Learning material recommendation methods, devices, and electronic equipment based on smart education platforms

Abstract

This disclosure presents a learning material recommendation method, device, and electronic device based on a smart education platform, relating to the field of educational technology. The method includes: acquiring a student's learning trajectory on a learning map; determining multiple knowledge point combinations and weak knowledge point combinations based on the trajectory from clusters of basic and key knowledge points; identifying target learning materials from a learning resource database based on the weak knowledge point combinations; designating these as target learning content; and generating recommendation information. Compared to existing technologies, this disclosure relies on a learning map to acquire a real trajectory, accurately extracting multiple knowledge point combinations, avoiding the limitations of traditional single-knowledge-point recommendations; it filters weak knowledge point combinations through error rate analysis, matches suitable materials, and generates recommendation information, solving the problems of scattered learning content and monotonous question types, helping students focus on core weaknesses, and significantly improving learning efficiency.

Description

Technical Field

[0001] This disclosure relates to the field of educational technology, and in particular to a method, apparatus, storage medium, and electronic device for recommending learning materials based on a smart education platform. Background Technology

[0002] Learning materials can be the sum of all information carriers and resources that students use to acquire knowledge, understand concepts, practice skills, complete assignments, and prepare for exams during the learning process.

[0003] Currently, the main approach is to identify the knowledge points tested based on students' history exam questions, recommend learning materials for each knowledge point to students, and then determine the learning content for students based on the learning materials to help them study the knowledge points tested.

[0004] However, this method can only recommend suitable learning materials for each knowledge point when recommending learning materials to students. As a result, the types of questions for each knowledge point that students learn from the learning materials are relatively limited. Furthermore, generating a large number of questions based on each knowledge point will also lead to students learning a relatively scattered knowledge point during the learning process, making it impossible for them to focus on the knowledge points, thus resulting in low learning efficiency. Summary of the Invention

[0005] In view of this, this disclosure provides a method, device, storage medium, and electronic device for recommending learning materials based on a smart education platform. The main purpose is to improve the existing technology that only determines the knowledge points to be tested based on students' past exam questions and then recommends learning materials for individual knowledge points. This results in students being exposed to a single type of question for each knowledge point, and because a large number of questions are generated based on a single knowledge point, students' learning process is fragmented and they cannot focus on the core related content, ultimately leading to low learning efficiency.

[0006] Firstly, this disclosure provides a learning material recommendation method based on a smart education platform, including:

[0007] The learning trajectory of students in the learning map is obtained, which is obtained by mapping basic knowledge point clusters and key knowledge point clusters in multiple learning grids arranged according to a predetermined relationship.

[0008] Based on the learning trajectory, multiple combinations of knowledge points learned by the student are determined from the basic knowledge point cluster and the key knowledge point cluster, and the student's weak knowledge point combinations are determined from the multiple knowledge point combinations.

[0009] Based on the aforementioned combination of weak knowledge points, target learning materials that meet the relevant conditions are determined from the learning resource database;

[0010] The target learning materials are identified as the student's target learning content, and recommendation information for the target learning content is generated. This recommendation information is used to recommend the target learning content to the student.

[0011] Optionally, the step of determining multiple combinations of knowledge points learned by the student from the basic knowledge point cluster and the key knowledge point cluster based on the learning trajectory, and determining the student's weak knowledge point combinations from the multiple knowledge point combinations, includes:

[0012] Based on the learning trajectory, the basic knowledge point cluster, and the key knowledge point cluster, a target knowledge point association network corresponding to the student is generated.

[0013] The target knowledge point combination is determined from the basic knowledge point cluster and the key knowledge point cluster based on the target knowledge point association network;

[0014] Determine the student's error rate for each of the multiple knowledge point combinations, and identify the knowledge point combinations with an error rate greater than an error rate threshold as the weak knowledge point combinations.

[0015] Optionally, the step of determining target learning materials that meet relevant conditions from the learning resource database based on the combination of weak knowledge points includes:

[0016] Determine the set of knowledge points covered by each learning resource in the learning resource database;

[0017] Based on the knowledge point coverage set, evaluate the relevance between the weak knowledge point combination and each learning material, and obtain the relevance data corresponding to each learning material;

[0018] The target learning materials are determined from the learning materials database based on the correlation data.

[0019] Optionally, determining the target learning material from the learning resource database based on the relevance data includes:

[0020] Based on the target knowledge point association network, the knowledge point association strength between the set of covered knowledge points of each learning material and the set of weak knowledge points is determined;

[0021] The correlation data is corrected based on the correlation strength of the knowledge points to obtain the target correlation data corresponding to each learning material;

[0022] Learning materials in the learning resource library whose target relevance data is greater than the relevance data threshold are identified as the target learning materials.

[0023] Optionally, generating the target knowledge point association network corresponding to the student based on the learning trajectory, the basic knowledge point cluster, and the key knowledge point cluster includes:

[0024] The target question bank is formed by extracting basic knowledge point clusters and key knowledge point clusters from student textbooks and generating questions of different difficulty levels and question types based on these basic knowledge point clusters and key knowledge point clusters.

[0025] Based on the relationship and strength of association between the knowledge points corresponding to each question in the target question bank, a knowledge point association network corresponding to the target question bank is generated.

[0026] Based on the learning trajectory, the target association strength between the target knowledge points learned by the student based on the learning map is generated.

[0027] Optionally, determining the target learning materials as the student's target learning content and generating recommendation information for the target learning content includes:

[0028] The target learning materials are defined as the target learning content for the students.

[0029] Based on the target learning content and the corresponding weak knowledge points in the target learning materials, recommendation information for the target learning content is generated.

[0030] Secondly, this disclosure provides a learning material recommendation device based on a smart education platform, comprising:

[0031] The acquisition module is configured to acquire the student's learning trajectory in the learning map, which is obtained by mapping basic knowledge point clusters and key knowledge point clusters in multiple learning grids arranged according to a predetermined relationship.

[0032] The determination module is configured to determine multiple combinations of knowledge points learned by the student from the basic knowledge point cluster and the key knowledge point cluster based on the learning trajectory, and to determine the student's weak knowledge point combinations from the multiple knowledge point combinations.

[0033] The determination module is also configured to determine target learning materials that meet relevant conditions from the learning resource library based on the combination of weak knowledge points;

[0034] The generation module is configured to determine the target learning materials as the target learning content for the student and generate recommendation information for the target learning content, the recommendation information being used to recommend the target learning content to the student.

[0035] Thirdly, this disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the learning assistance method described in the first aspect.

[0036] Fourthly, this disclosure provides an electronic device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the computer program to implement the learning assistance method described in the first aspect.

