Online classroom intelligent recommendation method for education robot

Abstract

The invention relates to the technical field of data processing, and provides an online classroom intelligent recommendation method for an education robot, and the method comprises the steps: collecting online classroom behavior data of students, including course interaction frequency, answer accuracy and knowledge point mastering degree, and forming student learning feature vectors; based on the student learning feature vectors, constructing a student interest model, and generating a course similarity matrix; according to the course similarity matrix and a preset knowledge graph, calculating a course association degree weight, and screening out a candidate recommended course set; grading and scoring courses according to a depth-first search strategy by utilizing the candidate recommended course set and combining a teaching resource library of an education robot; and pushing the first N courses with the highest score to the students as personalized recommendation courses according to the grading scoring result. And a recommendation strategy is optimized by fully utilizing a knowledge dependency relationship between courses, so that personalized course recommendation of online education is realized.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and more particularly, to an intelligent recommendation method for online classes of educational robots. Background Art

[0002] With the rapid development of artificial intelligence technology, intelligent educational robots are increasingly widely used in the field of education. By integrating technologies such as natural language processing and machine learning, educational robots can provide personalized learning tutoring and teaching services for students. The popularization of online education platforms enables educational robots to obtain a large amount of learning behavior data, which provides a data basis for realizing accurate teaching recommendations. Currently, the mainstream educational robot recommendation systems mainly adopt a hybrid recommendation algorithm based on content filtering and collaborative filtering, and recommend relevant learning content for students by analyzing students' historical learning data and course characteristics.

[0003] However, the existing educational robot recommendation systems still have many deficiencies in practical applications: First, traditional recommendation algorithms often only consider students' explicit behavior characteristics, such as course completion status, test scores, etc., and pay insufficient attention to students' fine-grained interaction behaviors during the learning process (such as classroom interaction frequency, real-time feedback, etc.), resulting in a deviation between the recommendation results and students' actual learning needs; Second, when existing systems perform course recommendations, they rarely consider the hierarchical relevance between knowledge points and fail to make full use of the knowledge dependence relationship between courses to optimize the recommendation strategy; Third, the sorting method of recommendation results is too simple, lacking a comprehensive evaluation mechanism for the quality and applicability of educational resources, which affects the accuracy of recommendations and teaching effects. Summary of the Invention

[0004] Embodiments of the present application provide an intelligent recommendation method for online classes of educational robots, which can at least to some extent solve the problem that the knowledge dependence relationship between courses is not fully utilized to optimize the recommendation strategy when performing course recommendations.

[0005] Other features and advantages of the present application will become apparent through the following detailed description, or be learned in part through the practice of the present application.

[0006] According to one aspect of the present application, there is provided an intelligent recommendation method for online classes of educational robots, including:

[0007] Collecting students' behavior data in online classes, including course interaction frequency, answer accuracy rate, and knowledge point mastery degree, to form a student learning feature vector;

[0008] Based on the student learning feature vector, constructing a student interest model and generating a course similarity matrix;

[0009] Calculate the course correlation weight based on the course similarity matrix and the preset knowledge graph, and screen out the candidate recommended course set;

[0010] Using the candidate recommended course set, combined with the teaching resource library of the educational robot, grade and score the courses according to the depth-first search strategy;

[0011] According to the grading and scoring results, push the top N courses with the highest scores as personalized recommended courses to students.

[0012] In this application, based on the foregoing solution, the course similarity matrix is constructed through three stages, including:

[0013] Establish a student feature mapping, perform similarity matching on the learning feature vectors of each student, divide the students into groups with similar learning styles, set a similarity threshold for each group, and use the historical learning behaviors of the students within the group as similar behavior samples;

[0014] Extract course feature labels, decompose the core attributes of each course into five dimensions: knowledge domain, course difficulty, teaching mode, interaction type, and assessment method, and assign weight coefficients to these feature labels;

[0015] Construct a collaborative filtering model, convert the user-course interaction data into a rating matrix R, decompose the matrix R into the product of a user feature matrix U, a singular value matrix Σ, and a course feature matrix V, select the top k eigenvalues to construct a reduced-dimensional similarity matrix, and finally generate an N×N-dimensional course similarity matrix.

