An information recommendation method, device, equipment and storage medium

Abstract

The application discloses an information recommendation method, device and equipment and a storage medium. The method comprises the following steps: obtaining student information, exercise information and knowledge information; clustering the student information, the exercise information and the knowledge information to obtain student clusters, exercise clusters and knowledge clusters; determining associated information according to the exercise clusters and the knowledge clusters; determining a mastery state of the knowledge information of students in the student clusters according to the interaction information, the student clusters and the associated information; and recommending exercises in the associated information to the students in the student clusters according to the mastery state. The method provided in the embodiment of the application can accurately recommend corresponding exercises according to the mastery state of students on knowledge, and can realize personalized exercise recommendation for students and improve the learning experience of students.

Description

Technical Field

[0001] This invention belongs to the field of artificial intelligence technology, and in particular relates to an information recommendation method, apparatus, device and storage medium. Background Technology

[0002] In teaching, recommending exercises to students in a personalized way can effectively improve their learning level. The key to recommending exercises to students in a personalized way is to accurately infer students' mastery of knowledge.

[0003] Traditional teaching methods generally rely on teachers' experience to roughly judge students' mastery of knowledge points.

[0004] Traditional teaching methods cannot accurately assess students' mastery of knowledge points. Summary of the Invention

[0005] This invention provides an information recommendation method, apparatus, device, and storage medium that can accurately determine students' mastery of knowledge and accurately recommend corresponding exercises based on students' mastery of knowledge, thereby achieving personalized exercise recommendations for students and improving their learning experience.

[0006] In a first aspect, embodiments of the present invention provide an information recommendation method, the method comprising:

[0007] Acquire student information, exercise information, and knowledge information. Exercise information includes exercises, interaction information between student information and exercises, and the correspondence between knowledge information and exercises.

[0008] The student information, exercise information, and knowledge information are clustered to obtain student clusters, exercise clusters, and knowledge clusters. Among them, students in the same student cluster interact with the same exercises in the exercise information, exercises in the same exercise cluster test the same knowledge points in the knowledge information, and a knowledge cluster includes a type of knowledge points in the knowledge information.

[0009] The association information is determined based on the exercise clusters and knowledge clusters. The association information includes the correspondence between exercise information and knowledge information, and the exercise information and knowledge information.

[0010] Based on interactive information, student clusters, and related information, determine the students' mastery of knowledge information within each student cluster;

[0011] Based on the student's understanding, recommend exercises from related information to students in the student cluster.

[0012] In one possible implementation, the mastery status of students within a student cluster regarding knowledge information is determined based on interaction information, student clusters, and association information, including:

[0013] Feature vectors are determined based on interaction information, student clusters, and association information;

[0014] Based on the feature vectors and memory network model, the mastery status of knowledge information of students in the student cluster is determined.

[0015] In one possible implementation, the mastery status of students within a student cluster regarding knowledge information is determined based on interaction information, student clusters, and association information, including:

[0016] Based on the interaction information and student clusters, identify the questions that students in the student clusters answered incorrectly;

[0017] Based on the incorrect answers and related information of students in the student cluster, determine the students' mastery of the knowledge information.

[0018] In one possible implementation, before acquiring student information, exercise information, and knowledge information, the method further includes:

[0019] Obtain exercise information, knowledge information, and original association information. The original association information includes original exercise information and original knowledge information, as well as the original correspondence between exercises and original knowledge information in the original exercise information.

[0020] Clustering exercise information and knowledge information yields the correspondence between exercises and knowledge information within the exercise information; where the correspondence represents the knowledge points examined by the exercises in the exercise information.

[0021] Based on the correspondence, the exercise information and knowledge information are added to the original associated information to obtain the associated information.

[0022] In one possible implementation, student information, exercise information, and knowledge information are clustered to obtain student clusters, exercise clusters, and knowledge clusters, including:

[0023] Based on the student information from the interaction, student clusters are obtained;

[0024] Based on the correspondence, the exercise information is clustered to obtain exercise clusters;

[0025] Knowledge information is clustered to obtain knowledge clusters.

[0026] In one possible implementation, student information, exercise information, and knowledge information are clustered to obtain student clusters, exercise clusters, and knowledge clusters, including:

[0027] The DMGC (Multi-Graph Clustering) algorithm model is used to cluster student information, exercise information, and knowledge information to obtain student clusters, exercise clusters, and knowledge clusters.

[0028] In one possible implementation, the mastery state includes the mastery state of each student in the student cluster for each type of knowledge point in the knowledge information;

[0029] Based on the student's understanding, recommend exercises from related information to students in the student cluster, including:

[0030] Based on the students' mastery of each type of knowledge point, the system recommends exercises corresponding to each type of knowledge point from the associated information to students in the student cluster.

[0031] Secondly, embodiments of the invention provide an information recommendation device, the device comprising:

[0032] The acquisition module is used to acquire student information, exercise information, and knowledge information. The exercise information includes exercises, the interaction information between student information and exercises, and the correspondence between knowledge information and exercises.

[0033] The clustering module is used to cluster student information, exercise information, and knowledge information to obtain student clusters, exercise clusters, and knowledge clusters. Students in the same student cluster interact with the same exercise information, exercises in the same exercise cluster test the same knowledge information, and a knowledge cluster includes a type of knowledge point from the knowledge information.