[0037] Using the above technical solution, this disclosure provides a learning material recommendation method, device, storage medium, and electronic device based on a smart education platform. This disclosure acquires a student's learning trajectory in a learning map; then, using the learning trajectory as the core, it extracts multiple knowledge point combinations encountered by the student from basic and key knowledge point clusters, and filters out weak knowledge point combinations with error rates exceeding a threshold; subsequently, it evaluates the relevance of each material's covered knowledge point set in the learning material database to the weak knowledge point combinations, and, if necessary, corrects the relevance data based on the target knowledge point association network to determine the target learning materials that meet the conditions; finally, it sets the target learning materials as the student's target learning content and generates recommendation information containing suitable matching information for weak knowledge point combinations, thus completing the accurate recommendation of learning materials. Compared with existing technologies, this disclosure obtains students' learning trajectories based on learning maps, accurately extracting multiple knowledge point combinations that students actually encounter, avoiding the limitations of relying solely on single knowledge point recommendations. By filtering weak knowledge point combinations through error rates, and matching target learning materials and generating recommendation information based on these weak combinations, it ensures that learning materials are highly compatible with students' weaknesses and clarifies learning directions through recommendation information. This effectively solves the problems of scattered learning content and monotonous question types in existing technologies, helping students focus on core weak content and significantly improving learning efficiency. Attached Figure Description

[0038] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0039] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0040] Figure 1 A flowchart illustrating a learning material recommendation method based on a smart education platform provided in an embodiment of this disclosure is shown.

[0041] Figure 2 A schematic diagram of a learning map example is shown;

[0042] Figure 3 A flowchart illustrating a learning material recommendation method based on a smart education platform provided in an embodiment of this disclosure is shown.

[0043] Figure 4 A schematic diagram of the structure of a learning material recommendation device based on a smart education platform provided in an embodiment of this disclosure is shown.

[0044] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure is shown. Detailed Implementation

[0045] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0046] To address the shortcomings of existing technologies that rely solely on historical exam questions to determine tested knowledge points and then recommend learning materials for individual points—resulting in students encountering only a limited range of question types and generating numerous questions based on single knowledge points—this embodiment provides a learning material recommendation method based on a smart education platform. Figure 1 As shown, the method includes:

[0047] Step 101: Obtain the student's learning trajectory on the learning map.

[0048] The learning map is obtained by mapping basic knowledge point clusters and key knowledge point clusters across multiple learning grids arranged according to a predetermined relationship.

[0049] In this embodiment of the disclosure, the learning map can be a learning map formed by mapping basic knowledge point clusters and key knowledge point clusters onto multiple learning grids in the order of textbook chapters, such as... Figure 2 As shown, basic knowledge point clusters can be identified by color marking the first information identifier, and key knowledge point clusters can be identified by color marking the second information identifier. Adjacent second information identifiers at a preset distance are selected to form key monitoring areas. Each learning grid corresponds to all the questions of a knowledge point cluster. The grid corresponding to the key knowledge point cluster is the key learning grid, which can be used to clearly present the learning order of knowledge points and the core monitoring scope.

[0050] For example, if we extract the basic knowledge point cluster Kbase={algebraic operations (a1), geometric proof (b1)} and the key knowledge point cluster Kimp={functions (c2), probability (c3)} from textbook A, we can map them into four learning grids according to the textbook chapters “a1→b1→c2→c3”, thus forming a learning map.

[0051] In this embodiment of the disclosure, the learning trajectory can be a dynamic record of the sequence of questions a student answers on a learning map, reflecting the actual knowledge points encountered by the student and their learning progress. For example, the learning trajectory extracted from the learning map can be represented as T=[C1,C2,…,Ci], where Ci is the combination of knowledge points tested in the i-th question (Ci belongs to Kbase∪Kimp).

[0052] For example, if student A's problem-solving order in the learning map is: problem a1 in learning grid 1 → problem a1+b1+c2 in learning grid 3 → problem b1 in learning grid 2 → problem b1+c3 in learning grid 4, then student A's learning trajectory can be extracted as T=[{a1},{a1,b1,c2},{b1},{b1,c3}].

[0053] The teaching material (which may be the teaching material A in this embodiment of the disclosure) refers to electronic teaching material content, including but not limited to structured text data such as textbooks, teaching outlines, lesson plans, and course handouts. The teaching material covers a complete description of subject knowledge points, such as trigonometric functions, calculus, chapter divisions, and hierarchical relationships of knowledge points in mathematics.

[0054] The history question bank (which can be the target question bank in this embodiment of the disclosure) refers to the collection of test questions accumulated by the online education platform in the past, including practice questions and exam questions for each subject, grade, and knowledge point, and records information such as the knowledge points, difficulty, and students' answers. For example, a middle school mathematics history question bank contains monthly, midterm, and final exam questions from the past five years.

[0055] Textbook texts can be obtained through electronic textbook interfaces provided directly by educational institutions, publishers, or third-party platforms. Secondly, if the textbook is in print, it can be scanned and converted into an editable text format. When obtaining the textbook text, the content also needs to be categorized and labeled. The titles, subheadings, knowledge points, formulas, charts, etc., of each chapter should be extracted and categorized into specific knowledge points or modules. Furthermore, the textbook content can be structured according to the teaching plan and syllabus to ensure consistency between the textbook content and the course objectives and syllabus.

[0056] Historical question banks can be obtained through data interfaces provided by educational institutions or question bank service providers. These question banks typically include various types of questions (multiple choice, fill-in-the-blank, short answer, etc.) and include answers, explanations, and annotations of the relevant knowledge points for each question. The acquired historical question bank should contain a large number of past exam questions, mock exams, and questions targeting different levels of difficulty and knowledge points. The question bank should be closely related to the textbook content, covering all important knowledge points and chapters in the textbook. When acquiring historical question banks, it is also necessary to tag the knowledge points involved in the questions, indicating the specific knowledge points involved in each question, and classify them according to their difficulty level, frequency of examination, and other attributes.

[0057] We extensively collect various textbook texts and historical question banks to build a rich learning resource library, meeting the personalized learning needs of different students and broadening their learning horizons.

[0058] Step 102: Based on the learning trajectory, determine multiple combinations of knowledge points learned by the student from the basic knowledge point cluster and the key knowledge point cluster, and determine the student's weak knowledge point combination from the multiple knowledge point combinations.

[0059] In this embodiment of the disclosure, the knowledge point combination can be a set of associations of basic knowledge points and key knowledge points that students encounter during their actual learning process, directly extracted from the learning trajectory. For example, the knowledge point combination Ci tested by the test question qi belongs to Kbase∪Kimp, where Kbase represents the basic knowledge point cluster and Kimp represents the key knowledge point cluster.