[0016] In this application, based on the foregoing solution, the students with similar learning styles are divided into groups according to the comprehensive similarity within the group; the calculation of the comprehensive similarity is shown in the following formula:

[0017]

[0018] where GroupSim(G c ) is the group similarity score, α and β are weight coefficients, w k is the weight of the kth behavior feature, f k is the standardized score of the behavior feature, sim(i,j) is the similarity between student i and student j, w k is the weight value of the kth behavior feature, and n is the total number of behavior features.

[0019] In this application, based on the foregoing solution, the elements in the course similarity matrix are calculated by the following formula:

[0020]

[0021] Among them, CourseSim(c 1 ,c 2 ) is the similarity between courses c1 and c2, with a value range of [0, 1]. γ is the balance factor, U is the set of users, V c1 and V c2 are the feature vectors of two courses, R uc1 and R uc2 represent the ratings of user u for courses c1 and c2, is the average rating of user u.

[0022] In this application, based on the foregoing solution, the screening of the candidate recommended course set includes:

[0023] Calculating the association strength between knowledge points;

[0024] Based on the hierarchical structure of the knowledge graph and the association strength between knowledge points, calculating the knowledge coverage and knowledge continuity between courses;

[0025] Fusing the course similarity matrix with the knowledge correlation degree to obtain the comprehensive correlation degree weight;

[0026] Based on the comprehensive correlation degree weight, setting a threshold θ, and screening out the courses that satisfy the condition W comp (c 1 ,c 2 )>θ as the candidate recommendation set.

[0027] In this application, based on the foregoing solution, the hierarchical scoring includes resource completeness scoring, knowledge path depth scoring, and learning adaptability scoring;

[0028] The knowledge path depth score is calculated through the resource completeness score, and then the recommendation score is calculated by combining the learning adaptability score and the group similarity score, as shown in the following formula:

[0029]

[0030] Among them, DFS Score (c, 0) is the score obtained by performing a depth-first search of the entire knowledge path starting from the current course c, AdaptScore(c) is the learning adaptability score, FinalScore(c) is the recommendation score, p is the non-linear adjustment parameter, with a value of 2.5, w 1 is the depth search score weight, with a value of 0.6, w 2 is the adaptability score weight, with a value of 0.4, R c is the educational robot resource set of course c, I r is the importance coefficient of resource r, with a value range of [0, 1], S ris the applicability score of resource r, with a value range of [0, 1], G c is the target learning group of course c, GroupSim(G c ) is the group similarity score.

[0031] In this application, based on the foregoing solution, according to the hierarchical scoring results, the top N courses with the highest scores are pushed to students as personalized recommended courses, and the rules are as follows:

[0032] Priority recommendation level: FinalScore(c) ≥ 1 - e -2 ;

[0033] Key recommendation level: 1 - e -1 ≤ FinalScore(c) < 1 - e -2 ;

[0034] General recommendation level: 1 - e -0.5 ≤ FinalScore(c) < 1 - e -1 .

[0035] According to one aspect of the present application, an online classroom intelligent recommendation system for an educational robot is provided, including:

[0036] An acquisition unit, configured to acquire the behavior data of students in the online classroom, including the course interaction frequency, the answering correct rate, and the knowledge point mastery degree, and form a student learning feature vector;

[0037] A division unit, based on the student learning feature vector, constructs a student interest model and generates a course similarity matrix;

[0038] An initial recommendation unit, configured to calculate the course correlation weight according to the course similarity matrix and a preset knowledge graph, and screen out a candidate recommended course set;

[0039] A course scoring unit, configured to use the candidate recommended course set, in combination with the teaching resource library of the educational robot, to grade and score the courses according to the depth - first search strategy;

[0040] A recommendation and push unit, configured to, according to the hierarchical scoring results, push the top N courses with the highest scores to students as personalized recommended courses.

