Intelligent teaching platform management method and system combined with virtual reality

Abstract

The invention relates to the technical field of education management, in particular to an intelligent teaching platform management method and system combined with virtual reality. An intelligent teaching platform management method combined with virtual reality comprises the following steps: S1, acquiring learning related data of an employee, and determining a first objective evaluation score and a second objective evaluation score which need to be calculated according to the learning related data of the employee; s2, obtaining a first objective evaluation score of the employee by using a double-engine subjective question intelligent review model according to the learning related data of the employee; and S3, obtaining a second objective evaluation score of the employee by using a two-dimensional integral quantization formula according to the learning related data of the employee. According to the invention, the real-time adjustment of the number and proportion of the guide type and cooperation type virtual characters in the virtual reality teaching environment is realized by analyzing the uncoiling and closing points of the employees and adopting a character number adjustment formula.

Description

Technical Field

[0001] The present invention relates to the technical field of education management, and particularly to a management method and system for an intelligent teaching platform combined with virtual reality. Background Art

[0002] With the in-depth promotion of the construction of intelligent power plants, a training platform with quantitative management as the core has become an important tool for enterprise talent cultivation. In the existing technology, the intelligent training platform has achieved cost reduction and management efficiency improvement through online learning, examinations, and a quantitative integral system. However, the existing solutions still have the following limitations: Most of the existing teaching platforms rely on a single evaluation index, and their evaluation results often cannot comprehensively reflect the comprehensive abilities and learning status of learners; Although the introduction of virtual reality technology has provided a new interaction method and immersive experience for the intelligent teaching platform, making the teaching scenarios more diverse, at present, the virtual reality teaching platform still lacks an effective management mechanism in teaching evaluation and environment regulation, and it is unable to dynamically optimize teaching content and interaction methods according to the performance of learners in the virtual environment, nor can it adjust the configuration in a timely manner to meet different learning needs.

[0003] Therefore, there is an urgent need to develop a management method for an intelligent teaching platform that can comprehensively collect and analyze multi-dimensional learning data of employees and realize adaptive learning environment regulation in combination with virtual reality technology, so as to break through the limitations of the existing evaluation mechanism and improve teaching effects and management efficiency. Summary of the Invention

[0004] In order to overcome the shortcoming of being difficult to adjust the configuration of virtual teaching roles in a timely manner to meet different learning needs, the present invention provides a management method and system for an intelligent teaching platform combined with virtual reality.

[0005] The technical implementation solution of the present invention is as follows: A management method for an intelligent teaching platform combined with virtual reality, comprising the following steps: S1: Obtain the learning-related data of employees, and determine the first objective evaluation score and the second objective evaluation score that need to be calculated according to the learning-related data of the employees; S2: Use a dual-engine subjective question intelligent grading model according to the learning-related data of the employees to obtain the first objective evaluation score of the employees; S3: Use a two-dimensional integral quantization formula according to the learning-related data of the employees to obtain the second objective evaluation score of the employees; S4: Perform weighted summation on the first objective evaluation score and the second objective evaluation score to obtain an objective evaluation score, and manage the employees and the teaching platform according to the objective evaluation score; S5: Perform auxiliary learning adjustment on employees according to the open-book learning integral and the closed-book learning integral of employees.

[0006] Preferably, obtaining the learning-related data of the employee, and determining the first objective evaluation score and the second objective evaluation score to be calculated according to the learning-related data of the employee, includes: obtaining the learning-related data of the employee, where the learning-related data of the employee includes the employee's answering data, the employee's effective learning duration, the quality of the employee's answers, the employee's learning status, and the employee's learning environment, and using a dual-engine subjective question intelligent grading model and a two-dimensional integral quantization formula according to the learning-related data of the employee to obtain the first objective evaluation score and the second objective evaluation score of teaching.

[0007] Preferably, obtaining the first objective evaluation score of the employee by using the dual-engine subjective question intelligent grading model according to the learning-related data of the employee includes: constructing a dual-engine subjective question intelligent grading model by using an intelligent word segmentation model and a fuzzy matching algorithm, inputting the employee's answering data into the dual-engine subjective question intelligent grading model to obtain the first objective evaluation score, training by using a Transformer model to obtain an intelligent word segmentation model, where the intelligent word segmentation model is used to perform semantic parsing and keyword extraction on the employee's answering data to obtain a word segmentation result; the fuzzy matching algorithm is based on a predefined rule library and a dynamic threshold adjustment mechanism, and is used to perform multi-dimensional matching on the word segmentation result, calculate the matching degree between the employee's answering data and the standard answering data, and obtain the first objective evaluation score of teaching.