[0060] In this embodiment of the disclosure, the weak knowledge point combination can be selected from multiple knowledge point combinations, and the combination in which the student's answer error rate exceeds a preset threshold can be used to identify the core content that the student needs to focus on reviewing.

[0061] For example, based on student A's learning trajectory T=[{a1},{a1,b1,c2},{b1},{b1,c3}]; if student A's answering situation is as follows: 1 out of 10 questions for knowledge point {a1} is answered incorrectly, 3 out of 8 questions for knowledge point combination {a1,b1,c2} are answered incorrectly, 1 out of 9 questions for knowledge point {b1} is answered incorrectly, and 3 out of 7 questions for knowledge point combination {b1,c3} are answered incorrectly, then the error rate for knowledge point {a1} is 10%, the error rate for knowledge point combination {a1,b1,c2} is 37.5%, the error rate for knowledge point combination {b1} is 11.1%, and the error rate for knowledge point combination {b1,c3} is 42.9%.

[0062] Furthermore, if the error rate threshold is set to 30%, then {a1,b1,c2} and {b1,c3} with an error rate > 30% can be identified as weak knowledge point combinations.

[0063] In this embodiment of the disclosure, a textbook knowledge point (i.e., a target knowledge point in this embodiment) refers to an independent teaching content unit in the textbook text, which is the smallest logical unit of subject knowledge, such as the Pythagorean theorem and the properties of quadratic function graphs in mathematics, metaphor in Chinese language, and the usage of function words in classical Chinese. Textbook knowledge points have clear semantic boundaries, which can be defined through the textbook table of contents or knowledge point tags; textbook knowledge points include knowledge point description text, related examples, exercises, and dependencies with other knowledge points.

[0064] In this embodiment of the disclosure, the level of difficulty refers to a quantitative rating of the cognitive ability required to master a knowledge point, which is divided into multiple levels (such as beginner, intermediate, and advanced). The rating criteria include the abstractness of the knowledge point, logical complexity, and the amount of related knowledge.

[0065] In this embodiment of the disclosure, the set of textbook knowledge points (i.e., the set of target knowledge points) is a subset composed of textbook knowledge points of the same difficulty level. For example, the elementary set contains all knowledge points of difficulty level 1, and the intermediate set contains knowledge points of difficulty level 2. The basic knowledge point cluster is a complete cluster composed of multiple sets of textbook knowledge points of different difficulty levels, used to cover the full difficulty spectrum of the subject knowledge system.

[0066] In this embodiment, natural language processing (NLP) technology is used to preprocess the textbook text. This process includes removing irrelevant content and segmenting and tagging each chapter, paragraph, and sentence in the textbook. Through deep semantic analysis of the text, all relevant textbook knowledge points are extracted. A machine learning model, combined with the contextual information of the knowledge points in the textbook, is used to classify each extracted knowledge point by difficulty level and assign a difficulty label to each knowledge point. Based on the assessed difficulty level, the knowledge points are divided into multiple categories, such as easy, medium, and difficult. Each category contains a set of knowledge points representing the textbook content at that difficulty level. Knowledge points of the same category are grouped by chapter or module to form basic knowledge point clusters. The knowledge points in each basic knowledge point cluster can be progressively arranged according to the course progress, topic relevance, and difficulty level, ensuring that students can start with simple knowledge points and gradually transition to more complex and profound content.

[0067] In this embodiment, grouping the target knowledge points according to their difficulty level helps students start from the basics and gradually accumulate knowledge, preventing confusion or frustration from encountering overly difficult concepts too early. This gradual approach helps students better understand and digest the textbook content, thereby improving learning outcomes.

[0068] Step 103: Based on the combination of weak knowledge points, determine the target learning materials that meet the relevant conditions from the learning material database.

[0069] In this embodiment, the relevant condition can be that the correlation between the set of knowledge points covered by the data and the combination of weak knowledge points is greater than or equal to a preset correlation threshold. The correlation can be calculated as Rel(R,k) = basic correlation value + the sum of correlation strength. By calculating the correlation between the data and the weak combination according to the correlation formula, target learning data that meets the correlation threshold can be selected.

[0070] For example, if the learning resource database contains 5 materials: R1 (SR={a1}, Algebra Basics), R2 (SR={b1}, Advanced Geometry), R3 (SR={c2}, Functions), R4 (SR={c3}, Probability), and R5 (SR={a1,c2}, Integrated Algebra and Functions), and Student A's weak knowledge points are {a1,b1,c2} and {b1,c3}, if the relevance threshold is set to 1.0, the weak points can be addressed... The correlation strength of the combination {a1, b1, c2} was calculated. The correlation strength of data R3 (SR including c2) Rel=1.0 (baseline value), and the correlation strength of data R5 (SR including a1, c2) Rel=1.0+wa1, c2=2.25, both ≥1.0. The correlation strength of the weak combination {b1, c3} was calculated: the correlation strength of R4 (SR including c3) Rel=1.0, ≥1.0. Therefore, R3, R5, and R4 can be determined as target learning materials that meet the conditions.

[0071] Step 104: Determine the target learning materials as the students' target learning content and generate recommendation information for the target learning content. The recommendation information is used to recommend the target learning content to the students.

[0072] In this embodiment, firstly, based on the student's weak knowledge point combination, target learning materials directly related to the weak knowledge point combination are extracted from the candidate materials, and then recommendation information is generated. The recommendation information may include the name of the target learning material and the appropriate weak knowledge point combination.

[0073] For example, if the target learning materials are R3, R5, and R4, and student A's weak knowledge points are {a1, b1, c2} and {b1, c3}, then the following can be extracted as target learning content: solutions to geometric function problems related to b1 from the function topic in R3 (fitting {a1, b1, c2}), comprehensive examples of algebraic operations and functions from R5 (fitting {a1, b1, c2}), and solutions to geometric probability problems related to b1 from the probability topic in R4 (fitting {b1, c3}).

[0074] For example, the recommended information could be: "Hello, based on your learning trajectory, you have a high error rate in questions involving combinations of {a1,b1,c2} and {b1,c3}. We recommend: 1. *Functions* (R3, correlation with c2 1.0), focusing on solving geometric function problems; 2. *Algebra and Functions Combined* (R5, correlation with c2 2.25), mastering the combined application of algebra and functions; 3. *Probability* (R4, correlation with c3 1.0), learning how to solve geometric probability problems. These materials can help you consolidate your foundation and key knowledge points simultaneously, improving your review efficiency."

