Intelligent teaching system and teaching method based on large language model

Abstract

The invention discloses an intelligent teaching system and teaching method based on a large language model, relates to the technical field of intelligent teaching, and solves the technical problems that deep analysis and utilization of teaching data are lacked, and teaching modes suitable for different subjects and different knowledge points are greatly different. According to the method, the input content of the teaching personnel is analyzed through the large language model, the teaching subject content is defined, and then according to the screened and matched teaching mode, the blindness of selecting the teaching mode only by experience traditionally is changed, the effect of evaluating the teaching mode by comprehensively considering the personnel score and the classroom participation degree is achieved, and the teaching efficiency is improved. Teaching records in the system are comprehensively classified in the same type, teaching modes and subject contents are closely associated, quick retrieval and screening are facilitated, the most suitable teaching mode is recommended for teaching personnel by means of vector cosine similarity calculation and the like, personalized teaching support is achieved, and personalized requirements of different teachers and students are met.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent teaching, and specifically provides an intelligent teaching system and teaching method based on a large language model. Background Art

[0002] With the continuous advancement of educational informatization, teaching systems are gradually developing towards intelligence and personalization. The traditional teaching mode faces the dilemma that it is difficult to accurately match teaching methods with teaching content and students' needs. In addition, with the rise of artificial intelligence technologies such as large language models, how to effectively apply them to teaching systems to improve teaching quality and efficiency has also become a current research hotspot.

[0003] According to the patent application with the publication number CN117437099B, an intelligent teaching system and teaching method based on a large language model are disclosed, which relate to the technical field of intelligent teaching. By using a knowledge graph to assist the large language model, intelligent agents are further realized, including an intelligent tutor module, an intelligent teaching assistant module, and an intelligent learning partner module. Before class, the intelligent teaching assistant module is used to realize intelligent auxiliary lesson preparation for teachers, dialogue-based autonomous preview for students, and personalized learning path recommendation; during class, the intelligent tutor module is used to realize guided teaching, and the intelligent learning partner discussion assistant is used to assist group discussions for teaching; after class, the intelligent teaching assistant module helps teachers generate targeted intelligent auxiliary evaluations and conduct intelligent auxiliary judgment on subjective questions to help students achieve personalized learning.

[0004] However, teachers lack a scientific basis when choosing teaching methods and mostly rely on subjective experience, making it difficult to accurately match according to specific teaching content and students' characteristics, resulting in uneven teaching effects. Some existing technologies lack in-depth analysis and utilization of teaching data. At the same time, there are significant differences in teaching methods applicable to different disciplines and different knowledge points. How to quickly and accurately screen out the optimal teaching method for specific teaching content has become an urgent problem to be solved in the education field. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present invention provides an intelligent teaching system and teaching method based on a large language model, which solves the problems of lack of in-depth analysis and utilization of teaching data and significant differences in teaching methods applicable to different disciplines and different knowledge points.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: An intelligent teaching system based on a large language model, comprising:

[0007] A teaching staff information collection module, configured to transmit the basic information of teaching staff to the personnel information analysis module, and the basic information includes the input content corresponding to the teaching staff and the historical scores of the teaching staff;

[0008] The personnel information analysis module is used to analyze the basic information transmitted by the teaching personnel information collection module. It uses a large language model to identify the input content of teaching personnel to obtain the content to be analyzed. At the same time, it classifies the teaching records to obtain a classification result, and screens the classification result based on the content to be analyzed to obtain a matching result, which is then transmitted to the teaching method analysis module;

[0009] The teaching method analysis module analyzes the teaching effect of the teaching methods in the matching results. It assigns a value to the change value of the personnel's grades to obtain a teaching grade assignment, and at the same time assigns a value to the change value of the personnel's participation times to obtain a classroom participation assignment. Then it sums the two to obtain a teaching effect value, and screens the teaching methods with a teaching effect value greater than the preset value to obtain preselected teaching methods, which are then transmitted to the comprehensive selection and processing module;

[0010] The comprehensive selection and processing module is used to analyze the preselected teaching methods, determine whether the preselected teaching methods exist in the historical usage records of teaching personnel, generate a matching result of existence or non-existence. For the existence matching result, it eliminates the same preselected teaching methods to obtain the screened teaching methods, and calculates the similarity between the screened teaching methods and the teaching personnel. Taking the screened teaching method with the greatest similarity as the standard, it generates teaching standard information;

[0011] For the non-existence matching result, it analyzes the historical content of the historical personnel corresponding to the preselected teaching methods, screens the historical personnel with the same input content as the teaching personnel and records them as the personnel to be analyzed, calculates the proportion of the number of the personnel to be analyzed, and at the same time combines the similarity for screening to generate teaching standard information, and transmits the teaching standard information to the teaching information output module.

