The content adaptation mechanism in traditional higher vocational course content recommendation technology relies on resource classification structure and label matching methods, and lacks structural integration of teaching task objectives and job responsibility logic, making the recommended content unable to cover the task completion ability requirements, resulting in a separation between the recommended path and the professional ability training goals. In terms of behavioral response monitoring, click frequency or access path is used as the benchmark for determining behavioral status, ignoring the structural dependence of the task stage on behavioral response, resulting in behavioral missing identification offset. In resource delivery control, a fixed strategy is adopted or the content completion rate is used as a rhythm reference indicator, without considering the differences in behavioral density and cognitive load, resulting in resource accumulation or push delays, reducing students' acceptance efficiency in the dynamic learning process, and limiting the adaptability and effectiveness in vocational education applications. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the existing technology and propose a method and system for recommending higher vocational course content based on big data analysis.

[0006] In order to achieve the above objectives, the present invention adopts the following technical solution: a method for recommending high-level vocational course content based on big data analysis, comprising the following steps:

[0007] S1: Obtain the task objective verbs, action result keywords, and cognitive level description phrases in the teaching task text of higher vocational courses, construct the course intention label structure tuple, and compare and calculate the matching degree with each knowledge point in the course outline to obtain the task-based knowledge point set;

[0008] S2: Calling the task-based knowledge point set, extracting the number of tasks, the number of operation processes, and the independent authority level in the job responsibility text, identifying the operational complexity and responsibility control scope of each responsibility item, delineating the job capability level boundary, and generating the job classification threshold;

[0009] S3: Calling the position grading threshold, identifying the task content of each position level, collecting each student's click, interaction, and submission records in the learning task execution log, comparing the behavior type preset in the task stage with the behavior log trigger status, marking untriggered behaviors, and obtaining silent behavior records;

[0010] S4: Based on the task-based knowledge point set and silent behavior records, extract the last trigger time, number of operations, and skill maintenance interval parameters of each knowledge point corresponding to the course content of the student, evaluate the degree of attenuation of the corresponding ability, adjust the recommendation priority of the corresponding course content, and obtain the recommendation list adjustment record.

[0011] As a further solution of the present invention, the task-based knowledge point set includes a task goal type label, a cognitive level label, and an achievement result type label. The job grading threshold is specifically a job ability level interval, a task level boundary parameter, and a course docking label. The silent behavior record includes the behavior untriggered stage number, the behavior type missing label, and the task stagnation duration parameter. The recommendation list adjustment record includes the knowledge point recommendation weight value, the compensation priority ranking value, and the ability decay mark item.

[0012] As a further solution of the present invention, the steps for obtaining the task-based knowledge point set are specifically as follows:

[0013] S111: Obtain task goal verbs, action result keywords, and cognitive level description phrases in the higher vocational course teaching task text, construct a course intention label structure tuple, and obtain a task semantic label set;

[0014] S112: Extract the title field of each knowledge point in the course outline based on the task semantic tag set, compare the semantic similarity between the keywords in each tag tuple and the words in the knowledge point title, calculate the tag coverage strength value, analyze the tag matching degree of each knowledge point, and obtain a tag coverage strength coefficient set;

[0015] S113: Calling the label coverage strength coefficient set, comparing the coverage strength value of each knowledge point title field in the course outline with the task label recognition threshold, dividing them into task-based knowledge points and non-task-based knowledge points, and obtaining a task-based knowledge point set.

[0016] As a further solution of the present invention, the steps for obtaining the job classification threshold are specifically as follows:

[0017] S211: Calling the task-based knowledge point set, extracting the number of tasks, the number of operation processes, and the independent authority level in the job responsibility text, structurally combining the process length, operation level, and authority level extracted from each job responsibility task, establishing a task capability indicator set, and obtaining a job task intensity parameter set;

[0018] S212: Analyze the capability level requirement boundary of each job responsibility based on the job task intensity parameter set, extract the operation process length, authority level, collaboration role weight value, number of job tasks, task operation granularity level, and course recommendation level, calculate the job task intensity level value, and combine the distribution range of the number of tasks within the job to obtain the level boundary switching condition and obtain the job capability level boundary coefficient group;

[0019] S213: Based on the job competency level boundary coefficient group, the course content structure tags in the task-based knowledge point set are called, and based on the matching relationship between the course content tag level and the boundary interval, the required course content is matched for the recommended path corresponding to the job competency level to obtain the job grading threshold.

[0020] As a further solution of the present invention, the step of obtaining the silent behavior record is specifically as follows:

[0021] S311: Call the position grading threshold, identify the task content of each position level, extract the task identification field and the task stage number field, and establish an index mapping between the position level and the stage task number to obtain the position task stage index set;

[0022] S312: Based on the job task stage index set, collect the click, interaction, and submission records of each student in the learning task execution log, extract the task binding behavior type and the actual behavior triggering status, calculate the student's behavior missing degree value in each task stage, and record the missing item type, task number, and behavior deviation degree to obtain the task stage behavior missing parameter set;

[0023] S313: Calculate the behavior missing duration of each student according to the task phase behavior missing parameter set, and obtain the silent behavior record.

[0024] As a further solution of the present invention, the step of obtaining the recommendation list adjustment record is specifically as follows:

[0025] S411: Extracting the last trigger time, operation times, and skill maintenance interval parameters of each student's knowledge point corresponding to the course content based on the task-based knowledge point set and the silent behavior record, and obtaining a knowledge point execution basic parameter set;

[0026] S412: Executing the basic parameter set for each knowledge point, comparing the last triggering time of each knowledge point with the skill maintenance interval parameter, calculating the ratio of the number of knowledge point operations deviating from the maintenance benchmark value, evaluating the attenuation degree of the target student's ability corresponding to each knowledge point, and obtaining an ability attenuation degree value;

[0027] S413: According to the capability attenuation degree value and the course content mapped to the knowledge point, the priority of the target course content in the recommendation list is adjusted, and a recommendation list adjustment record is obtained.