[0075] Compared with existing technologies, the embodiments of this disclosure obtain students' real trajectories based on learning maps, which can accurately extract multiple knowledge point combinations that students actually come into contact with, avoiding the limitations of traditional methods that rely solely on single knowledge point recommendations. By filtering weak knowledge point combinations through error rates, and matching target learning materials and generating recommendation information based on weak combinations, it ensures that learning materials are highly adapted to students' weaknesses, and clarifies the learning direction through recommendation information. This effectively solves the problems of scattered learning content and single question types in existing technologies, helps students focus on core weak content, and significantly improves learning efficiency.

[0076] When performing the task of "identifying multiple combinations of knowledge points learned by students from basic knowledge point clusters and key knowledge point clusters based on learning trajectories, and identifying combinations of students' weak knowledge points from these multiple knowledge point combinations," the following methods can be used, but are not limited to them: Figure 3 As shown, the method includes:

[0077] Step 201: Based on the learning trajectory, basic knowledge point clusters, and key knowledge point clusters, generate a target knowledge point association network corresponding to each student.

[0078] In this embodiment, the target knowledge point association network can be created by combining student learning trajectories with basic knowledge point clusters and key knowledge point clusters. The initial association strength in the initial knowledge point association network generated based on the target question bank is updated. After the association strength is updated, the initial knowledge point association network is updated to the target knowledge point association network. In the target knowledge point association network, the node set V = Kbase∪Kimp, where Kbase represents basic knowledge points, Kimp represents key knowledge points, and the edge weights represent the target association strength.

[0079] For example, if a target question bank is generated based on textbook A, it can contain questions corresponding to knowledge point combinations such as {a1, b1}, {a1, b1, c2}, and {b1, c3}. The initial strength between knowledge points in each knowledge point combination can be calculated separately. For example, the initial strength between knowledge points in the knowledge point combination {a1, b1} can be w. a1,c2≈0.99; The initial strength between knowledge points in the knowledge point combination {b1,c3} can be w. b1,c3 ≈0.75.

[0080] Furthermore, if student A's learning trajectory shifts from knowledge point combination {a1, b1} to knowledge point combination {a1, b1, c2} (with the addition of c2), the correlation strength w of knowledge point combination {a1, b1} can be determined based on the shift in the learning trajectory. a1,c2 Increasing by 0.05 updates the target association strength for student A to 1.04. If student A's learning trajectory shifts from knowledge point combination {b1} to knowledge point combination {b1,c3} (adding c3), the association strength w of knowledge point combination {b1,c3} can be determined based on the shift in the learning trajectory. b1,c3 If we increase it by 0.05, then the target association strength of student A can be updated to 0.8.

[0081] Furthermore, the initial strengths of the two target association strengths are replaced to form a target knowledge point association network, with nodes V={a1,b1,c2,c3} and edge weights equal to the updated target association strengths.

[0082] Step 202: Determine the combination of target knowledge points from the basic knowledge point cluster and the key knowledge point cluster based on the target knowledge point association network;

[0083] In this embodiment of the disclosure, the target knowledge point combination can be an important combination judged by the association strength, that is, from the target knowledge point association network, the knowledge point combination with an association strength greater than or equal to the screening threshold is selected.

[0084] For example, in the target knowledge point association network of student A, the strength of each combination is: w a1,c2 =1.04、w b1,c3 =0.8、w a1,b1 =0.9; set T1=0.85; filter out combinations with a strength greater than or equal to 0.85: knowledge point combination {a1,c2} has an association strength of 1.04, knowledge point combination {a1,b1} has an association strength of 0.9, and combined with the knowledge point combination {b1,c3} that has appeared in the learning trajectory (association strength 0.8, close to the threshold), the target knowledge point combinations are finally determined to be {a1,c2}, {a1,b1}, and {b1,c3}.

[0085] Optionally, the frequency of examination refers to how often a particular knowledge point appears in a historical question bank, i.e., how many times this knowledge point has been tested in past exams. In the context of the National College Entrance Examination (NCEE), the frequency of examination indicates that this knowledge point is frequently tested and may be a fundamental or important knowledge point. Answer accuracy refers to the percentage of correct answers a student gives to questions related to a particular knowledge point; it usually refers to the success rate a student achieves when encountering questions related to that knowledge point. A higher accuracy rate indicates a better grasp of that knowledge point. Correlation refers to the relevance between a particular knowledge point and other knowledge points. This correlation reflects the degree of connection between different knowledge points; for example, whether mastering one knowledge point affects the understanding of other knowledge points, or whether they frequently appear together in the same type of questions. High correlation indicates a strong interrelationship between these knowledge points in learning and exams.

[0086] Step 203: Determine the error rate of each knowledge point combination among multiple knowledge point combinations, and identify the knowledge point combinations with an error rate greater than the error rate threshold as weak knowledge point combinations.

[0087] In this embodiment of the disclosure, the error rate of each knowledge point in the target knowledge point combination is statistically analyzed, and then combinations with an error rate greater than a threshold error rate are selected and identified as weak knowledge point combinations.

[0088] For example, if student A's answer results include 2 questions wrong out of 8 questions corresponding to knowledge point combination {a1,c2}, 1 question wrong out of 9 questions corresponding to knowledge point combination {a1,b1}, and 3 questions wrong out of 7 questions corresponding to knowledge point combination {b1,c3}; and the error rate threshold is 30%.

[0089] Furthermore, the objective error rate was calculated to identify the weak knowledge point combinations: the error rate of knowledge point combination {a1,c2} was 25%, the error rate of knowledge point combination {a1,b1} was 11.1%, and the error rate of knowledge point combination {b1,c3} was 42.9%; {b1,c3} with an error rate > 30% was identified as the weak knowledge point combination.

[0090] When performing the task of “identifying target learning materials that meet relevant conditions from the learning resource database based on the combination of weak knowledge points”, the following methods can be used, but are not limited to these: determining the set of knowledge points covered by each learning material in the learning resource database; evaluating the relevance between the combination of weak knowledge points and each learning material based on the knowledge point coverage set, and obtaining the relevance data corresponding to each learning material; and identifying target learning materials from the learning resource database based on the relevance data.

[0091] In this embodiment of the disclosure, the set of covered knowledge points can be a set of basic and important knowledge points that are clearly marked in each learning material. For example, the set of covered knowledge points can be SR, and the set of covered knowledge points of the target learning material R5 is SR={a1,c2}.

