Online learner metacognitive ability processing method and system, electronic device and medium

Abstract

The present application discloses an online learner metacognitive ability processing method and system, electronic equipment and medium, which can be applied to the field of educational data processing technology. After constructing a learner metacognitive ability structure including learners' behavior sequences and behavior indicators, the present application labels the learner metacognitive ability structure with labels that characterize the learner's metacognitive ability level, and then processes the learner metacognitive ability structure with a preset graph convolutional neural network model trained with the labeled data to obtain a metacognitive ability evaluation result with high accuracy, and then deconstructs the metacognitive ability structure based on the evaluation result to obtain the behavioral pattern differences between learners with different metacognitive abilities and the contribution of individual learners' behavioral indicators to their metacognitive abilities, thereby obtaining a visualized interpretation result of the learner's metacognitive ability structure, which can provide effective data support for teaching activities so that teachers can make more scientific and reasonable educational decisions.

Description

Technical Field

[0001] The present application relates to the technical field of educational data processing, and in particular to a method and system for processing the metacognitive ability of online learners, an electronic device and a medium. Background Art

[0002] In related technologies, the rise of online learning enables learners to choose the appropriate time, place and learning content to learn according to their own needs. However, the asynchronous network learning environment makes the interaction between learners and teachers relatively limited, reduces the effective supervision of teachers on students' learning process, and affects students' willingness and learning effect of online autonomous learning. Metacognitive ability is considered to be a key factor in improving the interest and effect of online autonomous learning. Metacognition is a high-level thinking that can help learners think, understand and monitor their learning process. The level of its ability directly affects the effect of online autonomous learning. Its accurate evaluation is crucial for personalized teaching and effective learning. Many scholars have evaluated metacognitive ability based on online learning behavior. In online learning, metacognitive ability evaluation based on learning behavior has the advantages of authenticity, automaticity and real-time.

[0003] Existing metacognitive ability assessment methods based on learning behavior are mainly based on cognitive processes. Starting from the three stages of pre-task planning, monitoring and adjustment during execution, and post-execution evaluation and correction, the learning behaviors corresponding to the metacognitive skills or strategies adopted by learners in the cognitive process are quantified to establish a mapping relationship with metacognitive ability. Machine learning methods are used to train and classify quantified learning behaviors to achieve the effect of distinguishing the level of metacognitive ability. At the same time, process-based assessment methods can be subdivided into assessment based on the first behavior sequence and assessment based on the first behavior indicator. However, since the assessment process is only based on the first behavior sequence or the first behavior indicator, which is a single metacognitive representation content, the assessment results are not accurate, and thus cannot provide effective data support for teaching activities.

[0004] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the invention

[0005] The main purpose of the embodiments of the present application is to propose an online learner metacognitive ability processing method and system, electronic device and medium, which can effectively improve the accuracy of learner metacognitive ability evaluation results and provide a visual interpretation and analysis of the evaluation results, thereby providing effective data support for teaching activities.

[0006] To achieve the above purpose, one aspect of the embodiment of the present application proposes a method for processing the metacognitive ability of online learners, the method comprising the following steps:

[0007] Constructing a learner's metacognitive ability structure according to the learner's log file, and labeling the learner's metacognitive ability structure, wherein the learner's metacognitive ability structure includes a first behavior sequence and a first behavior indicator, and the label is used to characterize the learner's metacognitive ability level, and the labeled learner's metacognitive ability structure is used to train a preset graph convolutional neural network model;

[0008] Processing the learner's metacognitive ability structure through a trained preset graph convolutional neural network model to obtain a metacognitive ability evaluation result;

[0009] The learner's metacognitive ability structure is deconstructed according to the metacognitive ability assessment result to obtain a deconstruction result, wherein the deconstruction result includes the behavioral pattern differences between learners with different metacognitive abilities and the contribution of each of the first behavioral indicators in the individual learner to the metacognitive ability.

[0010] In some embodiments, constructing a learner's metacognitive ability structure based on the learner's log file includes:

[0011] Obtaining a log file of the learner's corresponding learning course;

[0012] Extracting behavioral patterns from the log file;

[0013] Cleaning the data in the log file according to the behavior rules;

[0014] The cleaned data is converted into the first behavior sequence according to the behavior occurrence time;

[0015] Constructing an adjacency matrix according to the first line sequence;

[0016] Extract the first behavior indicator from the cleaned data;

[0017] Using the first behavior indicator as a node feature;

[0018] The learner's metacognitive ability structure is constructed according to the adjacency matrix and the node features.