[0012] The teaching information output module is used to display the obtained teaching standard information to the corresponding operators.

[0013] As a further solution of the present invention, the specific method for the personnel information analysis module to obtain the matching result is as follows:

[0014] Obtain the input content in the basic information and input it into the large language model for analysis to obtain the content to be analyzed. Then obtain all the teaching records in the system, and classify the teaching records by the same type to obtain a classification result denoted as i, and i = 1, 2, …, j, where j represents the types corresponding to the classification results. At the same time, obtain the teaching methods corresponding to the classification result i and denote them as ic, and c = 1, 2, …, v, where v represents the number of teaching methods;

[0015] Then screen the classification result i based on the content to be analyzed to obtain a matching result, and at the same time generate matching result information.

[0016] As a further solution of the present invention, the specific method for the teaching method analysis module to obtain the teaching performance assignment is as follows:

[0017] Obtain all the personnel corresponding to teaching method c, obtain the pre- and post-performance change values of the personnel, calculate the sum of all the pre- and post-performance change values, calculate the corresponding average value at the same time, and match the obtained average value with the corresponding assignment interval to obtain the teaching performance assignment. And the assignment interval here is obtained based on different data tests.

[0018] As a further solution of the present invention, the specific method for the teaching method analysis module to obtain the classroom participation assignment is as follows:

[0019] Obtain the pre-teaching participation times and post-teaching parameter times of the personnel corresponding to teaching method c, calculate the corresponding parameter times difference, and assign a value to the parameter times difference in the same way as the personnel performance assignment method to obtain the classroom participation assignment.

[0020] As a further solution of the present invention, the specific method for the teaching method analysis module to obtain the preselected teaching method is as follows:

[0021] Sum up the obtained teaching performance assignment and classroom participation assignment to obtain the teaching effect value Gc corresponding to teaching method c. And the specific calculation formula is Gc = Fc×a 1 +Kc×a 2 , where Fc represents the teaching performance assignment, Kc represents the classroom participation assignment, a 1 and a 2 are both preset proportionality coefficients;

[0022] At the same time, compare the teaching effect value Gc corresponding to teaching method c with a preset value, and screen out the teaching methods corresponding to the teaching effect value Gc greater than the preset value, which are recorded as preselected teaching methods, and generate preselected teaching method information.

[0023] As a further solution of the present invention, the specific method for the comprehensive selection and processing module to generate a matching result (present or absent) is as follows:

[0024] Obtain the historical usage records of the teaching personnel, and judge whether there is a situation matching the preselected teaching method information in the historical usage records. If there is a corresponding preselected teaching method in the historical usage records, generate a present matching result; otherwise, generate an absent matching result.

[0025] As a further solution of the present invention, the specific method for the comprehensive selection and processing module to analyze the present matching result and generate teaching standard information is as follows:

[0026] Eliminate the same preselected teaching methods, and obtain the remaining preselected teaching methods, denoted as the screened teaching methods and recorded as a, where a = 1, 2, …, b, and b represents the number of screened teaching methods;

[0027] Obtain the basic information of the teaching staff, perform vector conversion on the input content in the basic information and the screened teaching methods, and calculate the cosine similarity between the two vectors, denoted as Ya;

[0028] At the same time, obtain all the cosine similarities Ya, and select the screened teaching method corresponding to the largest value of the cosine similarity Ya as the standard to generate teaching standard information.

