Fusion platform intelligent teaching method and system based on virtual simulation

Abstract

The invention provides a fusion platform intelligent teaching method and system based on virtual simulation, and the method comprises the steps: obtaining an operation behavior of a student at each interaction point in a virtual simulation scene, and forming an IFM data set; constructing an evaluation function, and comparing the actual performance data with the ideal performance value; forming a key evidence source data set based on actual performance data in the IFM data set; and constructing an optimization model containing data conflict items and consistency items of each key evidence source, generating teaching guidance data, and quantitatively evaluating the learning effect of students. According to the invention, students can perform operation exercises in a simulated real clinical scene, such as key nursing skills of intravenous injection, external chest compression and the like, and through multi-dimensional data collection including operation accuracy, answer conditions, interactive records and the like, more comprehensive learning evaluation is provided, so that scoring deviation caused by single data is avoided, and the learning efficiency is improved. And a rich information basis is provided for ability evaluation of students.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual simulation teaching, and specifically to an intelligent teaching method and system for a fusion platform based on virtual simulation. Background Art

[0002] In the traditional nursing teaching mode, students mainly acquire knowledge through book learning and teacher lectures. Although this method can help students master the necessary theoretical knowledge, it has obvious deficiencies in providing real clinical experiences and operation opportunities. This limitation is particularly prominent in modern nursing education, especially in cultivating students' clinical thinking and practical operation skills. For example, intravenous injection and external chest compression are basic skills that nursing students must master. Intravenous injection requires students to be able to accurately find veins, disinfect standardly, puncture, and inject drugs, while external chest compression requires students to be able to perform cardiopulmonary resuscitation quickly and accurately in emergency situations. These skills not only require students to have solid theoretical knowledge but also require them to be able to flexibly apply this knowledge in actual operations. The problem in the existing practical operations is that when students learn key nursing skills such as intravenous injection and external chest compression, they often lack practical operation experience, which directly affects students' ability to effectively apply theoretical knowledge to the actual clinical environment.

[0003] With the development of virtual simulation technology, taking the "Nursing Virtual Simulation Comprehensive Training System" developed by the School of Nursing of Shanghai Jiao Tong University as an example, the whole process of actual nursing work is reproduced by using the latest technical means such as computers, the Internet, virtual reality technology, and intelligent teaching equipment. Through the virtual simulation teaching system, students can carry out role-playing, skill operation, and decision-making exercises in a virtual environment, thus significantly improving their clinical thinking ability and practical operation skills.

[0004] Taking intravenous injection as an example, the rapid development of the virtual simulation teaching system can simulate a real ward environment, provide realistic patient models and operation tools. Students can practice operation steps such as disinfection, puncture, and injection in a virtual environment, and can obtain relevant information at any time during the operation, such as video explanations and animation demonstrations. Similarly, for the cultivation of external chest compression skills, the virtual simulation teaching system can simulate the patient's state in an emergency situation, allowing students to practice operation steps such as positioning, compression, and relaxation in a virtual environment to help students master standard operation techniques. In addition, there are specific application scenarios such as: palliative care training, in-hospital case simulation training, out-of-hospital trauma first aid training, nursing comprehensive skills training and other systems.

[0005] Although virtual simulation technology has brought significant progress to nursing teaching, there are still some deficiencies and challenges:

[0006] 1. Monitoring data is too single: The current system mainly relies on students' operation data in virtual simulation scenarios (such as operation accuracy and completion time) for scoring, ignoring the comprehensive consideration of multi-dimensional data such as students' answering status and interaction records during the learning process. As a result, the scoring results cannot fully reflect the students' actual learning effects and are prone to scoring bias;

[0007] 2. The existing system's scoring mechanism is based on preset standards and lacks dynamic adjustment capabilities. The weight distribution for different operation steps is fixed and cannot be dynamically optimized according to students' performance in different links. This static scoring mechanism cannot adapt to the learning progress and characteristics of different students, resulting in inaccurate scoring results. For example, when students perform intravenous injections in nursing teaching, the system only scores based on the accuracy and completion time of the operation when performing virtual simulation operations, ignoring the students' understanding and application of aseptic operation principles during the operation. This inaccurate scoring mechanism causes students to mistakenly believe that they have mastered the operation skills, but in fact, in actual operations, problems will arise due to ignoring aseptic operation principles, resulting in poor learning results. Summary of the invention

[0008] In view of the shortcomings of the prior art, the present invention aims to provide an intelligent teaching method and system for a fusion platform based on virtual simulation. In order to achieve the above-mentioned purpose, the present invention is implemented through the following technical scheme: an intelligent teaching method for a fusion platform based on virtual simulation, comprising the steps of:

[0009] In the virtual simulation teaching process, the operation behaviors of students at each interaction point are obtained, and the feedback content matching the interaction point is extracted according to the preset feedback logic, including standard operation prompts and non-standard operation prompts, to form an IFM data set;

[0010] Assign initial weights to the actual performance data I(i) in the IFM dataset that characterizes the student's operation behavior Construct the evaluation function: Compare the actual performance data I(i) with the ideal performance value I max For comparison, the actual performance data I(i) at least includes the student's operation accuracy in the virtual simulation scene. a , Completion time t 、Answer score q r 、Interaction Records r and understanding level s data, μ is a constant term;

[0011] Based on the actual performance data I(i), a set of key evidence source data {OA, OT, QR, IR, QS} is formed. For each key evidence source data in this set, its basic probability assignment function p(·) is defined. The initial trust degree of the key evidence source data is dynamically assigned by comparing the actual performance of the student with the ideal performance to calculate the consistency difference between different key evidence source data sets;

[0012] Construct an optimization model that includes the conflict items and consistency items of each key evidence source data,

[0013] Calculate the dynamic weight ω of each key evidence source data i , and through the formula: Obtain the corrected weight Substitute it into the evaluation function to generate teaching guidance data. In the formula, is the sum of the products of the initial weights and dynamic weights of all key evidence source data, ensuring that the sum of the weights is 1 after normalization. u and v are the balance coefficients of the conflict items and consistency items, and D(A i ,B i ) is the conflict value of the key evidence source data.

[0014] As the second aspect of the present invention, an intelligent teaching system for a virtual simulation-based fusion platform is also proposed, including: a memory and a processor. Among them, the memory includes an intelligent teaching system program for a virtual simulation-based fusion platform. When the intelligent teaching system program for a virtual simulation-based fusion platform is executed by the processor, the intelligent teaching method described above is implemented.

