Longitudinal assessment and early warning method, device and equipment for high-order thinking ability of student and medium

Abstract

The invention provides a longitudinal assessment and early warning method, device and equipment for high-order thinking ability of students and a medium, and relates to the technical field of education, and the method comprises the steps: improving a DINA model through employing a joint-cross load modeling method, constructing a longitudinal cognitive assessment model suitable for multi-source and multi-frequency features, and carrying out the longitudinal assessment and early warning of the high-order thinking ability of students. Processing the multi-source feature data of the student according to a longitudinal cognitive evaluation model to obtain a longitudinal evaluation result of the high-order thinking ability of the student; and constructing a prediction model suitable for multi-source and multi-frequency features based on a long-short term memory neural network, and processing a longitudinal evaluation result according to the prediction model to obtain a student high-order thinking ability prediction result at a future time point. According to the invention, comprehensive, objective and accurate evaluation and prediction of high-order thinking ability of students can be realized; potential problems can be recognized in advance, educators are helped to adjust teaching strategies in time, and therefore the teaching effect and the learning efficiency of students are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of educational technology, and particularly to a longitudinal evaluation and early warning method, device, equipment and medium for students' higher-order thinking ability. Background Art

[0002] At present, there are some evaluation methods for students' cognitive ability, such as traditional tests based on grades and simple questionnaires. Although these methods can provide certain evaluation information, they have obvious limitations. First, they mainly focus on students' knowledge mastery, ignoring multi-dimensional cognitive performances such as text expression ability and behavior sequence characteristics. Second, traditional evaluation methods are usually static, lacking continuous monitoring and analysis of the dynamic development of students' cognitive ability. In addition, existing methods are often not fine enough in feature extraction and selection, and it is difficult to eliminate low-correlated and redundant features, resulting in limitations in the accuracy and comprehensiveness of evaluation results. These problems limit the comprehensive evaluation and effective guidance of students' higher-order thinking ability. Summary of the Invention

[0003] The purpose of the present invention is to provide a longitudinal evaluation and early warning method, device, equipment and medium for students' higher-order thinking ability, which is used to solve problems such as the limitations in the accuracy and comprehensiveness of the evaluation results of existing evaluation methods for students' cognitive ability, can accurately evaluate the current situation of students' higher-order thinking ability, and can also timely discover and feedback potential problems in students' cognitive development, so as to provide more comprehensive and effective guidance.

[0004] To achieve the above purpose, in the first aspect, the present invention provides a longitudinal evaluation and early warning method for students' higher-order thinking ability, including:

[0005] Improve the DINA model by using the joint-cross loading modeling method to construct a longitudinal cognitive evaluation model applicable to multi-source and multi-frequency features, and process the multi-source feature data of students according to the longitudinal cognitive evaluation model to obtain the longitudinal evaluation results of students' higher-order thinking ability;

[0006] Construct a prediction model applicable to multi-source and multi-frequency features based on the long short-term memory neural network, and process the longitudinal evaluation results according to the prediction model to obtain the prediction results of students' higher-order thinking ability at future time points.

[0007] According to the longitudinal evaluation and early warning method for students' higher-order thinking ability provided by the present invention, the multi-source feature data of students includes task score data, text feature data, and behavior sequence feature data of students at multiple time points.

[0008] According to the longitudinal evaluation and early warning method for students' higher-order thinking ability provided by the present invention, the longitudinal evaluation results of students' higher-order thinking ability include the higher-order thinking ability levels, dimension mastery situations, and multi-dimensional features of student groups and individuals.

[0009] A longitudinal evaluation and warning method for students' higher-order thinking ability provided by the present invention, the multi-dimensional features include rational expression, safety awareness, and standardized behavior.

