Intelligent construction education platform course recommendation method and system

Abstract

This invention discloses a course recommendation method and system for an intelligent construction education platform, relating to the field of intelligent education information technology. The method includes constructing a knowledge graph to establish multi-level associations between tasks and knowledge points; analyzing assessment data, setting score thresholds to identify students' weak knowledge points, integrating knowledge point weakness weights and job weights to generate comprehensive knowledge point weights; generating course recommendation results based on these comprehensive knowledge point weights, updating score data during student assessments after course completion, and using feedback to optimize the recommendation logic in real time. The method described in this invention implements a multi-dimensional weight-driven course recommendation path construction mechanism; employing dynamic feedback and real-time scoring algorithms, it establishes a closed-loop logic of recommendation, assessment, and re-recommendation, effectively improving the personalization and job matching of recommendations, and solving the problems of untraceable knowledge points, weak job adaptability, and singular recommendation strategies in existing platforms.

Description

Technical Field

[0001] This invention relates to the field of intelligent education information technology, specifically to a method and system for intelligently building an education platform for course recommendation. Background Technology

[0002] With the continuous advancement of information technology, artificial intelligence, and educational big data processing capabilities, intelligent education platforms have been widely applied in vocational training, skills education, and engineering construction. In vocational colleges specializing in construction, industry training institutions, and in-house training systems of engineering companies, online platforms play a crucial role in course resource management, learning path planning, and skills assessment. To enhance course matching, most platforms integrate recommendation algorithms, constructing personalized recommendation mechanisms based on learning behavior data. Existing platforms primarily rely on static behavioral features such as click records, course completion rates, and page dwell time to build matching models, forming a relatively preliminary course recommendation framework.

[0003] The promotion of intelligent recommendation strategies has revealed several technical shortcomings. Most current platforms do not combine learners' mastery of knowledge points to conduct recommendation modeling, nor do they construct multi-level bidirectional associations between knowledge points and chapter tasks. They lack task recommendation paths driven by knowledge structure, and course recommendation results are mainly driven by course tags, without modeling based on the knowledge distribution logic within the course. Especially in architecture courses, the course structure often contains multiple cross-tasks, and the knowledge point coverage is complex. Existing platforms lack the ability to support task-level knowledge tracking, and the recommendation logic cannot effectively reflect the learning structure.

[0004] Recommendation systems underutilize assessment data. Most platforms fail to feed back chapter task completion status or scores into the recommendation logic. Recommendation paths lack responsiveness and adjustment to real-time learning outcomes. At the system architecture level, teaching content and job skill requirements are often stored separately without structural integration. This prevents the platform from dynamically adjusting course priorities based on job requirements, making it difficult for recommendation results to support competency-oriented learning objectives. The disconnect between course content and job skills is particularly evident in professional modules such as structural design and BIM modeling. The recommendation mechanism of intelligent education platforms still has room for improvement in terms of knowledge organization granularity, feedback timeliness, and job suitability. It is necessary to introduce course recommendation methods based on knowledge graph construction, dynamic score analysis, and job competency integration to improve recommendation accuracy and path intelligence. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by this invention is that existing intelligent construction education platform course recommendation methods suffer from problems such as coarse knowledge structure modeling, low recommendation accuracy, poor job suitability, and delayed feedback response. The invention also addresses how to dynamically identify knowledge point mastery based on branch task assessment data, adjust recommendation weights by integrating job skill requirements, and achieve real-time optimization and iteration of recommendation results.

