An intelligent recommendation system for learning resources based on data analysis

Abstract

The present invention relates to the technical field of learning resource recommendation, and specifically discloses a learning resource intelligent recommendation system based on data analysis, the system comprising: a behavior collection module for collecting user video learning data, exercise interaction data and learning rhythm data in real time; a real-time analysis module, comprising a knowledge point mastery assessment unit and a learning load assessment unit, the knowledge point mastery assessment unit being used to assess the user's mastery of each knowledge point, and the learning load assessment unit being used to assess the user's current learning pressure; the present invention widely collects various learning data through the behavior collection module, the real-time analysis module accurately assesses the knowledge point mastery and learning pressure, the resource recommendation module flexibly adjusts the exercise difficulty and subsequent learning sequence according to the analysis results, the learning resource recommendation is adapted to the user's personalization, and appropriate knowledge points and exercises are accurately recommended to the user, achieving the effect of significantly improving learning efficiency and pertinence.

Description

Technical Field

[0001] The present invention relates to the technical field of learning resource recommendation, and in particular to an intelligent learning resource recommendation system based on data analysis. Background Art

[0002] With the rapid development of information technology and the acceleration of the digital transformation of global education, online education has shown explosive growth. The popularization of 5G networks, the widespread application of smart terminal devices, and the maturity of cloud computing technology have provided solid technical support for the online transmission of educational resources. This has enabled the new learning model based on video courses and online exercises to break through the time and space limitations of traditional classrooms and become the mainstream learning model covering all fields such as education, vocational training, and higher education.

[0003] Traditional online education systems mostly use preset linear learning paths, where knowledge points and exercises are pushed in a fixed order. This results in a low match between the difficulty of the recommended knowledge points or exercises and the user's actual mastery level. This may cause users to repeat simple content they have already mastered, resulting in boredom, or users may be forced to challenge difficult content beyond their current capabilities, which reduces their learning enthusiasm due to difficulties in understanding. This not only ignores the differences in individual knowledge mastery, but also easily leads to users copying answers from each other through standardized paths, weakening learning outcomes. In addition, traditional online education systems lack dynamic assessment of user learning load and fail to consider the learning pressure behind behaviors such as taking too long to answer questions, frequently modifying answers, or unexpectedly switching knowledge points. As a result, the recommendation strategy fails to adapt to the learner's real-time status, affecting learning efficiency and experience. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent recommendation system for learning resources based on data analysis to solve the following technical problems:

[0005] How to solve the problem of insufficient personalized adaptation in traditional online education learning resource recommendations and the inability to accurately recommend appropriate knowledge points and exercises to users?

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A learning resource intelligent recommendation system based on data analysis, the system comprising:

[0008] Behavior collection module, used to collect user video learning data, exercise interaction data and learning rhythm data in real time;

[0009] A real-time analysis module includes a knowledge point mastery assessment unit and a learning load assessment unit. The knowledge point mastery assessment unit is used to assess the user's mastery of each knowledge point, and the learning load assessment unit is used to assess the user's current learning pressure.

[0010] The resource recommendation module adjusts the difficulty of the recommended exercises based on the data obtained from the analysis of the knowledge point mastery assessment unit; and adjusts the knowledge points and exercise sequence for subsequent learning based on the data obtained from the analysis of the knowledge point mastery assessment unit and the learning load assessment unit.

[0011] Furthermore, the process of evaluating the user's mastery of each knowledge point includes:

[0012]

[0013]

[0014]

[0015] The mastery evaluation coefficient of each knowledge point is obtained by analyzing and calculating formulas (1)-(3) ;

[0016] in, 、 is the first weight coefficient, The number of times the knowledge point is played back during the learning process of the learning video. is the playback attenuation coefficient, is the influencing factor of the learning time of the corresponding learning video of the knowledge point, is the correct answer rate of the exercises corresponding to the knowledge points, is the number of revisions during the answering process of the corresponding exercises of the knowledge point, is the modification number impact factor, The actual learning time of the learning video corresponding to the knowledge point, The reference learning time for the learning video corresponding to the knowledge point, The number of pauses during the learning process of the learning video corresponding to the knowledge point, is the suspension impact factor.