[0008] Preferably, obtaining the second objective evaluation score of the employee by using the two-dimensional integral quantization formula according to the learning-related data of the employee includes: inputting the employee's answering data into the two-dimensional integral quantization formula to obtain the second objective evaluation score, where the two-dimensional integral quantization formula is: ; In the formula, is the second objective evaluation score; is the th open-book learning integral; is the th weight coefficient of the open-book learning integral; is the th closed-book learning integral; is the th weight coefficient of the closed-book learning integral; The open-book learning integral formula is: ; In the formula, is the th open-book learning integral; is the employee's effective learning duration; is the weight coefficient of the effective learning duration; is the employee's learning status score; Rate the learning environment of employees; The formula for the closed-book learning score is: ; In the formula, is the th closed-book learning score; is the answer quality of the employee; is the answer quality weight coefficient; is the learning status score of the employee; is the learning environment score of the employee.

[0009] Preferably, the is the learning status score of the employee, including: obtaining the learning status score of the employee using the learning status score formula; and obtaining the learning environment score of the employee using the learning environment score formula, where the learning status score formula is: ; In the formula, is the learning status score of the employee; is the number of effective interactions in real answering; is the number of effective interactions in virtual reality answering; is the total learning duration in the real answering process; is the total learning duration in the virtual reality answering process; is the weight coefficient of the number of effective interactions in real answering; is the page switching weight between the real answering process and the virtual reality answering process; is the number of page switches in the real answering process; is the number of page switches in the virtual reality answering process; is the preset standard number of page switches.

[0010] Preferably, the is the learning environment score of the employee, including: The learning environment score formula is: ; In the formula, is the learning environment score of the employee; is the actual activity space of the employee during virtual reality answering; is the standard preset space of the employee during virtual reality answering; is the similarity between the employee's virtual reality learning environment and the employee's real learning environment; is the weight adjustment parameter.

[0011] Preferably, the auxiliary learning adjustment for employees based on the open-book learning scores and closed-book learning scores of employees includes: classifying the virtual characters in virtual reality into guiding characters and collaborative characters, where the guiding characters are teaching characters that adjust the explanation rhythm according to the employees' dynamics; the collaborative characters are trainee characters that simulate the behaviors of real trainees and participate in group discussions or ask questions. Adjust the proportions of the guiding characters and collaborative characters in virtual reality according to the open-book learning scores and closed-book learning scores of employees, and use the character number adjustment formula to adjust the total number of characters in the virtual reality learning environment for employees.

[0012] Preferably, the step of using the character number adjustment formula to adjust the total number of characters in the virtual reality learning environment for employees includes: the total number of characters includes the number of virtual characters and the number of real characters, and the character number adjustment formula is as follows: ; In the formula, is the total number of characters; is the minimum score adjustment value; is the maximum score adjustment value; is the critical value of; is the average value of the employee's historical open-book learning scores; is the average value of the employee's historical closed-book learning scores; is the adjustment parameter; is rounding up.

[0013] Preferably, the adjustment of the proportions of the guiding characters and collaborative characters in virtual reality according to the open-book learning scores and closed-book learning scores of employees includes: when the open-book learning score of an employee is greater than the closed-book learning score of the employee, increase the proportion of collaborative characters; when the open-book learning score of an employee is less than or equal to the closed-book learning score of the employee, increase the proportion of guiding characters.

[0014] Preferably, a smart teaching platform management system combined with virtual reality further includes: A data acquisition module, configured to acquire the learning-related data of employees, and determine the first objective evaluation score and the second objective evaluation score that need to be calculated according to the learning-related data of the employees; A first scoring acquisition module, configured to obtain the first objective evaluation score of an employee according to the learning-related data of the employee using a dual-engine subjective question intelligent grading model; A second scoring acquisition module, configured to obtain the second objective evaluation score of an employee according to the learning-related data of the employee using a two-dimensional integral quantization formula; A scoring acquisition management module, which is used to obtain an objective evaluation score by performing weighted summation on a first objective evaluation score and a second objective evaluation score, and manage employees and the teaching platform according to the objective evaluation score; An auxiliary learning adjustment module, which is used to perform auxiliary learning adjustment on employees according to the open-book learning score and the closed-book learning score of the employees.