[0007] To address the aforementioned technical problems, this invention provides the following technical solution: a course recommendation method for an intelligent construction education platform, comprising constructing a knowledge graph to complete multi-level associations between tasks and knowledge points; analyzing assessment data, setting score thresholds to identify students' weak knowledge points, integrating knowledge point weakness weights and job weights to generate comprehensive knowledge point weights; generating course recommendation results based on comprehensive knowledge point weights, updating score data in the assessment after students complete the course, and using feedback to optimize the recommendation logic in real time; the multi-level associations include many-to-many mapping of knowledge points and branch tasks, forming a bidirectional searchable index relationship, the association relationship supporting multiple coverage and cross-mapping between tasks and knowledge points in the graph, allowing one task to correspond to multiple knowledge points, and one knowledge point to be referenced by multiple tasks; integrating knowledge point weakness weights and job weights includes calculating the score ratio, comparing the score rate with a preset score threshold; if the score is lower than the target score, the student is considered to have a weak knowledge point. The threshold process involves extracting knowledge points corresponding to tasks and combining them with the job weights assigned to these knowledge points in the job competency model. A linear weighting method is used to merge these two types of weight data—knowledge point weakness weights and job weights—to generate a comprehensive knowledge point weight for recommendation ranking. Identifying students' weak knowledge points involves calculating the score rate of each side task based on the student's actual score and the preset total task score. This score rate is compared to a set score threshold. If the score is lower than the threshold, all knowledge points associated with the side task are marked as needing reinforcement. If the score is higher than the threshold, the knowledge point weakness weight is calculated based on the side task score and the task's proportion in the total score, forming a student knowledge point weakness profile. The linear weighting process involves setting preset and dynamic adjustment values ​​based on adjustable parameters for calculating the comprehensive knowledge point weights. These values ​​are adjusted according to job changes to construct a job-oriented weighted knowledge point recommendation priority list.

[0008] As a preferred embodiment of the course recommendation method for the intelligent construction education platform described in this invention, the construction of the knowledge graph includes: a preset knowledge point database; the construction of multiple knowledge point nodes using a tree-like classification structure; and the configuration of a unique identifier, name, category, set of prerequisite knowledge points, and job suitability weight for each knowledge point.

[0009] As a preferred embodiment of the course recommendation method for the intelligent construction education platform described in this invention, the multi-level association includes: based on the chapter task content already set in the intelligent construction education platform, performing a many-to-many mapping between multiple set branch tasks and related knowledge points in the knowledge point library, and maintaining a bidirectional searchable index relationship between chapters, tasks, and knowledge points in the graph.

[0010] As a preferred embodiment of the course recommendation method for the intelligent construction education platform described in this invention, the identification of students' weak knowledge points includes: calculating the score rate of the branch tasks based on the student's actual score after completing the branch tasks and the preset total score of the tasks, comparing it with a set score threshold; if it is lower than the score threshold, marking all knowledge points associated with the branch tasks as knowledge points to be strengthened; if it is higher than the score threshold, calculating the knowledge point weakness weight of the knowledge points to be strengthened according to the branch task score and the proportion of the task in the total score, forming a student's knowledge point weakness profile.

[0011] As a preferred embodiment of the course recommendation method for the intelligent construction education platform described in this invention, the integration of job weights includes: obtaining the target knowledge point weights of the current student's job from the enterprise job database, and linearly weighting the job weights and knowledge point weakness weights to obtain the comprehensive knowledge point weights of the knowledge points to be recommended.

[0012] As a preferred embodiment of the course recommendation method for the intelligent construction education platform described in this invention, the linear weighted processing includes setting a platform preset value and a dynamic adjustment value based on weight parameters, making corrections according to changes in job positions, and constructing a job-oriented weighted knowledge point recommendation priority list.

[0013] As a preferred embodiment of the course recommendation method for the intelligent construction education platform described in this invention, the generation of course recommendation results includes: retrieving all candidate courses from the course database that match the knowledge points corresponding to the weight of each comprehensive knowledge point, performing multi-dimensional priority sorting, calculating a comprehensive recommendation score, and establishing a candidate course set.

[0014] As a preferred embodiment of the course recommendation method for the intelligent construction education platform described in this invention, the calculation of the comprehensive recommendation score includes a weighted composite recommendation score based on three factors: the weight of comprehensive knowledge points, the global completion rate of the course in the platform's historical data, and the current student's learning preference value for the course, weighted according to a set ratio.

[0015] As a preferred embodiment of the course recommendation method for the intelligent construction education platform described in this invention, the real-time feedback optimization recommendation logic includes: after a student completes a recommended course, an online assessment is re-initiated for the side tasks associated with the course, and the student's score rate on the side tasks is calculated; if the score rate is higher than the platform's set score threshold, the knowledge points associated with the side tasks are removed from the list of areas to be strengthened and will not participate in the next round of recommendations; if the score rate is lower than the score threshold, the knowledge points are maintained in a high-priority state, the recommendation iteration logic is triggered, and the course type, difficulty level, and presentation method are changed according to the original recommendation record, and the recommendation results are re-output.