[0015] Compared with the prior art, the beneficial effects of the present invention are:

[0016] 1. The reverse scoring mechanism based on the DS evidence theory quantifies the fusion data. The key data such as the operation accuracy, completion time, answering situation, interaction participation degree, and feedback content understanding degree of the student in the virtual simulation environment are used as evidence sources. By defining the basic probability assignment function, it is transformed into a probability form. Subsequently, an evaluation index level is constructed, and an initial trust degree is assigned to each evidence source. By measuring the proportion of the common and opposing parts between the information of different evidence source data, an optimization model is constructed to dynamically adjust the weights of the evaluation indexes, and a multi-dimensional and dynamic quantification evaluation of the student's learning performance is carried out. Compared with the traditional evaluation method, it effectively solves the problem of the unreliability of a single evidence source and improves the objectivity and accuracy of the evaluation results;

[0017] 2. In the actual teaching environment, the present invention calibrates the feedback content (system judgment model) to ensure that it matches the actual teaching objectives and student performance, provides a reliable basis for subsequent evaluation, and then combines the key knowledge point information K to obtain the core content of teaching; according to the virtual simulation scene teaching demand information S, the specific requirements of the teaching scene are determined, and the interactive behavior of students in the virtual simulation process is obtained around the interactive point information A to form an IFM data set. After building a dynamic interactive feedback system, it is integrated into evaluation data to achieve multi-dimensional association and fusion data of teaching data. At this point, based on this fusion data, teachers can comprehensively evaluate students' learning performance, locate students' problems in mastering key knowledge points, operating virtual simulation scenes, and interactive feedback, and improve the quality and effect of virtual simulation teaching;

[0018] 3. In a virtual simulation teaching environment, the fusion data composed of multi-dimensional data such as students' learning behavior, interactive performance, and understanding of feedback content often have uncertainty, fuzziness, and information conflict problems. Traditional evaluation methods are difficult to conduct comprehensive and objective quantitative analysis on them. In order to further ensure the evaluation and accuracy of students' abilities, the present invention, by constructing a knowledge graph, associates teaching content, key knowledge points, virtual simulation scenario teaching requirements, and IFM data set elements to form a structured knowledge system, ensuring that every operation, interactive behavior, and response to feedback content of students in a virtual simulation environment are recorded and mapped to the corresponding nodes and relationships in the knowledge graph;

[0019] 4. Based on the optimized weights and comprehensive trust calculated by evidence theory, a comprehensive evaluation score is generated for each student in terms of the degree of mastery of key knowledge points, completion of virtual simulation scenario teaching requirements, accuracy of response to interaction points, degree of understanding of feedback content, and degree of participation, providing students with targeted improvement suggestions. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The disclosure of the present invention is described with reference to the accompanying drawings. It should be understood that the drawings are only for illustrative purposes and are not intended to limit the scope of protection of the present invention. In the drawings, the same reference numerals are used to refer to the same components. Among them:

[0021] Figure 1 A schematic diagram of a flow chart of an intelligent teaching method for a fusion platform based on virtual simulation proposed in one embodiment of the present invention;

[0022] Figure 2 It is a schematic diagram of the process structure of quantifying teaching elements based on the constructed structured knowledge graph when generating feedback content information proposed in one embodiment of the present invention;

[0023] Figure 3Schematic diagram of the main control panel in the user interface module of the integrated platform intelligent teaching system proposed in an embodiment of the present invention;

[0024] Figure 4 Schematic diagram of the main control panel in the user interface module for implementing the evaluation of students' learning effects proposed in an embodiment of the present invention. Detailed implementation manners

[0025] It is easy to understand that according to the technical solution of the present invention, without changing the essence of the present invention, those of ordinary skill in the art can propose various structural ways and implementation manners that can be mutually replaced. Therefore, the following detailed implementation manners and the accompanying drawings are only exemplary descriptions of the technical solution of the present invention, and should not be regarded as all of the present invention or as a limitation or restriction on the technical solution of the present invention.

[0026] The present invention will be further described in detail below with reference to the accompanying drawings, but it is not a limitation to the present invention.

[0027] For ease of understanding, the technical concept of the present invention is as follows: By collecting multi-dimensional operation data of students in the virtual simulation scenario in real time (such as operation accuracy, answer correct rate, interaction frequency, etc.), combined with the dynamic weight adjustment and system model update mechanism, accurately identify the weak links of students and generate targeted teaching guidance data.

[0028] The process is as follows: First, construct an IFM data set that binds interaction points and feedback content based on the teaching syllabus, and set the initial weight to reflect the preset importance of each knowledge point; Secondly, quantify the students' performance through an evaluation function, use the conflict metric DS evidence theory to analyze the consistency of multi-source data, and perform dynamic adjustment of weights in the students' operations: when a student's performance on a certain knowledge point is low or the data conflict is high, reduce its weight to weaken the impact, otherwise retain or increase the weight; At the same time, calculate the recognition difference ΔQ of the system model between normal and non-normal operations. When ΔQ is lower than the preset threshold, update the model parameters through the gradient descent method to optimize the accuracy of the system feedback prompt; Finally, combined with the semantic recommendation of the knowledge graph, dynamically push the calibrated feedback content (such as demonstration videos and error correction exercises to strengthen weak links) to students, forming a closed-loop optimization of evaluation, adjustment, feedback, and re-evaluation to ensure that the teaching strategy always adapts to the actual operations of students.

[0029] As Figure 1 shown, as an embodiment of the present invention, a virtual simulation-based integrated platform intelligent teaching method is proposed, including the following specific steps:

[0030] S1. Form an IFM data set:

[0031] Traverse each scenario requirement s in the teaching requirement information S of the virtual simulation scenario. After the student completes the teaching operation behavior or training batch, extract all the associated operation steps to form interaction point information: A = U s∈S {a | a is an operation step in s};

[0032] Use logistic regression or neural network to construct a pre-trained system classification model to form a feedback logic. According to the preset feedback logic, obtain multi-dimensional feedback content including the standard operation prompt p and non-standard operation prompt n output by the system at each interaction point, and form a feedback content set F. It should be noted that the logistic regression model uses the student's operation behavior, action execution process and results at the interaction point as input features, calculates the standard probability of the operation through the logistic regression algorithm, so as to judge whether the operation meets the standard requirements, and extracts the corresponding standard or non-standard operation prompts from the database according to this, and feedbacks them to the student to guide their learning;

[0033] Dynamically bind the interaction points and the feedback content to form an IFM data set: IFM = {(a, f) | a ∈ A, f ∈ F}. Implement the real-time feedback logic through coding. For example, when the student performs intravenous puncture, the system immediately compares the operation trajectory / behavior and triggers the corresponding feedback.