[0034] The determination module is used to determine the associated information based on the exercise cluster and the knowledge cluster. The associated information includes the correspondence between exercise information and knowledge information, and the exercise information and knowledge information. It is also used to determine the mastery status of students in the student cluster of knowledge information based on the interaction information, student cluster and associated information.

[0035] The recommendation module is used to recommend exercises from related information to students in a student cluster based on their mastery status.

[0036] Thirdly, embodiments of the present invention provide a data encryption device, the device comprising: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, it implements the method as described in the first aspect or any possible implementation thereof.

[0037] Fourthly, embodiments of the present invention provide a computer storage medium storing computer program instructions, which, when executed by a processor, implement the method as described in the first aspect or any possible implementation thereof.

[0038] The technical solutions provided by the embodiments of the present invention bring at least the following beneficial effects:

[0039] This invention first clusters student information, exercise information, and knowledge information to obtain student clusters, exercise clusters, and knowledge clusters. Then, based on the exercise clusters and knowledge clusters, it determines the association information, including the correspondence between exercise information and knowledge information. From the interaction information between student information and exercises, student clusters, and association information, it accurately determines the students' mastery status of knowledge information within each student cluster. After determining the students' mastery status, it can recommend exercises to students in each student cluster in a personalized way, accurately recommending corresponding exercises based on their mastery status, thereby improving students' learning experience. Attached Figure Description

[0040] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart illustrating an information recommendation method provided in an embodiment of the present invention;

[0042] Figure 2 This is a bipartite diagram formed by student information, exercise information, and knowledge information, as provided in an embodiment of the present invention.

[0043] Figure 3 This is a flowchart illustrating a method for determining a student's mastery of knowledge information according to an embodiment of the present invention.

[0044] Figure 4 This is a flowchart illustrating how an LSTM model is used to predict a student's mastery of knowledge information, according to an embodiment of the present invention.

[0045] Figure 5 This is a flowchart illustrating another method for determining a student's mastery of knowledge information provided in an embodiment of the present invention.

[0046] Figure 6 This is a flowchart illustrating a method for determining associated information provided in an embodiment of the present invention;

[0047] Figure 7 This is an information recommendation device provided in an embodiment of the present invention;

[0048] Figure 8 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0049] The features and exemplary embodiments of various aspects of the present invention will now be described in detail. To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only configured to explain the present invention and are not configured to limit the present invention. For those skilled in the art, the present invention can be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the present invention by illustrating examples of the invention.

[0050] It should be noted that, in this document, relational terms such as "first" and "second" are used merely 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. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0051] In teaching, recommending exercises to students in a personalized way can effectively improve their learning level. The key to recommending exercises to students in a personalized way is to accurately infer students' mastery of knowledge.

[0052] Traditional teaching methods generally rely on teachers' experience to roughly judge students' mastery of knowledge points.

[0053] Traditional teaching methods cannot accurately assess students' mastery of knowledge points.

[0054] This invention provides an information recommendation method, apparatus, device, and storage medium that can accurately determine students' mastery of knowledge and accurately recommend corresponding exercises based on students' mastery of knowledge, thereby achieving personalized exercise recommendations for students and improving their learning experience.

[0055] Currently, teachers can use devices that support online teaching to deliver lessons, assign and grade exercises, and students can use these devices to learn and complete assignments. These devices can include mobile phones, computers, or tablets.

[0056] The following will combine Figure 1 This invention provides a detailed description of an information recommendation method based on embodiments of the present invention.

[0057] like Figure 1 As shown, the method may include the following steps:

[0058] S110, obtain student information, exercise information and knowledge information. Exercise information includes exercises, interaction information between student information and exercises, and correspondence between knowledge information and exercises.

[0059] Student information may include student identifiers, such as student ID, account number, or name. Student information may also include information such as the student's grade level.

[0060] Interactive information may include whether a student answered a question correctly or incorrectly, and may also include information such as the time taken to answer the question.

[0061] When a problem tests one or more knowledge points, there is a correspondence between the problem and the knowledge points it tests.

[0062] Devices that support online teaching can acquire student information, exercise information, and knowledge information.

[0063] In one example, before using their mobile phones to learn online courses, students need to submit information such as their name and class. Students learn the knowledge points and answer questions that test their knowledge; the mobile phones then store the knowledge points they have learned and the questions they have answered.

[0064] In one example, the student's mobile phone forms a bipartite graph containing student information, exercise information, and knowledge information, such as... Figure 2 As shown. Figure 2 As shown, student information 201 includes at least one student, exercise information 202 includes at least one exercise, and knowledge information 203 includes at least one knowledge point. Arrows between students and exercises point to exercises answered by the student, generating interactive information. Arrows between exercises and knowledge points point to the knowledge points tested by each exercise; each exercise can simultaneously test multiple knowledge points from the knowledge information.

[0065] S120 clusters student information, exercise information, and knowledge information to obtain student clusters, exercise clusters, and knowledge clusters.

[0066] In this context, students in the same student cluster interact with the same exercises in the exercise information, and the exercises in the same exercise cluster test the same knowledge points in the knowledge information. Each knowledge cluster includes one type of knowledge point in the knowledge information.