[0029] As a further solution of the present invention, the specific method for the comprehensive selection processing module to analyze and generate teaching standard information for the non - matching results is as follows:

[0030] Calculate the cosine similarity between the preselected teaching method and the input content of the teaching staff. Then, obtain the historical personnel and corresponding historical content corresponding to the preselected teaching method, and compare the historical content with the input content of the teaching staff. Screen the historical personnel with the same input content as the teaching staff and record them as the personnel to be analyzed;

[0031] Calculate the proportion of the number corresponding to the personnel to be analyzed, and at the same time, compare the obtained proportion with the preset proportion value, and screen the preselected teaching methods corresponding to the proportion greater than the preset proportion value and record them as the teaching methods to be analyzed;

[0032] Then, obtain the similarity between the teaching methods to be analyzed and the teaching staff, and select the teaching method to be analyzed corresponding to the largest similarity as the standard to generate teaching standard information.

[0033] A smart teaching method based on a large - language model, which specifically includes the following steps:

[0034] Step S1: Obtain the input content of the teaching staff, identify the content to be analyzed through the large - language model, classify the teaching records to obtain the classification results, and screen the classification results with the content to be analyzed as the standard to obtain the matching results;

[0035] Step S2: Analyze the teaching effect of the teaching methods in the matching results. Assign a value to the change value of the personnel's grades to obtain the teaching performance assignment, assign a value to the change value of the personnel's participation times to obtain the classroom participation assignment. At the same time, sum the two to obtain the teaching effect value, and compare it with the preset value to screen the teaching methods to obtain the preselected teaching methods;

[0036] Step S3: Determine whether the preselected teaching method exists in the historical usage records of teaching staff, generate a matching result of existence or non-existence. For the case of an existing matching result, eliminate the same preselected teaching methods to obtain the screened teaching methods, calculate the similarity between the screened teaching methods and the teaching staff, select the screened teaching method with the highest similarity as the standard, and generate teaching standard information;

[0037] Step S4: For the case of no matching result, analyze the historical content of the historical personnel corresponding to the preselected teaching method, screen out the historical personnel with the same input content as the teaching staff and record them as the personnel to be analyzed, calculate the proportion of the number of personnel to be analyzed, and at the same time combine the similarity for screening to generate teaching standard information.

[0038] The present invention provides an intelligent teaching system and teaching method based on a large language model. Compared with the prior art, it has the following beneficial effects:

[0039] The present invention analyzes the input content of teaching staff through a large language model to clarify the teaching subject content, and then screens and matches the teaching methods accordingly, changing the blindness of the traditional method of only relying on experience to select teaching methods, and comprehensively considering the personnel's grades and classroom participation to evaluate the effect of teaching methods. It not only pays attention to the change value of grades before and after, but also comprehensively evaluates the teaching effect by calculating the difference in the number of classroom participations, etc.

[0040] Conduct a comprehensive classification of the same type of teaching records in the system, closely associate the teaching methods with the subject content, which is convenient for quick retrieval and screening. In the comprehensive selection and processing module, according to the historical usage records and input content of the teaching staff, analyze the cases of existing and non-existing matching results respectively. By calculating the cosine similarity of vectors and other methods, recommend the most suitable teaching method for the teaching staff to achieve personalized teaching support and meet the personalized needs of different teachers and students. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is the system principle block diagram of the present invention;

[0042] Figure 2 It is the step method diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0044] Example 1, please refer to Figure 1, this application provides an intelligent teaching system based on a large language model, including a teaching staff information collection module, a personnel information analysis module, a teaching method analysis module, a comprehensive selection and processing module, and a teaching information output module. At the same time, combined with Figure 1 It can be known that the above functional modules are connected by one-way electrical connection.

[0045] The teaching staff information collection module is used to obtain the basic information of the teaching staff, and at the same time transmit the basic information to the personnel information analysis module. The basic information of the teaching staff includes the input content corresponding to the teaching staff and the historical achievements of the teaching staff.

[0046] The personnel information analysis module is used to analyze the obtained basic information, obtain the input content in the basic information, and analyze the input content through the large language model to obtain the corresponding content to be analyzed. The content to be analyzed here represents the corresponding teaching subject content. For example, if the input is the explanation process of a math problem, then the input content will be classified to obtain the teaching subject - mathematics. Then, all teaching records in the system are obtained, and the teaching records are classified by the same type to obtain the classification result, denoted as i, and i = 1, 2,..., j, where j represents the type corresponding to the classification result, such as Chinese, mathematics, English, chemistry, and biology, etc. At the same time, the teaching method corresponding to the classification result i is obtained and denoted as ic, and c = 1, 2,..., v, where v represents the number of teaching methods;

[0047] Then, the classification result i is screened based on the content to be analyzed to obtain the matching result, and at the same time, the matching result information is generated and transmitted to the teaching method analysis module.