[0015] The present invention has the following advantages: 1. The present invention comprehensively collects and quantifies multiple indicators such as answering data, learning duration, answering quality, interaction behavior, learning status, and virtual learning environment during the learning process of employees, constructs a two-dimensional integral system that covers both open-book exams and closed-book exams, and realizes a comprehensive and objective evaluation of employees' learning situations. This evaluation method breaks through the traditional single evaluation mode, making the evaluation results more credible and referenceable; 2. Using a dual-engine intelligent subjective question grading model constructed by an intelligent word segmentation system based on the Transformer model and a fuzzy matching algorithm, it can automatically perform semantic parsing and keyword extraction on the open-ended questions answered by employees, accurately calculate the answering matching degree, improve the automation and accuracy of subjective question grading, thereby reducing the workload of manual grading and improving the overall grading efficiency; 3. By analyzing the open-book and closed-book scores of employees, this method uses a role quantity adjustment formula to realize real-time adjustment of the quantity and proportion of guiding and collaborative virtual roles in the virtual reality teaching environment. This dynamic regulation mechanism can provide personalized teaching guidance and interaction atmosphere according to the learning data feedback in different teaching scenarios, which not only optimizes the immersive experience of learners but also improves the overall teaching effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flowchart of the management method of the intelligent teaching platform combined with virtual reality of the present invention; Figure 2 It is a schematic structural diagram of the management system of the intelligent teaching platform combined with virtual reality of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] Referring to the embodiments herein means that the specific features, structures, or characteristics described in connection with the embodiments can be included in at least one embodiment of the present invention. The phrase appears in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0018] Embodiment 1: A management method of an intelligent teaching platform combined with virtual reality, as Figure 1 shown, includes the following steps: S1: Acquire learning-related data of an employee, and determine a first objective evaluation score and a second objective evaluation score to be calculated based on the learning-related data of the employee; The employee's learning-related data is obtained, wherein the employee's learning-related data includes the employee's answer data, the employee's effective learning time, the employee's answer quality, the employee's learning status and the employee's learning environment, and a dual-engine subjective question intelligent review model and a dual-dimensional integral quantification formula are used according to the employee's learning-related data to obtain a first objective evaluation score and a second objective evaluation score of the teaching.

[0019] It should be noted that the information of employees is collected from multiple angles through the preset data collection module. The learning-related data of employees include but are not limited to the following: Employee answer data: including answers to various questions completed by employees on the teaching platform, including multiple-choice questions, true-or-false questions and objective questions, as well as answers to subjective questions that require manual or intelligent scoring; Employee effective learning time: refers to the cumulative time that employees are in an effective learning state on the teaching platform. The teaching platform will eliminate abnormal interruptions or invalid time; Employee answer quality: quantitative scoring is performed based on the accuracy of employees' answers, completeness of answers and standardization; Employee learning status: assessed by examining the frequency of effective interactions of employees in the answering process, the total learning time in the real answering process and the total learning time in the virtual reality answering process; Employee learning environment: assessed by the similarity between the virtual reality and the real environment and the interactive space in which the employee is located. Comprehensive scoring is performed. After data collection is completed, the learning-related data of the employees obtained are preprocessed and normalized to ensure that the data of each dimension meets the requirements of subsequent operations. According to the pre-set calculation mode, the preset dual-engine subjective question intelligent review model and the two-dimensional integral quantification formula are used to process the data. The pre-processed employee answer data and other learning-related data are input one by one according to the project. For the first objective evaluation score, the embedded dual-engine subjective question intelligent review model is called to complete the scoring; for the second objective evaluation score, the open-book and closed-book scores of the employees are calculated according to the two-dimensional integral quantification formula. The various weight coefficients and scoring parameters involved in the formula are read from the pre-designed parameter library and matched with the actual collected data. Finally, the system integrates the two parts of the score to obtain an objective evaluation result reflecting the overall teaching effect, and uses this result as the basis for subsequent teaching platform management.

[0020] S2: using the dual-engine subjective question intelligent review model according to the learning-related data of the employee to obtain a first objective evaluation score of the employee; Build a dual - engine intelligent evaluation model for subjective questions using an intelligent word - segmentation model and a fuzzy - matching algorithm, and input the employee's answer data into the dual - engine intelligent evaluation model for subjective questions to obtain a first objective evaluation score. Train using a Transformer model to obtain an intelligent word - segmentation model, which is used to perform semantic parsing and keyword extraction on the employee's answer data to obtain a word - segmentation result; the fuzzy - matching algorithm is based on a predefined rule library and a dynamic threshold adjustment mechanism, used to perform multi - dimensional matching on the word - segmentation result, calculate the matching degree between the employee's answer data and the standard answer data, and obtain the first objective evaluation score for teaching.