[0019] In some embodiments, the training step of the preset graph convolutional neural network model includes:

[0020] quantifying the learner's metacognitive ability structure based on a graph matching algorithm;

[0021] Determining a label of the learner's metacognitive ability structure based on the quantitative results;

[0022] Dividing the learner's metacognitive ability structure after label determination into a training data set and a test data set;

[0023] Training the preset graph convolutional neural network model using the training data set;

[0024] Get the loss function during training;

[0025] Adjusting model parameters of the preset graph convolutional neural network model according to the loss function;

[0026] Testing the preset graph convolutional neural network model in the training process through the test data set to obtain test data, wherein the test data includes accuracy, precision, recall rate or F1 value;

[0027] Analyzing the classification performance of the preset graph convolutional neural network model during the training process according to the test results;

[0028] Determine that the classification performance meets preset requirements, and input the learner's metacognitive ability structure into the trained preset graph convolutional neural network model for classification.

[0029] In some embodiments, deconstructing the learner's metacognitive ability structure according to the metacognitive ability assessment result includes:

[0030] encoding the first behavior sequence in the learner's metacognitive ability structure;

[0031] generating a behavior transition diagram of the learner based on the encoding result;

[0032] generating the behavioral pattern differences between high metacognitive ability learners and low metacognitive ability learners based on the behavioral transformation diagram;

[0033] Fitting the first behavior indicator in the learner's metacognitive ability structure and the classification result to obtain a fitting result;

[0034] Perform Shapley value decomposition on the fitting result to obtain the contribution of the first behavior indicator of the learner.

[0035] In some embodiments, the quantifying the learner's metacognitive ability structure based on a graph matching algorithm includes:

[0036] Obtain a predefined second behavior sequence and a second behavior indicator;

[0037] constructing a benchmark structure according to the second behavior sequence and the second behavior indicator;

[0038] Obtaining a correlation matrix between the learner's metacognitive ability structure and the benchmark structure;

[0039] Acquire a matching matrix between the learner's metacognitive ability structure and the benchmark structure based on the association matrix;

[0040] The affinity score between the learner's metacognitive ability structure and the benchmark structure is calculated according to the association matrix and the matching matrix.

[0041] In some embodiments, the quantifying the learner's metacognitive ability structure based on a graph matching algorithm includes:

[0042] Obtain a predefined second behavior sequence and a second behavior indicator;

[0043] constructing a benchmark structure according to the second behavior sequence and the second behavior indicator;

[0044] The distance between the learner's metacognitive ability structure and the benchmark structure is calculated.

[0045] In some embodiments, the graph matching algorithm comprises a random walk-based graph matching algorithm or a spectral graph matching algorithm.

[0046] To achieve the above objectives, another aspect of the present application provides a system for processing metacognitive ability of online learners, the system comprising:

[0047] The first module is used to construct a learner's metacognitive ability structure according to the learner's log file, and label the learner's metacognitive ability structure, wherein the learner's metacognitive ability structure includes a first behavior sequence and a first behavior indicator, and the label is used to characterize the learner's metacognitive ability level. The labeled learner's metacognitive ability structure is used to train a preset graph convolutional neural network model;

[0048] The second module is used to process the learner's metacognitive ability structure through a trained preset graph convolutional neural network model to obtain a metacognitive ability evaluation result;

[0049] The third module is used to deconstruct the learner's metacognitive ability structure according to the metacognitive ability assessment result to obtain a deconstruction result, wherein the deconstruction result includes the behavioral pattern differences between learners with different metacognitive abilities and the contribution of each of the first behavioral indicators in the individual learner to the metacognitive ability.

[0050] To achieve the above objective, another aspect of an embodiment of the present application provides an electronic device, the electronic device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the above method when executing the computer program.

[0051] To achieve the above objective, another aspect of an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above method is implemented.

[0052] The embodiments of the present application include at least the following beneficial effects: The present application provides an online learner metacognitive ability processing method and system, electronic device and medium. The scheme constructs a learner metacognitive ability structure including the learner's behavior sequence and behavior indicators, and then labels the learner's metacognitive ability structure with labels that characterize the learner's metacognitive ability level, so as to train a preset graph convolutional neural network model with the labeled data, and then processes the learner's metacognitive ability structure with the trained preset graph convolutional neural network model to obtain a metacognitive ability evaluation result with high accuracy, and then deconstructs the metacognitive ability structure based on the evaluation result to obtain the behavioral pattern differences between learners with different metacognitive abilities and the contribution of learners' individual behavioral indicators to their metacognitive abilities, thereby obtaining a visualized interpretation result of the learner's metacognitive ability structure, which can provide effective data support for teaching activities so that teachers can make more scientific and reasonable educational decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a flow chart of a method for processing metacognitive ability of online learners provided in an embodiment of the present application;

[0054] Figure 2 is a schematic diagram of the learner's metacognitive ability structure provided by the embodiment of the present application;

[0055] Figure 3 It is a structural schematic diagram of a graph convolutional neural network model provided in an embodiment of the present application;

[0056] Figure 4 It is a schematic diagram of a metacognitive architecture based on an integrated double-loop system provided in an embodiment of the present application;

[0057] Figure 5 is a schematic diagram of a graph matching process provided in an embodiment of the present application;

[0058] Figure 6 is an example property graph provided in an embodiment of the present application;