[0092] In this embodiment of the disclosure, the relevance between weak knowledge points and each learning material is generated based on the set of covered knowledge points. Target learning materials suitable for combinations of weak knowledge points are then selected based on the relevance. The formula for calculating the relevance can be Formula 1, where R represents the learning material, k represents the weak knowledge point, k' represents all knowledge points, and Rel represents the sum of the association strengths from all knowledge points k' covered by the material to the weak knowledge point k.

[0093] (Formula 1)

[0094] For example, student A's weak knowledge points are combined as {a1, b1, c2} and {b1, c3}, where weak knowledge point k = c2 or c3; in the target knowledge point association network, w a1,c2 =1.04、w b1,c3 =0.8. Calculate the correlation:

[0095] If the weak knowledge point k=c2, the relevance Rel(R3, c2)=1.0+0=1.0, and the relevance Rel(R5, a1, c2)=1.0+w a1,c2 =1.0 + 1.04 = 2.04;

[0096] If the weak knowledge point k=c3, the relevance Rel(R4,c3)=1.0+0=1.0.

[0097] When performing the task of "determining target learning materials from the learning resource library based on relevance data", the following methods can be used, but are not limited to these: determining the knowledge point association strength between the set of covered knowledge points and the set of weak knowledge points for each learning material based on the target knowledge point association network; correcting the relevance data based on the knowledge point association strength to obtain the target relevance data for each learning material; and identifying the learning materials in the learning resource library whose target relevance data is greater than the relevance data threshold as target learning materials.

[0098] In this embodiment of the disclosure, the association strength between the knowledge points and weaknesses covered by each learning material is obtained from the target knowledge point association network. Based on the knowledge point association strength, the relevance data is corrected and calculated to obtain the target relevance data corresponding to each learning material. The target relevance data is then filtered, and learning materials with target relevance data greater than a relevant data threshold are identified as target learning materials.

[0099] Optionally, student A's weakness is K. weak ={c2,c3}, for the weak knowledge point c2, the relevance... However, the correlation of c2 itself lacks a defined definition. Formula 2 can be used to correct the correlation data based on the correlation degree of knowledge points. The correlation degree of data R to knowledge point k is defined as the sum of the correlation strengths between the knowledge points covered by the data and k. However, if k is in the data coverage, it is directly set to the maximum value. Formula 2 is as follows:

[0100] (Formula 2)

[0101] It should be noted that the relevance data determined based on Formula 2 is not continuous enough. Therefore, Formula 3 can be used to determine the relevance data based on the strength of associations in the network. Even if k is not in SR, as long as the knowledge points in SR are related to k, it is acceptable. Specific calculation methods include: if k′ has no direct edge to k, then w k′k =0. Therefore, for R3 (covering {c2}) and knowledge point c2, since self-association is usually not considered (i.e., wc2c2 is undefined), special handling is required: if k∈SR, then directly set a basic association value (e.g., 1.0); otherwise, follow Formula 2 above. Formula 3 is shown below in detail:

[0102] (Formula 3)

[0103] Optionally, based on Formula 3, the correlation data between weakness c2 and learning material R3 can be determined as follows: c2∈SR, Rel(R3,c2)=1+0=1.0;

[0104] Learning material R5 covers the knowledge point combination {a1, c2}. The correlation coefficient between the weak point c2 and learning material R5 is: c2∈SR, Rel(R5, c2)=1+w a1,c2 =1+1.25=2.25.

[0105] Optionally, for the weak point c3, the learning material R4 covers the knowledge point {c3}, and the relevance data Rel(R4,c3)=1.0; other materials have a low relevance if they are not directly related.

[0106] Optionally, a correlation threshold of 1.0 can be set to filter target materials whose correlation data between the target materials and the weaknesses is greater than 1.0, and these materials can be identified as target learning materials.

[0107] Furthermore, for weakness c2, learning material R5 (relevance 2.25) is recommended, and for weakness c3, learning material R4 (relevance 1.0) is recommended.

[0108] Optionally, the system will associate it with basic knowledge points. For example, c2 is a key knowledge point, but it is strongly associated with the basic knowledge point a1 (w a1,c2 =2.25), recommended learning material R5 (containing a1 and c2), which allows students to consolidate basic knowledge point a1 and learn key knowledge point c2.

[0109] When performing the task of "generating a network of target knowledge points for students based on learning trajectories, clusters of basic knowledge points, and clusters of key knowledge points," the following methods can be used, but are not limited to: extracting clusters of basic knowledge points and clusters of key knowledge points from students' textbooks, and generating a target question bank with different difficulty levels and question types based on these clusters; generating a network of knowledge points for the target question bank based on the relationships and strengths of relationships between the knowledge points corresponding to each question in the target question bank; and generating the target relationship strength between the target knowledge points that students learn based on the learning map, based on the learning trajectory.

[0110] In this embodiment of the disclosure, the target question bank can be based on the basic knowledge point cluster Kbase and the important knowledge point cluster Kimp, generating questions of different difficulties and question types, and each question is labeled with a knowledge point combination Ci, which can be used to construct an initial knowledge point association network.

[0111] Optionally, extracting knowledge point clusters specifically includes: using text mining, natural language processing, and semantic analysis techniques, dividing the textbook into chapters / units, extracting basic knowledge points and classifying them (such as level 1 difficulty a1, level 1 difficulty b1), forming Kbase;

[0112] Furthermore, by combining indicators such as the frequency of historical question bank examinations (Gj), the accuracy of answering questions (Gq), and the number of knowledge graph branches, a pre-set evaluation model is used to select and form Kimp (such as level 2 difficulty c2, level 2 difficulty c3).

[0113] Furthermore, Kbase's Level 1 difficulty knowledge points are matched with easy-difficulty questions, and Kimp's Level 2 knowledge points are matched with medium-difficulty questions; simple question types (multiple choice / fill-in-the-blank) are designed for single knowledge point combinations (such as {a1}), and complex question types (solution / proof) are designed for compound knowledge point combinations (such as {a1,b1,c2}).

[0114] In this embodiment of the disclosure, the knowledge point association network can be represented by the structure G=(V,E), which must have two core elements:

[0115] Node set V: consists of all knowledge points of the basic knowledge point Kbase and the important knowledge point Kimp (i.e., V=Kbase∪Kimp). Each node represents an independent knowledge point (such as a1, b1, c2, c3). Node attributes include the knowledge point level (first-level / second-level) and the number of knowledge graph branches.

[0116] Edge set E: Each edge eij connects two related knowledge points ki and kj, and the weight of the edge is the initial association strength w. ij The basis for determining the relationship is that ki and kj appear simultaneously in at least one knowledge point combination Ci in the target question bank (e.g., a1 and b1 form an edge e(a1,b1) because they both appear in {a1,b1}).