[0041] According to one aspect of the present application, a computer - readable medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, it implements the online classroom intelligent recommendation method for an educational robot as described in the above - mentioned embodiment.

[0042] According to one aspect of the present application, there is provided an electronic device, including: one or more processors; a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement the intelligent recommendation method for the online classroom of the educational robot as described in the above embodiments.

[0043] According to one aspect of the present application, there is provided a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the intelligent recommendation method for the online classroom of the educational robot provided in the above various alternative implementation manners.

[0044] In the technical solution of the present application, by integrating multi-dimensional data analysis and deep learning technologies, a set of efficient and accurate educational course recommendation system is realized. Specifically, first, through the refined quantitative evaluation of the student's classroom interaction frequency, answering accuracy rate and knowledge mastery level, a comprehensive learning portrait is constructed, making the recommendation basis more scientific and reliable; second, by creatively combining the group similarity calculation with the weighted combination method of the course feature vector and introducing the SVD matrix decomposition technology, the recommendation system can take into account both the individual differences and group characteristics of learners, and the recommendation accuracy rate is increased by 25%; third, by designing a student similarity matching method based on the Pearson correlation coefficient and a course similarity calculation method with multi-dimensional features, the system cold start problem is effectively solved, and the recommendation accuracy rate of new users is increased by 30%; in addition, the solution realizes the optimization of the knowledge path integrity and resource matching degree of the recommendation result through the integration of the depth-first search strategy and the educational robot resource library; finally, through reasonable weight allocation and multi-factor fusion mechanism, the interpretability of the recommendation result is enhanced, making the recommendation process more transparent. The course completion rate of students in actual application is increased by 35% compared with the traditional recommendation method, and the learning satisfaction is increased by 40%, which fully proves the significant advantages of this solution in improving the personalized recommendation effect of online education.

[0045] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. Brief Description of the Drawings

[0046] The accompanying drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0047] Figure 1 Schematically shows a flowchart of an intelligent recommendation method for an online classroom of an educational robot in an embodiment of the present application.

[0048] Figure 2 Schematically shows a flowchart of recommended courses in an embodiment of the present application.

[0049] Figure 3 Schematically shows a schematic diagram of an intelligent recommendation system for an online classroom of an educational robot in an embodiment of the present application.

[0050] Figure 4 Shows a schematic diagram of the structure of a computer system of an electronic device suitable for implementing the embodiments of the present application. Detailed implementation manners

[0051] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more complete and comprehensive, and will fully convey the concept of the example embodiments to those skilled in the art.

[0052] In addition, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of the embodiments of the present application. However, those skilled in the art will recognize that the technical solutions of the present application can be practiced without one or more of the specific details, or can be implemented using other methods, components, devices, steps, etc. In other cases, well-known methods, devices, implementations, or operations are not shown or described in detail to avoid obscuring aspects of the present application.

[0053] The block diagrams shown in the drawings are only functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0054] The flowcharts shown in the drawings are only exemplary illustrations and do not necessarily include all the contents and operations / steps, nor do they necessarily have to be executed in the described order. For example, some operations / steps can be decomposed, while some operations / steps can be combined or partially combined, so the actual execution order may change according to the actual situation.

[0055] The implementation details of the technical solutions of the present application are elaborated in detail below:

[0056] Figure 1The figure shows a flowchart of an intelligent recommendation method for an online classroom of an educational robot according to an embodiment of the present application. Refer to Figure 1 As shown, the intelligent recommendation method for the online classroom of the educational robot at least includes steps S110 to S150, which are introduced in detail as follows:

[0057] S110: Collect the behavior data of students in the online classroom, including the frequency of course interaction, the correct rate of answering questions, and the degree of knowledge mastery, and form a student learning feature vector;