[0059] Figure 7 is a diagram of an exemplary graph matching provided in an embodiment of the present application;

[0060] Figure 8 It is a behavior conversion diagram of high metacognitive ability provided by the embodiment of the present application;

[0061] Fig. 9 It is a behavior conversion diagram of low metacognitive ability provided by the embodiment of the present application;

[0062] Fig.10 is a schematic diagram of the indicator weight vector of the learner's metacognitive ability structure provided by the embodiment of the present application;

[0063] Fig.11 is a schematic diagram of the Shapley value decomposition result provided in an embodiment of the present application;

[0064] Fig.12 is a schematic diagram of the Shapley value of the learner corresponding to number 56 provided in the embodiment of the present application;

[0065] Fig.13 It is a structural diagram of the online learner metacognitive ability processing system provided by the embodiment of the present application;

[0066] Fig.14 It is a schematic diagram of the hardware structure of the electronic device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application, they are only examples of systems and methods consistent with some aspects of the embodiments of the present application.

[0068] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0069] Before describing the embodiments of the present application in detail, some nouns and terms involved in the embodiments of the present application are first described. The nouns and terms involved in the embodiments of the present application are subject to the following interpretations:

[0070] Attributed Graph0: Attributed Graph G = (V, E, A, W) is a type of Graph, where A = {a i |v i ∈V,a i ∈R k} is the attribute set of the node, and the node attribute value is a k-dimensional real vector, W = {w ij |(v i ,v j )∈E,w ij ∈R m} is the attribute set of the edge, and the edge attribute value is an m-dimensional real vector. The attribute set in the attribute graph is the key feature that distinguishes it from the graph. In particular, if w in the attribute graph ijis a scalar, then it is called a weighted graph. It is called an unweighted graph. The attribute graph in this application is an unweighted graph. The attribute graph is an extension of the graph, which has the structural information of the graph and additional attribute information. Objects in the real world and the relationships between objects generally contain attribute information, and the specific attribute set is determined by the problem and method being studied. For example, in a social network, if a node represents a person, the node contains attributes such as age, gender, and place of origin.

[0071] Graph Convolutional Network (GCN): GCN is a deep learning model for processing graph data. Unlike traditional convolutional neural networks (CNNs) that focus on processing data with regular structures (such as an image is a regular two-dimensional matrix with translation invariance), GCN is specifically used to process data with irregular structures, such as relational data such as social networks, knowledge graphs, and molecular structures.

[0072] The core idea of ​​GCN is to represent the nodes and edges in the graph structure as vectors or matrices, and extract features based on operations similar to convolutional neural networks, and then update the representation of each node through information transmission of neighboring nodes. After multiple rounds of information transmission and aggregation, the representation of each node will be continuously updated, so that the rich information of the node in the graph structure can be gradually captured, and the high-level features in the graph can be extracted, and then graph-related tasks such as node classification, edge classification, and graph classification can be performed based on the extracted features. The present invention subsequently uses the GCN algorithm to implement graph classification.

[0073] Metacognition Metacognition): Metacognition refers to the ability to recognize, monitor and regulate one's own cognitive process and cognitive state, that is, the cognition of cognition. The representation of metacognitive ability structure will be presented differently according to different metacognitive measurement methods. Traditional metacognition measurement methods are mainly carried out through survey scales, and its metacognition representation is determined according to the dimensions of metacognition, such as time management, seeking help, emotional regulation, effort monitoring, etc. The representation form is the total score or average score of the subscales under each dimension. The diary method is a record of events or observations by learners every day or at fixed time limits. Its metacognition representation is also determined according to the dimensions of metacognition, and the representation form can be a text vector. Multimodal learning analysis refers to the use of multiple modal data to represent metacognition, such as eye movement data, physiological data, behavioral data, etc., to achieve the mapping between data indicators and dimensions according to the "tripartite theory of metacognition" and the "three-stage model of self-regulation", and realize the representation of metacognition. For example, eye movement data indicators include eye saccades, blinking frequency, fixation times, etc.; physiological data include heart rate, skin electricity, brain waves, etc.; behavioral data mainly collects and analyzes learners' online learning platform log data, including login, video viewing, discussion participation, etc. The present invention mainly uses behavioral data to construct learners' metacognitive representation.

[0074] Graph Matching (GM): GM refers to matching the nodes in two graphs one by one, which requires not only high similarity of the matched nodes, but also high similarity of the connections between the nodes, that is, the edges. Graph matching attempts to map the nodes of one graph to the nodes of another graph, so that this node correspondence can minimize the difference between the two graphs under a certain metric. Commonly used graph matching algorithms include Reweighted Random Walk for Graph Matching (RRWM), Spectral Graph Matching (SM), and Neural Graph Matching (NGM).

[0075] Lag Sequential Analysis (LSA): LSA is mainly used to test the probability of learners performing one behavior after another behavior and whether it is statistically significant.