[0010] A longitudinal evaluation and warning method for students' higher-order thinking ability provided by the present invention, the longitudinal cognitive evaluation model includes a measurement model and a development model. The measurement model is used to model the task score data, text feature data, and behavioral sequence feature data of students at multiple time points, as well as the structural relationships between the higher-order thinking ability levels, dimension mastery situations, and multi-dimensional features of the student group and individuals based on the joint-cross loading modeling method to improve the DINA model; the development model is based on the measurement model and constructs a measurement model that integrates multiple time points by drawing on the idea of the latent growth model.

[0011] A longitudinal evaluation and warning method for students' higher-order thinking ability provided by the present invention, the measurement model includes a first-layer model and a second-layer model. The first-layer model is used to separately model the multi-source feature data of students, and the second-layer model is used to describe the relationship between the higher-order thinking ability of students and the multi-source feature data.

[0012] A longitudinal evaluation and warning method for students' higher-order thinking ability provided by the present invention, the prediction model includes an input module, a feature selection module, a feature integration module, a residual module, and an output module; the input module is used to input the longitudinal evaluation results and divide the longitudinal evaluation results into score data, mastery attribute data, and other feature data; the feature selection module is used to assign different weights to the score data, mastery attribute data, and other feature data respectively by using a dimension-based attention mechanism and generate a feature matrix; the feature integration module is used to extract deep features from the feature matrix using different convolutional layers to form a deep feature matrix, and perform feature fusion and linear transformation to obtain fused features; the residual module is used to perform two-layer long short-term memory neural network processing on the fused features, and then use a time-based attention mechanism to connect with residuals; the output module is used to process the output data of the residual module using two-layer fully connected layers and output the prediction results of the students' higher-order thinking ability at future time points.

[0013] In a second aspect, the present invention provides a longitudinal evaluation and warning device for students' higher-order thinking ability, including:

[0014] An evaluation unit, which is used to improve the DINA model by using the joint-cross loading modeling method, construct a longitudinal cognitive evaluation model applicable to multi-source and multi-frequency features, and process the multi-source feature data of students according to the longitudinal cognitive evaluation model to obtain the longitudinal evaluation results of the students' higher-order thinking ability;

[0015] An early warning unit, which is used to build a prediction model applicable to multi-source and multi-frequency features based on a long short-term memory neural network, and processes the longitudinal evaluation results according to the prediction model to obtain the prediction results of students' higher-order thinking ability at future time points.

[0016] In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, it implements the longitudinal evaluation and early warning method for students' higher-order thinking ability in the first aspect.

[0017] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, wherein when the computer program is executed by a processor, it implements the longitudinal evaluation and early warning method for students' higher-order thinking ability in the first aspect.

[0018] The technical solution of the present invention at least has the following technical effects:

[0019] The present invention provides a longitudinal evaluation and early warning method, device, equipment and medium for students' higher-order thinking ability, aiming to solve the deficiencies in the evaluation of students' higher-order thinking ability in the prior art. By integrating multi-source feature data (including task scores, text features, behavior sequence features, etc.), a longitudinal cognitive evaluation model is constructed, which can dynamically monitor and analyze the cognitive performance of students at multiple time points. It can not only accurately evaluate the current situation of students' higher-order thinking ability, but also through the prediction and early warning functions, timely discover and feedback potential problems in students' cognitive development, so as to provide more comprehensive and effective guidance. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0021] In the drawings:

[0022] Figure 1 is a flowchart of the longitudinal evaluation and early warning method for students' higher-order thinking ability of the present invention;

[0023] Figure 2 is a schematic diagram of the longitudinal cognitive evaluation model of the present invention;

[0024] Figure 3 is a schematic diagram of the measurement model of the present invention;

[0025] Figure 4 is a schematic diagram of the electronic device of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0027] The following will describe in detail some embodiments of the present invention with reference to the accompanying drawings. Without conflict, the embodiments and features in the following embodiments may be combined with each other.