[0067] Clustering can include:

[0068] 1. Cluster the student information based on the interaction information to obtain student clusters.

[0069] Students who have answered the same questions are grouped together to form student clusters. Within each student cluster, students interact with the same question information.

[0070] 2. Based on the correspondence, cluster the exercise information to obtain exercise clusters.

[0071] The exercises that test the same knowledge points are grouped together to form exercise clusters. Within each exercise cluster, the exercises test the same knowledge points.

[0072] 3. Cluster the knowledge information to obtain knowledge clusters.

[0073] Grouping knowledge points belonging to the same category within the knowledge information results in knowledge clusters. Each knowledge cluster contains one type of knowledge points from the knowledge information, and each type of knowledge point can include one or more knowledge points.

[0074] S130, determine the associated information based on the exercise cluster and the knowledge cluster.

[0075] The associated information includes the correspondence between exercise information and knowledge information, as well as the relationship between exercise information and knowledge information.

[0076] Based on the correspondence between the exercises and the knowledge points they test, the exercises are linked to the knowledge points, and the associated information is summarized.

[0077] In one example, the association information is represented by an association matrix, as shown below:

[0078] Exercises and Knowledge Point Association Matrix

[0079] Knowledge Point 1 Knowledge Point 2 Knowledge Point 3 Knowledge Point 4 Exercise 1 0 1 0 0 Exercise 2 1 0 1 0 Exercise 3 1 1 1 0

[0080] In the matrix above, 0 indicates that the exercise tested knowledge point 2, and 1 indicates that the exercise did not test knowledge point 3. For example, exercise 1 tested knowledge point 2, but did not test knowledge points 1, 3, and 4.

[0081] S140, based on the interaction information, student clusters, and association information, determine the students' mastery status of knowledge information within the student clusters.

[0082] By combining the students' answers to the exercises in the student clusters included in the interactive information with the knowledge points corresponding to the exercises in the associated information, it is possible to associate the exercises that students have answered with the knowledge points corresponding to the exercises. In addition, by combining the students' answers to the exercises, it is possible to accurately determine the students' mastery of the knowledge points.

[0083] S150, based on the student's mastery status, recommend exercises from related information to students in the student cluster.

[0084] A student's mastery of a knowledge point can be categorized as either "good" or "poor." When a student has a good mastery of a knowledge point, fewer practice questions related to that knowledge point in the associated information can be recommended to that student. Conversely, when a student has a poor mastery of a knowledge point, more practice questions related to that knowledge point in the associated information can be recommended to that student.

[0085] In this embodiment of the invention, student information, exercise information, and knowledge information are first clustered to obtain student clusters, exercise clusters, and knowledge clusters. Then, based on the exercise clusters and knowledge clusters, association information including the correspondence between exercise information and knowledge information is determined. Based on the interaction information between student information and exercises, student clusters, and association information, the mastery status of students in the student clusters of knowledge information is accurately determined. After determining the mastery status of students of knowledge information, exercises can be recommended to students in the student clusters in a personalized manner according to the mastery status, thereby improving the students' learning experience.

[0086] In one embodiment of the present invention, one method for determining a student's mastery of knowledge information in the above-described information recommendation method is as follows: Figure 3 As shown, the following steps may be included:

[0087] S310 determines the feature vector based on interaction information, student clusters, and association information.

[0088] Students in the same student cluster have answered the same exercises. Each student will either answer correctly or incorrectly for each exercise. The interactive information includes whether each student's answer for each exercise is correct or incorrect.

[0089] The correspondence relationship refers to the knowledge points tested in the exercises. When an exercise tests one or more knowledge points, there is a correspondence between the exercise and each knowledge point it tests.

[0090] The associated information includes the correspondence between exercise information and knowledge information, as well as the relationship between exercise information and knowledge information.

[0091] Specifically, the associated information includes the correspondence between exercises in the exercise information and knowledge points in the knowledge information, as well as the correspondence between knowledge points in the knowledge information and exercises in the exercise information.

[0092] Features represent interactive information and correspondences.

[0093] The initial embedding vector of the exercises that student j interacts with at time t is q. t =e t T ×Q, where vector e t Let Q represent the correspondence between knowledge points and exercises in the associated information, and let Q be the association matrix between exercises and knowledge points.

[0094] Let vector r t This represents the interaction information between the student and the exercises at time t.

[0095] e t and r t It can be a vector after dimensionality reduction.

[0096] For any student, at time t, the initial embedding vector q... t And the correspondence between knowledge points and exercises in the related information e t The determined eigenvector is x t =[q t ,r t ] T .

[0097] when e t and r t When it can be a dimension-reduced vector, it is determined by e. t and r t A defined eigenvector x t It is a vector after dimensionality reduction.

[0098] After the above process of determining the feature vector, the resulting feature vector is a vector after reducing the dimensionality of the student cluster and the associated information. The feature vector may include the associated information of exercise information and knowledge information, and may also include interaction information.

[0099] S320 determines the knowledge mastery status of students in a student cluster based on feature vectors and memory network models.

[0100] By inputting the feature vectors into the memory network model, we can obtain the student's mastery of knowledge information at the next time point.