[0028] As a further embodiment of the present invention, the method further comprises:

[0029] S5: Calling the recommendation list adjustment record, extracting the student's operation intensity parameters, resource call frequency, and stay period per unit time, identifying the learning behavior load state, and adjusting the planned cycle of the course content push task to generate a resource delivery adjustment result;

[0030] The resource delivery adjustment result includes the push cycle delay value, the task queue adjustment position, and the resource rhythm control parameters.

[0031] As a further solution of the present invention, the step of obtaining the resource allocation adjustment result is specifically as follows:

[0032] S511: Calling the recommendation list adjustment record, extracting the number of click operations, course resource call frequency, and page dwell period of the student in a unit time, and obtaining the original parameter set of the behavior load;

[0033] S512: Analyze the learning behavior activity of multiple students based on the original behavior load parameter set, the number of click operations, the resource call frequency, and the page dwell period, calculate the learning behavior load status, and obtain the learning behavior load status value;

[0034] S513: According to the learning behavior load state value and in combination with the current planned push time of each course content in the recommendation list, the planned cycle of the course content push task is adjusted to obtain the resource delivery adjustment result.

[0035] A system for recommending higher vocational course content based on big data analysis, wherein the system is used to implement the above-mentioned method for recommending higher vocational course content based on big data analysis, and the system comprises:

[0036] The course intention parsing module obtains the task objective verbs, action result keywords, and cognitive level description phrases in the teaching task text of the higher vocational course, constructs the course intention label structure tuple, compares the label structure tuple with each knowledge point in the course outline and calculates the matching degree, screens the knowledge points, and establishes a task-based knowledge point set;

[0037] The job capability analysis module extracts the number of tasks, the number of operation processes, and the independent authority levels in the job responsibility text based on the task-based knowledge point set, identifies the operational complexity and responsibility control scope of each responsibility item, delineates the job capability level boundaries, and generates job classification thresholds;

[0038] The behavioral silence diagnosis module identifies the task content of each job level based on the job classification threshold and collects the student's online learning task execution log data, including clicks, interactions, and submissions. It compares the behavior types preset in the task stage with the behavior log trigger status, marks untriggered behaviors, and obtains silent behavior records.

[0039] The knowledge decay assessment module extracts the last trigger time, number of operations, and skill maintenance interval parameters of each knowledge point corresponding to the course content of each student based on the task-based knowledge point set and silent behavior records. It calculates the decay degree of the skill corresponding to the knowledge point, dynamically adjusts the recommendation priority of the course content associated with the knowledge point, and obtains the adjustment record of the recommendation list.

[0040] The load status control module adjusts the records based on the recommendation list, extracts the students' operation intensity parameters, resource call frequency and stay period per unit time, identifies the students' learning behavior load status, and adjusts the planned cycle of the course content push task to generate resource delivery adjustment results.

[0041] Compared with the prior art, the advantages and positive effects of the present invention are:

[0042] In the present invention, by constructing a matching relationship between task labels and knowledge points, combining the job responsibility level division and learning behavior response status, the adaptation accuracy between course content and ability requirements is improved, and by utilizing the ability decay assessment and learning load adjustment mechanism, the push rhythm and content priority sorting of recommended content are optimized, making the course resource distribution more in line with the learning rhythm and job orientation, and enhancing the dynamic response ability of the recommendation path to individual status. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the main steps of the present invention;

[0044] Figure 2 Obtaining a flow chart for the task-based knowledge point set of the present invention;

[0045] Figure 3 A flowchart for obtaining the job classification threshold value of the present invention;

[0046] Figure 4 A flow chart for obtaining the silent behavior record of the present invention;

[0047] Figure 5 A flowchart for obtaining a record of adjusting a recommendation list of the present invention;

[0048] Figure 6 The present invention provides a flow chart for obtaining resource allocation adjustment results. DETAILED DESCRIPTION

[0049] 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.

[0050] 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.

[0051] See also Figure 1 The present invention provides a technical solution: a method for recommending high-level vocational course content based on big data analysis, comprising the following steps:

[0052] S1: Obtain the task objective verbs, action result keywords, and cognitive level description phrases in the teaching task text of higher vocational courses, construct the course intention label structure tuple, and compare and calculate the matching degree with each knowledge point in the course outline to obtain the task-based knowledge point set;

[0053] S2: Call the task-based knowledge point set to extract the number of tasks, the number of operation processes, and the independent authority level in the job responsibility text, identify the operational complexity and responsibility control scope of each responsibility item, define the job capability level boundary, and generate the job classification threshold;

[0054] S3: Call the job classification threshold to identify the task content of each job level, collect each student's click, interaction, and submission records in the learning task execution log, compare the behavior type preset in the task stage with the trigger status of the behavior log, mark untriggered behaviors, and obtain silent behavior records;

[0055] S4: Based on the task-based knowledge point set and silent behavior records, extract the last trigger time, operation times, and skill maintenance interval parameters of each knowledge point corresponding to the course content of the student, evaluate the degree of attenuation of the corresponding ability, adjust the recommendation priority of the corresponding course content, and obtain the recommendation list adjustment record;

[0056] S5: Call the recommendation list adjustment record, extract the student's operation intensity parameters, resource call frequency and stay period per unit time, identify the learning behavior load status, and adjust the planned cycle of the course content push task to generate resource delivery adjustment results.

[0057] The task-based knowledge point set includes task goal type labels, cognitive level labels, and achievement result type labels. The job classification threshold is specifically the job ability level range, task level boundary parameters, and course docking labels. The silent behavior record includes the behavior untriggered stage number, behavior type missing label, and task stagnation duration parameter. The recommendation list adjustment record includes the knowledge point recommendation weight value, compensation priority ranking value, and capability attenuation mark item. The resource delivery adjustment results include the push cycle delay value, task queue adjustment position, and resource rhythm control parameters.

