Intelligent learning plan making method based on large model

Abstract

The invention discloses an intelligent learning plan making method based on a large model, and the method comprises the steps: enabling a teacher to classify teaching content knowledge points, and determining the test proportion of each classification; constructing a test paper generation agent, and generating test paper by the test paper generation agent through knowledge point classification and proportion; the test score of each knowledge point of the student is counted according to the answer condition of the student; setting a specific gravity adjustment rule, and constructing a specific gravity adjustment agent; and according to a specific gravity adjustment result output by the specific gravity adjustment agent, the test paper generation agent generates personalized test paper. According to the method, different test contents can be customized in a short time according to the conditions of the students by means of the agents of the AI large model, so that the test contents are more beneficial to consolidating knowledge for the students and assisting in improving mastering of knowledge points.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent education and relates to a method for intelligently formulating a learning plan based on a large model. Background Art

[0002] With the development of AI large model technology, there are more and more mature large models on the market. In addition to the dialogue ability, AI large models are also equipped with tools based on AI large models such as agents and workflows, which play a huge auxiliary role in many work processes.

[0003] In the education industry, the teaching method is usually one-to-many or many-to-many, and the same method is adopted for different students, which may lead to different degrees of knowledge acceptance among students and large performance differences. With the development of AI large model technology, personalized customization can be assisted by AI large models, but there is currently a lack of relevant personalized customization methods. Summary of the Invention

[0004] In view of the above problems, the present invention proposes a method for intelligently formulating a learning plan based on a large model, which well solves the problems in the prior art.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A method for intelligently formulating a learning plan based on a large model, comprising:

[0007] The teacher classifies the knowledge points of the teaching content and determines the test ratio of each classification;

[0008] Build a test paper generation agent, and the test paper generation agent generates a test paper through knowledge point classification and ratio;

[0009] Statistical scores of each knowledge point of the student through the student's answer situation;

[0010] Set a ratio adjustment rule and build a ratio adjustment agent;

[0011] According to the ratio adjustment result output by the ratio adjustment agent, the test paper generation agent generates a personalized test paper.

[0012] Optionally, it further includes:

[0013] Build a questionnaire agent, generate a questionnaire through the questionnaire generation agent, and enable students to select the teaching style labels of the teacher;

[0014] Match the knowledge point classification with the teaching teacher;

[0015] Recommend courses on knowledge points according to the test score situation.

[0016] Optionally, the recommended knowledge point teaching courses according to the test score situation include:

[0017] Set the passing score line for the knowledge point test;

[0018] When the knowledge point test score is not less than the passing score line, recommend the teaching courses with the same teaching style label as the one selected by the student;

[0019] When the knowledge point test score is less than the passing score line, recommend the teaching courses with different teaching style labels from the one selected by the student.

[0020] Optionally, the proportion adjustment rule is:

[0021]

[0022] Among them, ω ι is the original proportion of the knowledge point, is the adjustment coefficient, s ι is the knowledge point test score.

[0023] Optionally, multiple groups of teaching style labels are set for the questionnaire, and each group includes two labels with opposite styles.

[0024] Optionally, the upper limit of the proportion adjustment is ω max :

[0025] ω max = -0.0625·α + 1.125

[0026] Among them, α is the total number of knowledge point classifications.

[0027] Optionally, the passing score line τ is:

[0028] τ = μ - κσ

[0029] Among them, μ is the average knowledge point test score, σ is the standard deviation, and κ is the knowledge point difficulty coefficient.

[0030] Optionally, establish a knowledge point library, update the knowledge point library according to the knowledge point test score situation, store the unmastered knowledge points, and delete the mastered knowledge points.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. By means of the intelligent agent of the AI large model, different test contents can be customized according to the situation of students in a short time, making the test contents more helpful for students to consolidate knowledge, and by comparing the teacher's label with the student's knowledge point test score, evaluating the absorption degree of students' knowledge points under different teaching styles in the teaching courses, and recommending the courses to assist in improving the mastery of knowledge points;

[0033] 2. Adjust the proportion of the next test based on the proportion of the original knowledge points and the test scores of the knowledge points, so that the knowledge points with lower test scores account for a larger proportion, enhance the test proportion of the knowledge with lower scores, and then increase the mastery of the knowledge points with lower scores. Moreover, the proportions of different knowledge points are dynamically adjusted to strengthen the memory of the knowledge points. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flowchart of the student test part of the embodiment of the present invention;