[0017] Furthermore, the process of evaluating the user's current learning stress includes:

[0018]

[0019] The user's current learning pressure load index is obtained by analyzing and calculating formula (4) ;

[0020] in, The actual time it takes for the user to complete the previous exercise. The reference answer time for the previous exercise is: The number of times the user modified their answers while answering a certain number of exercises. The number of exercises that the user has skipped while answering a certain number of exercises. is the impact factor for skipping exercises, To preset the number of references, The number of times the user switches knowledge points during the learning process without following the system recommended path. The total number of times the user switches knowledge points during the learning process. 、 、 is the second weight coefficient.

[0021] Furthermore, the process of adjusting the difficulty of the recommended exercises includes:

[0022]

[0023]

[0024] The recommendation difficulty is obtained by analyzing and calculating the formula (5)-(6) ;

[0025] in, is the initial benchmark difficulty corresponding to the knowledge point, is the difficulty adjustment coefficient, 、 、 Adjust the slope coefficient for difficulty, To preset the first knowledge point mastery evaluation coefficient, To preset the second knowledge point, master the evaluation coefficient.

[0026] Furthermore, the process of adjusting the difficulty of the recommended exercises further includes:

[0027] When the user completes a certain number of exercises of the same difficulty level in succession, the difficulty adjustment slope coefficient will be modified based on the user's performance;

[0028]

[0029] The revised difficulty adjustment slope coefficient is obtained by analyzing and calculating formula (7) ;

[0030] in, The correct rate of completing a certain number of exercises, To preset the basic accuracy, is the correction amplitude coefficient.

[0031] Furthermore, the process of adjusting the order of knowledge points and exercises for subsequent learning includes:

[0032] Based on the topological structure of the knowledge graph, the initial learning path is generated according to the dependency relationship of knowledge points;

[0033] The user's current learning pressure load index and the preset learning stress load index threshold Make a comparison;

[0034] like ,but

[0035]

[0036]

[0037]

[0038]

[0039] The adjusted learning path is obtained by analyzing and calculating formulas (8)-(11) ;

[0040] in, Score the coherence of the knowledge points in the learning path, Scoring for learning novelty, is a dynamic adjustment factor, is the mean of the mastery evaluation coefficients of each knowledge point in the learning path, is the preset critical value of learning pressure load index, 、 is the third weight coefficient, is the mastery evaluation coefficient corresponding to the i-th knowledge point in the learning path, is the initial benchmark difficulty corresponding to the i-th knowledge point, is the initial benchmark difficulty corresponding to the i-1th knowledge point, is the proportion of knowledge points in the learning path that the user has not learned. is the degree of repetition between the learning path and the user’s historical learning path, is the basic correction factor, The learning path to maximize the function output value;

[0041] like , then only knowledge points, and shorten the initial learning path length to half of its original length, Master the evaluation coefficient based on the preset knowledge points.

[0042] Furthermore, the process of evaluating the user's mastery of each knowledge point also includes:

[0043] Analyze the mastery evaluation coefficient of knowledge points that users have not learned;

[0044] Extract target knowledge points from the knowledge graph All pre-dependent knowledge points

[0045]

[0046] The mastery evaluation coefficient of the knowledge points that have not been learned is obtained by analyzing and calculating the formula (12) ;

[0047] in, Pre-dependent knowledge points The mastery evaluation coefficient, Pre-dependent knowledge points Evaluation coefficient for mastering knowledge points that have not been learned The influence weight of .

[0048] Beneficial effects of the present invention:

[0049] (1) The present invention widely collects various learning data through the behavior collection module, the real-time analysis module accurately evaluates the mastery of knowledge points and learning pressure, and the resource recommendation module flexibly adjusts the difficulty of exercises and the subsequent learning sequence according to the analysis results. The learning resource recommendation is personalized and adapted to the user, and appropriate knowledge points and exercises are accurately recommended to the user, achieving the effect of significantly improving learning efficiency and targeting, and solving the problem of insufficient personalized adaptation in the recommendation of learning resources in traditional online education, and being unable to accurately recommend appropriate knowledge points and exercises to users. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The present invention will be further described below with reference to the accompanying drawings.