[0058] See also Figure 2 ,The specific steps for obtaining the task-based knowledge point set are:

[0059] S111: Obtain task goal verbs, action result keywords, and cognitive level description phrases in the higher vocational course teaching task text, construct a course intention label structure tuple, and obtain a task semantic label set;

[0060] Through the specific examples of higher vocational course teaching task texts, we disassembled them and obtained three contents: task target verbs, behavioral result keywords, and cognitive level description phrases. Among them, the task target verbs called the standard verb library in Bloom's cognitive taxonomy for vocabulary mapping. For example, the "master" in the higher vocational software testing course task text "students can master the test case writing method" is mapped to "application". The behavioral result keywords are extracted from the course evaluation standard text. For example, "coverage" in "software test coverage reaches 90%" is the result keyword. The cognitive level description phrases call Bloom's six cognitive level definitions. For example, "analysis" in "students can analyze program code errors" is a level description word. According to the call and matching of the above three categories of vocabulary, the course task text is structured into tuples. The specific structure tuple examples are shown in Table 1:

[0061] Table 1 Task intent label structure tuple table

[0062] Task objective verb Behavior Result Keywords Cognitive level description phrases application Coverage analyze Evaluate Defect rate evaluate Identification Execution efficiency understand

[0063] Table 1 lists specific examples of task intention label structure tuples. Taking the first row of tuples as an example, the task goal verb is "application", the behavior result keyword is "coverage", and the cognitive level description phrase is "analysis". By constructing the label structure tuple of the entire course content in the above way, we get the task semantic label set.

[0064] S112: Based on the task semantic tag set, extract the title field of each knowledge point in the course outline, and compare the semantic similarity between the keywords in each tag tuple and the words in the knowledge point title using the formula:

[0065] ;

[0066] Calculate the label coverage strength value, analyze the label matching degree of each knowledge point, and obtain the label coverage strength coefficient set;

[0067] in, Indicates the The label coverage strength value of each knowledge point, Indicates the The target verb in the label tuple is the same as the The semantic similarity score of the knowledge point keywords, Indicates the matching score between the behavior result keyword and the knowledge point. Indicates the The label importance evaluation value of the label tuple, is the cognitive level value, Set the level of the task corresponding to this knowledge point. Indicates the total number of course task intent label tuples, Indicates the knowledge point number currently being processed. Indicates the The first knowledge point to match The sequence number of the tag tuple;

[0068] Call the structure tuple data in Table 1, and call the title field of each knowledge point in the course outline, such as "test case design", "functional defect analysis", "automated test script execution", etc., and compare the semantic similarity with the keywords "application", "coverage", "analysis" in the task label tuple one by one. For example, by calculating the cosine similarity of the word vector, the specific value is obtained. 、 At the same time, the course evaluation expert scoring data is called to obtain the label importance evaluation value ,Expert ratings are scored using the Likert 5-point rating method (1 is not important, 5 is very important).,For example, the expert evaluation value of the label tuple (application, coverage, analysis) is 4, is the cognitive level value of the current task label tuple. Based on Bloom's cognitive level, memory is 1, understanding is 2, application is 3, analysis is 4, evaluation is 5, and creation is 6. For example, the "analysis" of the above tuple is 4. Set the level of tasks corresponding to knowledge points through the task design document. For example, the corresponding cognitive level of "test case design" is 3. Now take the knowledge point "test case design" as an example and substitute the above parameters into the formula:

[0069] ;

[0070] set up , calculated by cosine similarity, , , , ,but:

[0071] ;

[0072] The label coverage strength value is calculated in the same manner for other knowledge points. This calculation is repeated for all knowledge points one by one, and the label coverage strength coefficients are aggregated to form a set. The label coverage strength value refers to the degree to which a knowledge point is semantically covered by multiple course intent label structure tuples within the course task system. This value is calculated by evaluating the semantic fit between the label verb, the result keyword, and the knowledge point title, combined with the consistency of the task cognitive level. This continuous measure represents the number of teaching tasks with which the current knowledge point is semantically associated, and whether the associations are structurally consistent with the teaching tasks. This helps the system identify content from all course knowledge points that truly possesses task-oriented attributes, constructing a task-based content set tailored to job competency requirements. This provides a clearly structured and fully labeled course foundation for subsequent job level mapping and path generation. A higher label coverage strength value indicates that the knowledge point can support a greater variety of teaching tasks, has a clearer task structure, and receives a higher priority in the recommended path. The formula effectively quantifies the strength of the association between knowledge point and task labels by comprehensively considering semantic similarity, label importance, and cognitive level gaps.

[0073] S113: calling the label coverage strength coefficient set, comparing the coverage strength value of each knowledge point title field in the course outline with the task label recognition threshold, dividing them into task-based knowledge points and non-task-based knowledge points, and obtaining a task-based knowledge point set;

[0074] The calculated values ​​are called, such as the coverage strength value of the knowledge point "test case design" is 1.55, the knowledge point "functional defect analysis" is 2.30, and the knowledge point "automated test script execution" is 0.95. The task label recognition threshold setting is determined based on the results of the previous experimental test. For example, the label coverage strength values ​​of 100 knowledge points are statistically analyzed. The coverage strength values ​​of the top 50% of the knowledge points are concentrated in the range of 1.20 to 2.50, so 1.20 is taken as the threshold. By comparing the coverage strength value of each knowledge point with the threshold, if the coverage strength value of the knowledge point "test case design" is 1.55, which is higher than the threshold of 1.20, it is determined to be a task-type knowledge point. The coverage strength value of "automated test script execution" is 0.95, which is lower than the threshold and is determined to be a non-task-type knowledge point. By comparing all knowledge points one by one in the above manner, the knowledge points above the threshold are divided into the task-type knowledge point set, and finally the task-type knowledge point set is obtained.