[0021] It should be noted that the teaching platform uses an intelligent word - segmentation model and a fuzzy - matching algorithm to build a dual - engine intelligent evaluation model for subjective questions. During the construction process, training is carried out using a Transformer model to obtain a high - precision intelligent word - segmentation model. The intelligent word - segmentation model is mainly used to perform semantic parsing and keyword extraction on the input employee answer data. The extracted word - segmentation result can accurately reflect the key technical terms, concepts, and logical associations in the employee's answer. Taking the employee's answer data as input data, input it into the completed dual - engine intelligent evaluation model for subjective questions to ensure that the data can be accessed and called between the engines of the model. In the fuzzy - matching part, based on a predefined rule library and a dynamic threshold adjustment mechanism, perform multi - dimensional matching on the obtained word - segmentation result. The core of this fuzzy - matching algorithm is to automatically calculate a quantified matching score by comparing the matching degree between the employee's answer data and the standard answer data, and thereby reflect the accuracy and integrity of the employee's subjective - question answer. Combine the results of the above two engines to obtain and output the first objective evaluation score of the employee.

[0022] S3: Use a two - dimensional integral quantization formula according to the employee's learning - related data to obtain the employee's second objective evaluation score; Input the employee's answer data into the two - dimensional integral quantization formula to obtain the second objective evaluation score, where the two - dimensional integral quantization formula is: ; In the formula, is the second objective evaluation score; is the th open - book learning integral; is the weight coefficient of the th open - book learning integral; is the th closed - book learning integral; is the weight coefficient of the ; Wherein, is the th open-book learning integral; is the effective learning duration of the employee; is the weight coefficient of the effective learning duration; is the learning status score of the employee; is the learning environment score of the employee; The formula for the closed-book learning integral is: ; Wherein, is the th closed-book learning integral; is the answer quality of the employee; is the weight coefficient of the answer quality; is the learning status score of the employee; is the learning environment score of the employee.

[0023] It should be noted that first, all relevant records are extracted from the pre-collected answer data of the employees. The answer data includes all information generated by the employees during open-book learning and closed-book examinations. After preprocessing the collected data, it will be input into the two-dimensional integral quantization formula in sequence. Among them is the effective learning duration of the employee, recording the total time when the employee is in an effective learning state on the teaching platform. 1. Original duration collection: Through the client or server-side log, record the total duration from when the employee enters the learning page (or starts the VR scene) to when they leave the page (or exit the VR), denoted as ; 2. Excluding invalid duration: Departure detection: When the user has no interaction (mouse movement, key pressing, head tracking, and voice detection) within a certain time (such as 30 consecutive seconds), it is considered in the "departure" state, and the accumulated duration is ; Hanging detection: If it is detected that the VR device or the client is in the background for a long time (such as minimized or out of focus), the corresponding duration will also be included in ; Excluding abnormal interruptions: In case of network interruption or system crash, the corresponding time period is excluded through log marking; 3. Calculating the effective duration: ; Among them, it should be ensured that , and lower limit truncation is performed on negative values or too small values (such as less than 5 seconds) to avoid interference from abnormal data; is the weight coefficient of the effective learning duration. Collect a large amount of learning data and final assessment scores or performance improvement indicators of employees. Using the effective learning duration, status, environment, and answer quality as features and performance as labels, train a linear regression or tree model. The linear coefficient or feature importance index fitted by the model is used as the weight coefficient of the effective learning duration; For the answering quality of employees, considering accuracy, completeness, and standardization comprehensively, weighted summation is performed to obtain the normalized answering quality. Among them, by automatically comparing the employee's answers with the standard answers (multiple-choice questions, true-false questions, and fill-in-the-blank questions), the ratio of the number of correct questions to the total number of questions is calculated to obtain the accuracy. Through the dual-engine intelligent grading model for subjective questions (word segmentation + fuzzy matching) or the semantic similarity model (such as BERT / SBERT similarity), the list of key points of the standard answer is extracted, and the proportion of key points hit in the employee's answer is checked to obtain the completeness. By comparing the employee's answer with the template format (paragraph structure and term list), the format symbol integrity is calculated, or regular rules and the keyword list supply rule base are used to check whether there are non-standard words and typos in the answer to obtain the standardization; is the weighting coefficient of the answering quality. According to the internal teaching experts, curriculum designers, HR, and data analysts in the factory, pairwise comparisons or anonymous scoring from 1 to 10 are made on the relative importance of the answering quality and other indicators (such as duration, environment, and status); after the feedback is summarized, the anonymous statistical results can be seen and the scoring can be corrected in the second round; the average value or weighted average of the final scoring is taken and normalized to obtain ; is the learning status score of the employee, which reflects the number of effective interactions of the employee during the actual answering process; is the learning environment score of the employee. This score reflects the comparison between the actual activity space where the employee is located during the virtual reality answering process and the preset standard space, as well as the environmental similarity. After multiplying all the processed open-book learning scores and closed-book learning scores by their respective weighting coefficients and then accumulating them, the final second objective evaluation score is obtained.