[0028] Embodiment 1

[0029] Please refer to Figure 1 , this embodiment provides a longitudinal evaluation and early warning method for students' higher-order thinking ability integrating multi-dimensional features, including:

[0030] Step 1: Improve the DINA model (Deterministic Inputs, Noisy “And” gate, cognitive diagnosis model) using the joint-cross loading modeling method to construct a longitudinal cognitive evaluation model (ML-DINA, Multimodal Longitudinal-DINA) applicable to multi-source and multi-frequency features. Process the multi-source feature data of students according to the longitudinal cognitive evaluation model to obtain the longitudinal evaluation results of students' higher-order thinking ability;

[0031] Specifically, the multi-source feature data of students includes task score data, text feature data, and behavioral sequence feature data of students at multiple time points. The longitudinal evaluation results of students' higher-order thinking ability include the higher-order thinking ability levels, dimension mastery situations, and multi-dimensional features (rational expression, safety awareness, standardized behavior, etc.) of student groups and individuals. By constructing a longitudinal cognitive evaluation model based on multi-source and multi-frequency features, processing multi-source feature data such as task scores, text features, and behavioral sequence features of students at multiple time points, calculating the higher-order thinking ability scores, dimension mastery situations, and current status and changes of multi-dimensional levels such as rational expression, safety awareness, and standardized behavior of student groups and individuals, so as to comprehensively evaluate the current status of the development of students' higher-order thinking ability.

[0032] Figure 2 is a schematic diagram of the longitudinal cognitive evaluation model. Figure 2 In, the higher-order thinking ability of students is the digital social responsibility of higher vocational students, θ is the digital social responsibility score (i.e., the ability level), α1 to α kLet \(\alpha\) be the potential attributes of digital social responsibility (i.e., digital social responsibility evaluation indicators or dimensions, such as \(\alpha_1\) representing "digital health" of digital social responsibility), \(\tau\) be the rational expression feature, \(\varepsilon\) be the safety awareness feature, \(\beta\) be the normative behavior feature, \(Y\) be the task score data, \(T\) be the text feature data corresponding to the rational expression feature, \(V\) be the text feature data corresponding to the safety awareness feature, and \(S\) be the behavior sequence feature data corresponding to the normative behavior feature.

[0033] This longitudinal cognitive assessment model includes a measurement model and a development model.

[0034] The measurement model is used to model the structural relationships between the feature data such as task scores, text features, and behavior sequence features of students at multiple time points and the higher-order thinking ability levels, dimension mastery (potential attributes), and multi-dimensional features (such as rational expression, safety awareness, and normative behavior) of the student group and individuals based on the joint-cross loading modeling method, improving the limitation of the traditional DINA model that cannot handle multi-source feature data. As Figure 3 shown, the measurement model includes a first-layer model and a second-layer model.

[0035] The first-layer model is mainly responsible for separately modeling the multi-source feature data of students, including task score data (such as answer results), text feature data, behavior sequence feature data, etc. Its core formula is:

[0036] \(P(Y ni = 1) = g i +(1 - s i - g i )\(\eta ni (1)

[0037] \(\logit(P(\alpha nk = 1)) = \gamma 1k \(\theta n - \gamma 0k (2)

[0038] where \(P\) represents the probability of correctly completing the task; \(\logit\) represents the probability of a student's mastery of a certain attribute; \(Y ni is the accuracy of student \(n\) answering question \(i\); \(g i is the guessing parameter of question \(i\); \(s i is the slip parameter of question \(i\); \(\eta ni is the ideal answer of student \(n\) to question \(i\); \(\Delta nk is the mastery of student \(n\) on attribute \(k\); \(\gamma 1k is the weight parameter of attribute \(k\); \(\theta n is the ability parameter of student \(n\); \(\gamma 0k is the threshold parameter of attribute \(k\).

[0039] Equations (1) and (2) are ability measurement models used to measure the relationship between students' higher-order thinking abilities and their response results. When introducing multi-modal data, new measurement models are introduced according to the differences in multi-modal data, and there are differences in the measurement models for each type of modal data.