[0075] See also Figure 3 , the specific steps for obtaining the job classification threshold are as follows:

[0076] S211: Calling the task-based knowledge point set, extracting the number of tasks, the number of operation processes, and the independent authority level from the job responsibility text, structurally combining the process length, operation level, and authority level extracted from each job responsibility task, establishing a task capability indicator set, and obtaining a job task intensity parameter set;

[0077] Call the acquired knowledge point data in the task semantic tag set, such as "test case design", "functional defect analysis", "performance test execution", etc., and disassemble the job responsibilities text item by item. For example, for the software test engineer job responsibilities text "responsible for software defect analysis and functional use case design, the authority level is level 3, and the number of operation processes is 5", then the number of tasks in the responsibility is called as 2 (defect analysis, use case design), the number of operation processes is 5, and the authority level is set to level 3. By calling multiple job responsibilities texts, the above data is extracted item by item. After the extraction is completed, the number of process calls is , authority level and number of tasks. The three data are combined and calculated according to the actual structure of the job requirements. Taking the software testing engineer position as an example, the task capability indicator structure is obtained by calculating (number of tasks 2, number of processes 5, authority level 3). The task intensity calculation process of this position is to multiply the number of tasks and the number of processes to get a value of 10, and then add it to the authority level value of 3 to get a value of 13, that is, the job task intensity value is 13. In the same way, calculate the task intensity values ​​of other job responsibilities one by one, summarize all the calculated job task intensity values, and form a job task intensity parameter set.

[0078] S212: Based on the job task intensity parameter set, analyze the capability level requirement boundary of each job task, extract the operation process length, authority level, collaboration role weight value, number of job tasks, task operation granularity level and course recommendation level, and use the formula:

[0079] ;

[0080] Calculate the job task intensity level value, combine it with the distribution range of the number of tasks within the job, obtain the level boundary switching condition, and obtain the job capability level boundary coefficient group;

[0081] in, For the The level intensity value of each job task, For the The length of the operation process of each job task, For the The authority level value of each job task, For the The first of the job tasks The count of collaborative roles, For the The first of the job tasks The weight of the collaboration role, For the The total number of tasks corresponding to each position, For the The operation granularity level value of each job task, For the The recommended level value of the course that matches each job task, Index for job task number, Index for collaboration role numbers, The total number of collaborative roles under a single job task;

[0082] The example values ​​of the task intensity parameters for the three positions of software test engineer, software development engineer, and project manager are called, which are 13, 16, and 21 respectively. By calling the specific task intensity parameters of the three positions, the operation process length (5, 6, and 7 respectively), authority level (3, 4, and 5 respectively), count value of collaborative roles (2, 3, and 4 respectively) and weight value of the corresponding role (senior role weight 0.8, intermediate role weight 0.5, junior role weight 0.3, and the average value is taken) of each position are further called one by one. The number of position tasks (2, 3, and 4 respectively), task operation granularity level (software test engineer level 3, development engineer level 4, project manager level 5), and course recommendation level (level 3, level 4, and level 5 respectively) are given in the form of Table 2:

[0083] Table 2 Job Responsibilities Parameters

[0084] Job Title Process length Permission Level Number of collaborative roles Average role weight Number of job tasks Operation granularity level Recommended level Software Test Engineer 5 3 2 0.65 2 3 3 Software Development Engineer 6 4 3 0.60 3 4 4 Project Manager 7 5 4 0.60 4 5 5

[0085] As shown in Table 2, the job task intensity level value is calculated by calling the parameters item by item. , taking the software test engineer position as an example, the call process length , permission level , Collaboration role count value , role weight , total number of tasks , Operation granularity level , recommended level , enter the formula to calculate:

[0086] ;

[0087] ;

[0088] Repeat the calculation for other positions. For example, the strength value of the software development engineer position is 45.0, and the strength value of the project manager position is 43.2. Further call the task intensity level values ​​of all positions, sort them according to the size of the task intensity value, and divide the level boundaries according to the previous experimental results. For example, the strength value greater than 45 is defined as the advanced ability level interval, the range between 40 and 45 is defined as the intermediate ability level interval, and the range below 40 is defined as the primary ability level interval. Based on this condition, the position ability level boundary coefficient group is defined. Among them, the job task intensity level value refers to a comprehensive indicator of a single job duty in a job under the cross-interaction of multiple dimensions such as task execution structure, authority requirements, collaboration load and course matching difficulty. It is used to express the strength of the ability level required for the task. This indicator reflects the internal complexity of the task itself, and reflects the degree of interference or adaptation cost on the learning path design through collaboration weight and course matching error. By grading and sorting different tasks according to this intensity value, the job ability level boundary is constructed. When the intensity values ​​of multiple tasks are in a similar range, they will be classified into the same level; when there is a clear gradient in the intensity value, it is used to define the level switching threshold and guide the level adaptation matching logic of the course content, thereby improving the matching accuracy of course stratification and ability fit in the recommended path. Through multi-dimensional calculations such as operation process, authority level, collaboration role and task operation granularity, and recommendation level, the refined boundary coefficient of the matching between job responsibilities and course content is obtained.

[0089] S213: Based on the job competency level boundary coefficient group, the course content structure tags in the task-based knowledge point set are called, and based on the matching relationship between the course content tag level and the boundary interval, the required course content is matched for the recommended path corresponding to the job competency level to obtain the job grading threshold;

[0090] The level boundaries corresponding to the three positions of software testing engineer, software development engineer and project manager calculated in the paragraph are called as junior, intermediate and senior respectively. The course content structure labels of the knowledge points in the task-based knowledge point set are further called, such as "test case design" is intermediate, "functional defect analysis" is intermediate, "performance test execution" is advanced, etc. The position ability level and the knowledge point label level are called for comparison one by one. Taking the software testing engineer as an example, its ability level is called as junior, and the knowledge points with the junior label level are matched. After performing the matching operation on each position, the course knowledge point content corresponding to each position ability level is obtained, and the position ability level and the matching knowledge point content are further bound, such as the software testing engineer corresponds to the "basic defect record" knowledge point, the software development engineer corresponds to the "test case design" knowledge point, and the project manager corresponds to the "performance test execution" knowledge point, so as to establish the position grading threshold.