[0051] Figure 1 This is a schematic block diagram of a learning resource intelligent recommendation system based on data analysis proposed by the present invention. DETAILED DESCRIPTION

[0052] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0053] See also Figure 1 As shown, in one embodiment, a learning resource intelligent recommendation system based on data analysis is provided, the system comprising:

[0054] A behavior collection module is used to collect user video learning data, exercise interaction data, and learning rhythm data in real time. The user video learning data includes but is not limited to viewing time, number of pauses, and playback frequency. The exercise interaction data includes but is not limited to answer accuracy, answer speed, and number of revisions. The learning rhythm data includes but is not limited to chapter switching frequency and knowledge point skipping.

[0055] A real-time analysis module includes a knowledge point mastery assessment unit and a learning load assessment unit. The knowledge point mastery assessment unit analyzes the collected relevant data to assess the user's mastery of each knowledge point. The learning load assessment unit analyzes the collected relevant data to assess the user's current learning pressure.

[0056] The resource recommendation module adjusts the difficulty of recommended exercises based on the data obtained from the knowledge point mastery assessment unit analysis. If the user has a good grasp of a certain knowledge point, more difficult exercises can be recommended. The module also adjusts the order of knowledge points and exercises for subsequent learning based on the data obtained from the knowledge point mastery assessment unit and the learning load assessment unit analysis, and reasonably arranges the knowledge points and exercises to be learned subsequently, making the learning process more scientific and efficient.

[0057] Through the above technical solution, this embodiment provides an intelligent recommendation system for learning resources based on data analysis. The system widely collects various learning data through the behavior collection module, and the real-time analysis module accurately evaluates the degree of mastery of knowledge points and learning pressure. The resource recommendation module flexibly adjusts the difficulty of exercises and subsequent learning sequence according to the analysis results. The learning resource recommendation is adapted to the user's personalization, and appropriate knowledge points and exercises are accurately recommended to users, achieving the effect of significantly improving learning efficiency and pertinence.

[0058] In one embodiment, the process of evaluating the user's mastery of each knowledge point includes:

[0059]

[0060]

[0061]

[0062] The mastery evaluation coefficient of each knowledge point is obtained by analyzing and calculating formulas (1)-(3) ;

[0063] in, 、 is the first weight coefficient, which reflects the relative importance of each factor to the mastery evaluation coefficient and can be preset through experience. The number of times a knowledge point is replayed during the learning process of the learning video. The number of replays reflects the difficulty of the user in understanding the video content. A high number of replays may indicate difficulty in understanding. The acquisition module collects the user's video learning behavior to obtain the information. is the replay attenuation coefficient, which can be obtained by fitting experimental data. As the number of replays increases, the influence of each replay on the knowledge point mastery evaluation gradually decreases. is the influencing factor of the learning time of the learning video corresponding to the knowledge point, which can be obtained by statistically analyzing the mastery of the knowledge points corresponding to learning videos of different lengths. The correct answer rate for the exercises corresponding to the knowledge point is the proportion of correct answers given by users when doing the exercises corresponding to the knowledge point. It can be calculated by recording the user's exercise answer results through the behavior collection module. The correct answer rate directly reflects the user's ability to apply the knowledge point. The number of revisions during the answering process of the corresponding knowledge point exercises can be obtained by recording the user's interactive behavior in answering questions. A large number of revisions may indicate that the user's grasp of the knowledge point is not solid enough. is the modification number influencing factor, which indicates the degree of influence of the modification number on the knowledge point mastery evaluation, and can be obtained through experimental data analysis. The actual learning time of the learning video corresponding to the knowledge point is the time the user actually spends on learning the video corresponding to the knowledge point. The user's video learning time is recorded by the behavior collection module, which reflects the user's energy invested in learning the knowledge point. The reference learning time for the learning video corresponding to the knowledge point is set based on general circumstances or teaching experience. The comparison between the actual learning time and the reference learning time can show the user's learning efficiency of the knowledge point. The number of pauses during the learning process of the learning video corresponding to the knowledge point is the number of times the user pauses when learning the video corresponding to the knowledge point. It is obtained by recording the user's video learning behavior. A high number of pauses may indicate that the user is confused during the learning process. is the pause influencing factor, which can be obtained through experimental data and used to adjust the impact of the number of pauses in the final evaluation coefficient.