[0040] The second-layer model is used to describe the relationship between students' higher-order thinking abilities and multi-source feature data. Its core formula is:

[0041] f(θ n ,α nk ,g i )=θ n (3)

[0042] Among them, the function f(θ n ,α nk ,g i ) represents how the latent ability θ of student n for a given question i affects the various features involved. The meaning here is that if only single-modal data is used, only θ is output n . If multi-modal data is introduced, additional formulas are introduced according to the data modality to expand the second-layer model. n

[0043] The development model is a measurement model that fuses multiple time points, constructed on the basis of the measurement model by drawing on the idea of the latent growth model, to evaluate the changes in the higher-order thinking abilities of student groups and individuals. For the higher-order thinking abilities of students, the group mean change is The group scale change is The individual change is The multi-dimensional characteristics of students' higher-order thinking abilities (such as rational expression, safety awareness

[0044] θ n =(θn1,…,θnT)′~MVN(μ,∑) (4)

[0045] Among them, θ n is the ability parameter of student n, and MVN indicates that a random vector follows a multivariate normal distribution with a mean of μ and a covariance matrix of ∑. Among them, the mean vector μ = (μ1,,...,μT)’, and the covariance matrix is:

[0046]

[0047] Taking the initial point and the comparison point, the measurement model assumes that the general latent ability at the first time point satisfies the standard normal distribution. Therefore, μ1 = 0 and σ1 2 =1; σ 1T is the covariance between the general latent abilities at time point 1 and time point T.

[0048] Step 2: Build a prediction model suitable for multi-source and multi-frequency features (Multimodal Data and Multifrequency Data-LSTM, MM-LSTM) based on the long short-term memory neural network (LSTM, Long Short-term Memory Networks), and process the longitudinal evaluation results according to the prediction model to obtain the prediction results of students' higher-order thinking ability at future time points.

[0049] Specifically, the prediction model suitable for multi-source and multi-frequency features includes an input module, a feature selection module, a feature integration module, a residual module, and an output module.

[0050] The input module is used to input the analysis data of ML-DINA, that is, the longitudinal evaluation results. The specific idea is as follows: According to the data type differences of the longitudinal evaluation results obtained by ML-DINA, they are divided into three types of feature data: score data, mastery attribute data, and other feature data, and input into the feature selection module. Its core formula is:

[0051] Xinput = {Xscore, Xmastery, Xother} (5)

[0052] Among them, Xscore, Xmastery, and Xother represent score data, mastery attribute data, and other feature data respectively, and Xinput is the longitudinal evaluation result.

[0053] The feature selection module is used to improve the feature recognition ability of LSTM. The specific idea is as follows: Adopt a dimension-based attention mechanism to filter different types of features and identify the influence degree of features on the model prediction results. Different weights are assigned to different features according to the influence degree, and a feature matrix is generated and input into the feature integration module. Its core formula is:

[0054] Wfeature = Attention(Xinput) (6)

[0055] Xselected = Xinput × Wfeature (7)

[0056] Among them, Attention represents the attention mechanism, Wfeature represents the weight, and Xselected represents the feature matrix.

[0057] The feature integration module is used to improve the feature extraction effect of LSTM. The specific idea is as follows: Use different convolutional layers to extract deep features from the feature matrix to form a deep feature matrix, and perform feature fusion and linear transformation to obtain the fused features and input them into the residual module. Its core formula is:

[0058] Fconv = Conv(Xselected) (8)

[0059] Fintegrated = Fusion(Fconv) (9)

[0060] Among them, Conv represents the convolutional layer operation, Fusion represents feature fusion, and Fintegrated represents the fused feature.