[0076] Shapley Value Decomposition: Shapley Value Decomposition is based on the combination of traditional regression equation decomposition and Shapley Value Method. It can decompose the inequality of the target variable into the contribution of its determinants and quantify it into the contribution of each regression variable to the inequality of the dependent variable. Its principle is to use the idea of ​​cooperative game to decompose the contribution of each regression variable to the difference of the dependent variable.

[0077] In related technologies, the rise of online learning enables learners to choose the appropriate time, place and learning content to learn according to their own needs. However, the asynchronous network learning environment makes the interaction between learners and teachers relatively limited, reduces the effective supervision of teachers on students' learning process, and affects students' willingness and learning effect of online autonomous learning. Metacognitive ability is considered to be a key factor in improving the interest and effect of online autonomous learning. Metacognition is a high-level thinking that can help learners think, understand and monitor their learning process. The level of its ability directly affects the effect of online autonomous learning. Its accurate evaluation is crucial for personalized teaching and effective learning. Many scholars have evaluated metacognitive ability based on online learning behavior. In online learning, metacognitive ability evaluation based on learning behavior has the advantages of authenticity, automaticity and real-time.

[0078] Existing metacognitive ability assessment methods based on learning behavior are mainly based on cognitive processes. Starting from the three stages of planning before task execution, monitoring and adjustment during execution, and evaluation and correction after execution, the learning behaviors corresponding to the metacognitive skills (planning, adjustment, monitoring, evaluation, etc.) or strategies (goal setting, seeking help, time management, etc.) adopted by learners in the cognitive process are quantified to establish a mapping relationship with metacognitive ability. Machine learning methods are used to train and classify quantified learning behaviors to achieve the effect of distinguishing the level of metacognitive ability. At the same time, process-based assessment methods can be subdivided into assessments based on behavioral sequences (focusing on the time sequence of student behaviors) and assessments based on behavioral indicators (focusing on quantitative indicators such as the frequency and interval of student behaviors). This assessment method is limited to a single metacognitive ability representation content (behavior sequence or behavior indicator), resulting in the loss of some metacognitive information, which directly affects the accuracy of prediction. In addition, the evaluation results are usually binary or multi-classified, and it is impossible to quantify the metacognitive abilities of individual learners. That is, the abilities of learners in the same category cannot be directly compared, and it is impossible to know which learning behaviors have made positive or negative contributions to the evaluation results of metacognitive abilities and the extent of the contributions. This is not conducive to students monitoring and adjusting their own learning status, and it is not conducive to teachers providing timely personalized guidance based on students' metacognitive status.

[0079] In view of this, an embodiment of the present application provides an online learner metacognitive ability processing method and system, electronic device and medium. After constructing a learner metacognitive ability structure including the learner's behavior sequence and behavior indicators, the embodiment annotates the learner's metacognitive ability structure with labels that characterize the learner's metacognitive ability level, trains a preset graph convolutional neural network model with the annotated data, and then processes the learner's metacognitive ability structure with the trained preset graph convolutional neural network model to obtain a metacognitive ability evaluation result with high accuracy. Based on the evaluation result, the metacognitive ability structure is deconstructed to obtain the behavioral pattern differences between learners with different metacognitive abilities and the contribution of learners' individual behavioral indicators to their metacognitive abilities, thereby obtaining a visualized interpretation result of the learner's metacognitive ability structure, which can provide effective data support for teaching activities so that teachers can make more scientific and reasonable educational decisions.

[0080] The method for processing the metacognitive ability of online learners provided in the embodiment of the present application relates to the field of educational data processing technology. The method for processing the metacognitive ability of online learners provided in the embodiment of the present application can be applied to a terminal, can be applied to a server, or can be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a car terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, and can also be configured to provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN) and cloud servers that provide basic cloud computing services such as big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application that implements the method for processing the metacognitive ability of online learners, etc., but is not limited to the above forms.

[0081] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network personal computers (Personal Computer, PC), minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0082] It should be noted that in each specific implementation of the present application, when it is necessary to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0083] Figure 1 is an optional flow chart of the method for processing the metacognitive ability of online learners provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps S110 to S130:

[0084] Step S110: construct a learner's metacognitive ability structure according to the learner's log file, and label the learner's metacognitive ability structure, wherein the learner's metacognitive ability structure includes a first behavior sequence and a first behavior indicator, and the label is used to characterize the learner's metacognitive ability level. The labeled learner's metacognitive ability structure is used to train a preset graph convolutional neural network model;

[0085] Step S120, processing the learner's metacognitive ability structure through the trained preset graph convolutional neural network model to obtain a metacognitive ability evaluation result;

[0086] Step S130, deconstructing the learner's metacognitive ability structure according to the metacognitive ability assessment result to obtain a deconstruction result, wherein the deconstruction result includes the behavioral pattern differences between learners with different metacognitive abilities and the contribution of each first behavioral indicator of the individual learner to the metacognitive ability.