[0064] Through the above technical solution, this embodiment discloses a method for evaluating the user's mastery of each knowledge point. By comprehensively and carefully considering factors such as the playback, duration, pause status of the learning video and the answer status of the exercises, the mastery evaluation coefficient is obtained by analysis and calculation using formulas (1)-(3), thereby achieving the benefit of more accurately evaluating the user's mastery of the knowledge point and achieving the effect of providing a more reliable basis for subsequent recommendations.

[0065] In one embodiment, the process of evaluating the user's current learning stress includes:

[0066]

[0067] The user's current learning pressure load index is obtained by analyzing and calculating formula (4) ;

[0068] in, The actual time the user took to complete the previous exercise is recorded through the behavior collection module. The reference time for answering the previous question is the reasonable answering time set according to the difficulty of the question and the answering performance of the average student. The comparison between the actual answering time and the reference time can reflect the user's perception of the difficulty of answering the question. The number of times a user modifies their answers while answering a certain number of exercises is obtained by recording the user's interactive behavior in answering questions. The number of questions that the user has skipped while answering a certain number of questions. This number is obtained by recording the user's answering behavior. The number of times the answer is modified and the number of questions skipped reflect the difficulty level of the user in the answering process. is the impact factor of skipped exercises, determined through experimental data, and used to adjust the role of the number of skipped exercises in the final index. The preset reference number is obtained based on experience. This is the number of times the user switches knowledge points in a non-system-recommended path during the learning process. This indicates the number of times the user switches knowledge points in a non-system-recommended order during the learning process. This can be obtained by recording the user's learning path. Switching not according to the recommended path may indicate that the user is confused about the current learning content or has their own learning needs. It is the total number of times the user switches knowledge points during the learning process. It is the cumulative number of times the user switches knowledge points during the entire learning process. It can be obtained by recording the user's learning path statistics. 、 、 is the second weight coefficient, which is the proportion of each factor when calculating the learning stress load index, and can be obtained through analysis of a large amount of experimental data.

[0069] Through the above technical solution, this embodiment provides a method for evaluating the user's current learning pressure. The method comprehensively considers factors such as the time spent on answering questions, the number of times the answer is modified, the number of skipped exercises, and the knowledge point switching situation, and uses formula (4) to analyze and calculate the learning pressure load index, thereby accurately evaluating the user's learning pressure.

[0070] In one embodiment, the process of adjusting the difficulty of the recommended exercises includes:

[0071]

[0072]

[0073] The recommendation difficulty is obtained by analyzing and calculating the formula (5)-(6) ;

[0074] in, The initial benchmark difficulty corresponding to the knowledge point is the basic difficulty set according to the characteristics of the knowledge point itself and the requirements of the syllabus. It can be set by education experts based on factors such as the complexity and importance of the knowledge point. is the difficulty adjustment coefficient, 、 、 The difficulty adjustment slope coefficient refers to the rate at which the recommended difficulty changes with the knowledge point mastery evaluation coefficient within different knowledge point mastery evaluation coefficient intervals. It can be obtained through experimental data fitting and is used to control the magnitude of changes in the recommended difficulty with the knowledge point mastery. To preset the first knowledge point mastery evaluation coefficient, To preset the second knowledge point mastery evaluation coefficient, 、 The mastery threshold standard is set based on experience. The difficulty of the recommended exercises is dynamically adjusted according to the comparison between the user's actual mastery and the preset mastery threshold standard. When the knowledge point mastery evaluation coefficient is higher than When the recommendation difficulty increases accordingly, The difficulty of recommendation will be reduced accordingly. When the knowledge point mastery evaluation coefficient is higher than The difficulty of recommendation will increase significantly.