[0101] In one example, taking into account the forgetting patterns of students during the learning process, a Long Short-Term Memory (LSTM) network is used to predict students' mastery of knowledge information.

[0102] The process of using an LSTM model to predict students' mastery of knowledge information is as follows: Figure 4 As shown.

[0103] If a student is learning online via mobile phone, and at time t-1, the student wants to use the mobile app to review knowledge points, the app needs to assess the student's current understanding of the knowledge points before recommending exercises based on that understanding. This assessment is based on the correspondence between knowledge points and exercises in the information at time t-1. t-1 Calculate the initial embedding vector q using the association matrix Q between exercises and knowledge points. t-1 , by q t-1Interaction information r between student A and the exercises t-1 Determine the eigenvector x at time t-1 t-1 , will x t-1 Inputting the LSTM model yields the mastery state vector of student A for K knowledge points in the knowledge cluster at time t-1. This represents the mastery status of student A over K knowledge points in the knowledge cluster at time t-1.

[0104] Here, student A represents any student in the student information. K represents the number of knowledge points, and K is greater than or equal to 1.

[0105] After a period of time, at time t, if a student wants to use a mobile app to review the knowledge points, the app needs to assess the student's current understanding of the knowledge points before recommending practice questions based on that understanding. This assessment is based on the correspondence between knowledge points and practice questions in the information at time t. t The initial embedding vector q is calculated using the association matrix Q between exercises and knowledge points, and then the embedding vector q is derived from q and the interaction information r between student A and the exercises. t Determine the eigenvector x at time t t , will x t Inputting the LSTM model into the feature vector x at time t t Input the LSTM model and... Inputting the LSTM model yields the proficiency vector of student A at time t for K knowledge points within the knowledge cluster. This represents student A's mastery status of K knowledge points in the knowledge cluster at time t.

[0106] After a period of time, at time t+1, if the student wants to use the mobile app to review the knowledge points, the app needs to assess the student's current understanding of the knowledge points before recommending practice questions based on the student's current level of mastery. This assessment is based on the correspondence between knowledge points and practice questions in the information at time t+1. t+1 Calculate the initial embedding vector q using the association matrix Q between exercises and knowledge points. t+1 , by q t+1 Interaction information r between student A and the exercises t+1 Determine the eigenvector x at time t+1 t+1 , will x t+1 Input the LSTM model and input the feature vector x at time t+1. t+1 Input the LSTM model and... Inputting the LSTM model yields the mastery state vector of student A for K knowledge points in the knowledge cluster at time t+1. This represents the mastery status of student A over K knowledge points in the knowledge cluster at time t+1.

[0107] like Figure 4 As shown, the LSTM model is used to fit the nonlinear characteristics of students' memory of knowledge points and the forgetting pattern of knowledge points over time during the learning process. Based on the students' historical mastery of knowledge points and their historical answers to exercises, the LSTM model is used to predict the students' future mastery of knowledge points.

[0108] The parameters of the LSTM model are learned using stochastic gradient descent. By continuously minimizing the loss function, the deviation between the output and the true value is continuously reduced, thereby enabling accurate prediction of students' mastery of knowledge.

[0109] The method provided in this invention can determine feature vectors based on interaction information, student clusters, and association information. The feature vectors are then input into a memory network model. The memory network model can accurately calculate students' mastery of knowledge points based on the forgetting patterns of students during the learning process, thereby improving the accuracy of assessing students' mastery of knowledge points and enabling more targeted recommendations of exercises to students.

[0110] In one embodiment of the present invention, another method for determining a student's mastery of knowledge information is as follows: Figure 5 As shown, the following steps may be included:

[0111] S510, based on the interaction information and the student cluster, identify the exercises in the student cluster that the students answered incorrectly.

[0112] Students in the same student cluster have answered the same exercises. Each student will either answer correctly or incorrectly for each exercise. The interactive information includes whether each student's answer for each exercise is correct or incorrect.

[0113] Based on the interaction information, extract the exercises that each student answered incorrectly from the student cluster.

[0114] S520, based on the questions answered incorrectly by students in the student cluster and related information, determines the students' mastery status of knowledge information in the student cluster.

[0115] When a problem tests one or more knowledge points, there is a correspondence between the problem and the knowledge points it tests.

[0116] The associated information includes the correspondence between exercise information and knowledge information, as well as the relationship between knowledge information and exercise information.

[0117] Specifically, the associated information includes the correspondence between exercises in the exercise information and knowledge points in the knowledge information, as well as the correspondence between knowledge points in the knowledge information and exercises in the exercise information.

[0118] When a student answers a question incorrectly, the relevant knowledge point corresponding to the question is identified from the associated information, indicating that the student has a poor grasp of that knowledge point.

[0119] The method provided in this invention can determine a student's mastery of knowledge points based on the exercises the student answered incorrectly, providing a basis for personalized exercise recommendations to students.

[0120] In the embodiments provided by this invention, one method for determining associated information in the above-described information recommendation method is as follows: Figure 6 As shown, the following steps are included before obtaining student information, exercise information, and knowledge information:

[0121] S610, Obtain exercise information, knowledge information, and original association information. The original association information includes original exercise information and original knowledge information, as well as the original correspondence between the exercises in the original exercise information and the knowledge points in the original knowledge information.