[0016] Another objective of this invention is to provide an intelligent construction education platform course recommendation system that can solve the problems of weak knowledge point structure expression, unclear task association, and inability to accurately track students' knowledge mastery in current course recommendation systems by constructing a knowledge graph and realizing multi-level association between tasks and knowledge points.

[0017] As a preferred embodiment of the intelligent construction education platform course recommendation system of the present invention, it includes: a knowledge graph construction module, a weak knowledge point identification module, a job matching and integration module, a multi-dimensional course recommendation module, and a feedback and iterative optimization module; the knowledge graph construction module is used to read the knowledge point database and chapter task information during initial course updates, construct a multi-level many-to-many mapping graph structure, and generate a bidirectional searchable index between tasks, chapters, and knowledge points for subsequent recommendation and traceability; the weak knowledge point identification module is used to determine whether a student's knowledge point is to be strengthened based on a comparison of the score with a score threshold after the student completes a side task, calculate the weakness weight based on the task score ratio, and update the student's knowledge point weakness profile. The job matching and integration module is used to call the job weights set for knowledge points in the job competency model, linearly weight the job weights and the knowledge point weakness weights, output the comprehensive knowledge point weights, and dynamically adjust the weight parameters according to job changes; the multi-dimensional course recommendation module is used to filter candidate courses from the course library according to the comprehensive knowledge point weights, sort them by weighted scores based on knowledge point matching degree, course completion rate, and student type preference, filter out courses below the set score threshold, and output the recommendation list; the feedback and iterative optimization module is used to trigger a reassessment after the student completes the recommended course. If the score rate is higher than the score threshold, the mastery status is updated; if it is not up to standard, the course type is switched to enter the next round of recommendation, until the standard is met and the upper limit of the round is reached.

[0018] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program as steps to implement a course recommendation method for an intelligent construction education platform.

[0019] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of a method for recommending courses on an intelligent construction education platform.

[0020] The beneficial effects of this invention are as follows: The intelligent construction education platform course recommendation method provided by this invention constructs a knowledge graph to complete multi-level associations between tasks and knowledge points, realizing a many-to-many mapping from chapters and sub-tasks to knowledge points. This enables the recommendation logic to accurately locate the knowledge point mastery structure, improving the knowledge tracking ability and path interpretability of the recommendation. By analyzing assessment data, identifying students' weak knowledge points, and setting score thresholds to integrate job weights, a weight fusion mechanism from dynamic student performance profiles to job skill requirements is realized, making the recommendation ranking more aligned with job competency requirements and enhancing the career orientation of the recommended content. By generating course recommendation results based on the identified weak knowledge points of students and using feedback to optimize the recommendation logic in real time, a real-time iterative recommendation process driven by assessment results is realized, ensuring that the recommendation path is continuously adjusted and optimized, improving the timeliness of recommendation response and the accuracy of course reinforcement. This invention achieves better results in terms of knowledge structure modeling accuracy, job orientation matching ability, and dynamic recommendation feedback mechanism. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is an overall flowchart of a course recommendation method for an intelligent construction education platform provided in Embodiment 1 of the present invention.

[0023] Figure 2 This is a flowchart illustrating the role interaction and processing of a course recommendation system for an intelligent construction education platform provided in Embodiment 2 of the present invention.

[0024] Figure 3 This is a schematic diagram of an intelligent construction education platform course recommendation system provided in Embodiment 2 of the present invention. Detailed Implementation

[0025] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0026] Example 1, referring to Figure 1 As an embodiment of the present invention, a method for recommending courses on an intelligent construction education platform is provided, comprising:

[0027] S1: Construct a knowledge graph to establish multi-level connections between tasks and knowledge points.

[0028] Furthermore, constructing a knowledge graph includes: pre-setting a knowledge point base; constructing multiple knowledge point nodes using a tree-like classification structure; and configuring each knowledge point with a unique identifier, name, category, set of prerequisite knowledge points, and job suitability weight.

[0029] It should be noted that the multi-level association includes mapping multiple branch tasks and related knowledge points in the knowledge point library to each other based on the chapter task content set in the intelligent construction education platform, and maintaining a bidirectional searchable index relationship between chapters, tasks and knowledge points in the graph.