[0075] Through the above technical solution, this embodiment adopts a method for adjusting the difficulty of recommended exercises. The method cleverly combines the initial baseline difficulty of knowledge points and the user's knowledge point mastery evaluation coefficient, and uses formulas (5)-(6) to analyze and calculate the recommended difficulty. This can achieve the purpose of recommending exercises of appropriate difficulty based on the user's actual level. The appropriate difficulty system can effectively improve learning effects.

[0076] In one embodiment, the process of adjusting the difficulty of the recommended exercises further includes:

[0077] When the user completes a certain number of exercises of the same difficulty level in succession, the difficulty adjustment slope coefficient will be modified based on the user's performance;

[0078]

[0079] The revised difficulty adjustment slope coefficient is obtained by analyzing and calculating formula (7) ;

[0080] in, The accuracy rate of a certain number of completed exercises is the proportion of correct answers when a user completes a certain number of exercises of the same difficulty level in a row. It can be calculated by recording the user's answer results. To preset the basic accuracy rate, a reasonable accuracy rate can be set according to the teaching objectives and the learning situation of general students. After the user completes a certain number of exercises of the same difficulty, the difficulty adjustment slope coefficient can be further modified according to the accuracy rate of the questions. If the accuracy rate is higher than the preset basic accuracy rate, it means that the user has a good grasp of the exercises of this difficulty, and the difficulty adjustment range can be appropriately increased; otherwise, the difficulty adjustment range can be reduced to make the difficulty adjustment of subsequent exercises more flexible and accurate. is the correction amplitude coefficient, which is the coefficient of the correction amplitude of the control difficulty adjustment slope coefficient and can be obtained through experimental data analysis. is the original difficulty adjustment slope coefficient, , Adjust the tilt factor for the modified difficulty.

[0081] Through the above technical solution, this embodiment provides a method for further adjusting the difficulty of recommended exercises. The method flexibly corrects the difficulty adjustment slope coefficient based on the user's performance in continuously completing exercises of the same difficulty, which can more flexibly and accurately adjust the difficulty of exercises, better adapt to changes in user learning ability, and further effectively improve learning effects.

[0082] In one embodiment, the process of adjusting the order of knowledge points and exercises for subsequent learning includes:

[0083] Based on the topological structure of the knowledge graph, the initial learning path is generated according to the dependencies between knowledge points. The knowledge graph shows the logical relationships and dependencies between knowledge points. Generating the initial learning path based on these relationships can ensure that users learn in a scientific order and avoid knowledge gaps.

[0084] The user's current learning pressure load index and the preset learning stress load index threshold Make a comparison;

[0085] like ,but

[0086]

[0087]

[0088]

[0089]

[0090] The adjusted learning path is obtained by analyzing and calculating formulas (8)-(11) ;

[0091] in, Scoring the coherence of knowledge points in the learning path is used to measure the rationality of the learning path and ensure the systematic learning of knowledge. Scoring learning novelty is used to increase the novelty of learning paths and stimulate users' learning interest. It is a dynamic adjustment factor used to make the adjustment of learning path more flexible and adapt to the learning status of different users. is the mean of the mastery evaluation coefficients of each knowledge point in the learning path, which is obtained by averaging the mastery evaluation coefficients of each knowledge point in the learning path. The critical value of the learning pressure load index is preset. It is obtained by statistically analyzing the learning effect of students under different learning pressures. It is used to judge whether the user's learning pressure is too high, and thus decide how to adjust the learning path. 、 is the third weight coefficient. When calculating the adjusted learning path, the proportion of the knowledge point coherence score and the learning novelty score can be obtained through a large amount of experimental data settings. It is used to balance the role of coherence and novelty in learning path adjustment. is the mastery evaluation coefficient corresponding to the i-th knowledge point in the learning path, is the initial benchmark difficulty corresponding to the i-th knowledge point, is the initial benchmark difficulty corresponding to the i-1th knowledge point, This is the percentage of knowledge points in the learning path that the user has not learned. It is calculated by comparing the knowledge points in the learning path with the knowledge points that the user has learned. It is used to control the amount of new knowledge learned when adjusting the learning path. The degree of repetition between the learning path and the user's historical learning path is calculated by comparing the learning path with the user's historical learning path. It is used to avoid excessive repetition of learning paths and improve learning efficiency. It is the basic correction factor, which can be obtained through experimental data setting and used to optimize the adjustment results of the learning path. In order to maximize the function output value, a learning path is selected. When the user's learning pressure is within an appropriate range, the adjusted learning path is calculated by comprehensively considering factors such as knowledge point coherence, learning novelty, knowledge point mastery evaluation coefficient, and difficulty through formulas (8)-(11). This ensures that the learning path can ensure both knowledge coherence and novelty, while also taking into account the user's knowledge mastery and difficulty adaptation.