[0122] Original knowledge information can include knowledge points that students have already learned, and original exercise information can include exercises corresponding to the original knowledge information.

[0123] The original association information may include the original exercise information, the original knowledge information, and the original correspondence between the exercises in the original exercise information and the knowledge points in the original knowledge information.

[0124] As students continuously learn new knowledge and answer new questions, the original association information may not include the knowledge information that students are currently learning and the question information that they are currently answering, so the original association information needs to be updated.

[0125] S620, cluster the exercise information and knowledge information to obtain the correspondence between the exercises and knowledge information in the exercise information.

[0126] Among them, the correspondence represents the knowledge points corresponding to the knowledge points examined in the exercise information.

[0127] The exercises that test the same knowledge points are grouped together to form exercise clusters. Within each exercise cluster, the exercises test the same knowledge points.

[0128] Grouping knowledge points belonging to the same category within the knowledge information results in knowledge clusters. Each knowledge cluster contains one type of knowledge points from the knowledge information, and each type of knowledge point can include one or more knowledge points.

[0129] Based on the obtained information, each cluster of exercises is matched with the knowledge points it examines, thus obtaining the correspondence between the exercises and the knowledge information in the exercise information.

[0130] S630, based on the correspondence, add the exercise information and knowledge information to the original associated information to obtain the associated information.

[0131] The exercise information and knowledge information are added to the original associated information according to the corresponding relationship, so as to update the original associated information and obtain the associated information.

[0132] In one example, the original association information can be represented by the original association matrix. The association matrix of the exercises and knowledge points to be estimated is then estimated using maximum likelihood estimation based on the prior distribution of the original association matrix of the original exercises and knowledge points. Finally, through continuous iterative operations, an updated association matrix of exercises and knowledge points is obtained.

[0133] Let a represent the knowledge points annotated by experts, and b represent the knowledge points obtained from clustering. This indicates that in the i-th exercise, a is more important than b, where i is a positive integer and i is greater than or equal to 1.

[0134] Define a partial order relation: Where M is a positive integer, and M is greater than or equal to 1.

[0135] Define the problem-knowledge point association matrix to be estimated as follows: Where M represents the number of exercises, K represents the number of knowledge points (K is greater than or equal to 1), and R represents the correlation matrix. The elements in the array are real numbers.

[0136] Then in the matrix In the i-th exercise, the probability that a is more important than b is:

[0137]

[0138] Wherein, λ represents the parameter that controls the distinguishability between knowledge points.

[0139] The process of maximum likelihood estimation is as follows:

[0140]

[0141] Where, σ 2 Let K denote the variance, j denote the j-th knowledge point, j is less than or equal to K, and j is a positive integer.

[0142] The method provided in this invention can update the original association information to obtain the updated association information, thereby accurately determining the student's mastery of knowledge information based on the updated association information, and thus recommending personalized exercises to the student.

[0143] In one embodiment of the present invention, student information, exercise information, and knowledge information are clustered to obtain student clusters, exercise clusters, and knowledge clusters, including:

[0144] The DMGC (Multi-Graph Clustering) algorithm model is used to cluster student information, exercise information, and knowledge information to obtain student clusters, exercise clusters, and knowledge clusters.

[0145] According to the DMGC algorithm, the bipartite graphs of student information and exercise information are clustered to obtain student clusters, and students in each student cluster have answered the same exercises.

[0146] Based on the DMGC algorithm, the bipartite graphs of exercise information and knowledge information are clustered to obtain exercise clusters. Exercises in the same exercise cluster examine the same knowledge points, and the association information between the bipartite graphs of exercise information and knowledge information is also obtained.

[0147] According to the DMGC algorithm, knowledge information is clustered to obtain knowledge clusters, and the same knowledge cluster includes the same type of knowledge points.

[0148] Specifically, the DMGC clustering algorithm first uses an autoencoder network to learn... Figure 2 The embedding representation of nodes (student nodes, exercise nodes, and knowledge point nodes) is then used. A minimum entropy-based clustering strategy is employed to cluster nodes in the embedding space of each graph, while simultaneously introducing two regularizers to utilize both intra-graph and cross-graph dependencies. Furthermore, an attention mechanism is developed to learn cross-graph clustering associations to improve clustering performance.

[0149] The method provided in this invention can cluster student information, exercise information, and knowledge information using the DMGC algorithm to obtain student clusters, exercise clusters, and knowledge clusters, providing a basis for improving the accuracy of determining students' mastery of knowledge information.

[0150] In one embodiment of the present invention, the mastery status includes the mastery status of each type of knowledge point in the knowledge information by the students in the student cluster;

[0151] Based on the student's understanding, recommend relevant exercises from the associated information to students in the student cluster, specifically including:

[0152] Based on the students' mastery of each type of knowledge point, the system recommends exercises corresponding to each type of knowledge point from the associated information to students in the student cluster.

[0153] The method provided in this invention can recommend personalized exercises to students in a student cluster based on their mastery of each type of knowledge point in the knowledge information.

[0154] This invention also provides an information recommendation device 700, which may include an acquisition module 710, a clustering module 720, a determination module 730, and a recommendation module 740. Wherein:

[0155] The acquisition module 710 is used to acquire student information, exercise information, and knowledge information. The exercise information includes exercises, interaction information between student information and exercises, and the correspondence between knowledge information and exercises.