[0087] In the embodiment of the present application, the behavior sequence includes signing in (SI), checking in (Ci), setting goals (SG), viewing material (VM), participating in discussion (PiD), taking notes (TN), watching video (WV), viewing calendar (VC), participating in test (PiT), submitting work (SW), reviewing material (RM), reviewing notes (RN), reviewing video (RV), reviewing test (RT), and logging out (Lo).

[0088] The behavioral indicators include learning time (LT), learning location (LL), spacing learning (SL), time change (TC), number of platform visits (NoPV), goal setting times (GST), open note times (ONT), add note times (ANT), note length (NL), delete note times (DNT), number of notes (NoN), discussion duration (DD), number of posts (NoP), number of views (NoVP), view calendar times (VCT), view material times (VMT), test times (TT), test duration (TD), test range (TR), watch video times (WVT), watch video duration (WVT), and so on. VideoDuration (WVD), Pause Video Duation (PVD), View WorkTimes (VWT), Work Quality (WQ), Review Test Times (RTT), Review Test Duration (RTD), Review Video Times (RVT), Review Video Duration (RVD), Review Material Times (RMT), Review Material Duration (RMD), Review Note Times (RNT), Review Note Duration (RND), Learn on Time (LoT), Learn in Advance (LiA), Distributed Learning (DLT),DL), test lag interval (TLI), time investment (TI), browsing behavior (BB), eye movement behavior (EMB), micro interaction (MI), electroencephalogram (EEG), task completion rate (TCR).

[0089] In an embodiment of the present application, the process of constructing a learner's metacognitive ability structure based on the learner's log file can be to extract the behavioral patterns in the learner's log file by data mining, automatically clean the log file, extract the first behavior sequence and the first behavior indicator from the cleaned data, save the first behavior indicator in the form of a matrix, and save the first behavior sequence in the form of an adjacency matrix, and then combine the two to generate an attribute graph as the learner's metacognitive ability structure. For example, taking a learner's online learning course as an example, the processing process of this step includes but is not limited to the following:

[0090] Step 1: Select an online learning course, obtain the learner's log file from the online learning platform, and clean the log data according to the behavioral patterns recorded in the log file: if the module is "test", if the specific behavior is "viewing the course module", it can be deleted (the student simply viewed it and did not take the test).

[0091] Step 2: Based on the cleaned data, process it into the first behavior sequence according to the time when the behavior occurred: such as "login" → "goal setting" → "view handouts" → "watch video" → "participate in discussion" → "participate in test" → "..." → "log out", etc. Save the behavior as a node, and represent the connection relationship between behaviors as a sparse matrix E∈R in COO format 2×n , which is the adjacency matrix.

[0092] Step 3: According to the metacognitive ability representation framework and the learning behavior indicator mapping method, extract the first behavior indicator and save it as a vector V i ∈R d , as the attribute value of the node (i.e., node feature or behavior indicator), the feature vectors of all nodes are stored as the node feature matrix N∈R n×d .

[0093] Step 4: Combine the adjacency matrix, nodes, and node features of steps 2 and 3 to finally form an attribute graph (i.e., the learner's metacognitive ability structure), such as Figure 2As shown in the figure, each node represents a learning behavior, the node attributes represent the indicators covered by the learning behavior, and the lines between the nodes represent the order of the learning behavior.

[0094] In the embodiment of the present application, the preset graph convolutional neural network model includes but is not limited to a graph convolutional neural network model and a graph attention neural network model. Figure 3 Taking the graph convolutional neural network model shown in the figure as an example, the training process of the preset graph convolutional neural network model includes but is not limited to the following steps:

[0095] Quantify the learners' metacognitive ability structure based on graph matching algorithm;

[0096] Determine the label of learners’ metacognitive ability structure based on the quantitative results;

[0097] Divide the learner's metacognitive ability structure after labeling into a training data set and a test data set;

[0098] Perform binary classification training on the preset graph convolutional neural network model through the training data set;

[0099] Get the loss function during training;

[0100] Adjust the model parameters of the preset graph convolutional neural network model according to the loss function;

[0101] The preset graph convolutional neural network model in the training process is tested through the test data set to obtain test data. The test data includes accuracy, precision, recall or F1 value;

[0102] Analyze the classification performance of the preset graph convolutional neural network model during training based on the test results;

[0103] When the classification performance meets the preset requirements, it is determined that the training of the preset graph convolutional neural network model is completed.

[0104] In the embodiment of the present application, the process of quantifying the learner's metacognitive ability structure based on the graph matching algorithm includes but is not limited to the following steps:

[0105] Obtaining a predefined second behavior sequence and a second behavior indicator, and constructing a benchmark structure according to the second behavior sequence and the second behavior indicator;

[0106] Obtain the correlation matrix between the learner's metacognitive ability structure and the benchmark structure, and obtain the matching matrix between the learner's metacognitive ability structure and the benchmark structure based on the correlation matrix;

[0107] The affinity scores between the learner's metacognitive ability structure and the benchmark structure are calculated based on the association matrix and matching matrix.