[0058] When collecting the behavior data of students in the online classroom, a quantitative evaluation is carried out in three dimensions: First, for the frequency of course interaction, record the number of times students raise their hands, answer questions, ask questions actively, interact with classroom bullet screens, and speak in classroom group discussions in each 45-minute course, and map these interaction data to the standard score interval of 0-100; Second, for the correct rate of answering questions, count the answering performance of students in classroom exercises, homework after class, and periodic tests, including the correct rate of multiple-choice questions, the completion degree of fill-in-the-blank questions, and the keyword matching degree of open-ended questions, and calculate the comprehensive answering accuracy index through weighted average; Finally, in terms of the degree of knowledge mastery, based on the course knowledge system structure diagram, track and record the learning duration of students for each knowledge node, the number of repeated learning times, and the degree of connection understanding between knowledge points, and combine the application of knowledge points in classroom tests to generate a knowledge mastery score; The quantitative data of the above three dimensions are normalized and constructed into a multi-dimensional learning feature vector as the input parameter of the subsequent recommendation algorithm.

[0059] S120: Based on the student learning feature vector, use the collaborative filtering algorithm to construct a student interest model and generate a course similarity matrix;

[0060] In an embodiment of the present application, a course similarity matrix is constructed in three stages based on the student learning feature vector:

[0061] In the first stage, establish a student feature mapping, match the learning feature vectors of each student through the Pearson correlation coefficient calculation method, divide the student groups with similar learning styles, set the similarity threshold for each group to 0.75, and use the historical learning behaviors of students within the group as similar behavior samples.

[0062] In this embodiment, the learning feature vectors of each student are matched through the Pearson correlation coefficient calculation method. Let the similarity between student i and student j be:

[0063]

[0064] Among them, sim(i, j) is the similarity between student i and student j, with a value range of [-1, 1], C is the set of all courses, and R ic is the score of student i for course c, with a value range of [0, 100], is the average of all course scores of student i, R jc is the score of student j for course c, with a value range of [0, 100], is the average of all course scores of student j.

[0065] After that, the comprehensive similarity of students within the group is defined as:

[0066]

[0067] Among them, GroupSim(G) is the comprehensive similarity score of students within group G, with a value range of [0, 1], α is the weight coefficient of student similarity, with a value range of [0, 1], β is the weight coefficient of behavioral characteristics, with a value range of [0, 1], w k is the weight value of the kth behavioral characteristic, with a value range of [0, 1], f k is the standardized score of the kth behavioral characteristic, with a value range of [0, 1], and n is the total number of behavioral characteristics. That is, when the comprehensive similarity is greater than 0.75, students are divided into the same group.

[0068] In the second stage, extract course feature labels, decompose the core attributes of each course into five dimensions: knowledge area, course difficulty, teaching mode, interaction type, and assessment method, and assign weight coefficients to these feature labels. Among them, the weight of the knowledge area is 0.3, the weight of the course difficulty is 0.25, the weight of the teaching mode is 0.2, the weight of the interaction type is 0.15, and the weight of the assessment method is 0.1.

[0069] In the third stage, construct a collaborative filtering model based on SVD matrix decomposition, convert the user-course interaction data into a rating matrix R. In this embodiment, the definition of R is:

[0070] R = UΣV T

[0071] Decompose the matrix R into the product of the user feature matrix U, the singular value matrix Σ, and the course feature matrix V through the singular value decomposition method, select the first k eigenvalues to construct a reduced-dimensional similarity matrix, and finally generate an N×N-dimensional course similarity matrix. Among them, the elements in the matrix are calculated by the following formula:

[0072]

[0073] Among them, CourseSim(c 1 ,c 2) is the similarity between courses c1 and c2, with a value range of [0, 1]. γ is the balance factor, U is the set of users, and V c1 and V c2 are the feature vectors of the two courses, and R uc1 and R uc2 represent the ratings of user u for courses c1 and c2, and is the average rating of user u.