[0091] See also Figure 4 The specific steps for obtaining silent behavior records are as follows:

[0092] S311: Call the job classification threshold, identify the task content of each job level, extract the task identification field and the task stage number field, and establish an index mapping between the job level and the stage task number to obtain the job task stage index set;

[0093] Taking the software testing positions of "junior, intermediate, and senior" as an example, the call is made by extracting the task identification field and task stage number field involved in the specific task content under each position level to achieve one-by-one decomposition. For example, the "task identification field" in the task content corresponding to the junior position is "defect record", and the task stage number field is "stage 1"; the task identification field corresponding to the intermediate position is "test case design", and the task stage number field is "stage 2"; the task identification field corresponding to the senior position is "performance testing", and the task stage number field is "stage 3". The index mapping operation of the position level and the stage task number is further performed. By calling the position level data and the task stage number field data one by one, taking the junior position "stage 1" as an example, the "junior" position level is called and mapped to "junior-stage 1". The corresponding relationship between other position levels and stage numbers is obtained through the above mapping operation, as shown in Table 3:

[0094] Table 3 Mapping table of job level and stage task number

[0095] Job Level Mission phase number Mapping Index primary Phase 1 Primary - Stage 1 intermediate Phase 2 Intermediate - Stage 2 advanced Phase 3 Advanced - Stage 3

[0096] As shown in Table 3, all job levels and stage task number mapping indexes are established one by one, and all mapping index entries are summarized to form a job task stage index set.

[0097] S312: Based on the job task stage index set, collect the click, interaction, and submission records of each student in the learning task execution log, extract the task binding behavior type and the actual behavior triggering status, and use the formula:

[0098] ;

[0099] Calculate the student's behavior missing degree value at each task stage, record the missing item type, task number, and behavior deviation degree, and obtain the behavior missing parameter set for the task stage;

[0100] in, For the Students in the The behavior loss degree value of each task stage, For the Students in the In the mission phase The item should execute the behavior state value, For the Students in the In the mission phase The actual completion behavior status value of the item, For the Students in the In the mission phase The behavior impact weight value of the behavior item, The number of behavior items preset in each task stage, For the Students in the The behavior trigger delay time value in each task stage, The behavior trigger tolerance time constant set for the system, Index for student number, is the task phase number index, Index the behavior items.

[0101] By calling "Primary-Stage 1" in Table 3 as an example, the student number is Taking the student in stage 1 as an example, the three types of records of the student's click, interaction, and submission in the learning task execution log are called. Among them, the click behavior includes the page visit action; the interaction behavior includes the discussion area message action; the submission behavior includes the task submission action. Each behavior record is called separately and the task binding behavior type is extracted. For example, the click behavior is bound to "page visit", the interaction behavior is bound to "message interaction", and the submission behavior is bound to "task submission". The actual behavior triggering status data of the student is extracted one by one, among which the completed click behavior ( ), unfinished interactive behavior ( ), unfinished submission behavior ( ); The preset behavior states of the task stage are set to click 1, interaction 1, and submission 1 ( The influence weight of each behavior is set based on expert scoring, for example, the weight of page visit is 0.3, the weight of message interaction is 0.4, and the weight of task submission is 0.3; further calling the behavior trigger delay time, the actual delay of the student is , the system behavior trigger tolerance time constant is set to ; Substitute the above parameters into the formula for calculation:

[0102] ;

[0103] The calculation process is:

[0104] ;

[0105] The same method was used to calculate the behavioral absence degree values ​​for other students at each task stage. Each missing behavior item was recorded as "interaction behavior" or "submission behavior," with the task number "stage 1" and the behavioral deviation degree as 0.135. This yielded a set of task stage behavior absence parameters. The behavior absence degree value quantifies the degree of deviation in a student's response from the pre-set behavioral items during a task stage. A higher value indicates a lower rate of expected behavioral execution and a more pronounced behavioral gap, indicating a "silent" or "interrupted" student's engagement, responsiveness, or task status at the current task node. This indicator, serving as the core metric for determining "silent behavior," is directly used to identify behavioral gaps within a stage, determine learning discontinuities, and serve as trigger conditions for behavioral activation or input for recommended task rebound control parameters. It is the core numerical result in the silent behavior identification mechanism, ensuring dynamic comparability of behavior monitoring across task granularity, time granularity, and response structure, and providing a data foundation for tracking student learning progress and adjusting course content recommendations. The formula comprehensively calculates the difference between the pre-set and actual completed behavioral states, along with weights and delays, to quantify the degree of behavioral absence.

[0106] S313: Calculate the behavior missing duration of each student based on the behavior missing parameter set of the task phase and obtain the silent behavior record;

[0107] By student number For example, the behavior missing degree value of stage 1 is called as 0.135, and the standard behavior trigger time preset in the task stage is further called. For example, the standard trigger time of the interactive behavior in stage 1 is 5 minutes, and the standard trigger time of the submission behavior is 10 minutes. The actual completion time of the interactive behavior and the submission behavior in the student's actual behavior log is called. Assume that the student's interactive behavior is actually completed in 12 minutes, and the submission behavior is completed in 15 minutes; call the preset standard time and the actual completion time respectively and perform subtraction operations one by one. The missing duration of the interactive behavior is the actual time 12 minutes minus the standard time 5 minutes, and the missing duration is 7 minutes; the missing duration of the submission behavior is the actual time 15 minutes minus the standard time 10 minutes, and the missing duration is 5 minutes; execute the above missing duration calculation process one by one, and obtain the behavior missing duration data of other students in other stages by repeating the above calculation, and summarize the missing duration record data to form a silent behavior record.