[0034] S2. Based on the IFM data set, construct an evaluation function for teaching guidance:

[0035]

[0036] In the formula, is the initial weight set based on the teaching syllabus, representing the preset importance of each evidence source. μ is a constant term that controls the weight decrease and / or increase rate. I(i) is the actual performance data of the student in the virtual simulation scenario teaching, and I max is the ideal performance value.

[0037] S3. Collect the key evidence source data set:

[0038] Determine the evidence sources for evaluating the student's learning effect from the IFM data set. Based on the evidence sources, define evaluation indicators for quantifying the student's actual performance data respectively. For each evaluation indicator, collect the corresponding key evidence source data to form a key evidence source data set.

[0039] S4. Synthesize the key evidence source data and perform conflict measurement:

[0040] Construct an evaluation index level: α = {α proficient, α good, α mastered, α basic, α not mastered}. Assign an initial trust degree to each key evidence source data, and obtain the consistency difference between different key evidence source data sets by measuring the proportion of the common and / or opposing parts between different key evidence source data sets;

[0041] S5. Build an optimization model, and combine the actual performance data I of the students with the conflict measure to calculate the dynamic weight of each key evidence source data. After correcting the initial weight based on this dynamic weight, return to step S2, and finally generate teaching guidance data to specifically quantify the performance of students in various parts of the virtual simulation scenario teaching.

[0042] It can be understood that the traditional virtual simulation system has obvious defects in evaluating students' operations. Taking intravenous injection as an example, it only scores based on operation accuracy and completion time, ignoring the understanding, application of the aseptic operation principle, as well as the investigation of decision-making and problem-solving abilities, resulting in single data and being unable to comprehensively reflect the actual abilities of students. Moreover, the scoring mechanism of the existing system is static and the weight is fixed, making it difficult to adapt to the learning progress and characteristics of different students, and the scoring accuracy is poor. For example, when a student performs poorly in a high-weight knowledge point, the high weight of this knowledge point will overly affect the overall evaluation, making the scoring result unable to truly reflect the learning effect of the student. Therefore, the present invention proposes to build a more comprehensive virtual simulation teaching scenario. While the system records operation accuracy, it collects multi-dimensional data such as answering situation and interaction records to achieve a more comprehensive evaluation of learning effects. In addition, a dynamic weight adjustment mechanism is introduced to optimize the weight allocation according to the performance of students in different links, ensuring the balance and accuracy of the evaluation, thereby improving the evaluation quality and teaching effect of virtual simulation teaching.

[0043] In an embodiment of the present invention, step S1 includes:

[0044] S1-1. Traverse all knowledge points in the teaching syllabus T, and evaluate the relevance between each chapter and section directory t in the teaching syllabus T and the target course information G that students need to master by constructing a teaching relevance evaluation function R (t,g) , R (t,g) =a×CR (t,g) +b×SR (t,g) +c×AR (t,g) , and extract the key knowledge point information K that is closely related to the target course information G to provide a knowledge basis for subsequent virtual simulation teaching, where In the formula, CR (t,g) is the teaching content relevance score, used to measure the matching degree between the content of the chapter and section directory t and the target course g, SR (t,g) is the teaching skill relevance score, used to measure the matching degree between the skills involved in the chapter and section directory t and the skills required for the target course g, AR (t,g)It is for the relevance score of teaching applications, used to measure the actual application degree of the chapter directory t mastered by students in achieving the target course g. a, b, and c are respectively the relevance evaluation weight coefficients, used to adjust the importance of different teaching relevance factors in the total teaching relevance evaluation. θ is a preset threshold, used to determine the relevance strength between the chapter directory t and the target course g.

[0045] S1-2. Based on the extracted key knowledge point information K, construct virtual simulation scenario teaching requirement information S for each key knowledge point k, which at least includes the information O of the teaching-related operation requirements to be determined and the information G of the target course. S = {(k, o, g) | k ∈ K}, where o is the operation requirement related to the key knowledge point k, and g is the target course related to the key knowledge point k. It can be understood that by determining the virtual simulation scenario teaching requirement information S for each key knowledge point k, each key knowledge point that students need to learn can obtain effective teaching guidance in virtual simulation teaching.

[0046] S1-3. Traverse each scenario requirement s in the virtual simulation scenario teaching requirement information S, and determine all the operation steps a in the scenario requirement s to form interaction point information A. A = U s∈S {a | a is the operation step in s}. It can be understood that the interaction point is the specific operation step or node for students to interact with the system, and it is all the operation steps a included in each scenario requirement s in the virtual simulation scenario teaching requirement information S. These operation steps a are the key actions or decision points that students need to execute in the virtual simulation environment. The system evaluates students' mastery of key knowledge points by monitoring students' operation behaviors at these interaction points, such as the accuracy of operations and the completion time.

[0047] At the same time, based on the pre-trained model, determine and obtain the corresponding feedback content set F for each interaction point. F = {(p, n) | p is the standard operation prompt, n is the non-standard operation prompt} to enable the system to provide specific feedback when students execute operation steps, thereby enhancing the learning effect.

[0048] S1-4. Dynamically bind each interaction point a to its corresponding feedback content f to form an IFM data set. IFM = {(a, f) | a ∈ A, f ∈ F}. Implement real-time feedback logic through coding. When students perform intravenous puncture, the system immediately compares their operation trajectory with the standard path and triggers the corresponding feedback. When constructing the IFM data set, the main purposes of extracting the feedback content are as follows: Real-time feedback: By binding the interaction point to the feedback content to form an IFM data set, when students execute operations, the system immediately compares students' operations with the standard path / behavior according to the mapping relationship in the data set, triggers and provides the corresponding feedback content. At the same time, for students with non-standard operations, the system provides targeted non-standard operation prompts to help correct errors.