[0117] For example, generating a knowledge point association network may specifically include:

[0118] Traverse all questions in the target question bank and extract Ci for each question; for any two knowledge points ki and kj, if there exists at least one knowledge point combination Ci that contains both ki and kj, then determine that there is a relationship between the two and add an edge eij to the network.

[0119] Furthermore, the knowledge point association network containing the knowledge point association strength can be obtained through Formula 3. Formula 4 is shown below in detail:

[0120] (Formula 4)

[0121] In Formula 4, w ij (t) α can represent the association strength at time t (initial value is 0); α can represent the historical weight decay coefficient (0 < α < 1), for example, 0.6, which represents the decay of historical association; β, γ, and η can represent adjustment coefficients; N ij (t) This can represent the deadline t, and the number of questions in the question bank that simultaneously contain knowledge points ki and kj; N i (t) N j (t) It can represent the deadline t, and the number of questions containing knowledge points ki or kj; R i (t) R j (t) It can represent the deadline t, which includes the number of branches of knowledge point ki or kj in the knowledge graph.

[0122] Optionally, based on the above-mentioned knowledge point relationships and relationship strengths, a knowledge point relationship network is generated with knowledge points as nodes and relationship strengths as edge weights.

[0123] In this embodiment, the target association strength is an updated value of the initial association strength wij, which is based on the student's individual learning trajectory. The learning trajectory can be dynamic data recording the sequence of questions a student answers during the learning process, T=[C1,C2,…,Ci], where Ci is the knowledge point combination for the i-th question.

[0124] Furthermore, for each set of learning trajectory transfer relationships, common knowledge points and newly added knowledge points are identified, and the target association strength can be increased by a fixed increment δ (δ is an empirical value, such as 0.05, which can be adjusted according to the difficulty of the subject).

[0125] For example, student A's learning trajectory T=[{a1},{a1,b1,c2},{b1},{b1,c3}]; increment δ=0.05; initial association strength: w a1,c2 =0.99、w b1,c3 =0.75、w a1,b1 =1.28、w b1,c2 =1.34;

[0126] Learning trajectory transfer relationship 1: From knowledge point set {a1} to knowledge point set {a1,b1,c2}, common knowledge point a1, new knowledge points b1, c2: correlation strength w a1,b1 =1.28 + 0.05 = 1.33; Correlation strength w a1,c2 =0.99 + 0.05 = 1.04;

[0127] Learning trajectory transfer relationship 2: Transfer from knowledge point set {a1,b1,c2} to knowledge point set {b1}, common knowledge point b1, no new knowledge points are added (Ccurr={b1} is a subset of Cprev={a1,b1,c2}), no intensity needs to be updated;

[0128] Learning trajectory transfer relationship 3: from knowledge point set {b1} to knowledge point set {b1,c3}, common knowledge point b1, newly added knowledge point c3: correlation strength w b1,c3 =0.75 + 0.05 = 0.8;

[0129] Strength not involving transfer: Association strength w b1,c2 =1.34 (no corresponding transfer relationship), keep the initial value;

[0130] Updated target association strength: association strength w a1,b1 =1.33, Correlation Strength w a1,c2 =1.04, Correlation Strength w b1,c2 =1.34, Correlation Strength w b1,c3 =0.8, this target association strength can be used to construct a target knowledge point association network specific to student A.

[0131] When determining the target learning materials as the student's target learning content and generating recommended information for the target learning content, the following methods can be used, but are not limited to: determining the target learning materials as the student's target learning content; and generating recommended information for the target learning content based on the combination of the target learning content and the corresponding weak knowledge points.

[0132] In this embodiment of the disclosure, the target learning content can be specific content (such as example analysis and formula derivation) extracted from the target learning materials that is directly related to the combination of weak knowledge points, excluding basic content that is unrelated to the weak points.

[0133] For example, if the target learning material is R5, its content includes knowledge point a1 (basic operations), knowledge point c2 (function definition), a comprehensive example problem of knowledge point a1+c2, and a geometry problem of knowledge point c2; if student A's weak combination is {a1, b1, c2}, and the core weakness is the application of a1+c2, excluding irrelevant content such as knowledge point a1 (basic operations) and c2 (function definition), the core content can be extracted as: a comprehensive example problem of knowledge point a1 and knowledge point c2 (covering the a1c2 relationship) and a geometry problem of knowledge point c2 (covering the c2b1 relationship), which will be the target learning content.

[0134] In this embodiment, the recommendation information may include the name of the target learning material, the target learning content, the appropriate weak point combination, and learning suggestions, which can be used to guide students to understand the recommendation logic and learn efficiently. The recommendation information generation rules may include: integrating the target learning content and weak point information to generate recommendation information with relevance criteria, while adding learning suggestions (such as learning order).

[0135] For example, if Student A's target learning content includes comprehensive examples of knowledge points a1 and c2, and geometry problems related to knowledge point c2, in learning material R5; and Student A's weak combination {a1, b1, c2} has an error rate of 37.5%, the generated recommendation information is: Hello! Based on your problem-solving data, you have a weakness in problems involving knowledge point combination {a1, b1, c2} (error rate 37.5%). We recommend core material A (learning material R5, with a correlation strength of 2.04 with knowledge point c2, and a corrected correlation strength of 3.05). It is suggested that you prioritize learning the comprehensive examples of knowledge points a1 and c2, and then learn the geometry problems related to knowledge point c2.

[0136] This embodiment of the disclosure acquires textbook text and historical question banks, extracts and combines textbook knowledge points to form basic knowledge point clusters, uses a preset model to filter out important knowledge point clusters, and then generates target question banks of different difficulty levels. These are then mapped onto a learning map to generate students' learning trajectories and recommend suitable review materials. Based on students' learning progress and knowledge mastery, it can recommend review materials suitable for their current learning stage, effectively solving the problem of lacking continuity in review materials and ensuring that review materials match students' learning status.

[0137] Compared with existing technologies, the embodiments of this disclosure obtain students' real trajectories based on learning maps, which can accurately extract multiple knowledge point combinations that students actually come into contact with. This avoids the limitations of traditional methods that rely solely on single knowledge point recommendations. By filtering weak knowledge point combinations through error rates, and matching target learning materials and generating recommendation information based on these weak combinations, it ensures that learning materials are highly compatible with students' weaknesses and clarifies the learning direction through recommendation information. At the same time, it solves the problem of inaccurate weak point positioning in traditional methods by combining target knowledge point association networks with error rate filtering. By correcting the relevance data, it eliminates the bias in the material filtering process, effectively solving the problems of scattered learning content and single question types in existing technologies. It recommends suitable learning content to students, helps them focus on core weak content, and significantly improves learning efficiency.