[0108] See also Figure 5 , the specific steps for obtaining the recommendation list adjustment record are as follows:

[0109] S411: Based on the task-based knowledge point set and the silent behavior record, extract the last trigger time, operation times, and skill maintenance interval parameters of each knowledge point corresponding to the course content of the student, and obtain the basic parameter set for knowledge point execution;

[0110] Call the task-based knowledge point data obtained in the previous article, such as "test case design", "functional defect analysis", etc., and the silent behavior record data such as student number The missing time in "Test Case Design" is 7 minutes. The actual data of the student learning log is called to extract the specific timestamp of the most recent time that the student accessed and executed the course content of each knowledge point. For example, the last trigger of the "Test Case Design" knowledge point was 14:00 on December 20, 2024, and the "Functional Defect Analysis" was 09:30 on December 21, 2024. The actual number of operations of each knowledge point is called, for example, the number of operations of the "Test Case Design" knowledge point is 3 times, and the number of operations of the "Functional Defect Analysis" knowledge point is 2 times. The knowledge point skill maintenance interval parameters specified in the industry skill standards are called, for example, the maintenance interval of "Test Case Design" is 7 days, and the maintenance interval of "Functional Defect Analysis" is 10 days. The specific parameter data of the above calls and extractions are shown in Table 4:

[0111] Table 4 Basic parameters for knowledge point execution

[0112] Knowledge point name Last trigger time Number of operations Skill maintenance interval (days) Test case design 2024-12-2014:00 3 7 Functional defect analysis 2024-12-2109:30 2 10

[0113] As shown in Table 4, parameter calls and data are summarized for each knowledge point and organized into a basic parameter set for knowledge point execution.

[0114] S412: Based on the knowledge point execution basic parameter set, the last trigger time of each knowledge point is compared with the skill maintenance interval parameter, the ratio of the knowledge point operation count deviation from the maintenance baseline value is calculated, and the degree of attenuation of the target student's ability corresponding to each knowledge point is evaluated to obtain the ability attenuation degree value;

[0115] Call the data in Table 4, for the knowledge point of "Test Case Design", by calling the current evaluation date 12:00 on December 28, 2024 and the last trigger time 14:00 on December 20, 2024, one by one to calculate the time difference, the trigger interval is about 8 days, further call the skill maintenance interval parameter 7 days, perform the subtraction operation between the trigger interval and the skill maintenance interval, calculate the excess maintenance interval value is 1 day, by calling the operation number 3 times and calling the maintenance benchmark value specified in the skill standard 4 times, calculate the proportion of the operation number deviation from the maintenance benchmark value for each knowledge point, with (4 times -3 times) / 4 times is calculated as 0.25 (25%); in the same way, the "Functional Defect Analysis" knowledge point is called. The current date 12:00 on December 28, 2024 and 09:30 on December 21, 2024 are subtracted to 7 days and 3 hours, which is about 7.13 days. The maintenance interval is 10 days, and the subtraction is calculated as -2.87 days, which does not exceed the maintenance interval. The operation is called twice, and the baseline value is maintained 3 times. The deviation from the baseline value is calculated as (3-2) / 3, which is about 0.33 (33%); the deviation ratio values ​​calculated for all knowledge points are summarized to obtain the capability attenuation value.

[0116] S413: According to the ability decay degree value and the course content mapped to the knowledge point, the priority of the target course content in the recommendation list is adjusted, and a recommendation list adjustment record is obtained;

[0117] The attenuation value of "Test Case Design" is called to be 25%, and the attenuation value of "Functional Defect Analysis" is called to be 33%. The order of knowledge points in the course recommendation list is called one by one. For example, the recommended order of knowledge points in the current recommendation list is: "Functional Defect Analysis" is in sequence number 2, and "Test Case Design" is in sequence number 4. After calling the attenuation value data, the comparison operation is performed one by one. Taking the attenuation value of 25% corresponding to the recommendation level priority threshold set to 30% as an example, by judging that 25% is lower than the 30% threshold, the corresponding course recommendation priority is not adjusted; the attenuation value of 33% exceeds the 30% threshold, and the original sorting position of the corresponding knowledge point "Functional Defect Analysis" is called to be sequence number 2, and it is adjusted to a higher priority, that is, sequence number 1. The above comparison and judgment operations are performed item by item, and the position adjustment of the knowledge points in the recommendation list is completed according to the attenuation value to form a recommendation list adjustment record.

[0118] See also Figure 6 The specific steps for obtaining resource allocation adjustment results are as follows:

[0119] S511: Call the recommendation list adjustment record, extract the number of click operations, course resource call frequency, and page dwell period of the student in a unit time, and obtain the original parameter set of the behavior load;

[0120] By calling the student number The course recommendation sequence data after adjustment, such as the knowledge point "Functional Defect Analysis" adjusted to position 1, and "Test Case Design" adjusted to position 4, is further extracted. The student's operation log on the course platform is further extracted, and the specific click operation number data, course resource call frequency data, and page stay period data within a unit time (1 hour as a unit, such as the period from 10:00 to 11:00 on December 28, 2024) are called one by one. For example, the number of click operations within 1 hour is 15 times, the resource call frequency is 5 times, and the average page stay period is 180 seconds. The data within other time periods are called and extracted one by one in a similar way. Examples are shown in Table 5:

[0121] Table 5 Behavior load original parameters

[0122] Student ID Time period Click actions Resource call frequency Page dwell time (seconds) 1 2024-12-2810:00~11:00 15 5 180 1 2024-12-2811:00~12:00 12 3 240

[0123] As shown in Table 5, the original parameter set of behavioral load is obtained by calling the data records of each student in each unit time one by one.