[0049] In one embodiment of the present invention, in S2, is the initial weight set based on the teaching syllabus, and the initial weights in the evaluation function and the actual performance data I(i) at least include:

[0050] w ki is the weight of the i-th key knowledge point, indicating the importance of this key knowledge point in teaching, and I(ki) is the performance data of the student on the i-th key knowledge point;

[0051] w sj is the weight of the teaching requirement of the j-th virtual simulation scenario, indicating the importance of this scenario requirement in teaching, and I(sj) is the performance data of the student on the teaching requirement of the j-th virtual simulation scenario;

[0052] w ak is the weight of the k-th interaction point, indicating the importance of this interaction point in teaching, and I(ak) is the performance data of the student on the k-th interaction point;

[0053] w f'l is the weight of the l-th calibrated feedback content, indicating the importance of this feedback content in teaching, and I(f'l) is the performance data of the student on the l-th calibrated feedback content.

[0054] In one embodiment of the present invention, the specific steps of S3 proposed include:

[0055] S3-1. Determine the evidence sources for evaluating the learning effect of students from the IFM data set: It can be understood that the evidence sources at least include the operation data of students in the virtual simulation scenario (such as operation accuracy, completion time), answering situation, and interaction records. The specific operation process is as follows:

[0056] S3-11. Define the evaluation indicators for quantifying the performance of students in the virtual simulation teaching system, including, during the virtual simulation teaching process, when students perform operations at each interaction point: the key knowledge point mastery degree E ki related to the performance of students on key knowledge points, which is matched with the operation skill mastery degree EOA in the virtual simulation teaching scenario; the virtual simulation scenario teaching requirement completion degree E sj related to the performance of students on the teaching requirements of the virtual simulation scenario, which is matched with the operation skill completion degree EOT in the virtual simulation teaching scenario; the interaction point response accuracy E ak related to the performance of students at the interaction point, which is matched with the interaction participation degree EIR in the virtual simulation teaching scenario; the calibrated feedback content understanding degree E f'l, which matches the knowledge mastery correct rate EQR and the participation degree E in the feedback content in the virtual simulation teaching scenario ifm .

[0057] S3-12, according to the evaluation index, collect the following key evidence source data:

[0058] Operation data: including the operation accuracy o of students in the virtual simulation scenario a (such as in intravenous injection, record the needle insertion angle and depth), completion time o t (such as in intravenous injection, record the total time from when the student starts disinfection to the completion of intravenous injection); answering situation q r : including the scoring situation of students in the relevant feedback content (such as in intravenous injection, the score of students in the drug knowledge quiz), and interaction record i r : including the interaction participation degree of students in the virtual simulation scenario (such as in intravenous injection practice, the interaction frequency between students and simulated patients) and the understanding degree q of students on the feedback content s (such as in intravenous injection practice, the understanding and application degree of students on the operation feedback).

[0059] In an embodiment of the present invention, the specific steps of S4 proposed include:

[0060] S4-1. Define the evaluation index levels: α = {α proficient, α good, α mastered, α basic, α not mastered}, to evaluate the skill mastery degree of students in performing teaching operations (such as intravenous injection operations) in the virtual simulation teaching system. For each key evidence source data collected, assign an initial trust degree, so as to obtain the learning effect of students on each evaluation index by comprehensively considering multiple evidence sources, that is,

[0061] For the operation accuracy set OA, OA = {o a1 , o a2 ,..., o an}, completion time set OT, OT = {o t1 , o t2 ,..., o tn}, interaction record set IR, IR = {i r1 , i r2 ,..., i rn}, answering score set QR, QR = {q r1 , q r2 ,..., q rn} and understanding degree set QS, QS = {q s1 , q s2 ,..., q sn}, respectively define the basic probability assignment function p OA , pOT , p IR , p QR , p QS , where \(i\in\{n, o\}\) ai , o ti , i ri , q ri , q si are the accuracy of the \(i\)-th operation step, the operation completion time, the \(i\)-th interaction record, the correct rate of the \(i\)-th question, and the answering speed of the \(i\)-th question, respectively; where

[0062] where \(\alpha\) is the evaluation index level, \(w\) OAi is the weight of the \(i\)-th operation step / behavior, \(P\) rOAi (\(\alpha\)) is the probability determined by comparing the performance of the student in the \(i\)-th operation step / behavior with the ideal performance; where \(w\) OTj is the weight of the \(j\)-th operation step / behavior, \(P\) rOTj (\(\alpha\)) is the probability determined by comparing the time spent by the student in the \(i\)-th operation step / behavior with the expected time; where \(w\) IRk is the weight of the \(k\)-th interaction point, \(P\) rIRk (\(\alpha\)) is the probability determined by comparing the performance of the student at the \(k\)-th interaction point with the ideal performance; where \(w\) QRl is the weight of the correct rate of the \(l\)-th question, \(P\) rQRl (\(\alpha\)) is the probability determined by comparing the correct rate of the student answering the \(l\)-th question with the ideal performance; where \(w\) QSz is the weight of the answering speed of the \(z\)-th question, \(P\) rQSz (\(\alpha\)) is the probability determined by comparing the time spent by the student answering the \(z\)-th question with the expected time.

[0063] S4-2. Obtain the consistency difference between different key evidence source data sets. The specific process is as follows:

[0064] First, based on the obtained basic probability assignment functions \(p\) OA , \(p\) OT , \(p\) IR , \(p\) QR , \(p\) QS , merge them to resolve the conflicts in different key evidence source data sets and generate the global confidence:

[0065]

[0066] where \(P\) i (A) and \(P\)j (B) represents the basic probability assignment of two independent key evidence source data i and j to two key evidence source data sets A and B, indicating the confidence of this key evidence source data in this key evidence source data set. The value range is between 0 and 1. A∪B = C means that the intersection of key evidence source data sets A and B is key evidence source data set C, that is, the part that belongs to both A and B constitutes C. ∑ A∪B＝C P i (A)·P j (B) represents calculating the sum of the products of their basic probabilities for all combinations where the intersection of A and B is C, that is, combining the support degrees of two key evidence source data for key evidence source data set C. means that the intersection of key evidence source data sets A and B is an empty set, that is, the two key evidence source data sets are completely non - intersecting and have no common part. means calculating the sum of the products of their basic probabilities for all combinations where the intersection of A and B is an empty set, that is, quantifying the conflict degree between two key evidence source data. Among them, the confidence after synthesis is biased towards the key evidence source data with less conflict.

[0067] Secondly, through the following formula, calculate the consistency difference between different key evidence source data sets for dynamically correcting the initial weight: The higher the conflict value D(A,B), the greater the contradiction between key evidence source data sets, and the weights of relevant key evidence source data need to be reduced.