[0108] In this embodiment, the predefined process can be defined by the teacher, who can Figure 4 The metacognitive architecture based on the integrated double-loop system is defined as shown in Figure 1. Figure 5 As shown in the figure, the correlation matrix between the learner's metacognitive ability structure and the benchmark structure is obtained. (Affinity Matrix). Figure 6 and Figure 7 As an example, n1 and n2 represent the attribute graph and graph matching diagram of Figure 6 The number of nodes and Figure 7 The association matrix encodes the similarity between the two graphs (learner metacognitive ability structure and benchmark structure), including node similarity and edge similarity, where both node similarity and edge similarity take the node tensor and edge tensor as input and then calculate their similarity using the inner product. The diagonal elements of the association matrix K represent the similarity between nodes, and the non-diagonal elements represent the similarity between edges. Based on the association matrix K, the Neural Graph Matching (NGM) neural network is used to obtain the matching matrix of the two structures. The matching matrix X encodes the matching results and is the decision variable in the graph matching problem, where X i,a =1 means Figure 6 Node 1 and Figure 7 The node a matches, X i,a =0 means Figure 6 Node 1 and Figure 7 Node a does not match.

[0109] Based on the matching relationship X and the correlation matrix K, the affinity score between the learner's metacognitive ability structure and the reference structure is calculated using formula (1):

[0110]

[0111] in, It means converting the matrix X into a column vector, then transposing the column vector into a row vector, and then performing a dot product between the row vector and the matrix K and the column vector vec(X), and finally obtaining a score. This score can reflect the degree of match between the learner's metacognitive ability structure and the reference structure, and can be regarded as the score of the learner's metacognitive ability structure. The higher the score, the higher the learner's metacognitive ability.

[0112] In the embodiment of the present application, the calculation result of the metacognitive ability can also be evaluated. Specifically, the evaluation can be performed using formula (2):

[0113]

[0114] Among them, H represents the metacognitive ability structure score set of high metacognitive learners, L represents the metacognitive ability structure score set of low metacognitive learners, M represents the sample number of high metacognitive learners, and N represents the sample number of low metacognitive learners. The accuracy rate obtained according to the calculation formula is 90.15%, indicating that the calculation of the learner metacognitive ability structure score proposed in this application can better reflect the level of metacognitive ability.

[0115] In other embodiments, the present application performs a quantification process on the learner's metacognitive ability structure based on a graph matching algorithm, and can also use a graph edit distance algorithm to obtain the distance between the learner's metacognitive ability structure and the reference structure, and measure the gap between the learner and the target level based on the distance to indicate which learner has a better metacognitive level. Specifically, after obtaining a predefined second behavior sequence and a second behavior indicator, a benchmark structure can be constructed according to the second behavior sequence and the second behavior indicator, and then the distance between the learner's metacognitive ability structure and the benchmark structure can be calculated, and the distance can be used for marking, thereby completing the quantification process of the learner's metacognitive ability structure.

[0116] In an embodiment of the present application, after the learner's metacognitive ability structure is quantified, the label of each data can be determined based on the quantification result. Among them, the label annotation can be annotated by the results of the metacognitive questionnaire. After completing the data annotation, the training data set is input into the graph convolutional neural network model in small batches and multiple times to train the model. During the training process, different model parameters are adjusted to optimize the model performance. Specifically, the graph convolution layer of the model in this embodiment uses the ReLU activation function to introduce nonlinearity, and adds a dropout layer (with a dropout probability of 0.3) to prevent overfitting. The final fully connected layer uses global mean pooling and log_softmax function to convert the output of the neural network into a probability distribution, and uses the cross entropy loss function to calculate the difference between the probability distribution and the true label, and uses the Adam optimizer to perform gradient update and optimization on the model parameters.

[0117] At the same time, the model is tested and evaluated using the test data set. Specifically, this embodiment inputs the test data set into the graph convolutional neural network model in small batches and multiple times to obtain the binary classification results of the learner's metacognitive ability structure, and uses accuracy, precision, recall and F1-Score to evaluate the classification performance of the graph convolutional neural network model.

[0118] In the embodiment of the present application, the graph convolutional neural network model implemented in the present application is compared with the existing method in terms of evaluation accuracy and other indicators of high metacognitive ability and low metacognitive ability, and the results are shown in Table 1:

[0119] Table 1

[0120]

[0121]

[0122] It can be seen from Table 1 that the model provided by the implementation of this application has high evaluation index values ​​for both high metacognitive ability and low metacognitive ability.

[0123] In the embodiment of the present application, the process of deconstructing the learner's metacognitive ability structure according to the classification results includes but is not limited to the following steps:

[0124] Code the first behavior sequence in the learner's metacognitive ability structure; for example, code "set goals" as "SG", code "participate in discussions" as "PiD", etc.