[0048] Suppose the input content in the basic information received by the personnel information analysis module is a detailed explanation of the "derivation process of the quadratic formula for finding roots". After analyzing this input content through the large language model, it is clear that the teaching subject of the content to be analyzed is mathematics. 2. Teaching record classification: After the system obtains all teaching records, it classifies them into Chinese, mathematics, English, chemistry, and biology, these j = 5 categories according to the subject. Under the mathematics subject (assuming i = 2 represents mathematics), the teaching methods are the lecture method (2 1 ), the group discussion method (2 2 ), and the multimedia demonstration method (2 3 ), that is, v = 3;

[0049] Taking the content to be analyzed, "Derivation Process of the Quadratic Formula", as the standard, the classification results are screened. It is found that the lecture method, group discussion method, and multimedia demonstration method under the mathematics discipline may all be applicable to the teaching of this content. Therefore, matching result information containing these matching results is generated and transmitted to the teaching method analysis module for further analysis of which teaching method is the most suitable in the follow-up.

[0050] Teaching method analysis module, which is used to analyze the obtained matching result information, obtain the matching result information, and at the same time analyze the teaching effects corresponding to the teaching methods in the matching result information, and the analysis of the teaching effects is comprehensively analyzed through the corresponding personnel scores and classroom participation degrees;

[0051] The specific method for analyzing the personnel scores is as follows: Obtain all the personnel corresponding to teaching method c, and obtain the change value of the scores before and after for the personnel. The change value of the scores before and after represents the difference between the in-class test scores before the implementation of this teaching method and the in-class test scores after the implementation of this teaching method, and calculate the sum value of all the change values of the scores before and after. At the same time, calculate the corresponding average value, and match the obtained average value with the corresponding assignment interval to obtain the teaching score assignment. Here, the assignment interval is obtained through different data tests. Specifically, assume that these 5 students are Student A, Student B, Student C, Student D, and Student E respectively, and the corresponding in-class test scores are shown in the in-class test score table before and after adopting the project-based learning method as follows:

[0052] Adding up the change values of the scores before and after for these 5 students, the sum value is obtained as 41 points. Further calculating the corresponding average value gives 8.2 points. Through different data tests, the corresponding assignment interval is obtained, as shown in the following assignment interval table:

[0053]

[0054] According to the obtained average value and the assignment interval table for matching, since the calculated average value is 8.2 points and it is within this assignment interval, the teaching effect of this teaching method (project-based learning method) can be "general effect", and at the same time, the corresponding assignment 3 is obtained.

[0055] Then analyze the classroom participation degree, obtain the number of participation times before teaching and the number of participation times after teaching for the personnel corresponding to teaching method c, and calculate the difference in the number of participation times. Similarly to the assignment method of personnel scores, assign a value to the difference in the number of participation times to obtain the classroom participation assignment;

[0056] The obtained teaching performance assignment and classroom participation assignment are summed up to obtain the teaching effect value Gc corresponding to the teaching method c, and the specific calculation formula is Gc=Fc×a 1 +Kc×a 2 , where Fc represents the teaching performance assignment, Kc represents the class participation assignment, and a 1 and a 2 They are all preset proportional coefficients, and the specific values ​​are set by the operator;

[0057] At the same time, the teaching effect value Gc corresponding to the teaching method c is compared with a preset value, and the specific value of the preset value is set by the operator, and the teaching method corresponding to the teaching effect value Gc greater than the preset value is selected and recorded as the pre-selected teaching method, and the information corresponding to the pre-selected teaching method is obtained to obtain the pre-selected teaching method information;

[0058] Then the obtained pre-selected teaching method information is transmitted to the comprehensive selection processing module.