[0068] Based on this, it can be understood that by calculating the differences between each key evidence source data, the consistency or conflict of this key evidence source data in evaluating students' learning effects can be determined. In specific implementation, if the value of the difference D(A,B) is small, it indicates that the evaluation results of different key evidence source data for students' learning effects are relatively consistent, indicating that the performance of students in each operation step, interaction point, and question answer is highly consistent with the ideal performance. On the contrary, if the value of D(A,B) is large, it indicates that there is inconsistency or conflict between different key evidence source data, indicating that there are large differences in students' performance in different aspects, or the evaluation criteria are applied inconsistently between different evidence sources, and the relevant weights need to be dynamically reduced. Therefore, through the above method, the system can identify which key evidence source data provides similar evaluation results and which need further investigation and analysis, so as to provide a basis for students' operation feedback.

[0069] In an embodiment of the present invention, the specific steps of S5 include:

[0070] S5 - 1. Construct an optimization model, and combine the actual performance data I(i) of students with the conflict metric to calculate the dynamic weight of each key evidence source data. The specific process is as follows:

[0071] First, define the optimization objective and constraints respectively:

[0072] The optimization objective is to minimize the weighted sum of squares of data conflicts among key evidence sources and maximize the consistency among data sets of different key evidence sources: And

[0073] Constraints:

[0074] Secondly, construct the revised optimization model:

[0075] Taking the conflict value D(A i , B i ) of key evidence source data i to be minimized and the consistency to be maximized, construct the optimization model:

[0076]

[0077] In the formula, ω i is the dynamic weight of key evidence source data i. The smaller the value of D(A i , B i ), the lower the conflict. The larger the value of 1 - D(A i , B i ), the higher the consistency. u and v are the balance coefficients of conflict and consistency. Among them, the weight is reduced when the conflict term dominates, and the weight is preserved when the consistency term dominates;

[0078] Introduce the Lagrange multiplier λ with normalization constraints and construct the Lagrangian function:

[0079]

[0080] Take the partial derivative of ω i and solve to obtain the weight calculation formula:

[0081]

[0082] Substitute the normalization constraints and solve for λ by the Newton - Raphson iterative method:

[0083]

[0084] After arrangement:

[0085]

[0086] The solution is obtained:

[0087]

[0088] In the formula, the denominator

[0089] The λ obtained by using the iterative method is used to calculate the weight ω of each key evidence source data i through the weight calculation formula i , and non-negativity verification is performed: if v(1 - D(A i , B i )) + λ ≥ 0, then ω i ≥ 0. The u and v balance coefficients are adjusted inversely. When the value of D(A i , B i ) is larger, the conflict is higher. When the value of 2uD(A i , B i ) is larger and ω i is smaller, the goal of reducing the weight for high conflict or low performance is achieved to correct the initial weight.

[0090] Based on the above technical concept, it can be understood that since there is a corresponding relationship between the weights calculated in the DS evidence theory and the various performance parts of the teaching guidance data, therefore, by directly mapping the weights calculated in the DS evidence theory to the various performance parts of the teaching guidance data and correcting the initial weights, the teaching guidance data can be obtained, so that the evaluation results of students are more objective and accurate, ensuring that the weights of each evaluation dimension are reasonably allocated when evaluating the learning effectiveness of students.

[0091] Based on the above technical concept, the specific process of correcting the initial weight to obtain the teaching guidance data is as follows:

[0092] First, the dynamic weight ω of the key evidence source data i obtained is used to i adjust the initial weight set based on the teaching syllabus to obtain the corrected weight

[0093]

[0094] In the formula, represents the importance of the key evidence source data before adjustment, is the sum of the products of the initial weights and dynamic weights of all key evidence source data, ensuring that the sum of the weights is 1 after normalization;

[0095] Secondly, the corrected weight is substituted into the evaluation function to obtain the teaching guidance data. According to the evaluation results and the requirements of the target course, the learning effectiveness of students is evaluated to improve the learning efficiency and teaching quality.

[0096] In an embodiment of the present invention, in the specific process of constructing the IFM data set, when performing teaching relevance evaluation: it is preferable to use the Python language in combination with the Pandas data processing library and the Scikit-learn machine learning library to construct the teaching relevance evaluation function R(t,g) , by collecting data of syllabus T and target course information G, the initial values of CR (t,g) , SR (t,g) , AR (t,g) are determined by using content analysis method and expert scoring method, and the Apriori algorithm is used to optimize the weight coefficients a, b, c, and the correlation score between each chapter and section t and the target course g is calculated. According to the preset threshold θ, the key knowledge point information K is screened out. When constructing the teaching requirement information of the virtual simulation scenario, preferably based on the Unity3D virtual simulation development platform, the teaching requirement information S of the virtual simulation scenario is constructed for each key knowledge point k. Among them, a scene is created in Unity3D, and the operation requirement information O and the target course information G related to the key knowledge point k are defined using C# scripts, forming a data structure of S = {(k, o, g)|k ∈ K}. At the same time, when extracting the interactive point information and determining the feedback content, each scene requirement s is traversed in Unity3D, and all operation steps a of the students in the scene are captured through the event trigger and operation detector in the scene, forming the interactive point information A, and a standard operation prompt p and a non-standard operation prompt n are designed for each interactive point in the UI system of Unity3D to construct the feedback content set F. Finally, the interactive point a and the feedback content f are dynamically associated through data binding technology in Unity3D to form the IFM data set, and the real-time feedback logic is written using C# scripts to achieve instant feedback.