[0030] It should be noted that the tree-like classification structure includes, based on the knowledge system of the construction engineering field, dividing knowledge points layer by layer according to the field direction to form a classification path with hierarchical logic. This includes using BIM technology as the root node of the knowledge graph to identify the professional direction to which the course belongs; setting modeling specifications as standard process sub-nodes at the root node to aggregate key technical rules for construction; and further refining assessable skill nodes under the standard process sub-nodes to associate course tasks, assessment standards, and students' knowledge mastery status.

[0031] It should also be noted that setting modeling standards includes standardizing the naming format in tasks, evaluating the accuracy and traceability of students' task completion based on the parameters and constraints in the tasks, standardizing the modeling operation methods and parameter configuration structures in course tasks, supporting the precise association between tasks and knowledge points, and establishing the logic for assessing and recommending students' parametric modeling abilities.

[0032] S2: Analyze the assessment data, set a score threshold to identify the trainees' weak knowledge points, integrate the weight of knowledge point weakness and the weight of the position, and generate a comprehensive knowledge point weight.

[0033] Furthermore, identifying students' weak knowledge points involves calculating the score rate of the side tasks based on the student's actual score after completing the side tasks and the preset total score of the tasks. This score rate is then compared with a set score threshold. If the score rate is lower than the threshold, all knowledge points associated with the side tasks are marked as knowledge points to be strengthened. If the score rate is higher than the threshold, the knowledge point weakness weight is calculated based on the side task score and the proportion of the task in the total score, thus forming a profile of the student's weak knowledge points.

[0034] Furthermore, the expression for calculating the score rate of the side quest is:

[0035]

[0036] Among them, S task S represents the student's side quest score rate. actual S represents the student's actual score in this side quest. total This represents the set total score for the side task. Based on the aforementioned side task score rate, a score threshold is determined. If a student's side task score rate is less than 60%, the knowledge point associated with that task is marked as needing reinforcement. Furthermore, based on the correlation between the task score rate and the knowledge point, the weight of the knowledge point's weakness is calculated, expressed as:

[0037]

[0038] Among them, W weak This represents the weight of a student's weakness in a particular knowledge point, where n represents the number of branch tasks associated with that knowledge point, and i represents the index of the branch task. This represents the student's score rate for the i-th side task. P represents the total score set for the i-th side quest for the student. total This represents the total score for the chapter containing the knowledge point. The "level of not mastering" (i.e., the score level) is calculated by weighted summing of all related sub-tasks for that knowledge point. Multiplying the weight of the task in the chapter yields the overall level of mastery of that knowledge point. The higher the weight, the more the knowledge point needs to be strengthened, making it a priority target for reinforcement in the course recommendation system. Through the analysis of side task scores, the system achieves a response time of seconds. If a student fails to master a certain knowledge point in the assessment, the system will immediately push related courses after the student submits the assessment, saving the update cycle.

[0039] It should be noted that the integration of knowledge point weakness weights and job weights includes calculating the score ratio and comparing the score rate with a preset score threshold. If the score is lower than the threshold, the knowledge points corresponding to the task are extracted, and combined with the job weights set for the knowledge points in the job competency model, the two types of weight data, knowledge point weakness weights and job weights, are merged in a linear weighting manner to generate a comprehensive knowledge point weight for recommendation ranking.

[0040] It should be noted that linear weighting includes setting platform preset values ​​and dynamic adjustment values ​​based on adjustable parameters for calculating the weights of comprehensive knowledge points, making corrections according to changes in job positions, and constructing a job-oriented weighted knowledge point recommendation priority list.

[0041] It should also be noted that constructing a job-oriented weighted knowledge point recommendation priority list includes receiving the current trainee's job identification information returned by job identification, extracting the set of knowledge point weights for the corresponding job from the job competency model, creating a one-to-one mapping between each knowledge point and the job weight, and obtaining the comprehensive knowledge point weight of the knowledge points to be recommended, represented as:

[0042] W total =α·W weak +β·W position

[0043] Among them, W total W represents the overall recommendation weight of the knowledge point to be recommended. weak W represents the weight of a student's weakness in a particular knowledge point. position This represents the importance weight of the target position for this knowledge point. α and β are adjustable parameters used to control the fusion ratio between the two, typically satisfying α + β = 1. Before the overall recommendation weight of the knowledge point to be recommended, W is... weak W position The values ​​of each are divided by W. weak W position The maximum value within the statistical calculation range is normalized to map data of different dimensions to dimensionless values ​​in the interval [0,1]. If a certain knowledge point has high requirements for the job (W... position (Large) and the students' grasp is weak (W) weak (Large), then the comprehensive knowledge point weight W total The higher the score, the higher the ranking in the recommendation list. This integrated mechanism ensures that course recommendations not only reflect the current skill gaps of learners but also the actual skill requirements of enterprise positions. Through dual calibration of knowledge point networks and job weights, the accuracy of recommendations is significantly improved, leading to increased learner satisfaction with recommended courses.

[0044] S3: Based on the weight of comprehensive knowledge points, generate course recommendation results, update the score data in the assessment after students complete the course, and use feedback to optimize the recommendation logic in real time.

[0045] Furthermore, generating course recommendation results includes retrieving all candidate courses from the course database that match the knowledge points corresponding to the weight of each comprehensive knowledge point, performing multi-dimensional priority ranking, calculating a comprehensive recommendation score, and establishing a candidate course set.

[0046] It should be noted that the comprehensive recommendation score is calculated by weighting three factors according to a set ratio: the weight of comprehensive knowledge points, the global completion rate of the course in the platform's historical data, and the current student's learning preference value for the course.

[0047] It should be noted that the recommendation rating model is expressed as follows:

[0048] S course =γ·W total +δ·R comp +ε·P type

[0049] Among them, S course This represents the overall score for course recommendations, used for ranking. (W) total R represents the overall recommendation weight of the knowledge point to be recommended. comp This represents the average completion rate of the course over the platform's history, reflecting course quality and acceptance. (P) type γ represents the current student's preference coefficient for this type of course, δ represents the weight parameter for knowledge point relevance, ε represents the weight parameter for course completion rate, and ε represents the weight parameter for course type preference, satisfying γ + δ + ε = 1. Information from these three dimensions is integrated into a single score. Before calculating the comprehensive course recommendation score, W... total R comp P type The values ​​of each are divided by W. total R comp P type The statistical calculation performs normalization within the sample range, mapping data of different dimensions to dimensionless values ​​in the [0,1] interval. The system then calculates S for all candidate courses. course Then, sort by score and select the top few items as the recommendation results output. This ensures that the course recommendations are not only based on accurate matching of knowledge needs, but also take into account the acceptability of courses and user preferences, thus constructing a dynamic sorting logic driven by multiple factors.

[0050] It should also be noted that the real-time feedback optimization recommendation logic includes the following steps: after a student completes a recommended course, an online assessment is re-initiated for the associated side tasks, and the student's score rate on the side tasks is calculated. If the score rate is higher than the platform's set score threshold, the knowledge points associated with the side tasks are removed from the list of areas to be reinforced and will not participate in the next round of recommendations. If the score rate is lower than the score threshold, the knowledge points are maintained as high-priority, triggering the recommendation iteration logic. Based on the original recommendation record, the course type, difficulty level, and presentation method are changed, and the recommendation results are re-output.

[0051] Example 2, refer to Figures 2-3 As an embodiment of the present invention, an intelligent construction education platform course recommendation system is provided, including a knowledge graph construction module 100, a weak knowledge point identification module 200, a job adaptation and integration module 300, a multi-dimensional course recommendation module 400, and a feedback and iterative optimization module 500.

[0052] X1: The knowledge graph construction module 100 includes a knowledge point base management submodule 101 and a chapter and task mapping submodule 102.

[0053] It should be noted that the knowledge point base management submodule 101 is used to preset a multi-level knowledge point structure in the field of construction engineering. It uses a tree-like classification structure to configure the unique identifier, name, category, prerequisites, and job suitability weight of each knowledge point, thereby achieving a standardized representation of knowledge content. The chapter and task mapping submodule 102 is used to associate knowledge points with tasks in a many-to-many manner according to the chapters and branch task content set by the platform, and to build a bidirectional searchable index relationship between tasks and knowledge points.

[0054] It should also be noted that the knowledge graph construction module 100 initialization is used to establish a structural graph between tasks and knowledge points, which is the basis for subsequent analysis of students' abilities and generation of recommended courses.