[0156] Clustering module 720 is used to cluster student information, exercise information and knowledge information to obtain student clusters, exercise clusters and knowledge clusters. Among them, students in the same student cluster interact with the same exercise information, exercises in the same exercise cluster test the same knowledge information, and a knowledge cluster includes a type of knowledge point in the knowledge information.

[0157] The determination module 730 is used to determine the association information based on the exercise cluster and the knowledge cluster. The association information includes the correspondence between exercise information and knowledge information, and the exercise information and knowledge information. It is also used to determine the mastery status of students in the student cluster of knowledge information based on the interaction information, the student cluster and the association information.

[0158] Recommendation module 740 is used to recommend exercises from related information to students in a student cluster based on their mastery status.

[0159] The device provided in this embodiment of the invention first clusters student information, exercise information, and knowledge information to obtain student clusters, exercise clusters, and knowledge clusters. Then, based on the exercise clusters and knowledge clusters, it determines the association information, including the correspondence between exercise information and knowledge information. Based on the interaction information between student information and exercises, student clusters, and association information, it accurately determines the students' mastery status of knowledge information in the student clusters. After determining the students' mastery status of knowledge information, it can recommend exercises to students in the student clusters in a personalized way according to the mastery status, thereby improving the students' learning experience.

[0160] In one embodiment of the present invention, the determining module 730 may be specifically used for:

[0161] Feature vectors are determined based on interaction information, student clusters, and association information, with the feature values ​​representing the interaction information and correspondences.

[0162] Based on the feature vectors and memory network model, the mastery status of knowledge information of students in the student cluster is determined.

[0163] Students in the same student cluster have answered the same exercises. Each student will either answer correctly or incorrectly for each exercise. The interactive information includes whether each student's answer for each exercise is correct or incorrect.

[0164] When a problem tests one or more knowledge points, there is a correspondence between the problem and the knowledge points it tests.

[0165] The associated information includes the correspondence between exercise information and knowledge information, as well as the relationship between exercise information and knowledge information.

[0166] Specifically, the associated information includes the correspondence between exercises in the exercise information and knowledge points in the knowledge information, as well as the correspondence between knowledge points in the knowledge information and exercises in the exercise information.

[0167] The device provided in this invention can determine feature vectors based on interaction information, student clusters, and association information, and input the feature vectors into a memory network model. The memory network model can accurately calculate the students' mastery of knowledge points based on the forgetting patterns of students during the learning process, thereby improving the accuracy of assessing students' mastery of knowledge points and enabling more targeted recommendation of exercises to students.

[0168] In one embodiment of the present invention, the determining module 730 may further be specifically used for:

[0169] Based on the interaction information and student clusters, identify the questions that students in the student clusters answered incorrectly;

[0170] Based on the incorrect answers and related information of students in the student cluster, determine the students' mastery of the knowledge information.

[0171] Students in the same student cluster have interacted with the same exercise information, generating interactive information, which includes whether the student answered the exercise correctly or incorrectly.

[0172] When a problem tests one or more knowledge points, there is a correspondence between the problem and the knowledge points it tests.

[0173] The associated information includes the correspondence between exercise information and knowledge information, as well as the relationship between knowledge information and exercise information.

[0174] Specifically, the associated information includes the correspondence between exercises in the exercise information and knowledge points in the knowledge information, as well as the correspondence between knowledge points in the knowledge information and exercises in the exercise information.

[0175] The device provided in this invention can determine a student's mastery of knowledge points in the knowledge information based on the exercises that the student answered incorrectly, thus providing a basis for personalized exercise recommendations to students.

[0176] In one embodiment of the present invention, the acquisition module 710 can also be used to acquire exercise information, knowledge information and original association information. The original association information includes original exercise information and original knowledge information, as well as the original correspondence between the exercises in the original exercise information and the original knowledge information.

[0177] The clustering module 720 can also be used to cluster exercise information and knowledge information to obtain the correspondence between exercise information and knowledge information.

[0178] Among them, the correspondence represents the knowledge points corresponding to the knowledge points examined in the exercise information.

[0179] The information recommendation device 700 may also include an adding module 750, which adds exercise information and knowledge information to the original associated information according to the correspondence, so as to obtain associated information.

[0180] The device provided in this embodiment of the invention can update the original association information to obtain the updated association information, thereby accurately determining the student's mastery of knowledge information based on the updated association information, and thus recommending exercises to the student in a personalized manner.

[0181] In one embodiment of the present invention, the clustering module 720 can be specifically used for:

[0182] Based on the student information from the interaction, student clusters are obtained;

[0183] Based on the correspondence, the exercise information is clustered to obtain exercise clusters.

[0184] Knowledge information is clustered to obtain knowledge clusters.

[0185] The device provided in this embodiment of the invention can cluster student information, exercise information, and knowledge information to obtain student clusters, exercise clusters, and knowledge clusters, providing a basis for improving the accuracy of determining students' mastery of knowledge information.

[0186] In one embodiment of the present invention, the clustering module 720 can also be specifically used to cluster student information, exercise information and knowledge information using the multi-graph clustering DMGC algorithm model to obtain student clusters, exercise clusters and knowledge clusters.