[0024] Using the learning status score formula, the learning status score of the employee is obtained; and using the learning environment score formula, the learning environment score of the employee is obtained. Among them, the learning status score formula is: ; In the formula, is the learning status score of the employee; is the number of effective interactions in the real answering; is the number of effective interactions in the virtual reality answering; is the total learning duration in the real answering process; is the total learning duration in the virtual reality answering process; is the weighting coefficient of the number of effective interactions in the real answering; is the weight of the number of page switches in the real answering process and the virtual reality answering process; is the number of page switches in the real answering process; is the number of page switches in the virtual reality answering process; is the preset standard number of page switches.

[0025] It should be noted that is the effective number of interactions in real answering, reflecting the interaction frequency between employees and the platform or teachers during the traditional answering process; is the effective number of interactions in virtual reality answering, reflecting the effective number of interactions of employees in the virtual answering environment; is the total learning duration during real answering; is the total learning duration during virtual reality answering, used to normalize the number of interactions to ensure the comparability of scoring; is the number of page switches during real answering; is the number of page switches during virtual reality answering, and the sum of the two is used to quantify the interference factors during the learning process; is the preset standard number of page switches, used as a reference value under normal learning conditions; where is the weight coefficient of the effective number of interactions in real answering. Usually, the effective interactions in the real environment are more difficult to simulate, and at the same time, they can better reflect the employees' real-time understanding and feedback of the content. Therefore, ; For the interactions in the existing VR environment, due to technical controllability, immersion error, and hardware differences, their quality should adopt ; Among them In the small stage, the weights of real and virtual interactions are close, and the balance between the two can be flexibly explored; in the large stage, the weight of real interaction continues to increase, but the weight of virtual interaction tends to saturate, avoiding the overemphasis of the virtual interaction part and distorting the evaluation results.

[0026] The learning environment scoring formula is: ; In the formula, is the learning environment score of the employee; is the actual activity space of the employee during virtual reality answering; is the standard preset space of the employee during virtual reality answering; is the similarity between the employee's virtual reality learning environment and the employee's real learning environment; is the weight adjustment parameter.

[0027] It should be noted that collecting the data of the actual activity space of employees during answering reflects the actual space range where employees can move freely in the real environment during virtual reality answering; The standard preset space for employees during virtual reality answering questions, representing the space standard required for the real environment during normal virtual reality answering questions. 1. Scenario requirement analysis: Set by the system administrator or training designer during the scenario deployment phase through site measurement and experimental verification. The specific process is to clarify the minimum activity range according to the course type (such as mechanical principle demonstration, assembly simulation, and fire evacuation drill); refer to industry standards or safety specifications to give the space lower limit; 2. On-site measurement and verification: Conduct measurement and on-site research in the training site or the VR area in employees' homes to ensure that the site at least meets the preset space. Let several test users demonstrate according to the key points of the course actions, and check whether there is no collision and no lag under the standard preset space; 3. Parameter entry: Enter the parameters; The similarity between the virtual reality learning environment of employees and their real learning environment, obtained by calculating through comparing environmental characteristics (such as light, sound, and layout). First, conduct data collection. 1. Visual characteristics, real environment: Collect multiple panoramic images through an environmental camera or the user's mobile phone; virtual environment: Capture the rendered images (panoramic or multi-perspective) of the corresponding scene in VR; 2. Sound characteristics, real environment: Use a microphone to record the background sound (volume, reverberation duration, and spectral distribution); virtual environment: The environmental audio track output by the system (also collect the spectral and reverberation indices); 3. Spatial characteristics, real environment: Obtain the spatial geometry (wall distance, door and window positions, and furniture layout) through an indoor mapping instrument or a depth camera; virtual environment: The scene geometry data provided by the engine (collision boxes, static object distribution, and spatial voxelization description); Use a pre-trained deep network (such as ResNet, Vision-Transformer) to extract the image feature vectors; Extract the Mel-spectrogram after calculating the short-time Fourier transform (STFT), and then obtain the sound characteristics through dimensionality reduction by a convolutional network or PCA; Voxelize the three-dimensional geometry to a unified resolution grid, and perform two-dimensional or three-dimensional convolutional encoding on the occupied voxels to obtain the spatial characteristics; Concatenate or weight and fuse the visual, acoustic, and layout features, and finally use the cosine similarity to calculate the visual, acoustic, and layout similarities respectively and then do a weighted average to finally obtain the similarity between the virtual reality learning environment of employees and their real learning environment; Finally, perform normalization processing on the output results to ensure that the obtained learning environment score matches other scoring items, so as to occupy a reasonable weight in the overall multi-dimensional integration system; The actual activity space for employees during virtual reality question answering refers to the physical space volume or area that employees actually occupy and freely move in during virtual reality (VR) question answering or interaction scenarios. It reflects the degree required by employees in the real environment to complete VR learning, including the behavioral categories of walking, pole-rounding interaction, and arm waving. It is obtained through the following steps: 1. Trajectory data collection: Sample the 3D coordinates of the headset and controller at the start of learning, with a sampling frequency usually of 30–90Hz; record all position points during the entire effective interaction period. 2. Spatial boundary reconstruction: Convex hull algorithm: Calculate the three-dimensional convex hull of the trajectory point set to obtain the minimum convex polyhedron; Bounding box: Or directly use the axis-aligned or orientation-aligned minimum bounding box as an approximate boundary. 3. Volume calculation: Calculate the volume of the convex hull or OBB; or measure the actual space size reserved as the actual activity space for employees during virtual reality question answering. is a weight adjustment parameter. By inviting several test users to learn under various combinations of real environments and VR scenarios, and having an expert team score the comfort or satisfaction, align the learning environment scoring formula with the comfort scores marked by the experts, define a loss function, and use gradient descent to iteratively solve for the optimal value on the offline samples. .