[0097] In an embodiment of the present invention, to ensure the effective implementation of the IFM data set and achieve the purpose of improving the teaching effect and the learning experience of students, after each student completes a teaching operation behavior or a training batch, it is also necessary to improve the model's recognition ability of the students' standard and non-standard operations to ensure that the feedback content set F matches the actual performance data of the students. The specific process is as follows:

[0098] The first step is to define the standard operation prompt information in the set based on the obtained feedback content set F, denoted as {(D, P valid , C target )} and the non-standard operation prompt information, denoted as {(D, P invalid , C error )}, where D is the teaching element of various information and data related to the teaching content. For example, in intravenous injection training, it includes the description of key teaching elements such as "syringe", "vein", and "disinfectant", or in chest compressions training, it includes the description of key teaching content or operation objects such as "chest compressions", "compression frequency", and "compression depth". P validIt is a standardized operation step prompt message, which provides standardized operation guidance for students based on the teaching syllabus T and the target course g. For example, in intravenous injection training, the standardized operation step is "disinfect the intravenous area first, and then insert the syringe", or in chest compressions training, the standardized operation step is "ensure that the compression frequency is 100 - 120 times per minute, and the depth is to depress the sternum by 5 - 6 cm", C target For the specific operation objectives related to the teaching element description information D and the standardized operation step prompt information P valid For example, in intravenous injection training, the operation objective is "ensure that the drug is injected into the vein in a standardized manner", or in chest compressions training, the operation objective is "ensure effective cardiac pumping". P invalid It is non-standardized operation step relationship information to help students identify and avoid non-standardized operations in virtual simulation operations. It includes non-standardized operation prompt guidance pointed out to students based on the teaching syllabus T and the target course g. For example, in intravenous injection training, the non-standardized operation step is "insert the syringe directly without disinfection", or in chest compressions training, the non-standardized operation steps are "too fast or too slow compression frequency, insufficient or excessive compression depth", C error For the operation objectives that are contradictory to the standardized operation objective C target It represents the subsequent reactions caused by non-standardized operation prompts. For example, in intravenous injection training, "injection failure, drug spillage", or in chest compressions training, "insufficient cardiac pumping, further medical intervention is required".

[0099] Second step, set the model update trigger condition:

[0100] By comparing the difference in recognition accuracy ΔQ between the system output of standardized operation prompts and non-standardized operation prompts after the student performs teaching operations, it is judged whether the model needs to be updated. ΔQ = QS(C target ∣D,P valid ) - FE target (C target ∣D,P invalid ). When ΔQ is lower than the preset threshold θ, it indicates that the current model has insufficient recognition of students' standardized and non-standardized operations, and the parameters need to be optimized to improve the recognition accuracy.

[0101] It should be noted that QS(C target ∣D,P valid ) is the recognition accuracy under standardized operations, In the formula, f j (D,P valid ) is the j-th feature extracted from the teaching element D and the standardized operation prompt P valid , such as operation accuracy, time deviation, is the weight of the j-th item in the standardized operation prompt P valid , obtained by training based on historical data, is the Sigmoid function, which is used to map the linear combination to a probability value. In specific implementation, it is preferred to use a pre-trained classification model (such as logistic regression, neural network), input the student operation data (such as disinfection range coverage rate, operation time), and output the target achievement probability, FE target (C target ∣D,P invalid ) is the target error recognition accuracy rate under non-standard operations, which is used to calculate the occurrence probability of each possible error result C error ,

[0102]

[0103] In the formula, reflects the severity of the k-th type of error, is the feature weight corresponding to the k-th type of error, which is obtained by training based on historical data, f j (D,P invalid ) is the j-th feature extracted from the teaching element D and the non-standard operation prompt P invalid , such as the wrong disinfection wiping direction.

[0104] In the third step, the gradient descent method is used to adjust the model parameter Φnew to minimize the difference loss between the standard operation recognition accuracy rate and the non-standard operation recognition accuracy rate:

[0105]

[0106] In the formula, G j is the recognition accuracy rate of the j-th item under standard operations, H j is the recognition accuracy rate of the j-th item under non-standard operations, α j is the confidence weight of different key evidence source data items.

[0107] In the fourth step, update the model: In the formula, η is the learning rate, is the gradient of the loss function with respect to the model parameter Φnew.

[0108] In the fifth step, based on the updated model, regenerate the calibrated feedback content set F′, F'={(p',n')∣

[0109] p'=Model new (D,P valid ),n'=Model new (D,P invalid )}.

[0110] It should be noted that in this process, a Python programming environment is preferably adopted, and the updated model is loaded using the TensorFlow or PyTorch deep learning framework. At the same time, the Pandas data processing library is integrated to operate the reading and preprocessing of data, ensuring that the student operation data can be input in the format required by the model.

[0111] Then, read the student's standard operation data D_Pvalid and non-standard operation data D_Pinvalid from the database or data file. (These data should include various operation records of the student in the virtual simulation scenario, such as the disinfection range coverage rate and operation time, and need to be preprocessed according to the model input requirements, such as normalization and encoding operations).

[0112] Input the preprocessed standard operation data D_Pvalid into the updated model, call the prediction function of the model, and obtain the standard operation feedback content p′ = Model new (D, Pvalid), and generate positive feedback that meets the teaching objectives. Input the preprocessed non-standard operation data D_Pinvalid into the updated model, and also call the prediction function of the model to obtain the non-standard operation feedback content n′ = Model new (D, Pinvalid). At this time, the model generates targeted corrective feedback based on the learned non-standard operation characteristics and impacts to help students recognize and correct incorrect operations. Integrate the generated standard operation feedback p′ and non-standard operation feedback n′ into the calibrated feedback content set F′ = {p′, n′}, and store it in a structured manner or directly display it to students in the virtual simulation teaching system so that students can obtain operation feedback in a timely and accurate manner and improve the learning effect.

[0113] To facilitate the understanding of the model update concept proposed by the present invention, the following data example is given:

[0114] First, after the student completes the teaching operation or training batch, obtain the standard and non-standard operation prompt information in the feedback content set F. The data is shown in Table 1:

[0115] Set Name Description Collection D of Teaching Element Descriptions Key teaching contents or operation objects such as "syringe", "vein", "disinfectant solution" <![CDATA[Collection P of correct operation step tips valid > For example, "Disinfect the vein area first, then insert the syringe" <![CDATA[Error operation step prompt set P invalid > For example, "Insert the syringe directly without disinfection" <![CDATA[Correct operation target set C target > For example, "Ensure that the drug is correctly injected into the vein" <![CDATA[Error operation target set C error > For example, "Injection failure, drug spillage"

[0116] Table 1

[0117] Secondly, calculate the recognition accuracy rate, calculate the recognition accuracy rate under the standard and non-standard operation prompts. For the convenience of calculation, take QS(C target |D, P valid ) = 0.90, FE target (C target |D, P invalid ) = 0.50;

[0118] Next, the updated model parameters are compared by defining the difference metric ΔQ. ΔQ = 0.90 - 0.50 = 0.40, which is less than the preset model update threshold of 0.5. Then, the system training model parameters are updated. For ease of calculation, when the system training model is updated, there are two operation steps, namely, α j = 0.6 or 0.4, to standardize the operation recognition accuracy G j (D, P valid ) = 0.8 or 0.7, and the non-standard operation recognition accuracy H j (D, P invalid ) = 0.2 or 0.3.