[0138] Furthermore, as Figure 1 and Figure 3 The specific implementation of the method shown in this embodiment provides a learning material recommendation device based on a smart education platform, such as... Figure 4 As shown, the device includes: an acquisition module 31, a determination module 32, and a generation module 33.

[0139] The acquisition module 31 is configured to acquire the student's learning trajectory in the learning map, which is obtained by mapping basic knowledge point clusters and key knowledge point clusters in multiple learning grids arranged according to a predetermined relationship.

[0140] The determination module 32 is configured to determine multiple combinations of knowledge points learned by the student from the basic knowledge point cluster and the key knowledge point cluster based on the learning trajectory, and to determine the student's weak knowledge point combination from the multiple knowledge point combinations.

[0141] Module 32 is also configured to determine target learning materials that meet relevant conditions from the learning resource database based on the combination of weak knowledge points;

[0142] The generation module 33 is configured to identify the target learning materials as the target learning content for students and generate recommendation information for the target learning content, which is used to recommend the target learning content to students.

[0143] In some embodiments of this disclosure, the generation module 33 is specifically configured to generate a target knowledge point association network corresponding to the student based on the learning trajectory, the basic knowledge point cluster, and the key knowledge point cluster.

[0144] In some embodiments of this disclosure, the determining module 32 is specifically configured to determine a combination of target knowledge points from the basic knowledge point cluster and the key knowledge point cluster based on the target knowledge point association network.

[0145] In some embodiments of this disclosure, the determining module 32 is specifically configured to determine the error rate of the student for each of the multiple knowledge point combinations, and to determine the knowledge point combinations with an error rate greater than the error rate threshold as weak knowledge point combinations.

[0146] In some embodiments of this disclosure, the determining module 32 is specifically configured to: determine the set of knowledge points covered by each learning material in the learning material library; evaluate the relevance between the combination of weak knowledge points and each learning material based on the knowledge point coverage set, and obtain the relevance data corresponding to each learning material; and determine the target learning material from the learning material library based on the relevance data.

[0147] In some embodiments of this disclosure, the determining module 32 is specifically configured to determine the knowledge point association strength between the covered knowledge point set and the weak knowledge point set of each learning material based on the target knowledge point association network; correct the relevance data based on the knowledge point association strength to obtain the target relevance data corresponding to each learning material; and determine the learning materials in the learning material library whose target relevance data is greater than the relevant data threshold as target learning materials.

[0148] In some embodiments of this disclosure, the generation module 33 is specifically configured to extract basic knowledge point clusters and key knowledge point clusters from student textbooks, and generate a target question bank consisting of questions of different difficulty levels and question types based on the basic knowledge point clusters and key knowledge point clusters; generate a knowledge point association network corresponding to the target question bank based on the association relationship and association strength between the knowledge points corresponding to each question in the target question bank; and generate the target association strength between the target knowledge points learned by the student based on the learning map based on the learning trajectory.

[0149] In some embodiments of this disclosure, the determining module 32 is specifically configured to determine the target learning materials as the student's target learning content.

[0150] In some embodiments of this disclosure, the generation module 33 is specifically configured to generate recommendation information for the target learning content based on the combination of weak knowledge points corresponding to the target learning content and the target learning materials.

[0151] It should be noted that other corresponding descriptions of the functional units involved in the learning content determination device based on the smart education platform provided in this embodiment can be found in the following references. Figure 1 and Figure 3 The corresponding description in [the document] will not be repeated here.

[0152] Compared with existing technologies, this disclosure obtains students' real trajectories based on learning maps, accurately extracting multiple knowledge point combinations that students actually encounter. This avoids the limitations of traditional methods that rely solely on single knowledge point recommendations. By filtering weak knowledge point combinations through error rates, and matching target learning materials and generating recommendation information based on these weak combinations, it ensures that learning materials are highly compatible with students' weaknesses and clarifies learning directions through recommendation information. Furthermore, by combining target knowledge point association networks with error rate filtering, it solves the problem of inaccurate weak point identification in traditional methods. By correcting the relevance data, it eliminates biases in the material selection process, effectively addressing the problems of scattered learning content and monotonous question types in existing technologies. It recommends suitable learning content to students, helping them focus on core weak areas and significantly improving learning efficiency.

[0153] Based on the above, Figure 1 and Figure 3 Accordingly, this embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. Figure 1 and Figure 3 The method shown.

[0154] Based on this understanding, the technical solution disclosed herein can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, or portable hard drive) and includes several instructions to cause a computer device (such as a personal computer, server, or network device) to execute the methods of various implementation scenarios of this disclosure.

[0155] like Figure 5 The diagram shown is a hardware structure schematic of an electronic device according to the present invention, comprising:

[0156] At least one processor 401; and,

[0157] Memory 402 is communicatively connected to at least one processor 401; wherein,

[0158] The memory 402 stores instructions that can be executed by at least one processor, which enables the at least one processor to perform the learning aid method as described above.

[0159] Figure 5 Take a processor 401 as an example.

[0160] The electronic device may also include an input device 403 and a display device 404.

[0161] The processor 401, memory 402, input device 403, and display device 404 can be connected via a bus or other means. Figure 5 Taking the example of a connection between China and Israel via a bus.

[0162] Memory 402, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the learning assistance method in the embodiments of this disclosure, for example, Figure 1 and Figure 3 The method flow is shown. The processor 401 executes various functional applications and data processing by running non-volatile software programs, instructions, and modules stored in the memory 402, thereby implementing the learning assistance method in the above embodiments.

[0163] Memory 402 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the learning-aiding method, etc. Furthermore, memory 402 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 402 may optionally include memory remotely located relative to processor 401, and these remote memories may be connected via a network to the apparatus performing the learning-aiding method. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0164] The input device 403 can receive user clicks and generate signal inputs related to user settings and function control of the learning assistance method. The display device 404 may include a display screen or other display device.

[0165] One or more modules are stored in memory 402, and when run by one or more processors 401, the learning assistance method in any of the above method embodiments is executed.

[0166] Optionally, the aforementioned physical devices may also include a user interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The user interface may include a display screen, input units such as a keyboard, etc., and optional user interfaces may also include USB interfaces, card reader interfaces, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0167] Those skilled in the art will understand that the physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0168] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the aforementioned physical device, supporting the operation of information processing programs and other software and / or programs. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software in the information processing physical device.