[0092] like , then only knowledge points, and shorten the initial learning path length to half of its original length, This is an evaluation coefficient for the basic mastery of preset knowledge points, which is obtained based on experience. When the user's learning pressure is too great, the number of knowledge points in the learning path is reduced, and only the knowledge points that are well mastered are retained, thereby reducing the user's learning burden and avoiding learning fatigue due to excessive pressure.

[0093] Through the above technical solution, this embodiment discloses a method for adjusting the order of subsequent learning knowledge points and exercises. The method generates an initial learning path based on the knowledge graph, adjusts it in different ways according to the comparison results of the learning pressure load index and the critical value, and reasonably arranges the learning order according to the user's learning pressure and knowledge mastery, thereby achieving the effect of improving learning efficiency and reducing the learning burden.

[0094] The process of evaluating the user's mastery of each knowledge point also includes:

[0095] Analyze the mastery evaluation coefficient of knowledge points that users have not learned;

[0096] Extract target knowledge points from the knowledge graph All pre-dependent knowledge points

[0097]

[0098] The mastery evaluation coefficient of the knowledge points that have not been learned is obtained by analyzing and calculating the formula (12) ;

[0099] in, Pre-dependent knowledge points The mastery evaluation coefficient of the user's prior dependent knowledge points for the target knowledge point can be calculated using the previous formulas (1)-(3), which is used to infer the mastery possibility of the unlearned knowledge points based on the mastery of the prior knowledge points. Pre-dependent knowledge points Evaluation coefficient for mastering knowledge points that have not been learned The influence weight can be determined by analyzing the dependency strength between knowledge points in the knowledge graph, and is used to reflect the importance of different prerequisite knowledge points when calculating the mastery evaluation coefficient of unlearned knowledge points;

[0100] For knowledge points that users have not learned, their mastery evaluation coefficient is evaluated by considering their mastery of the preceding dependent knowledge points. This is because knowledge is coherent and dependent. A good mastery of the preceding knowledge points may be more beneficial for the understanding and mastery of subsequent unlearned knowledge points.

[0101] Through the above technical solution, this embodiment discloses a method for evaluating the user's mastery of unlearned knowledge points. The method extracts pre-dependent knowledge points from the knowledge graph and uses formula (12) to analyze and calculate the mastery evaluation coefficient of the unlearned knowledge points, thereby more comprehensively evaluating the user's knowledge mastery and providing a more accurate basis for recommending unlearned knowledge points.