[0061] The residual module is used to predict the future development status of students' higher-order thinking ability. The specific idea is as follows: The fused feature input to the feature integration module is processed by two layers of LSTM. Then, a time-based attention mechanism is adopted to enhance the recognition ability of the impact of time on the change of students' higher-order thinking ability, and the residual is used for connection to solve the problem of information loss in LSTM and enhance the prediction effect of the model. Its core formula is:

[0062] Hlstm = LSTM(Fintegrated) (10)

[0063] Hresidual = Hlstm + Fintegrated (11)

[0064] Among them, Hlstm represents the processing of two layers of LSTM; Hresidual represents the residual connection.

[0065] The output module is used to transform the output data of the residual module and output the prediction result. The specific idea is as follows: Two fully connected layers are used to process the output data generated by the residual module, and the prediction result of the students' higher-order thinking ability at future time points is output, including the level of students' higher-order thinking ability, the mastery of attributes, and other features, etc. Its core formula is:

[0066] Youtput = FC(Hresidual) (12)

[0067] Among them, FC represents the fully connected layer, and Youtput represents the prediction result.

[0068] Embodiment 2

[0069] Based on the same inventive concept, this embodiment provides a longitudinal evaluation and early warning device for students' higher-order thinking ability. This device corresponds to the method of Embodiment 1. This device includes:

[0070] An evaluation unit, which is used to improve the DINA model by using the joint-cross load modeling method, construct a longitudinal cognitive evaluation model applicable to multi-source and multi-frequency features, and process the multi-source feature data of students according to the longitudinal cognitive evaluation model to obtain the longitudinal evaluation result of students' higher-order thinking ability;

[0071] An early warning unit is used to construct a prediction model applicable to multi-source and multi-frequency features based on a long short-term memory neural network, and process the longitudinal evaluation results according to the prediction model to obtain the prediction results of students' higher-order thinking abilities at future time points.

[0072] Embodiment 3

[0073] Figure 4 An example of a schematic structural diagram of an electronic device is shown as Figure 4 shown. The electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840. Among them, the processor 810, the communication interface 820, and the memory 830 complete mutual communication through the communication bus 840. A computer program that can run on the processor 810 is stored on the memory 830. When the processor 810 executes the computer program, the longitudinal evaluation and early warning method for students' higher-order thinking abilities in Embodiment 1 is implemented.

[0074] Embodiment 4

[0075] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the longitudinal evaluation and early warning method for students' higher-order thinking abilities in Embodiment 1 is implemented.

[0076] It should be noted that the computer-readable storage medium in this embodiment may be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0077] In this embodiment, the computer-readable storage medium can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. In this embodiment, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable storage medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0078] The above computer-readable storage medium can be written in one or more programming languages or combinations thereof for executing the computer program of this embodiment. The above programming languages include object-oriented programming languages - such as Java, Python, C++, and also include conventional procedural programming languages - such as C language or similar programming languages. The program can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0079] In summary, a method, apparatus, device, and medium for longitudinal evaluation and early warning of students' higher-order thinking ability proposed by the present invention constructs a longitudinal cognitive evaluation model by integrating multi-source feature data, and can comprehensively and dynamically evaluate students' higher-order thinking ability. Compared with traditional evaluation methods, the present invention not only improves the accuracy and comprehensiveness of evaluation, but also can dynamically monitor the development and change of students' higher-order thinking ability, providing more timely and effective feedback. Experimental data shows that the method of the present invention has significantly improved the evaluation accuracy and is significantly better than existing static evaluation techniques. In addition, through the prediction and early warning functions, the present invention can identify potential problems in advance, helping educators adjust teaching strategies in a timely manner, thereby significantly improving teaching effects and students' learning efficiency. Through the longitudinal cognitive evaluation model and machine learning technology, a comprehensive, objective, and accurate evaluation and prediction of students' higher-order thinking ability are realized.

[0080] Other embodiments of the present invention will be readily apparent to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the invention following the general principles of the invention and including known common knowledge or conventional technical means in the technical field not disclosed by the present invention. It should be understood that the present invention is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present invention is only limited by the appended claims.