[0055] X2: The weak knowledge point identification module 200 includes a score rate calculation submodule 201 and a knowledge point weakness analysis submodule 202.

[0056] It should be noted that the score rate calculation submodule 201 is used to collect the scores of trainees in the side tasks, and calculate the score rate of the side tasks by comparing the actual score of the side task with the set total score of the side task; the knowledge point weakness analysis submodule 202 is used to compare the score rate with the preset score threshold. If it is lower than the score threshold, the corresponding knowledge point is marked as an item to be strengthened, and a weighted sum is performed to generate the knowledge point weakness weight.

[0057] It should also be noted that the weak knowledge point identification module 200 is based on the constructed graph structure, analyzes the student's performance in real time, and outputs a personalized knowledge mastery graph, providing the original weight basis for subsequent job matching and course ranking.

[0058] X3: The job matching and integration module 300 includes the job skills extraction sub-module 301 and the comprehensive knowledge point weight calculation sub-module 302.

[0059] It should be noted that the job skills extraction submodule 301 is used to extract the knowledge points required for the trainee's job and their job weights from the enterprise job competency database; the comprehensive knowledge point weight calculation submodule 302 is used to linearly weight the job weights and the knowledge point weakness weights, and generate comprehensive knowledge point recommendation weights based on the proportion parameters set by the platform.

[0060] It should also be noted that the Job Matching Integration Module 300 integrates educational content with job requirements, making the recommendation logic not only based on the student's current abilities but also taking into account the actual hiring standards of enterprises, thereby improving the job matching degree of the recommended courses.

[0061] X4: The multi-dimensional course recommendation module 400 includes a candidate course screening sub-module 401 and a recommendation score calculation sub-module 402.

[0062] It should be noted that the candidate course screening submodule 401 is used to extract the set of all courses that match the knowledge points to be recommended from the course database; the recommendation score calculation submodule 402 is used to integrate three factors: comprehensive knowledge point weight, course historical completion rate and student course type preference, and calculate the recommendation score for each course based on the weighted scoring model.

[0063] It should also be noted that the multi-dimensional course recommendation module 400 generates the final recommendation list and is a key execution module for the transition from knowledge to courses, forming a closed loop with the feedback and iterative optimization module 500.

[0064] X5: Feedback and Iterative Optimization Module 500 includes Reassessment Submodule 501 and Recommended Path Update Submodule 502.

[0065] It should be noted that the reassessment submodule 501 is used to reassess the related tasks after the student completes the recommended course and update the score data; the recommendation path update submodule 502 is used to remove the relevant knowledge points with a score rate higher than the set score threshold from the list of areas to be strengthened; if the score is not met, the course type, difficulty or content strategy is changed and the next round of recommendations is triggered.

[0066] It should also be noted that the feedback and iterative optimization module 500 enables the adaptive evolution of the recommendation mechanism and is the part of the dynamic recommendation closed loop, ensuring that the recommendation process is continuously corrected and optimized.

Claims

1. A method for recommending courses on an intelligent construction education platform, characterized in that, include: Construct a knowledge graph to establish multi-level connections between tasks and knowledge points; Analyze the assessment data, set scoring thresholds to identify trainees' weak knowledge points, and integrate the weights of knowledge point weakness and job position to generate a comprehensive knowledge point weight; Based on the weight of comprehensive knowledge points, course recommendation results are generated, and the score data is updated in the assessment after students complete the course. The recommendation logic is optimized in real time by using feedback. Multi-level associations include many-to-many mappings of knowledge points and branch tasks, forming bidirectional searchable index relationships. These relationships support multiple coverages and cross-mappings between tasks and knowledge points in the graph, allowing one task to correspond to multiple knowledge points and one knowledge point to be referenced by multiple tasks. The method integrates the weighting of knowledge point weaknesses and job position, calculates the score ratio, and compares the score rate with the preset score threshold. If the score is below the threshold, the knowledge points corresponding to the task are extracted. Combined with the job weights set for the knowledge points in the job competency model, the two types of weight data, knowledge point weakness weight and job weight, are merged in a linear weighting method to generate a comprehensive knowledge point weight for recommendation ranking. Identifying students' weak knowledge points includes: Based on the student's actual score in completing the side task and the preset total score, the score rate of the side task is calculated and compared with the set score threshold. If it is lower than the score threshold, all knowledge points associated with the side task are marked as knowledge points to be strengthened. If the score exceeds the threshold, the weight of the knowledge point weakness is calculated based on the score of the side task and the proportion of the task in the total score, forming a profile of the student's knowledge point weakness. Linear weighting includes, Based on adjustable parameters that calculate the weight of comprehensive knowledge points, the platform sets preset values ​​and dynamic adjustment values, and adjusts them according to changes in job positions to construct a job-oriented weighted knowledge point recommendation priority list.