[0059] A comprehensive selection processing module, which is used to analyze the obtained pre-selected teaching method information, obtain the corresponding historical usage records of the teaching staff, and determine whether there is a match with the pre-selected teaching method information in the historical usage records, and generate a matching result or a non-matching result. Specifically, if the corresponding pre-selected teaching method exists in the historical usage records, a matching result is generated, otherwise a non-matching result is generated, and then the two are analyzed respectively;

[0060] Analyze the generated matching results, eliminate the same pre-selected teaching methods, and obtain the remaining pre-selected teaching methods as screening teaching methods, and label them as a, and a=1, 2, ..., b, where b represents the number of screening teaching methods. Then obtain the basic information of the teaching staff, and quantize the input content in the basic information. At the same time, perform vector conversion on the quantized input content, and perform the same processing on the screening teaching methods to obtain the corresponding vector, and then calculate the cosine similarity between the two vectors and record it as Ya;

[0061] For example, for screening teaching method 1 (traditional lecture-based teaching), the proportion of the lecture method is 100% (quantified value is 1), the frequency of multimedia use is low (assuming it is used once a week, at most 5 times a week, quantified value is 0.2), and the expected proportion of students' autonomous learning time is 10% (quantified value is 0.1). Construct the vector [1, 0.2, 0.1], and after normalization, it is [0.0833, 0.167, 0.083]. For screening teaching method 2 (inquiry-based teaching method based on group discussion), the proportion of group discussion and inquiry-based learning is 80% (quantified value is 0.8), the frequency of multimedia use is moderate (used 3 times a week, quantified value is 0.6), and the expected proportion of students' autonomous learning time is 30% (quantified value is 0.3). Construct the vector [0.8, 0.6, 0.3], and after normalization, it is [0.615, 0.462, 0.231].

[0062] Simultaneously obtain all cosine similarities Ya, and select the screening teaching method corresponding to the largest cosine similarity Ya value as the standard to generate teaching standard information, and then transmit it to the teaching information output module;

[0063] Analyze the generated non-matching results, calculate the cosine similarity between the preselected teaching method and the content input by the teaching staff, and sort them from largest to smallest according to the cosine similarity. Then obtain the historical personnel corresponding to the preselected teaching method, and at the same time obtain the historical content corresponding to the historical personnel, and compare the historical content with the content input by the teaching staff to screen the historical personnel whose content is the same as that input by the teaching staff. Here, the same content means that it belongs to the same type as the content input by the current teaching staff. Taking the above input "derivation process of the quadratic formula for finding roots" as an example, it is analyzed that the content is for mathematics teaching, then the corresponding historical personnel whose historical content is for mathematics teaching are obtained, denoted as the personnel to be analyzed, and calculate the proportion of the number of the personnel to be analyzed, and at the same time compare the obtained proportion of the number with the preset proportion value. The specific value of the preset proportion value is set by the operator according to experience, and screen the preselected teaching methods corresponding to the proportion of the number greater than the preset proportion value and denote them as the teaching methods to be analyzed;

[0064] Then obtain the similarity between the teaching method to be analyzed and the teaching staff, and select the teaching method to be analyzed corresponding to the largest similarity as the standard to generate teaching standard information, and at the same time transmit it to the teaching information output module.

[0065] The teaching information output module is used to display the obtained teaching standard information to the corresponding teaching staff.

[0066] Example 2, please refer to Figure 2 , this application provides an intelligent teaching method based on a large language model, and this method specifically includes the following steps:

[0067] Step S1, obtaining the input content of the teaching staff, and obtaining the content to be analyzed through the large language model recognition, and classifying the teaching records to obtain the classification results, and screening the classification results based on the content to be analyzed as the standard to obtain the matching results, and the specific processing method is the same as the processing method of the personnel information analysis module in the first embodiment;

[0068] Step S2, analyzing the teaching effect of the teaching method in the matching result, assigning the teaching result value by assigning the change value of the personnel's score, assigning the class participation value by assigning the change value of the personnel's participation times, and summing the two to obtain the teaching effect value, and comparing with the preset value to screen the teaching method to obtain the pre-selected teaching method, and the specific processing method is the same as the processing process of the teaching method analysis module in Example 1;