[0028] S4: Obtain the objective evaluation score by performing a weighted sum of the first objective evaluation score and the second objective evaluation score, and manage the employees and the teaching platform according to the objective evaluation score; S5: Perform auxiliary learning adjustment on the employees according to the open-book learning points and the closed-book learning points of the employees.

[0029] The virtual characters in virtual reality are divided into guiding characters and collaborative characters. The guiding characters are teaching characters that adjust the explanation rhythm according to the employees' dynamics; the collaborative characters are trainee characters that simulate the behaviors of real trainees and participate in group discussions or ask questions. Adjust the proportions of the guiding characters and collaborative characters in the virtual reality learning environment of the employees according to the open-book learning points and the closed-book learning points of the employees, and use the character number adjustment formula to adjust the total number of characters in the virtual reality learning environment of the employees.

[0030] It should be noted that in the intelligent teaching platform integrating virtual reality, in order to simulate the interaction and guidance effects in real teaching, virtual characters are divided into two categories: guiding characters: such characters mainly undertake teaching responsibilities and can dynamically adjust the explanation rhythm according to the integral fluctuations of employees during the learning process. Guiding characters provide real-time explanations and prompts for employees through voice and animation display methods, making the teaching process smoother and more personalized; collaborative characters: such characters simulate the behaviors of real students, participate in group discussions, ask questions or interact, and help employees obtain a richer interactive experience in the virtual reality environment, enhancing team collaboration ability and comprehensive quality; when making auxiliary learning adjustments, the open-book learning scores and closed-book learning scores of employees are statistically analyzed and compared. When the open-book learning score of an employee is significantly higher than the closed-book learning score, it is considered that the employee performs well in independent thinking and real-time interaction. At this time, the proportion of collaborative characters is increased to stimulate the enthusiasm of employees to participate in discussions and solve problems cooperatively; when the open-book learning score of an employee is lower than or equal to the closed-book learning score, it is mainly considered that the employee needs more guidance and explanations. At this time, the proportion of guiding characters is increased to provide more targeted teaching explanations and operation prompts for employees; the average values of open-book and closed-book scores are calculated based on the historical score data of employees, and the total number of characters is determined by using a preset critical value and adjustment parameter, so that the number of virtual characters (including guiding and collaborative characters) can not only meet the teaching interaction needs but also avoid information overload or interference.

[0031] The total number of characters includes the number of virtual characters and the number of real characters, and the character number adjustment formula is as follows: ; In the formula, is the total number of characters; is the minimum integral adjustment value; is the maximum integral adjustment value; is 's critical value; is the average value of the employee's historical open-book learning scores; is the average value of the employee's historical closed-book learning scores; is the adjustment parameter; is rounding up.