[0102] The above is a detailed description of an embodiment of the present invention. However, the content described is only a preferred embodiment of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A learning resource intelligent recommendation system based on data analysis, characterized in that: The system comprises: Behavior collection module, used to collect user video learning data, exercise interaction data and learning rhythm data in real time; A real-time analysis module includes a knowledge point mastery assessment unit and a learning load assessment unit. The knowledge point mastery assessment unit is used to assess the user's mastery of each knowledge point, and the learning load assessment unit is used to assess the user's current learning pressure. The resource recommendation module adjusts the difficulty of recommended exercises based on the data obtained from the knowledge point mastery assessment unit analysis; and adjusts the order of knowledge points and exercises for subsequent learning based on the data obtained from the knowledge point mastery assessment unit and the learning load assessment unit analysis; The process of evaluating the user's current learning pressure includes: ； The user's current learning pressure load index is obtained by analyzing and calculating formula (4) ; in, The actual time it takes for the user to complete the previous exercise. The time it takes to answer the reference question corresponding to the previous exercise is: The number of times the user modified their answers while answering a certain number of exercises. The number of exercises that the user has skipped while answering a certain number of exercises. is the impact factor for skipping exercises, To preset the number of references, The number of times the user switches knowledge points during the learning process without following the system recommended path. The total number of times the user switches knowledge points during the learning process. 、 、 is the second weight coefficient; The process of adjusting the order of knowledge points and exercises for subsequent learning includes: Based on the topological structure of the knowledge graph, the initial learning path is generated according to the dependency relationship of knowledge points; The user's current learning pressure load index and the preset learning stress load index threshold Make a comparison; like ,but ； ； ； ； The adjusted learning path is obtained by analyzing and calculating formulas (8)-(11) ; in, Score the coherence of the knowledge points in the learning path, Scoring for learning novelty, is a dynamic adjustment factor, For each knowledge point, the mastery evaluation coefficient is: is the mean of the mastery evaluation coefficients of each knowledge point in the learning path, is the preset critical value of learning pressure load index, 、 is the third weight coefficient, is the mastery evaluation coefficient corresponding to the i-th knowledge point in the learning path, is the initial benchmark difficulty corresponding to the i-th knowledge point, is the initial benchmark difficulty corresponding to the i-1th knowledge point, is the proportion of knowledge points in the learning path that the user has not learned. is the degree of repetition between the learning path and the user’s historical learning path, is the basic correction factor, The learning path to maximize the function output value; like , then only knowledge points, and shorten the initial learning path length to half of its original length, Master the evaluation coefficient based on the preset knowledge points.

2. The intelligent learning resource recommendation system based on data analysis according to claim 1, characterized in that: The process of evaluating the user's mastery of each knowledge point includes: ； ； ； The mastery evaluation coefficient of each knowledge point is obtained by analyzing and calculating formulas (1)-(3) ; in, 、 is the first weight coefficient, The number of times the knowledge point is played back during the learning process of the learning video. is the playback attenuation coefficient, is the influencing factor of the learning time of the corresponding learning video of the knowledge point, is the correct answer rate of the exercises corresponding to the knowledge points, is the number of revisions during the answering process of the corresponding exercises of the knowledge point, is the modification number impact factor, The actual learning time of the learning video corresponding to the knowledge point, The reference learning time for the learning video corresponding to the knowledge point, The number of pauses during the learning process of the learning video corresponding to the knowledge point, is the suspension impact factor.

3. The intelligent learning resource recommendation system based on data analysis according to claim 2, characterized in that: The process of adjusting the difficulty of the recommended exercises includes: ； ； The recommendation difficulty is obtained by analyzing and calculating the formula (5)-(6) ; in, is the initial benchmark difficulty corresponding to the knowledge point, is the difficulty adjustment coefficient, 、 、 Adjust the slope coefficient for difficulty, To preset the first knowledge point mastery evaluation coefficient, To preset the second knowledge point, master the evaluation coefficient.

4. The intelligent learning resource recommendation system based on data analysis according to claim 3, characterized in that: The process of adjusting the difficulty of the recommended exercises further includes: When the user completes a certain number of exercises of the same difficulty level in succession, the difficulty adjustment slope coefficient will be modified based on the user's performance; ； The revised difficulty adjustment slope coefficient is obtained by analyzing and calculating formula (7) ; in, The correct rate of completing a certain number of exercises, To preset the basic accuracy, is the correction amplitude coefficient.

5. The intelligent learning resource recommendation system based on data analysis according to claim 4 is characterized in that: The process of evaluating the user's mastery of each knowledge point also includes: Analyze the mastery evaluation coefficient of knowledge points that users have not learned; Extract target knowledge points from the knowledge graph All pre-dependent knowledge points ; ； The mastery evaluation coefficient of the knowledge points that have not been learned is obtained by analyzing and calculating the formula (12) ; in, Pre-dependent knowledge points The mastery evaluation coefficient, Pre-dependent knowledge points Evaluation coefficient for mastering knowledge points that have not been learned The influence weight of .

Images