[0069] Step S3, determine whether the pre-selected teaching method exists in the historical usage record of the teaching staff, generate a matching result of existence or non-existence, and for the matching result, eliminate the same pre-selected teaching method to obtain the screening teaching method, and calculate the similarity between the screening teaching method and the teaching staff, select the screening teaching method with the largest similarity as the standard, generate teaching standard information, and the specific processing method is the same as the processing method of the comprehensive selection processing module in Example 1;

[0070] Step S4: For those that do not have a matching result, analyze the historical content of the historical personnel corresponding to the pre-selected teaching method, select the historical personnel with the same content as the teaching personnel input as the personnel to be analyzed, and calculate the proportion of the number of personnel to be analyzed. At the same time, select based on the similarity to generate teaching standard information. The specific processing method is the same as the processing method of the comprehensive selection processing module in Example 1.

[0071] Some of the data in the above formulas are calculated by taking their numerical values, and are not substituted into parameter units for calculation. At the same time, the contents not described in detail in this specification belong to the existing technologies known to those skilled in the art.

[0072] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An intelligent teaching system based on a large language model, characterized in that: include: The personnel information analysis module is used to analyze the basic information transmitted by the teaching personnel information collection module, identify the input content of the teaching personnel through the large language model to obtain the content to be analyzed, and classify the teaching records to obtain the classification results, and filter the classification results based on the content to be analyzed to obtain the matching results, and transmit them to the teaching method analysis module; The teaching method analysis module analyzes the teaching effect of the teaching method in the matching result, obtains the teaching result value by assigning the change value of the personnel's score, and obtains the class participation value by assigning the change value of the personnel's participation times, and sums the two to obtain the teaching effect value, and selects the teaching method with a teaching effect value greater than the preset value to obtain the pre-selected teaching method, and transmits it to the comprehensive selection processing module; The comprehensive selection processing module is used to analyze the pre-selected teaching methods, determine whether the pre-selected teaching methods exist in the historical usage records of the teaching staff, generate a matching result of existence or non-existence, and for the matching results, eliminate the same pre-selected teaching methods to obtain the screening teaching methods, and calculate the similarity between the screening teaching methods and the teaching staff, select the screening teaching method with the greatest similarity as the standard, and generate teaching standard information; For those that do not have a matching result, the historical content of the historical personnel corresponding to the pre-selected teaching method is analyzed, and the historical personnel with the same content as the teaching personnel input are screened out as the personnel to be analyzed, and the proportion of the number of personnel to be analyzed is calculated. At the same time, the screening is performed based on the similarity, and the teaching standard information is generated and transmitted to the teaching information output module.

2. According to claim 1, a smart teaching system based on a large language model is characterized in that: It also includes a teaching staff information collection module and a teaching information output module; The teaching staff information collection module is used to transmit the basic information of the teaching staff to the staff information analysis module, and the basic information includes the input content corresponding to the teaching staff and the historical achievements of the teaching staff; The teaching information output module is used to display the acquired teaching standard information to the corresponding operator.

3. The intelligent teaching system based on a large language model according to claim 1, characterized in that: The specific method for the personnel information analysis module to obtain the matching result is: The input content in the basic information is obtained and input into the large language model for analysis to obtain the content to be analyzed. Then all the teaching records in the system are obtained and the teaching records are classified into the same type. The classification result is recorded as i, and i=1, 2, ..., j, where j represents the type corresponding to the classification result. At the same time, the teaching method corresponding to the classification result i is obtained and recorded as ic, and c=1, 2, ..., v, where v represents the number of teaching methods; Then, the classification result i is screened based on the content to be analyzed to obtain matching results, and matching result information is generated at the same time.

4. The intelligent teaching system based on a large language model according to claim 1, characterized in that: The specific method for the teaching method analysis module to obtain the teaching performance value is: Get all the personnel corresponding to teaching method c, and get the change values ​​of the corresponding scores before and after, and calculate the sum of all the change values ​​of the scores before and after, and calculate the corresponding mean at the same time, and match the obtained mean with the corresponding assignment interval to get the teaching score assignment, and the assignment interval here is obtained based on different data tests.

5. The intelligent teaching system based on a large language model according to claim 1, characterized in that: The specific method for the teaching method analysis module to obtain the classroom participation value is: Get the number of participants before teaching and the number of parameters after teaching corresponding to teaching method c, and calculate the corresponding difference in parameter times. Similarly, assign the difference in parameter times in the same way as the assignment method of personnel scores to obtain the class participation assignment.