[0187] The device provided in this embodiment of the invention can cluster student information, exercise information and knowledge information through the DMGC algorithm to obtain student clusters, exercise clusters and knowledge clusters, providing a basis for improving the accuracy of determining students' mastery of knowledge information.

[0188] In one embodiment of the present invention, the mastery status includes the mastery status of each type of knowledge point in the knowledge information by the students in the student cluster.

[0189] The recommendation module 740 can be specifically used to recommend exercises corresponding to each type of knowledge point in the associated information to students in the student cluster based on their mastery status of each type of knowledge point.

[0190] The device provided in this embodiment of the invention can recommend personalized exercises to students in a student cluster based on their mastery of each type of knowledge point in the knowledge information.

[0191] The information recommendation device provided in this embodiment of the invention performs the following: Figure 1 The steps in the method shown can accurately recommend corresponding exercises based on students' mastery of knowledge, achieving the technical effect of personalized exercise recommendations and improving students' learning experience. For the sake of brevity, they will not be described in detail here.

[0192] Figure 8 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention is shown.

[0193] An electronic device may include a processor 801 and a memory 802 storing computer program instructions.

[0194] Specifically, the processor 801 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of the present invention.

[0195] Memory 802 may include mass storage for data or instructions. For example, and not limitingly, memory 802 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 802 may include removable or non-removable (or fixed) media. Where appropriate, memory 802 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 802 is non-volatile solid-state memory. In a particular embodiment, memory 802 includes read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0196] The processor 801 reads and executes computer program instructions stored in the memory 802 to achieve... Figures 1 to 6 Any of the information recommendation methods in the illustrated embodiments.

[0197] In one example, the electronic device may also include a communication interface 803 and a bus 810. For example, Figure 8As shown, the processor 801, memory 802, and communication interface 803 are connected through bus 810 and complete communication with each other.

[0198] The communication interface 803 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of the present invention.

[0199] Bus 810 includes hardware, software, or both, that couples components of an electronic device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 810 may include one or more buses. Although specific buses are described and illustrated in embodiments of the invention, the invention contemplates any suitable bus or interconnect.

[0200] This electronic device can execute the information recommendation method in the embodiments of the present invention, thereby achieving a combination Figure 1 The information recommendation method described.

[0201] Furthermore, in conjunction with the information recommendation methods in the above embodiments, this invention can be implemented using a computer storage medium. This computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the information recommendation methods described in the above embodiments.

[0202] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0203] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the required tasks. The programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0204] It should also be noted that the exemplary embodiments mentioned in this invention describe methods or systems based on a series of steps or apparatus. However, this invention is not limited to the order of the steps described above; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0205] The above description is merely a specific embodiment of the present invention. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. An information recommendation method, characterized in that, The method includes: Acquire student information, exercise information, and knowledge information, wherein the exercise information includes exercises, interaction information between student information and the exercises, and the correspondence between knowledge information and the exercises; The student information, the exercise information, and the knowledge information are clustered to obtain student clusters, exercise clusters, and knowledge clusters; wherein, students in the same student cluster interact with the same exercises in the exercise information, exercises in the same exercise cluster examine the same knowledge points in the knowledge information, and a knowledge cluster includes a type of knowledge points in the knowledge information. The association information is determined based on the exercise cluster and the knowledge cluster. The association information includes the correspondence between the exercise information and the knowledge information, the exercise information and the knowledge information. The association information is obtained by maximum likelihood estimation of the association matrix between the original exercises and the original knowledge points using the prior distribution of the original association matrix between the original exercises and the original knowledge points. Based on the interaction information, the student cluster, and the association information, determine the student's mastery status of the knowledge information within the student cluster; Based on the mastery status, recommend exercises from the associated information to the students in the student cluster; The step of determining the knowledge information mastery status of each student in the student cluster based on the interaction information, the student cluster, and the association information includes: The feature vector is determined based on the interaction information, the student cluster, and the association information; Wherein, the initial embedding vector of the exercises that interact with student j at time t is... Vector e t Let Q be the correspondence between knowledge points and exercises in the associated information, and let Q be the association matrix between the exercises and knowledge points; let vector r t This represents the interaction information between the student and the exercises at time t. For any student, at time t, the initial embedding vector q represents the interaction information between the student and the exercises. t The correspondence between knowledge points and exercises in the associated information e t Determine the feature vector as ; Based on the feature vector and memory network model, determine the mastery status of the knowledge information of each student in the student cluster; including: determining the feature vector x at time t-1. t-1 Inputting an LSTM model yields the student's mastery state vector for K knowledge points within a knowledge cluster at time t-1. t-1 The feature vector x at time t t Input the LSTM model and... t-1 Inputting an LSTM memory network model yields a student's proficiency vector for K knowledge points within a knowledge cluster at time t. t , t This represents the student's mastery status of K knowledge points in the knowledge cluster at time t, where, t-1 This represents the student's mastery status of K knowledge points in the knowledge cluster at time t-1; The step of clustering the student information, the exercise information, and the knowledge information to obtain student clusters, exercise clusters, and knowledge clusters includes: The multi-graph clustering (DMGC) algorithm model is used to cluster the student information, the exercise information, and the knowledge information to obtain the student cluster, the exercise cluster, and the knowledge cluster. The LSTM model is used to fit the nonlinear characteristics of students' memory of knowledge points over time and the forgetting pattern of knowledge points during the learning process. The LSTM model is also used to predict students' future mastery of knowledge points based on their historical mastery of knowledge points and their historical answers to exercises.