[0119] Calculate the loss function L(Φ) = 0.6×(0.8 - 0.2) 2 + 0.4×(0.7 - 0.3) 2 = 0.28,

[0120] Calculate the parameter update amount: For each item J, the gradient is The total gradient is: At this time, the learning rate η is set to 0.5, Obtain the model parameter Φ new = Φ old + 0.52. Thus, the model parameters can be updated according to the performance of the model under standard and non-standard operation prompts.

[0121] In an embodiment of the present invention, after the model is updated, by constructing a knowledge graph structure, the generation of feedback content information is realized. The specific process is as follows:

[0122] First, based on the knowledge points, operation steps, and feedback content extracted from the teaching syllabus, a teaching element set D T = {t1, t2,..., ti} is formed, and the logical relationship R = {r1, r2,..., rj} between the teaching element sets is defined. In the formula, each rj represents a relationship class;

[0123] Secondly, it is defined that each teaching element ti is represented as a node vi, and each relationship rj is represented as an edge eij connecting two nodes, obtaining the information of the nodes and edges as V and E respectively: V = {v1, v2,..., vn} and E = {eij∣eij∈R, vi, vj∈V}, so as to map the teaching elements and relationships into a graph structure to form a semantic recommendation;

[0124] Thirdly, the knowledge graph structure is dynamically adjusted according to the student operation data to improve the recommendation accuracy;

[0125] Next, use the NLP model to parse the student operation logs and feedback content, extract the keyword U representing semantic information, and map the keyword to the knowledge graph structure node to generate a candidate resource set: Candidate resource = v j |sim(u, v j )>0.7, u∈U}, where sim is the similarity calculation formula, and v j is the knowledge graph node;

[0126] Finally, when ΔQ is lower than the preset threshold θ, update the model parameter Φ new , generate calibration feedback, and insert it into the knowledge graph node v j , adjust the associated edge weight, and based on the updated knowledge graph and weight, push the resources related to the calibration feedback as the feedback content set F'.

[0127] Example: Table 2 is the structured knowledge graph data table shown:

[0128]

[0129] Table 2

[0130] Using natural language processing technology, identify and extract keywords and phrases. At the same time, in the graph query stage, match the extracted keywords with the nodes in the knowledge graph to identify the teaching elements directly related to the student's operations. Finally, in the system recommendation stage, recommend the most relevant learning resources for the student based on the search results, such as video tutorials, operation demonstrations, or additional reading materials, to help the student effectively master the required knowledge.

[0131] Such as Figure 3 As the second aspect of the present invention, a virtual simulation-based integrated platform intelligent teaching system is proposed, including a memory and a processor. Among them, the memory includes a virtual simulation-based integrated platform intelligent teaching system program. When the virtual simulation-based integrated platform intelligent teaching system program is executed by the processor, it implements the intelligent teaching method in the above embodiments.

[0132] Specifically, when implemented, the virtual simulation-based integrated platform intelligent teaching system program includes:

[0133] A virtual simulation module for creating a highly realistic learning environment so that students can perform operation exercises in a simulated real clinical scenario;

[0134] A graph retrieval module for constructing a knowledge graph and designing an algorithm, and providing learning resource recommendations according to the student's learning progress;

[0135] An interaction and feedback module for providing instant feedback and guidance to students by constructing interaction point information and feedback content information;

[0136] The data collection and analysis module is responsible for collecting multi-dimensional data such as operation data, answering situation, and interaction records, and providing support for the evaluation and scoring module;

[0137] The evaluation and scoring module analyzes multi-source data based on the DS evidence theory, collaborates with the data collection and analysis module, calculates the comprehensive score of students and generates feedback;

[0138] The user interface module provides a friendly operation interface for users to ensure the usability and manageability of the system.

[0139] The technical scope of the present invention is not limited to the content described above. Those skilled in the art can make various deformations and modifications to the above embodiments without departing from the technical idea of the present invention, and these deformations and modifications shall fall within the protection scope of the present invention.

Claims

1. An intelligent teaching method for a fusion platform based on virtual simulation, characterized in that: Including the steps: During the virtual simulation teaching process, obtain the operation behaviors that occur when students perform operations at each interaction point, and extract the feedback content matching the interaction point from the preset feedback logic, including standard operation prompts and non-standard operation prompts, to form an IFM data set; Configure an initial weight for the actual performance data I(i) that represents the student's operation behavior in the IFM data set Construct an evaluation function: Compare the actual performance data I(i) with the ideal performance value I max The actual performance data I(i) at least includes the operation accuracy o of the student in the virtual simulation scenario a , completion time o t , answering score q r , interaction record i r and understanding level q s Data, μ is a constant term; Based on the actual performance data I(i), form a key evidence source data set {OA, OT, QR, IR, QS}, define the basic probability assignment function p(·) for each key evidence source data in this set, and dynamically allocate the initial trust degree of the key evidence source data by comparing the difference probability determined by the student's actual performance and the ideal performance, so as to calculate the consistency difference between different key evidence source data sets; Construct an optimization model that includes the conflict items and consistency items of each key evidence source data; Calculate the dynamic weight ω of each key evidence source data i , and through the formula: Obtain the corrected weight Substitute it into the evaluation function to generate teaching guidance data. In the formula, is the sum of the products of the initial weights and dynamic weights of all key evidence source data, ensuring that the sum of the weights is 1 after normalization. u and v are the balance coefficients of the conflict term and the consistency term. D(A i , B i ) is the conflict value of the key evidence source data.

2. The intelligent teaching method of the fusion platform based on virtual simulation according to claim 1, characterized in that: Before constructing the optimization model, First, it is necessary to define the optimization objective and constraints to minimize the weighted sum of squares of the conflicts of each key evidence source data and maximize the consistency between different key evidence source data sets. Among them, The defined optimization objective is: and The defined constraints are as follows: Then, taking the minimum of the conflict value D(A i ,B i ) of the key evidence source data i and maximizing the consistency, an optimization model is constructed.

3. The intelligent teaching method of the fusion platform based on virtual simulation according to claim 1, characterized in that: The specific process of allocating the initial trust degree for each of the key evidence source data is as follows: First, define the evaluation indicators for quantifying the student's actual performance data I(i), and construct the evaluation indicator levels: α = {α proficient, α good, α mastered, α basic, α not mastered}; Secondly, for the operation accuracy set OA, the completion time set OT, the interaction record set IR, the answering score set QR, and the understanding degree set QS, the basic probability assignment functions p OA , p OT , p IR , p QR , p QS , Among them, In the formula, α is the evaluation index level, w OAi is the weight of the i-th operation step, P rOAi (α) is the difference probability determined by comparing the actual performance of the student in the i-th operation step with the ideal performance.