[0125] Generate a behavior transition diagram for learners based on the encoding results; among them, the behavior transition diagram for high metacognitive ability is as follows Figure 8 As shown in the figure, the behavioral transformation diagram of low metacognitive ability is as follows Fig. 9 As shown. Figure 8 and Fig. 9 It can be seen that learners with high metacognitive ability have more associations between behaviors than learners with low metacognitive ability, such as:

[0126] ① “Set goals” after “reviewing the handouts” (this means that people with high metacognitive abilities will set appropriate goals based on the task content);

[0127] ② After “submitting your work”, click “check the handouts” (it means that people with high metacognitive ability will review the content of their work based on the handouts);

[0128] ③ “Check the handouts” after “reviewing the test” (indicating that people with high metacognitive ability will re-learn the handouts based on the wrong test questions);

[0129] ④ "Participate in discussions" after "setting goals" (this means that after setting goals, people with high metacognitive abilities will enter the discussion area to understand other people's goals and then adjust their own goals).

[0130] The differences in behavioral patterns between learners with high metacognitive abilities and learners with low metacognitive abilities were analyzed based on the behavior transition diagram.

[0131] The first behavior indicator and classification results in the learner's metacognitive ability structure are fitted, and the overall weight of the first behavior indicator is obtained according to the fitting results. In this step, the learner's behavior indicator and classification results can be saved in a table, and the behavior indicator and evaluation results can be fitted by logistic regression. The overall weight of the node in the learner's metacognitive ability structure is obtained according to the fitting equation, such as Fig.10As shown. Fig.10 In the figure, the horizontal axis represents different features, and the vertical axis represents the weight vector of the class. The darker the color, the greater the weight of the behavior indicator in the class.

[0132] The fitting results are decomposed by Shapley value to obtain the contribution of the learner's first behavior indicator. Since the above step only calculates the overall weight of each category, it is impossible to determine which characteristics of individual learners contribute more to their learner metacognitive ability structure. Therefore, this step further performs Shapley value decomposition to obtain Fig.11 The results shown are used to obtain the eigenvalues ​​that contribute more to the learner's metacognitive ability structure at the individual level. Specifically, Fig.12 It can be seen that the contribution of learners' behavioral indicators to the structure of their metacognitive ability. Among them, learner 56's number of times he viewed lecture notes (VMT) and number of posts (NoP) made a greater positive contribution to his metacognitive ability, while the number of platform visits (NoPV) and the number of tests (TT) made a greater negative contribution to his metacognitive ability. This shows that the number of platform visits and the number of tests are limiting factors for the improvement of the learner's metacognitive ability. Teachers need to design strategies for learner 56 to improve the number of platform visits and the number of tests for personalized teaching, so as to improve his metacognitive ability level.

[0133] Reference Fig.13 In an embodiment of the present application, a system for processing metacognitive ability of online learners is provided, the system comprising:

[0134] The first module 1310 is used to construct a learner metacognitive ability structure according to the learner's log file, and label the learner metacognitive ability structure, wherein the learner metacognitive ability structure includes a first behavior sequence and a first behavior indicator, and the label is used to characterize the learner's metacognitive ability level. The labeled learner metacognitive ability structure is used to train a preset graph convolutional neural network model;

[0135] The second module 1320 is used to process the learner's metacognitive ability structure through the trained preset graph convolutional neural network model to obtain a metacognitive ability evaluation result;

[0136] The third module 1330 is used to deconstruct the learner's metacognitive ability structure according to the metacognitive ability assessment results to obtain the deconstruction results, wherein the deconstruction results include the behavioral pattern differences between learners with different metacognitive abilities and the contribution of each first behavioral indicator of the individual learner to the metacognitive ability.

[0137] It can be understood that the contents of the above method embodiments are all applicable to the present system embodiments, the functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0138] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned online learner metacognitive ability processing method when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0139] It can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0140] See also Fig.14 , Fig.14 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0141] The processor 1410 may be implemented by a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0142] The memory 1420 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1420 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 1420, and the processor 1410 calls and executes the online learner metacognitive ability processing method of the embodiment of this application;

[0143] Input / output interface 1430, used to implement information input and output;

[0144] Communication interface 1440, used to realize communication interaction between the device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0145] Bus 1450 , which transmits information between various components of the device (e.g., processor 1410 , memory 1420 , input / output interface 1430 , and communication interface 1440 );

[0146] The processor 1410 , the memory 1420 , the input / output interface 1430 , and the communication interface 1440 are connected to each other in communication within the device via a bus 1450 .

[0147] The embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned method for processing the metacognitive ability of online learners is implemented.

[0148] It can be understood that the contents of the above method embodiments are all applicable to the present storage medium embodiments, the functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0149] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0150] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0151] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0152] The system embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0153] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0154] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0155] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0156] In the several embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of systems or units, which can be electrical, mechanical or other forms.