[0124] S512: Analyze the learning behavior activity of multiple students based on the original behavior load parameter set, the number of click operations, the resource call frequency, and the page dwell period, calculate the learning behavior load status, and obtain the learning behavior load status value;

[0125] Taking the data of student No. 1 in Table 5 as an example, the parameters of click operation times, resource call frequency and page stay period are called, and the numerical analysis operation of the parameters is performed one by one. The click operation times such as 15 times and 12 times are compared with the set click activity threshold of 10 times respectively. More than 10 times are judged as high activity, and less than or equal to 10 times are judged as low activity. Therefore, the time period from 10:00 to 11:00 is high activity, and the time period from 11:00 to 12:00 is high activity; the resource call frequency is called 5 times and 3 times respectively, and the resource call frequency activity threshold is 3 times. More than 3 times are high activity, and less than or equal to 3 times are low activity. The former is judged as high activity and the latter is low activity; the page stay period is called 180 seconds and 240 seconds, and the call stay period is called 180 seconds and 240 seconds respectively. The standard threshold for the cycle is 200 seconds. The cycle length is compared with the threshold. A cycle shorter than 200 seconds is determined to be a lower dwelling cycle, and a cycle longer than 200 seconds is determined to be a higher dwelling cycle. Therefore, the former is a lower cycle, and the latter is a higher cycle. The above numerical comparison and analysis are performed one by one, and the parameters of high activity are recorded as 1, and low activity is recorded as 0. The above values ​​are accumulated one by one. For example, the total activity from 10:00 to 11:00 is 2 (high clicks, high resource calls), and the dwelling cycle value is low, so the overall activity value is high; the activity value from 11:00 to 12:00 is 1 (high clicks, low resource calls), the dwelling cycle is high, and the overall activity value is low. The behavioral load status values ​​of all students in each period are calculated item by item to obtain the learning behavior load status value.

[0126] S513: According to the learning behavior load state value and the current planned push time of each course content in the recommendation list, the planned cycle of the course content push task is adjusted to obtain the resource delivery adjustment result;

[0127] The load status data of student number 1 is called. For example, the behavior load status value from 10:00 to 11:00 is high, and the load status value from 11:00 to 12:00 is low. The current planned push time of each course content in the recommended list is further called. For example, the current push time of the knowledge point "Functional Defect Analysis" is 11:00 on December 28, 2024, and the knowledge point "Test Case Design" is 12:00 on December 28, 2024. By comparing the behavior load status value with the corresponding push time one by one, the subsequent push content corresponding to the period with higher behavior load status value from 10:00 to 11:00 is adjusted. The content originally scheduled to be pushed at 11:00 is pushed in advance to 10:30, and the course push content corresponding to the period with lower behavior load status value from 11:00 to 12:00 is postponed from 12:00 to 12:30. The above push plan cycle adjustment calculation operations are completed one by one. The actual push time after adjustment is shown in Table 6:

[0128] Table 6 Resource allocation adjustment results

[0129] Knowledge point name Original push time Adjusted push time Functional defect analysis 2024-12-28-11:00 2024-12-28-10:30 Test case design 2024-12-28-12:00 2024-12-28-12:30

[0130] As shown in Table 6, all adjusted push times are summarized to form the resource delivery adjustment result.

[0131] A system for recommending higher vocational course content based on big data analysis is provided. The system is used to implement the above-mentioned method for recommending higher vocational course content based on big data analysis. The system includes:

[0132] The course intention parsing module obtains the task objective verbs, action result keywords, and cognitive level description phrases in the teaching task text of the higher vocational course, constructs the course intention label structure tuple, compares the label structure tuple with each knowledge point in the course outline and calculates the matching degree, screens the knowledge points, and establishes a task-based knowledge point set;

[0133] The job competency analysis module, based on a set of task-based knowledge points, extracts the number of tasks, the number of operational processes, and the level of independent authority in the job description. It identifies the operational complexity and scope of each responsibility item, defines the boundaries of job competency levels, and generates job classification thresholds.

[0134] The behavioral silence diagnosis module identifies the task content of each job level based on the job classification threshold and collects students' online learning task execution log data, including clicks, interactions, and submissions. It compares the behavior types preset in the task stage with the behavior log trigger status, marks untriggered behaviors, and obtains silent behavior records.

[0135] The knowledge decay assessment module extracts the last trigger time, number of operations, and skill maintenance interval parameters of each knowledge point corresponding to the course content of each student based on the task-based knowledge point set and silent behavior records. It calculates the decay degree of the skill corresponding to the knowledge point, dynamically adjusts the recommendation priority of the course content associated with the knowledge point, and obtains the adjustment record of the recommendation list.

[0136] The load status control module adjusts the records based on the recommendation list, extracts the students' operation intensity parameters, resource call frequency and stay period per unit time, identifies the students' learning behavior load status, and adjusts the planned cycle of the course content push task to generate resource delivery adjustment results.

[0137] 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 method for recommending higher vocational course content based on big data analysis, characterized in that: The following steps are involved: S1: Obtain the task objective verbs, action result keywords, and cognitive level description phrases in the teaching task text of higher vocational courses, construct the course intention label structure tuple, and compare and calculate the matching degree with each knowledge point in the course outline to obtain the task-based knowledge point set; S2: Calling the task-based knowledge point set, extracting the number of tasks, the number of operation processes, and the independent authority level in the job responsibility text, identifying the operational complexity and responsibility control scope of each responsibility item, delineating the job capability level boundary, and generating the job classification threshold; The steps for obtaining the job classification threshold are as follows: S211: Calling the task-based knowledge point set, extracting the number of tasks, the number of operation processes, and the independent authority level in the job responsibility text, structurally combining the process length, operation level, and authority level extracted from each job responsibility task, establishing a task capability indicator set, and obtaining a job task intensity parameter set; S212: Analyze the capability level requirement boundary of each job responsibility based on the job task intensity parameter set, extract the operation process length, authority level, collaboration role weight value, number of job tasks, task operation granularity level, and course recommendation level, calculate the job task intensity level value, and obtain the job capability level boundary coefficient group based on the intensity distribution and gradient characteristics in combination with the distribution range of the number of tasks within the job; S213: Based on the job competency level boundary coefficient group, the course content structure tags in the task-based knowledge point set are called, and based on the matching relationship between the course content tag level and the boundary interval, the required course content is matched for the recommended path corresponding to the job competency level to obtain the job grading threshold; S3: Calling the position grading threshold, identifying the task content of each position level, collecting each student's click, interaction, and submission records in the learning task execution log, comparing the behavior type preset in the task stage with the behavior log trigger status, marking untriggered behaviors, and obtaining silent behavior records; S4: Based on the task-based knowledge point set and silent behavior records, extract the last trigger time, number of operations, and skill maintenance interval parameters of each knowledge point corresponding to the course content of the student, evaluate the degree of attenuation of the corresponding ability, adjust the recommendation priority of the corresponding course content, and obtain the recommendation list adjustment record.