4. The intelligent teaching method for a virtual simulation-based fusion platform according to claim 1 or 3, characterized in that: When constructing the optimization model, the specific process of obtaining the consistency difference between different key evidence source data sets is as follows: First, based on the obtained basic probability assignment functions p OA , p OT , p IR , p QR , p QS , they are merged to resolve the conflicts in the data sets of different key evidence sources and generate the global confidence: where P i (A) and P j (B) are the basic probability assignments of two independent key evidence source data i and j to two key evidence source data sets A and B, respectively; Secondly, the consistency difference between different key evidence source data sets is calculated through the following formula: The higher the conflict value D(A,B), the greater the contradiction between the key evidence source data sets, and the weights of the relevant key evidence source data need to be reduced.

5. The intelligent teaching method of the fusion platform based on virtual simulation according to claim 2, characterized in that: After constructing the optimization model, By introducing the Lagrange multiplier λ with a normalization constraint, the Lagrangian function L is constructed, and the dynamic weight ω is calculated i , and the specific process is as follows: Construct the Lagrangian function L: For ω i Take the partial derivative and solve to obtain the weight calculation formula: Substitute the normalization constraint and solve for λ by the Newton-Raphson iteration method; The λ obtained by using the iterative method is used to calculate the weight ω of each key evidence source data i through the weight calculation formula i , and non-negativity verification is performed: if v(1 - D(A i , B i )) + λ ≥ 0, then ω i ≥ 0. The u and v balance coefficients are adjusted inversely. When the value of D(A i , B i ) is larger, the conflict is higher. When the value of 2uD(A i , B i ) is larger and ω i is smaller, the goal of reducing the weight for high conflict or low performance is achieved to correct the initial weight.

6. The intelligent teaching method for a virtual simulation-based fusion platform according to claim 1, characterized in that: In the virtual simulation teaching system, the preset feedback logic is generated by a logistic regression model trained based on historical data, and after each student completes a teaching operation behavior or a training batch, it is also necessary to judge the recognition ability of the pre-trained model for the student's standard and non-standard operations to ensure that the feedback content matches the student's actual performance data. The specific process is as follows: First, set the model update trigger condition: By comparing the recognition accuracy difference ΔQ between the standard operation prompt and the non-standard operation prompt output by the model after the student has a teaching operation behavior, judge whether the model needs to be updated. When ΔQ is lower than the preset threshold θ, it indicates that the current model has insufficient recognition of the student's standard and non-standard operations, and the parameters need to be optimized to improve the recognition accuracy; Secondly, the gradient descent method is used to adjust the model parameters Φ new , so as to minimize the difference loss between the recognition accuracy of the canonical operation and the recognition accuracy of the non-canonical operation: where G j is the recognition accuracy rate of the j-th item under standard operations, and H j is the recognition accuracy rate of the j-th item under non-standard operations, and α j is the confidence weight of different key evidence source data items. Again, update the model: where η is the learning rate, is the gradient of the loss function with respect to the model parameters Φ new ; Finally, regenerate the calibrated feedback content based on the updated model.

7. The intelligent teaching method for a virtual simulation-based fusion platform according to claim 6, characterized in that: After the model is updated, generate the feedback content information by constructing a knowledge graph structure. The specific process is as follows: First, form a teaching element set DT = {t1, t2,..., ti} based on the teaching syllabus of virtual simulation teaching, and define the logical relationship R = {r1, r2,..., rj} between the teaching element sets. In the formula, each rj represents a relationship class; Secondly, define that each teaching element ti is represented as a node vi, and each relationship rj is represented as an edge eij connecting two nodes, obtaining the information of nodes and edges as V and E respectively: V = {v1, v2,..., vn} and E = {eij | eij ∈ R, vi, vj ∈ V}, so as to map teaching elements and relationships into a graph structure and form a semantic recommendation; Thirdly, dynamically adjust the knowledge graph structure according to students' operation data to improve the recommendation accuracy; Next, use the NLP model to parse the student operation logs and feedback content, extract the keyword U that represents semantic information, and map the keyword to the knowledge graph structure node to generate a candidate resource set: Candidate resource = v j |sim(u, v j )>0.7, u ∈ U}, where sim is the similarity calculation formula, v j is the knowledge graph node; Finally, when ΔQ is lower than the preset threshold θ, update the model parameters Φ new , generate calibration feedback, and insert the knowledge graph node v j , adjust the weights of the associated edges, and based on the updated knowledge graph and weights, push the resources related to the calibration feedback as feedback content.

8. The intelligent teaching method based on a virtual simulation integration platform according to claim 1, wherein: During the virtual simulation teaching process, each of the interaction points is a specific operation step or node for students to interact with the system, determined based on each scenario requirement s in the virtual simulation scenario teaching requirement information S, where The virtual simulation scenario teaching requirement information S is generated by constructing a teaching relevance evaluation function R (t,g) , and filtering out key knowledge points k related to the teaching objective course G to be mastered from the teaching syllabus T of virtual simulation teaching, where The teaching relevance evaluation function R (t,g) is: R (t,g) = a × CR (t,g) + b × SR (t,g) + c × AR (t,g) , where CR (t,g) is the teaching content relevance score, SR (t,g) is the teaching skill relevance score, AR (t,g) is the teaching application relevance score, and a, b, and c are the relevance evaluation weight coefficients respectively.

9. The intelligent teaching method based on a virtual simulation integration platform according to claim 2, wherein: The evaluation indicators include the degree of mastery of key knowledge points E when students operate at each interaction point during the virtual simulation teaching process ki , the completion degree E of the teaching requirements of the virtual simulation scenario sj , the response accuracy E of the interaction point ak , the degree of understanding E of the feedback content f'l , and the participation degree E ifm .

10. An intelligent teaching system for a fusion platform based on virtual simulation, characterized in that: It includes: A memory and a processor, wherein the memory includes an intelligent teaching system program for a virtual simulation integration platform, and when the intelligent teaching system program for a virtual simulation integration platform is executed by the processor, it implements the intelligent teaching method according to any one of claims 1 to 9.

Images