[0169] Through the above description of the embodiments, those skilled in the art can clearly understand that this disclosure can be implemented using software plus necessary general-purpose hardware platforms, or it can be implemented using hardware. By applying the solution of this embodiment, compared with the prior art, this embodiment obtains the student's actual trajectory based on the learning map, accurately extracting multiple knowledge point combinations that the student actually encounters. This avoids the limitations of traditional methods that rely solely on single knowledge point recommendations. It filters weak knowledge point combinations through error rate screening, and matches target learning materials and generates recommendation information based on these weak combinations. This ensures that the learning materials are highly compatible with the student's weaknesses and clarifies the learning direction through the recommendation information. Simultaneously, it solves the problem of inaccurate weak point positioning in traditional methods by combining target knowledge point association networks with error rate screening. By correcting the relevance data, it eliminates biases in the material screening process, effectively solving the problems of scattered learning content and single question types in the prior art. It recommends suitable learning content to students, helping them focus on core weak content and significantly improving learning efficiency.

[0170] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0171] The above are merely specific embodiments of this disclosure, enabling those skilled in the art to understand or implement this disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A learning material recommendation method based on a smart education platform, characterized by, The method comprises the following steps: acquiring a learning track of a student in a learning map, the learning map being obtained by mapping a basic knowledge point cluster and a key knowledge point cluster in a plurality of learning grids arranged according to a predetermined relationship; determining a plurality of knowledge point combinations learned by the student from the basic knowledge point cluster and the key knowledge point cluster based on the learning track, and determining a weak knowledge point combination of the student from the plurality of knowledge point combinations, comprising: generating a target knowledge point association network corresponding to the student based on the learning track, the basic knowledge point cluster and the key knowledge point cluster; determining the target knowledge point combination from the basic knowledge point cluster and the key knowledge point cluster based on the target knowledge point association network; determining an error rate of each knowledge point combination in the plurality of knowledge point combinations for the student, and determining a knowledge point combination with an error rate greater than an error rate threshold in the plurality of knowledge point combinations as the weak knowledge point combination; the generating of the target knowledge point association network corresponding to the student based on the learning track, the basic knowledge point cluster and the key knowledge point cluster comprises: extracting the basic knowledge point cluster and the key knowledge point cluster from a student textbook, and generating a question bank composed of questions of different difficulty levels and different question types according to the basic knowledge point cluster and the key knowledge point cluster; generating a knowledge point association network corresponding to the question bank based on the association relationship and the association strength between the knowledge points corresponding to each question in the question bank; generating a target association strength between target knowledge points learned by the student based on the learning map based on the learning track; wherein the generating of the target association strength between the target knowledge points learned by the student based on the learning map based on the learning track comprises: determining a knowledge point transfer relationship between the target knowledge point combinations based on the learning track, and determining the association strength increment corresponding to the target knowledge point combinations based on the transfer relationship; determining the target association strength according to the association strength increment and the association strength; determining target learning materials meeting relevant conditions from a learning material library according to the weak knowledge point combination, comprising: determining a covered knowledge point set of each learning material in the learning material library; evaluating the relevance between the weak knowledge point combination and each learning material based on the knowledge point covered set, to obtain relevance data corresponding to each learning material; determining the target learning materials from the learning material library according to the relevance data; determining the target learning materials as target learning content of the student, and generating recommendation information of the target learning content, the recommendation information being used to recommend the target learning content to the student.

2. The method of claim 1, wherein, the determining of the target learning materials from the learning material library according to the relevance data comprises: determining a knowledge point association strength between the covered knowledge point set of each learning material and the weak knowledge point set based on the target knowledge point association network; correcting the relevance data based on the knowledge point association strength to obtain target relevance data corresponding to each learning material; Determine the target learning material as the target learning content of the student.

3. The method of claim 1, wherein, The target learning material is determined as the target learning content of the student, and recommendation information of the target learning content is generated, including: Determine the target learning material as the target learning content of the student. Generate the recommendation information of the target learning content based on the target knowledge point combination corresponding to the target learning content and the weak knowledge point combination. 4.A learning material recommendation device based on a smart education platform, characterized by, Including: The acquisition module is configured to acquire a learning trajectory of a student in a learning map, the learning map being obtained by mapping a basic knowledge point cluster and a key knowledge point cluster in a plurality of learning grids arranged according to a predetermined relationship; The determination module is configured to determine a plurality of knowledge point combinations learned by the student from the basic knowledge point cluster and the key knowledge point cluster based on the learning trajectory, and determine a weak knowledge point combination of the student from the plurality of knowledge point combinations, including: generating a target knowledge point association network corresponding to the student based on the learning trajectory, the basic knowledge point cluster and the key knowledge point cluster; determining the target knowledge point combination from the basic knowledge point cluster and the key knowledge point cluster based on the target knowledge point association network; determining the error rate of each knowledge point combination in the plurality of knowledge point combinations, and determining the knowledge point combination with an error rate greater than an error rate threshold in the plurality of knowledge point combinations as the weak knowledge point combination; The target knowledge point association network corresponding to the student is generated based on the learning trajectory, the basic knowledge point cluster and the key knowledge point cluster, including: extracting the basic knowledge point cluster and the key knowledge point cluster from the student's textbook, and generating a question bank composed of questions of different difficulty levels and different types according to the basic knowledge point cluster and the key knowledge point cluster; generating a knowledge point association network corresponding to the question bank based on the association relationship and the association strength between the knowledge points corresponding to each question in the question bank; generating a target association strength between the target knowledge points learned by the student based on the learning map based on the learning trajectory; Wherein, the target association strength between the target knowledge points learned by the student based on the learning map is generated based on the learning trajectory, including: determining the knowledge point transfer relationship between the target knowledge point combinations based on the learning trajectory, determining the association strength increment corresponding to the target knowledge point combinations based on the transfer relationship; determining the target association strength according to the association strength increment and the association strength; The determination module is further configured to determine the target learning material meeting the relevant conditions from the learning material library according to the weak knowledge point combination, including: determining a knowledge point coverage set of each learning material in the learning material library; evaluating the relevance between the weak knowledge point combination and each learning material based on the knowledge point coverage set, obtaining the relevance data corresponding to each learning material; determining the target learning material from the learning material library according to the relevance data; The generating module is configured to determine the target learning material as the target learning content of the student, and generate recommendation information of the target learning content, the recommendation information being used to recommend the target learning content to the student.

5. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by a processor, implements the method of any one of claims 1 to 3.

6. An electronic device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, The processor implements the method of any one of claims 1 to 3 when executing the computer program.

7. A computer program product comprising a computer program, characterized in that, The computer program, when executed by a processor, implements the method of any one of claims 1 to 3. The computer program, when executed by a processor, implements the method of any one of claims 1 to 3.

Images