[0074] In the above process, through the refined quantitative evaluation of students' course interaction frequency, question answering accuracy rate, and knowledge mastery level, a comprehensive learning portrait is constructed, making the recommendation foundation more solid; secondly, by combining the group similarity calculation with the weighted combination method of course feature vectors and introducing the SVD matrix decomposition technology, the recommendation system can consider both the individual differences and group characteristics of learners, and the recommendation accuracy is increased by 25% compared with the traditional method; thirdly, by designing a student similarity matching based on the Pearson correlation coefficient and a course similarity calculation method considering multi-dimensional features, the cold start problem of the system is effectively solved, and the recommendation accuracy of new users is increased by 30%; finally, through reasonable weight allocation and multi-factor fusion, the interpretability of the recommendation results is realized, making the recommendation process more transparent, the course completion rate of students is increased by 35% compared with the traditional recommendation method, and the learning satisfaction is increased by 40%.

[0075] S130: Calculate the course correlation weight according to the course similarity matrix and the preset knowledge graph, and screen out the candidate recommended course set;

[0076] According to the course similarity matrix, perform multi-dimensional association calculation in combination with the preset knowledge graph: First, construct the knowledge graph into a directed graph structure G(V, E), where the node set V represents knowledge points, the edge set E represents the association relationship between knowledge points, and the association strength between knowledge points is represented by the edge weight w ij as follows:

[0077] w ij = λ 1 ·PreReq(i, j)+λ 2 ·CoOccur(i, j)

[0078] where, w ij represents the association strength from knowledge point i to knowledge point j, PreReq(i, j) represents the prerequisite degree of knowledge point i for j, CoOccur(i, j) represents the knowledge co-occurrence frequency, and λ 1 and λ 2 are the weight coefficients.

[0079] After that, based on the hierarchical structure of the knowledge graph, calculate the knowledge coverage rate and knowledge continuity between courses, and define the course knowledge association degree:

[0080]

[0081] Among them, CourseRel(c 1 ,c 2 ) is the knowledge correlation degree between courses c1 and c2, K c1 and K c2 respectively represent the sets of knowledge points contained in courses c 1 and c 2 , |K c1 | and |K c2 | respectively represent the numbers of knowledge points of courses c 1 and c 2 ;

[0082] After that, the course similarity matrix is fused with the knowledge correlation degree to obtain the comprehensive correlation weight:

[0083] W comp (c 1 ,c 2 ) = α W ·CourseSim(c 1 ,c 2 ) + β W ·CourseRel(c 1 ,c 2 ) + γ W

[0084] ·TimeFactor(c 1 ,c 2 )

[0085] Among them, W comp (c 1 ,c 2 ) is the comprehensive correlation degree of courses c1 and c2, TimeFactor(c 1 ,c 2 ) represents the course time sequence relationship factor, α W , β W , γ W are weight coefficients;

[0086] Finally, based on the comprehensive correlation weight, a threshold θ is set to screen out the courses that satisfy the condition W comp (c 1 ,c 2 ) > θ as, and arrange them in descending order of the correlation weight, and select the top M courses to form the final candidate recommendation set candidate recommendation course set.

[0087] S140: Using the candidate recommended course set, combined with the teaching resource library of the educational robot, grade and score the courses according to the depth-first search strategy;

[0088] As Figure 2 shown, in an embodiment of the present application, integrating the teaching resource library of the educational robot by using the above course similarity matrix specifically includes:

[0089] Classify the teaching resources in the educational robot resource library according to the teaching function dimension, including: basic teaching resources (videos, courseware), interactive learning resources (exercises, quizzes), experimental operation resources (simulation environment, control programs), and extended learning resources (cases, projects);

[0090] Evaluate the overall quality and completeness of the course teaching resources in combination with the course similarity matrix, and quantify them through scoring:

[0091]

[0092] Among them, ResourceScore(c) is the resource completeness score, c is the currently evaluated course, n is the total number of resource types (such as videos, exercises, etc.), r i is the actual quantity of the i-th type of resource, R i is the standard configuration quantity of the i-th type of resource, α i is the importance weight of the i-th type of resource, q i is the quality coefficient of the i-th type of resource, with a value range of [0, 1], S c is the set of courses with significant similarity to course c, CourseSim(c, s) 2 is the similarity between course c and course s.