2. The method for recommending high-vocational course content based on big data analysis according to claim 1, characterized in that: The task-based knowledge point set includes a task goal type label, a cognitive level label, and an achievement result type label. The job classification threshold is specifically a job capability level interval, a task level boundary parameter, and a course docking label. The silent behavior record includes the behavior untriggered stage number, the behavior type missing label, and the task stagnation duration parameter. The recommendation list adjustment record includes the knowledge point recommendation weight value, the compensation priority ranking value, and the ability decay mark item.

3. The method for recommending high vocational course content based on big data analysis according to claim 1 is characterized in that: The steps for obtaining the task-based knowledge point set are specifically as follows: S111: Obtain task goal verbs, action result keywords, and cognitive level description phrases in the higher vocational course teaching task text, construct a course intention label structure tuple, and obtain a task semantic label set; S112: Extract the title field of each knowledge point in the course outline based on the task semantic tag set, compare the semantic similarity between the keywords in each tag tuple and the words in the knowledge point title, calculate the tag coverage strength value, analyze the tag matching degree of each knowledge point, and obtain a tag coverage strength coefficient set; S113: Calling the label coverage strength coefficient set, comparing the coverage strength value of each knowledge point title field in the course outline with the task label recognition threshold, dividing them into task-based knowledge points and non-task-based knowledge points, and obtaining a task-based knowledge point set.

4. The method for recommending high vocational course content based on big data analysis according to claim 1, characterized in that: The specific steps for obtaining the silent behavior record are: S311: Call the position grading threshold, identify the task content of each position level, extract the task identification field and the task stage number field, and establish an index mapping between the position level and the stage task number to obtain the position task stage index set; S312: Based on the job task stage index set, collect the click, interaction, and submission records of each student in the learning task execution log, extract the task binding behavior type and the actual behavior triggering status, calculate the student's behavior missing degree value in each task stage, and record the missing item type, task number, and behavior deviation degree to obtain the task stage behavior missing parameter set; S313: Calculate the behavior missing duration of each student based on the task phase behavior missing parameter set, and obtain the silent behavior record.

5. The method for recommending high-vocational course content based on big data analysis according to claim 4 is characterized in that: The steps for obtaining the recommendation list adjustment record are specifically as follows: S411: Extracting the last trigger time, operation times, and skill maintenance interval parameters of each student's knowledge point corresponding to the course content based on the task-based knowledge point set and the silent behavior record, and obtaining a knowledge point execution basic parameter set; S412: Executing the basic parameter set for each knowledge point, comparing the last triggering time of each knowledge point with the skill maintenance interval parameter, calculating the ratio of the number of knowledge point operations deviating from the maintenance benchmark value, evaluating the attenuation degree of the target student's ability corresponding to each knowledge point, and obtaining an ability attenuation degree value; S413: According to the capability attenuation degree value and the course content mapped to the knowledge point, the priority of the target course content in the recommendation list is adjusted, and a recommendation list adjustment record is obtained.

6. The method for recommending high-vocational course content based on big data analysis according to claim 1 is characterized in that: The method further comprises: S5: Calling the recommendation list adjustment record, extracting the student's operation intensity parameters, resource call frequency, and stay period per unit time, identifying the learning behavior load state, and adjusting the planned cycle of the course content push task to generate a resource delivery adjustment result; The resource delivery adjustment result includes the push cycle delay value, the task queue adjustment position, and the resource rhythm control parameters.

7. The method for recommending high-vocational course content based on big data analysis according to claim 6 is characterized in that: The steps for obtaining the resource allocation adjustment result are specifically as follows: S511: Calling the recommendation list adjustment record, extracting the number of click operations, course resource call frequency, and page dwell period of the student in a unit time, and obtaining the original parameter set of the behavior load; S512: Analyze the learning behavior activity of multiple students based on the original behavior load parameter set, the number of click operations, the resource call frequency, and the page dwell period, calculate the learning behavior load status, and obtain the learning behavior load status value; S513: According to the learning behavior load state value and in combination with the current planned push time of each course content in the recommendation list, the planned cycle of the course content push task is adjusted to obtain the resource delivery adjustment result.

8. A high-vocational course content recommendation system based on big data analysis, characterized in that: The system is used to implement the method for recommending higher vocational course content based on big data analysis according to any one of claims 1 to 7, and the system includes: The course intention parsing module obtains the task objective verbs, action result keywords, and cognitive level description phrases in the teaching task text of the higher vocational course, constructs the course intention label structure tuple, compares the label structure tuple with each knowledge point in the course outline and calculates the matching degree, screens the knowledge points, and establishes a task-based knowledge point set; The job capability analysis module extracts the number of tasks, the number of operation processes, and the independent authority levels in the job responsibility text based on the task-based knowledge point set, identifies the operational complexity and responsibility control scope of each responsibility item, delineates the job capability level boundaries, and generates job classification thresholds; The behavioral silence diagnosis module identifies the task content of each job level based on the job classification threshold and collects the student's online learning task execution log data, including clicks, interactions, and submissions. It compares the behavior types preset in the task stage with the behavior log trigger status, marks untriggered behaviors, and obtains silent behavior records. The knowledge decay assessment module extracts the last trigger time, number of operations, and skill maintenance interval parameters of each knowledge point corresponding to the course content of each student based on the task-based knowledge point set and silent behavior records. It calculates the decay degree of the skill corresponding to the knowledge point, dynamically adjusts the recommendation priority of the course content associated with the knowledge point, and obtains the adjustment record of the recommendation list. The load status control module adjusts the records based on the recommendation list, extracts the students' operation intensity parameters, resource call frequency and stay period per unit time, identifies the students' learning behavior load status, and adjusts the planned cycle of the course content push task to generate resource delivery adjustment results.

Images