2. The course recommendation method for the intelligent construction education platform as described in claim 1, characterized in that: The construction of the knowledge graph includes A pre-defined knowledge point base is constructed using a tree-like classification structure to create multiple knowledge point nodes. Each knowledge point is configured with a unique identifier, name, category, set of prerequisite knowledge points, and job suitability weight.

3. The course recommendation method for the intelligent construction education platform as described in claim 2, characterized in that: The multi-level association includes, Based on the chapter tasks already set in the intelligent construction education platform, multiple branch tasks and related knowledge points in the knowledge point library are mapped in a many-to-many manner, maintaining a bidirectional searchable index relationship between chapters, tasks, and knowledge points in the graph.

4. The course recommendation method for the intelligent construction education platform as described in claim 3, characterized in that: The generated course recommendation results include, Retrieve all candidate courses from the course database that match the knowledge points corresponding to the weight of each comprehensive knowledge point, sort them by multi-dimensional priority, calculate the comprehensive recommendation score, and establish a set of candidate courses.

5. The course recommendation method for the intelligent construction education platform as described in any one of claims 4, characterized in that: The calculation of the comprehensive recommendation score includes, The recommendation score is a composite score that combines three factors: the weight of the knowledge points, the overall completion rate of the course in the platform's historical data, and the current student's learning preference value for the course, weighted according to a set ratio.

6. The course recommendation method for the intelligent construction education platform as described in claim 5, characterized in that: The logic for real-time optimization of recommendations using feedback includes... After the student completes the recommended course, an online assessment is launched again for the side tasks associated with the course, and the student's score rate on the side tasks is calculated. If the score rate is higher than the platform's set score threshold, the knowledge points associated with the side quest will be removed from the list of quests to be strengthened and will no longer participate in the next round of recommendations. If the score rate is lower than the score threshold, the knowledge point will be kept in a high-priority state, triggering the recommendation iteration logic. Based on the original recommendation record, the course type, difficulty level, and presentation method will be changed, and the recommendation results will be re-output.

7. A course recommendation system for an intelligent construction education platform, employing the course recommendation method for an intelligent construction education platform as described in any one of claims 1 to 6, characterized in that: It includes a knowledge graph construction module (100), a weak knowledge point identification module (200), a job matching and integration module (300), a multi-dimensional course recommendation module (400), and a feedback and iterative optimization module (500). The knowledge graph construction module (100) is used to read the knowledge point base and chapter task information during the initialization of course updates, construct a multi-level many-to-many mapping graph structure, and generate a bidirectional searchable index between tasks, chapters and knowledge points for subsequent recommendation and tracing. The weak knowledge point identification module (200) is used to determine whether a knowledge point needs to be strengthened based on the comparison between the score and the score threshold after the student completes the side task, calculate the weakness weight in combination with the task score ratio, and update the student's knowledge point weakness profile. The job adaptation and integration module (300) is used to call the job weights set for knowledge points in the job competency model, perform linear weighting on the job weights and knowledge point weakness weights, output the comprehensive knowledge point weights, and dynamically adjust the weight parameters according to job changes. The multi-dimensional course recommendation module (400) is used to filter candidate courses from the course library according to the comprehensive knowledge point weight, sort them by weighted score based on knowledge point matching degree, course completion rate and student type preference, and output a recommendation list after filtering out courses below the set score threshold. The feedback and iterative optimization module (500) is used to trigger a reassessment after the student completes the recommended course. If the score rate is higher than the score threshold, the mastery status is updated. If the requirements are not met, switch to a different course type and proceed to the next round of recommendations until the requirements are met and the maximum number of rounds is reached.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent construction education platform course recommendation method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the course recommendation method for the intelligent construction education platform as described in any one of claims 1 to 6.

Images