2. The method according to claim 1, characterized in that, Determining the knowledge mastery status of each student in the student cluster based on the interaction information, the student cluster, and the association information includes: Based on the interaction information and the student cluster, determine the questions in which the students in the student cluster answered incorrectly; Based on the questions answered incorrectly by the students in the student cluster and the associated information, the mastery status of the knowledge information of the students in the student cluster is determined.

3. The method according to claim 1, characterized in that, Prior to acquiring student information, exercise information, and knowledge information, the method further includes: Obtain the exercise information, the knowledge information, and the original association information, wherein the original association information includes the original exercise information and the original knowledge information, as well as the original correspondence between the exercises in the original exercise information and the original knowledge information; The exercise information and the knowledge information are clustered to obtain the correspondence between the exercises in the exercise information and the knowledge information; wherein, the correspondence represents that the exercises in the exercise information examine the corresponding knowledge points in the knowledge information; Based on the correspondence, the exercise information and the knowledge information are added to the original association information to obtain the association information.

4. The method according to claim 1, characterized in that, The step of clustering the student information, the exercise information, and the knowledge information to obtain student clusters, exercise clusters, and knowledge clusters includes: Based on the interaction information, the student information is clustered to obtain the student cluster; Based on the correspondence, the exercise information is clustered to obtain the exercise cluster; The knowledge information is clustered to obtain the knowledge cluster.

5. The method according to any one of claims 1 to 4, characterized in that, The mastery status includes the mastery status of each student in the student cluster for each type of knowledge point in the knowledge information; The step of recommending exercises from the associated information to students in the student cluster based on their mastery status specifically includes: Based on the mastery status of each type of knowledge point, the corresponding exercises for each type of knowledge point in the associated information are recommended to the students in the student cluster.

6. An information recommendation device, characterized in that, The device includes: The acquisition module is used to acquire student information, exercise information, and knowledge information. The exercise information includes exercises, interaction information between student information and the exercises, and the correspondence between knowledge information and the exercises. The clustering module is used to cluster the student information, the exercise information, and the knowledge information to obtain student clusters, exercise clusters, and knowledge clusters; wherein, students in the same student cluster interact with the same exercise information, exercises in the same exercise cluster examine the same knowledge information, and a knowledge cluster includes a type of knowledge point from the knowledge information; The determination module is used to determine association information based on the exercise cluster and the knowledge cluster. The association information includes the correspondence between the exercise information and the knowledge information, and the exercise information and the knowledge information. The association information is obtained by maximum likelihood estimation of the association matrix between the original exercises and the original knowledge points using the prior distribution of the original association matrix between the original exercises and the original knowledge points. The module is also used to determine the mastery status of the students in the student cluster regarding the knowledge information based on the interaction information, the student cluster, and the association information. The recommendation module is used to recommend exercises from the associated information to the students in the student cluster based on their mastery status. The step of determining the knowledge information mastery status of each student in the student cluster based on the interaction information, the student cluster, and the association information includes: The feature vector is determined based on the interaction information, the student cluster, and the association information; Wherein, the initial embedding vector of the exercises that interact with student j at time t is... Vector e t Let Q be the correspondence between knowledge points and exercises in the associated information, and let Q be the association matrix between the exercises and knowledge points; let vector r t This represents the interaction information between the student and the exercises at time t. For any student, at time t, the initial embedding vector q represents the interaction information between the student and the exercises. t The correspondence between knowledge points and exercises in the associated information e t Determine the feature vector as ; Based on the feature vector and memory network model, determine the mastery status of the knowledge information of each student in the student cluster; including: determining the feature vector x at time t-1. t-1 Inputting an LSTM model yields the student's mastery state vector for K knowledge points within a knowledge cluster at time t-1. t-1 The feature vector x at time t t Input the LSTM model and... t-1 Inputting an LSTM memory network model yields a student's proficiency vector for K knowledge points within a knowledge cluster at time t. t , t This represents the student's mastery status of K knowledge points in the knowledge cluster at time t, where, t-1 This represents the student's mastery status of K knowledge points in the knowledge cluster at time t-1; The clustering module is further configured to use the Multi-Graph Clustering (DMGC) algorithm model to cluster the student information, the exercise information, and the knowledge information to obtain the student cluster, the exercise cluster, and the knowledge cluster. The determining module is also used to fit the nonlinear characteristics of students' memory of knowledge points and the forgetting pattern of knowledge points over time during the learning process using the LSTM model, and to predict students' future mastery of knowledge points based on students' historical mastery of knowledge points and their historical answers to exercises using the LSTM model.

7. An electronic device, characterized in that, The device includes: a processor and a memory storing computer program instructions; the processor, when executing the computer program instructions, implements the information recommendation method as described in any one of claims 1-5.

8. A computer storage medium, characterized in that, The computer storage medium stores computer program instructions, which, when executed by a processor, implement the information recommendation method as described in any one of claims 1-5.

Images