[0035] Figure 2 is a flowchart of the course recommendation part of the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] 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. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0037] Please refer to Figure 1 , which discloses an intelligent learning plan formulation method based on a large model according to an embodiment of the present invention, including:

[0038] The teacher classifies the knowledge points of the teaching content and determines the test proportion of each classification;

[0039] Build a test paper generation intelligent agent, and the test paper generation intelligent agent generates a test paper through knowledge point classification and proportion;

[0040] Statistically calculate the test scores of each knowledge point of the student based on the student's answer situation;

[0041] Set a proportion adjustment rule and build a proportion adjustment intelligent agent;

[0042] According to the proportion adjustment result output by the proportion adjustment intelligent agent, the test paper generation intelligent agent generates a personalized test paper.

[0043] Specifically, by classifying and setting the proportion of knowledge points, when the test paper generation intelligent agent generates a test paper, different score proportions of tests are set for different knowledge points. After the test, the mastery of the knowledge points by the student is obtained through the scores of each knowledge point, and the proportion of the next test is adjusted according to the mastery situation, and different test contents are output for each student's situation.

[0044] In this way, by means of the agents of the large AI model, different test contents can be customized according to the situation of students in a short time, making the test contents more helpful for students to consolidate knowledge and assisting in improving their mastery of knowledge points.

[0045] In some feasible ways, the proportion adjustment rule is as follows:

[0046]

[0047] Among them, ω ι is the original knowledge point proportion, is the adjustment coefficient, s l is the test score of the knowledge point, is the sum of the products of different original knowledge point proportions and the adjustment coefficient.

[0048] It should be understood that by adjusting the proportion of the next test through the original knowledge point proportion and the test score of the knowledge point, the knowledge points with lower test scores will occupy a larger proportion, enhancing the test proportion of the knowledge with lower scores, and then increasing the mastery degree of the knowledge points with lower scores. Moreover, the proportions of different knowledge points are in dynamic adjustment to strengthen the memory of the knowledge points.

[0049] When new knowledge points appear, by setting a basic proportion and normalizing it with the adjusted proportion of the knowledge points:

[0050]

[0051] Among them, ω1 represents the adjusted proportion of the original knowledge points, ∑ω j represents the sum of the adjusted proportions of the original knowledge points, ω x represents the basic proportion set for the new knowledge point. When calculating the adjusted proportion of the new knowledge point, ω1 is ω x .

[0052] Furthermore, the upper limit of proportion adjustment is ω max :

[0053] ω max =-0.0625·α + 1.125

[0054] Among them, α is the total number of knowledge point classifications.

[0055] It should be understood that by setting the upper limit of proportion adjustment, the unlimited increase of the proportion of unmastered knowledge points is restricted, and the compression of the proportions of other knowledge points is reduced. When the proportion of a certain knowledge point reaches the upper limit, the adjusted proportion is the upper limit value, and the proportions of other knowledge points are readjusted proportionally.

[0056] The method for setting the upper limit of specific gravity adjustment is as follows: through the total number of knowledge point classifications α, set the maximum allowable specific gravity for a single knowledge point, construct a regression equation through at least two total numbers of knowledge point classifications α and the maximum allowable specific gravity for a single knowledge point, and make adjustments.

[0057] As a specific implementation manner of an intelligent learning plan formulation method based on a large model provided by the application, it further includes:

[0058] Construct a questionnaire intelligent agent, generate a questionnaire through the questionnaire generation intelligent agent, and enable students to select the teaching style labels of teachers;

[0059] Match the knowledge point classifications with the teaching teachers;

[0060] Recommend knowledge point teaching courses according to the test score situation.

[0061] It should be understood that in the learning plan, in addition to tests, it also includes the absorption degree of knowledge points in the usual teaching courses. By establishing the style labels of students for the teaching teachers in the questionnaire, when recommending courses, recommend according to the style labels to improve students' mastery of knowledge points.

[0062] In some feasible ways, there are multiple groups of teaching style labels for the questionnaire, and each group includes two labels with opposite styles. Through labels with opposite styles, such as serious and humorous, strict and lenient, etc., the style differentiation is made more obvious, which is convenient for classification and recommendation.