Claims

1. A method for longitudinal assessment and early warning of students' higher-order thinking ability, characterized in that: include: The DINA model was improved by using the joint-cross-load modeling method, and a longitudinal cognitive assessment model suitable for multi-source and multi-frequency characteristics was constructed. The multi-source characteristic data of students were processed according to the longitudinal cognitive assessment model to obtain the longitudinal assessment results of students' higher-order thinking ability. A prediction model suitable for multi-source and multi-frequency features is constructed based on a long short-term memory neural network, and the longitudinal assessment results are processed according to the prediction model to obtain prediction results of students' higher-order thinking ability at future time points.

2. The method for longitudinal assessment and early warning of students' higher-order thinking ability according to claim 1 is characterized in that: The multi-source feature data of the students include task score data, text feature data, and behavior sequence feature data of the students at multiple time points.

3. The method for longitudinal assessment and early warning of students' higher-order thinking ability according to claim 2 is characterized in that: The longitudinal assessment results of students' higher-order thinking ability include the higher-order thinking ability level, dimensional mastery and multi-dimensional characteristics of student groups and individuals.

4. The method for longitudinal assessment and early warning of students' higher-order thinking ability according to claim 3 is characterized in that: The multi-dimensional characteristics include rational expression, safety awareness, and standardized behavior.

5. The method for longitudinal assessment and early warning of students' higher-order thinking ability according to claim 3 is characterized in that: The longitudinal cognitive assessment model includes a measurement model and a development model. The measurement model is used to model the structural relationship between students' task score data, text feature data, behavior sequence feature data at multiple time points and the higher-order thinking ability level, dimensional mastery, and multi-dimensional characteristics of student groups and individuals based on the joint-cross-load modeling method to improve the DINA model; the development model is a measurement model that integrates multiple time points and is constructed on the basis of the measurement model and draws on the idea of ​​the potential growth model.

6. The method for longitudinal assessment and early warning of students' higher-order thinking ability according to claim 5 is characterized in that: The measurement model includes a first-layer model and a second-layer model. The first-layer model is used to separately model the multi-source feature data of the students, and the second-layer model is used to describe the relationship between the students' high-order thinking ability and the multi-source feature data.

7. The method for longitudinal assessment and early warning of students' higher-order thinking ability according to claim 5 is characterized in that: The prediction model includes an input module, a feature selection module, a feature integration module, a residual module and an output module; the input module is used to input the longitudinal evaluation results and divide the longitudinal evaluation results into score data, mastery attribute data and other feature data; The feature selection module is used to use a dimension-based attention mechanism to assign different weights to the score data, the master attribute data and the remaining feature data, and generate a feature matrix; the feature integration module is used to use different convolutional layers to extract deep features from the feature matrix to form a deep feature matrix, and perform feature fusion and linear transformation to obtain fused features; The residual module is used to process the fusion features using a two-layer long short-term memory neural network, and then use residuals to connect them based on the time attention mechanism; the output module is used to process the output data of the residual module using two fully connected layers, and output the prediction results of students' high-order thinking ability at future time points.

8. A device for longitudinal assessment and early warning of students' higher-order thinking ability, characterized in that: include: An evaluation unit, used to improve the DINA model by using the joint-cross-load modeling method, to construct a longitudinal cognitive evaluation model suitable for multi-source and multi-frequency features, to process the multi-source feature data of the students according to the longitudinal cognitive evaluation model, and to obtain a longitudinal evaluation result of the students' higher-order thinking ability; The early warning unit is used to construct a prediction model suitable for multi-source and multi-frequency features based on a long short-term memory neural network, process the longitudinal evaluation results according to the prediction model, and obtain the prediction results of students' higher-order thinking ability at future time points.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for longitudinal assessment and early warning of students' higher-order thinking ability as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for longitudinal assessment and early warning of students' higher-order thinking ability as described in any one of claims 1 to 7 is implemented.

Images