[0093] After that, based on the pre-constructed knowledge graph, use the depth-first search strategy to evaluate the integrity and coherence of the course knowledge path, so as to evaluate the position importance and knowledge inheritance integrity of the course in the knowledge system:

[0094]

[0095] Among them, DFS Score (c, d) is the knowledge path depth score, d is the current search depth, Child(c) is the set of subsequent courses of course c, δ is the basic depth decay factor, with a value of 0.85, CourseRel(c, v) is the knowledge correlation degree between course c and course v, and v is the subsequent course node.

[0096] After that, integrate the learner's knowledge background, course difficulty, and the comprehensive correlation degree weight calculated above to evaluate the matching degree between the course and the learner, and ensure the accuracy of personalized recommendation:

[0097]

[0098] Among them, AdaptScore(c) is the learning adaptability score, D c is the difficulty coefficient of course c, with a value range of [0, 1], D s is the current ability level of the learner, with a value range of [0, 1], λ A is the difficulty sensitivity parameter, with a value of 2.5, K c is the set of knowledge points included in course c, K s is the set of knowledge points mastered by the learner, Match(K c , K s ) is the knowledge point matching degree function, c pre is the most recently completed course by the learner, μ is the correlation influence factor, with a value of 1.5, W comp (c, c pre ) is the comprehensive correlation weight.

[0099] Finally, based on the recommendation for a single student, it is spread to the entire student group to achieve mass recommendation:

[0100]

[0101] Among them, FinalScore(c) is the recommendation score, p is the non-linear adjustment parameter, with a value of 2.5, w 1 is the depth search score weight, with a value of 0.6, w 2 is the adaptability score weight, with a value of 0.4, R c is the educational robot resource set of course c, I r is the importance coefficient of resource r, with a value range of [0, 1], S r is the applicability score of resource r, with a value range of [0, 1], G c is the target learning group of course c, GroupSim(G c ) is the group similarity score.

[0102] S150: According to the hierarchical scoring results, the top N courses with the highest scores are pushed to students as personalized recommended courses.

[0103] Specifically, the division is carried out according to the following rules:

[0104] Priority recommendation level: FinalScore(c) ≥ 1 - e -2 .

[0105] Key recommendation level: 1 - e -1 ≤ FinalScore(c) < 1 - e -2 .

[0106] General recommendation level 1-e -0.5 ≤FinalScore(c)<1-e -1 。

[0107] In an embodiment of the present application, by integrating multi-dimensional data analysis and deep learning technologies, an efficient and accurate educational course recommendation system is implemented. Specifically, first, through the refined quantitative evaluation of students' classroom interaction frequency, answer accuracy rate, and knowledge mastery level, a comprehensive learning portrait is constructed, making the recommendation basis more scientific and reliable; second, by creatively combining the calculation of group similarity with the weighted combination method of course feature vectors and introducing the SVD matrix decomposition technology, the recommendation system can take into account both the individual differences and group characteristics of learners, and the recommendation accuracy is increased by 25%; third, by designing a student similarity matching method based on the Pearson correlation coefficient and a course similarity calculation method with multi-dimensional features, the cold start problem of the system is effectively solved, and the recommendation accuracy of new users is increased by 30%; in addition, through the integration of the depth-first search strategy and the educational robot resource library, the integrity of the knowledge path and the resource matching degree of the recommendation results are optimized; finally, through reasonable weight allocation and multi-factor fusion mechanism, the interpretability of the recommendation results is enhanced, making the recommendation process more transparent. The course completion rate of students in actual application is increased by 35% compared with the traditional recommendation method, and the learning satisfaction is increased by 40%, which fully demonstrates the significant advantages of this solution in improving the personalized recommendation effect of online education.

[0108] The following introduces the device embodiments of the present application, which can be used to execute the online classroom intelligent recommendation method of the educational robot in the above embodiments of the present application. It can be understood that the device can be a computer program (including program code) running in a computer device, for example, the device is an application software; the device can be used to execute the corresponding steps in the method provided by the embodiments of the present application. For the details not disclosed in the device embodiments of the present application, please refer to the embodiments of the online classroom intelligent recommendation method of the educational robot in the above of the present application.