[0063] Furthermore, recommending knowledge point teaching courses according to the test score situation includes:

[0064] Set the passing line for knowledge point test scores;

[0065] When the knowledge point test score is not less than the passing line, recommend the teaching course with the same teaching style label as the one selected by the student;

[0066] When the knowledge point test score is less than the passing line, recommend the teaching course with a different teaching style label from the one selected by the student.

[0067] It should be understood that through the style labels, the teaching methods of teachers are labeled, and the mastery degree of students under the teaching style of this label is evaluated through the knowledge point test scores. When the mastery degree is poor, recommend courses with other label teaching styles for them to change.

[0068] In some feasible ways, the passing line τ is:

[0069] τ = μ - κσ

[0070] Wherein, μ is the average knowledge point test score, σ is the standard deviation, and κ is the knowledge point difficulty coefficient.

[0071] It should be understood that by setting the average knowledge point test score and the standard deviation, the passing line is set to conform to the learning situations of more students, and by setting the knowledge point difficulty coefficient, according to the difficulty of the knowledge point, the passing line of this knowledge point conforms to the required mastery level. Among them, the value of κ is selected by the teacher according to the actual knowledge point, 0.5 < κ < 1.5. Among them, a value of 1 indicates that the knowledge point is a knowledge point of conventional difficulty. The smaller the value, the simpler it is, and the larger the value, the more difficult it is.

[0072] Furthermore, it also includes establishing a knowledge point library, updating the knowledge point library according to the knowledge point test score situation, storing the unmastered knowledge points, and deleting the mastered knowledge points.

[0073] It should be understood that by establishing a knowledge point library and saving the unmastered knowledge points, students can more intuitively observe their own mastery situations and more easily find the unmastered knowledge points for learning. And the knowledge point library is dynamically updated according to the score situation, so that the knowledge points that are forgotten after being mastered can also enter the knowledge point library during the update, strengthening the memory points.

[0074] Among them, the unmastered knowledge points can be determined by setting a passing line according to the test score.

[0075] Although the present invention has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for intelligently formulating a learning plan based on a large model, characterized in that: include: Teachers classify the teaching content and knowledge points, and determine the test weight of each category; Construct an intelligent test paper generation agent, which generates test papers by classifying and weighting knowledge points; According to the students' answers, the test scores of each knowledge point are counted; Set weight adjustment rules and build a weight adjustment agent; According to the weight adjustment results output by the weight adjustment agent, the test paper generation agent generates a personalized test paper.

2. According to the method of intelligently formulating a learning plan based on a large model according to claim 1, it is characterized in that: Also includes: Construct a questionnaire agent, and use the questionnaire generation agent to generate a questionnaire, so that students can choose the teacher's teaching style label; Classify knowledge points and match them with the instructors; Based on the test scores, knowledge point teaching courses are recommended.

3. The method for intelligently formulating a learning plan based on a large model according to claim 2, characterized in that: According to the test scores, the recommended knowledge point courses include: Set the passing score for knowledge point tests; If the knowledge point test score is not less than the passing score, we recommend courses with the same teaching style label as the one selected by the student. When the knowledge point test score is lower than the passing score, a course with a different teaching style label from the one selected by the student is recommended.

4. The method for intelligently formulating a learning plan based on a large model according to claim 1, characterized in that: The weight adjustment rule is: Among them, ω l is the proportion of original knowledge points, is the adjustment coefficient, s l Test scores for knowledge points.

5. The method for intelligently formulating a learning plan based on a large model according to claim 3, characterized in that: The teaching style labels of the questionnaire are provided with multiple groups, and each group includes two labels with opposite styles.

6. The method for intelligently formulating a learning plan based on a large model according to claim 4, characterized in that: The upper limit of the specific gravity adjustment is ω max : oh max =-0.0625·α+1.125 Among them, α is the total number of knowledge point categories.

7. The method for intelligently formulating a learning plan based on a large model according to claim 3, characterized in that: The passing line τ is: τ=μ-κσ Among them, μ is the average knowledge point test score, σ is the standard deviation, and κ is the knowledge point difficulty coefficient.

8. The method for intelligently formulating a learning plan based on a large model according to claim 4, characterized in that: Establish a knowledge point database, update the knowledge point database according to the knowledge point test scores, store the knowledge points that have not been mastered, and delete the knowledge points that have been mastered.

Images