[0032] It should be noted that when is less than or equal to , the linear term is greater than or equal to , and the function takes a linear value; when is greater than , the linear term is less than , and the function remains ; ensure that the total number of characters is within the preset range. In this embodiment is the minimum integral adjustment value and at the same time is the lowest total number of roles; is the maximum integral adjustment value and at the same time is the highest total number of roles; is the critical value of the critical integral value of

[0033] When the open-book learning integral of an employee is greater than the closed-book learning integral of the employee, the proportion of collaborative roles is increased; when the open-book learning integral of the employee is less than or equal to the closed-book learning integral of the employee, the proportion of guiding roles is increased.

[0034] Embodiment 2: On the basis of Embodiment 1, a smart teaching platform management system integrating virtual reality, as Figure 2 shown, further includes: A data acquisition module, configured to acquire learning-related data of an employee, and determine a first objective evaluation score and a second objective evaluation score that need to be calculated according to the learning-related data of the employee; A first scoring acquisition module, configured to use a dual-engine subjective question intelligent marking model to obtain a first objective evaluation score of the employee according to the learning-related data of the employee; A second scoring acquisition module, configured to use a two-dimensional integral quantization formula to obtain a second objective evaluation score of the employee according to the learning-related data of the employee; A scoring acquisition management module, configured to perform weighted summation on the first objective evaluation score and the second objective evaluation score to obtain an objective evaluation score, and manage the employee and the teaching platform according to the objective evaluation score; An auxiliary learning adjustment module, configured to perform auxiliary learning adjustment on the employee according to the open-book learning integral and the closed-book learning integral of the employee.

[0035] The technical principle of the embodiments of the present invention has been described above in conjunction with specific embodiments. These descriptions are only for explaining the principle of the embodiments of the present invention and cannot be interpreted in any way as a limitation on the protection scope of the embodiments of the present invention. Based on the explanations herein, those skilled in the art can think of other specific embodiments of the embodiments of the present invention without creative labor, and these embodiments will fall within the protection scope of the embodiments of the present invention.

Claims

1. A management method for an intelligent teaching platform integrating virtual reality, characterized in that It includes the following steps: S1: Obtain the learning-related data of the employee, and determine the first objective evaluation score and the second objective evaluation score to be calculated according to the learning-related data of the employee; S2: Use the dual-engine subjective question intelligent grading model according to the learning-related data of the employee to obtain the first objective evaluation score of the employee; S3: Use the two-dimensional integral quantization formula according to the learning-related data of the employee to obtain the second objective evaluation score of the employee; S4: Perform weighted summation on the first objective evaluation score and the second objective evaluation score to obtain the objective evaluation score, and manage the employee and the teaching platform according to the objective evaluation score; S5: Perform auxiliary learning adjustment on the employee according to the open-book learning score and the closed-book learning score of the employee.

2. The management method of an intelligent teaching platform combined with virtual reality according to claim 1, characterized in that, The obtaining of the learning-related data of the employee and determining the first objective evaluation score and the second objective evaluation score to be calculated according to the learning-related data of the employee includes: obtaining the learning-related data of the employee, where the learning-related data of the employee includes the employee's answering data, the employee's effective learning duration, the employee's answering quality, the employee's learning status, and the employee's learning environment, and using the dual-engine subjective question intelligent grading model and the two-dimensional integral quantization formula according to the learning-related data of the employee to obtain the first objective evaluation score and the second objective evaluation score of the teaching.

3. A management method for a smart teaching platform integrating virtual reality according to claim 1, characterized in that, The using of the dual-engine subjective question intelligent grading model according to the learning-related data of the employee to obtain the first objective evaluation score of the employee includes: constructing a dual-engine subjective question intelligent grading model using an intelligent word segmentation model and a fuzzy matching algorithm, and inputting the employee's answering data into the dual-engine subjective question intelligent grading model to obtain the first objective evaluation score. Use the Transformer model for training to obtain the intelligent word segmentation model, and the intelligent word segmentation model is used to perform semantic parsing and keyword extraction on the employee's answering data to obtain the word segmentation result; the fuzzy matching algorithm is based on a predefined rule library and a dynamic threshold adjustment mechanism, and is used to perform multi-dimensional matching on the word segmentation result, calculate the matching degree between the employee's answering data and the standard answering data, and obtain the first objective evaluation score of the teaching.