[0109] Figure 3 The block diagram of the online classroom intelligent recommendation system of the educational robot according to an embodiment of the present application is shown.

[0110] Figure 4 The structural schematic diagram of the computer system of the electronic device suitable for implementing the embodiments of the present application is shown.

[0111] It should be noted that the computer system of the electronic device in this embodiment is only an example, and should not bring any limitations to the functions and usage ranges of the embodiments of the present application.

[0112] In this embodiment, the computer system includes a central processing unit 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory 402 or the program loaded from the storage section 408 into the random access memory 403, such as executing the online class intelligent recommendation method of the educational robot described in the above embodiment. In the random access memory 403, various programs and data required for system operation are also stored. The central processing unit 401, the read-only memory 402, and the random access memory 403 are connected to each other via a bus 404. An input / output interface 405 is also connected to the bus 404.

[0113] The following components are connected to the input / output interface 405: an input section 406 including a keyboard, a mouse, etc.; an output section 407 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the input / output interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed so that a computer program read from it can be installed into the storage section 408 as needed.

[0114] Specifically, according to the embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments of the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication section 409, and / or installed from the removable medium 411. When the computer program is executed by the central processing unit 401, various functions defined in the system of the present application are executed.

[0115] It should be noted that the computer-readable medium shown in the embodiments of the present application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, and this computer-readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted by any suitable medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0116] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. Among them, each block in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, and the combination of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0117] The units involved in the embodiments described in this application can be implemented in software or in hardware, and the described units can also be provided in a processor. Among them, the names of these units do not, in some cases, constitute a limitation on the units themselves.

[0118] According to one aspect of the present application, there is provided a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in the above various alternative implementation manners.

[0119] As another aspect, the present application further provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or may exist alone without being assembled into the electronic device. The above computer-readable medium carries one or more programs, and when the one or more programs are executed by an electronic device, the electronic device implements the online classroom intelligent recommendation method of the educational robot described in the above embodiments.

[0120] It should be noted that although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of the two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0121] Through the description of the above embodiments, those skilled in the art can easily understand that the example embodiments described herein can be implemented in software or in a manner of software combined with necessary hardware. Therefore, the technical solutions according to the embodiments of the present application 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, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (such as a personal computer, a server, a touch terminal, or a network device, etc.) to execute the methods according to the embodiments of the present application.

[0122] After considering the specification and practicing the embodiments disclosed herein, those skilled in the art will readily conceive of other embodiments of the present application. The present application is intended to cover any variations, uses, or adaptations of the present application, which follow the general principles of the present application and include well-known knowledge or conventional technical means in the technical field not disclosed in the present application.

[0123] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. An online classroom intelligent recommendation method for an educational robot, characterized in that: include: Collect students' online classroom behavior data, including course interaction frequency, answer accuracy, and knowledge mastery, to form students' learning feature vectors; Based on the student learning feature vector, a student interest model is constructed to generate a course similarity matrix; According to the course similarity matrix and the preset knowledge graph, the course association weight is calculated to screen out a set of candidate recommended courses; Using the candidate recommended course set and combining it with the teaching resource library of the educational robot, the courses are graded and scored according to the depth-first search strategy; Based on the grading results, the top N courses with the highest scores are pushed to students as personalized recommended courses.

2. The online classroom intelligent recommendation method of an educational robot according to claim 1, characterized in that: The course similarity matrix is ​​constructed in three stages, including: Establish student feature mapping, perform similarity matching on each student's learning feature vector, divide students into groups with similar learning styles, set a similarity threshold for each group, and use the historical learning behaviors of students in the group as similar behavior samples; Extract course feature labels, decompose the core attributes of each course into five dimensions: knowledge domain, course difficulty, teaching mode, interaction type, and assessment method, and assign weight coefficients to these feature labels; A collaborative filtering model is constructed to transform the user-course interaction data into a rating matrix R. The matrix R is decomposed into the product of the user feature matrix U, the singular value matrix Σ and the course feature matrix V. The first k eigenvalues ​​are selected to construct the reduced-dimensional similarity matrix, and finally an N×N-dimensional course similarity matrix is ​​generated.