6. The intelligent teaching system based on a large language model according to claim 1, characterized in that: The specific method for the teaching method analysis module to obtain the pre-selected teaching method is: The obtained teaching performance assignment and classroom participation assignment are summed up to obtain the teaching effect value Gc corresponding to the teaching method c, and the specific calculation formula is Gc=Fc×a1+Kc×a2, where Fc represents the teaching performance assignment, Kc represents the classroom participation assignment, and a1 and a2 are both preset proportional coefficients; At the same time, the teaching effect value Gc corresponding to the teaching method c is compared with the preset value, and the teaching method corresponding to the teaching effect value Gc greater than the preset value is selected and recorded as the pre-selected teaching method, and the pre-selected teaching method information is generated.

7. The intelligent teaching system based on a large language model according to claim 1, characterized in that: The specific method for the comprehensive selection processing module to generate the existence or non-existence of a matching result is: Obtain the historical usage records corresponding to the teaching staff, and determine whether there is a match with the pre-selected teaching method information in the historical usage records. If the corresponding pre-selected teaching method exists in the historical usage records, a matching result is generated, otherwise a non-matching result is generated.

8. The intelligent teaching system based on a large language model according to claim 1, characterized in that: The specific method of the comprehensive selection processing module analyzing the existing matching results to generate teaching standard information is as follows: Eliminate the same pre-selected teaching methods, and obtain the remaining pre-selected teaching methods, which are recorded as screening teaching methods and a, and a=1, 2, ..., b, where b represents the number of screening teaching methods; Obtain the basic information of the teaching staff, perform vector conversion on the input content and the screening teaching method in the basic information, and calculate the cosine similarity between the two vectors, recorded as Ya; At the same time, all cosine similarities Ya are obtained, and the screening teaching method corresponding to the maximum cosine similarity Ya value is selected as the standard to generate teaching standard information.

9. The intelligent teaching system based on a large language model according to claim 1, characterized in that: The specific method of the comprehensive selection processing module analyzing the non-existent matching results to generate teaching standard information is: Calculate the cosine similarity between the pre-selected teaching method and the content input by the teaching staff, then obtain the historical personnel and the corresponding historical content corresponding to the pre-selected teaching method, and compare the historical content with the content input by the teaching staff, and select the historical personnel with the same content as the teaching staff input as the personnel to be analyzed; The number ratio corresponding to the personnel to be analyzed is calculated, and the obtained number ratio is compared with the preset ratio value, and the pre-selected teaching methods corresponding to the number ratio greater than the preset ratio value are selected and recorded as the teaching methods to be analyzed; Then, the similarity between the teaching method to be analyzed and the teaching staff is obtained, and the teaching method to be analyzed corresponding to the maximum similarity is selected as the standard to generate teaching standard information.

10. A smart teaching method based on a large language model, executed by a smart teaching system based on a large language model as claimed in any one of claims 1 to 9, characterized in that: The method specifically comprises the following steps: Step S1, obtaining the input content of the teaching staff, and obtaining the content to be analyzed through the large language model recognition, and classifying the teaching records to obtain the classification results, and screening the classification results based on the content to be analyzed as the standard to obtain the matching results; Step S2, analyzing the teaching effects of the teaching methods in the matching results, assigning a teaching score value by assigning a change value of the personnel's score, assigning a class participation value by assigning a change value of the personnel's participation times, and summing the two to obtain a teaching effect value, and comparing it with a preset value to screen the teaching methods to obtain a pre-selected teaching method; Step S3, judging whether the pre-selected teaching method exists in the historical usage record of the teaching staff, generating a matching result of existence or non-existence, and for the matching result, eliminating the same pre-selected teaching method to obtain the screening teaching method, and calculating the similarity between the screening teaching method and the teaching staff, selecting the screening teaching method with the greatest similarity as the standard, and generating the teaching standard information; Step S4: For those without matching results, analyze the historical contents of historical personnel corresponding to the pre-selected teaching method, select historical personnel with the same content as the teaching personnel input as personnel to be analyzed, calculate the proportion of the number of personnel to be analyzed, and screen them based on the similarity to generate teaching standard information.

Images