4. A management method for a smart teaching platform combined with virtual reality according to claim 1, characterized in that The using of the two-dimensional integral quantization formula according to the learning-related data of the employee to obtain the second objective evaluation score of the employee includes: inputting the employee's answering data into the two-dimensional integral quantization formula to obtain the second objective evaluation score, where the two-dimensional integral quantization formula is: ； Wherein, is the second objective evaluation score; is the th open-book learning integral; is the weight coefficient of the th open-book learning integral; is the th closed-book learning integral; is the weight coefficient of the th closed-book learning integral; The open-book learning score formula is: ； In the formula, is the th open-book learning score; is the effective learning duration of the employee; is the weight coefficient of the effective learning duration; is the learning status score of the employee; is the learning environment score of the employee; The closed-book learning score formula is: ； Wherein, is the th closed-book learning score; is the answer quality of the employee; is the answer quality weight coefficient; is the learning status score of the employee; is the learning environment score of the employee.

5. A management method for an intelligent teaching platform combined with virtual reality according to claim 4, characterized in that, The learning status score for employees includes: obtaining the learning status score of employees using the learning status scoring formula; and obtaining the learning environment score of employees using the learning environment scoring formula, where the learning status scoring formula is: ； Wherein, is the learning status score of the employee; is the number of effective interactions in the real answering; is the number of effective interactions in the virtual reality answering; is the total learning duration in the real answering process; is the total learning duration in the virtual reality answering process; is the weight coefficient of the number of effective interactions in the real answering; is the page switching times weight in the real answering process and the virtual reality answering process; is the number of page switches in the real answering process; is the number of page switches in the virtual reality answering process; is the preset standard number of page switches.

6. A management method for a smart teaching platform integrating virtual reality according to claim 4, characterized in that, The said rate the learning environment of employees, including: The learning environment scoring formula is: ； In the formula, is the learning environment score of the employee; is the actual activity space when the employee takes a virtual reality quiz; is the standard preset space when the employee takes a virtual reality quiz; is the similarity between the employee's virtual reality learning environment and the employee's real learning environment; is the weight adjustment parameter.

7. A management method for a smart teaching platform integrating virtual reality according to claim 1, characterized in that, Auxiliary learning adjustment is performed on employees according to their open-book learning scores and closed-book learning scores, including: classifying virtual characters in virtual reality into guiding characters and collaborative characters, where the guiding characters are instructional characters that adjust the explanation rhythm according to employees; the collaborative characters are learner characters that simulate the behaviors of real learners, participate in group discussions or ask questions, adjusting the proportions of the guiding characters and collaborative characters in virtual reality according to the open-book learning scores and closed-book learning scores of employees, and using a character number adjustment formula to adjust the total number of characters in the virtual reality learning environment of employees.

8. A management method for a smart teaching platform integrating virtual reality according to claim 7, characterized in that, And using the character number adjustment formula to adjust the total number of characters in the virtual reality learning environment of employees, including: the total number of characters includes the number of virtual characters and the number of real characters, where the character number adjustment formula is: ； Wherein, is the total number of roles; is the minimum integral adjustment value; is the maximum integral adjustment value; is the critical value of; is the average value of the employee's historical open-book learning points; is the average value of the employee's historical closed-book learning points; is the adjustment parameter; is rounding up.

9. A management method for an intelligent teaching platform combined with virtual reality according to claim 7, characterized in that, Adjusting the proportions of the guiding characters and collaborative characters in virtual reality according to the open-book learning scores and closed-book learning scores of employees, including: when the open-book learning score of an employee is greater than the closed-book learning score of the employee, increasing the proportion of collaborative characters; when the open-book learning score of an employee is less than or equal to the closed-book learning score of the employee, increasing the proportion of guiding characters.

10. A management system for an intelligent teaching platform integrating virtual reality, according to the method for managing an intelligent teaching platform integrating virtual reality as described in any one of claims 1-9, characterized in that it further Including: A data acquisition module, configured to acquire learning-related data of employees, and determine a first objective evaluation score and a second objective evaluation score that need to be calculated according to the learning-related data of the employees; A first scoring acquisition module, configured to obtain the first objective evaluation score of an employee by using a dual-engine subjective question intelligent grading model according to the learning-related data of the employee; A second scoring acquisition module, configured to obtain the second objective evaluation score of an employee by using a two-dimensional integral quantization formula according to the learning-related data of the employee; A scoring acquisition management module, configured to obtain an objective evaluation score by performing weighted summation on the first objective evaluation score and the second objective evaluation score, and manage the employees and the teaching platform according to the objective evaluation score; An auxiliary learning adjustment module, configured to perform auxiliary learning adjustment on employees according to their open-book learning scores and closed-book learning scores.

Images