[0157] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0158] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0159] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method of each embodiment of the present application. The aforementioned storage medium includes: various media that can store programs, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.

[0160] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A method for processing online learners' metacognitive ability, characterized in that: The method comprises the following steps: Constructing a learner's metacognitive ability structure according to the learner's log file, and labeling the learner's metacognitive ability structure, wherein the learner's metacognitive ability structure includes a first behavior sequence and a first behavior indicator, and the label is used to characterize the learner's metacognitive ability level, and the labeled learner's metacognitive ability structure is used to train a preset graph convolutional neural network model; Processing the learner's metacognitive ability structure through a trained preset graph convolutional neural network model to obtain a metacognitive ability evaluation result; The learner's metacognitive ability structure is deconstructed according to the metacognitive ability assessment result to obtain a deconstruction result, wherein the deconstruction result includes the behavioral pattern differences between learners with different metacognitive abilities and the contribution of each of the first behavioral indicators in the individual learner to the metacognitive ability.

2. The method according to claim 1, characterized in that The step of constructing a learner's metacognitive ability structure according to the learner's log file includes: Obtaining a log file of the learner's corresponding learning course; Extracting behavioral patterns from the log file; Cleaning the data in the log file according to the behavior rules; The cleaned data is converted into the first behavior sequence according to the behavior occurrence time; Constructing an adjacency matrix according to the first line sequence; Extract the first behavior indicator from the cleaned data; Using the first behavior indicator as a node feature; The learner's metacognitive ability structure is constructed according to the adjacency matrix and the node features.

3. The method according to claim 1, characterized in that The training steps of the preset graph convolutional neural network model include: quantifying the learner's metacognitive ability structure based on a graph matching algorithm; Determining a label of the learner's metacognitive ability structure based on the quantitative results; Dividing the learner's metacognitive ability structure after label determination into a training data set and a test data set; Performing binary classification training on the preset graph convolutional neural network model using the training data set; Get the loss function during training; Adjusting model parameters of the preset graph convolutional neural network model according to the loss function; Testing the preset graph convolutional neural network model in the training process through the test data set to obtain test data, wherein the test data includes accuracy, precision, recall rate or F1 value; Analyzing the classification performance of the preset graph convolutional neural network model during the training process according to the test data; When the classification performance meets the preset requirements, it is determined that the training of the preset graph convolutional neural network model is completed.

4. The method according to claim 1, characterized in that: Deconstructing the learner's metacognitive ability structure according to the metacognitive ability assessment result includes: encoding the first behavior sequence in the learner's metacognitive ability structure; generating a behavior transition diagram of the learner based on the encoding result; generating the behavioral pattern differences between high metacognitive ability learners and low metacognitive ability learners based on the behavioral transformation diagram; Fitting the first behavior indicator in the learner's metacognitive ability structure and the metacognitive ability evaluation result to obtain a fitting result; Perform Shapley value decomposition on the fitting result to obtain the contribution of the first behavior indicator of the learner.

5. The method according to claim 3, characterized in that: The quantification of the learner's metacognitive ability structure based on a graph matching algorithm includes: Obtain a predefined second behavior sequence and a second behavior indicator; constructing a benchmark structure according to the second behavior sequence and the second behavior indicator; Obtaining a correlation matrix between the learner's metacognitive ability structure and the benchmark structure; Acquire a matching matrix between the learner's metacognitive ability structure and the benchmark structure based on the association matrix; The affinity score between the learner's metacognitive ability structure and the benchmark structure is calculated according to the association matrix and the matching matrix.

6. The method according to claim 3, characterized in that The quantification of the learner's metacognitive ability structure based on a graph matching algorithm includes: Obtain a predefined second behavior sequence and a second behavior indicator; constructing a benchmark structure according to the second behavior sequence and the second behavior indicator; The distance between the learner's metacognitive ability structure and the benchmark structure is calculated.

7. The method according to any one of claims 5 or 6, characterized in that: The graph matching algorithm includes a random walk-based graph matching algorithm or a spectral graph matching algorithm.

8. A metacognitive ability processing system for online learners, characterized in that: The system comprises: The first module is used to construct a learner's metacognitive ability structure according to the learner's log file, and label the learner's metacognitive ability structure, wherein the learner's metacognitive ability structure includes a first behavior sequence and a first behavior indicator, and the label is used to characterize the learner's metacognitive ability level. The labeled learner's metacognitive ability structure is used to train a preset graph convolutional neural network model; The second module is used to process the learner's metacognitive ability structure through a trained preset graph convolutional neural network model to obtain a metacognitive ability evaluation result; The third module is used to deconstruct the learner's metacognitive ability structure according to the metacognitive ability assessment result to obtain a deconstruction result, wherein the deconstruction result includes the behavioral pattern differences between learners with different metacognitive abilities and the contribution of each of the first behavioral indicators in the individual learner to the metacognitive ability.

9. An electronic device, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images