3. The online classroom intelligent recommendation method of the educational robot according to claim 2 is characterized in that: The division of students with similar learning styles into groups is based on the comprehensive similarity of students in the group; the calculation of the comprehensive similarity is shown in the following formula: Among them, GroupSim(G c ) is the group similarity score, α and β are weight coefficients, w k is the weight of the kth behavior feature, f k is the standardized score of the behavioral feature, sim(i,j) is the similarity between student i and student j, and w k is the weight value of the kth behavior feature, and n is the total number of behavior features.

4. The online classroom intelligent recommendation method of the educational robot according to claim 2 is characterized in that: The elements in the course similarity matrix are calculated by the following formula: Among them, CourseSim(c1,c2) ​​is the similarity between courses c1 and c2, with a value range of [0,1], γ is the balance factor, U is the user set, V c1 and V c2 are the feature vectors of the two courses, R uc1 and R uc2 represents the ratings of user u on courses c1 and c2, is the average rating of user u.

5. The online classroom intelligent recommendation method of an educational robot according to claim 1, characterized in that: The screening of the candidate recommended course set includes: Calculate the association strength between knowledge points; Based on the hierarchical structure of the knowledge graph and the strength of association between knowledge points, the knowledge coverage and knowledge continuity between courses are calculated; The course similarity matrix is ​​integrated with the knowledge relevance to obtain the comprehensive relevance weight; Based on the comprehensive correlation weight, set the threshold θ and filter out the items that meet the condition W comp The courses with (c1,c2)>θ are selected as the candidate recommendation set.

6. The online classroom intelligent recommendation method of the educational robot according to claim 5, characterized in that: The grading includes resource completeness scoring, knowledge path depth scoring and learning adaptability scoring; The knowledge path depth score is calculated by the resource completeness score, and then the recommendation score is calculated by combining the learning fitness score and the group similarity score, as shown in the following formula: Among them, DFS Score (c,0) is the score obtained by depth-first searching the entire knowledge path starting from the current course c, AdaptScore(c) is the learning fitness score, FinalScore(c) is the recommended score, p is the nonlinear adjustment parameter, with a value of 2.5, w1 is the depth search score weight, with a value of 0.6, w2 is the adaptability score weight, with a value of 0.4, R c A collection of educational robotics resources for course c, I r is the importance coefficient of resource r, ranging from [0,1], S r is the applicability score of resource r, ranging from [0,1], G c For the target learning group of course c, GroupSim(G c ) is the group similarity score.

7. The online classroom intelligent recommendation method of an educational robot according to claim 5, characterized in that: Based on the grading results, the top N courses with the highest scores are pushed to students as personalized recommended courses. The rules are as follows: Priority recommendation: FinalScore(c)≥1-e -2 ; Key Recommendation: 1-e -1 ≤FinalScore(c)<1-e -2 ; General recommendation level: 1-e -0.5 ≤FinalScore(c)<1-e -1 .

8. An online classroom intelligent recommendation system for educational robots, characterized in that: include: The collection unit is used to collect students' online classroom behavior data, including course interaction frequency, answer accuracy, and knowledge point mastery, to form students' learning feature vectors; Dividing the unit, constructing a student interest model based on the student learning feature vector, and generating a course similarity matrix; The initial recommendation unit is used to calculate the course relevance weight according to the course similarity matrix and the preset knowledge graph, and screen out a set of candidate recommended courses; A course scoring unit, used to use the candidate recommended course set, combined with the teaching resource library of the educational robot, to grade and score the courses according to a depth-first search strategy; The recommendation push unit is used to push the top N courses with the highest scores as personalized recommended courses to students based on the grading results.

9. A computer readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the online classroom intelligent recommendation method for an educational robot as described in any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, enables the one or more processors to implement the online classroom intelligent recommendation method for the educational robot as described in any one of claims 1 to 7.

Images