A method and device for intelligent question setting based on deep learning model

The intelligent question-setting method using a deep learning model addresses the issue of random question selection by employing a decision tree model to tailor test papers to individual students' needs, effectively increasing the coverage of knowledge gaps to 96.72%.

CN118279101BActive Publication Date: 2024-12-06苏州国舜网络安全有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410321726.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-03-20
Publication Date
2024-12-06
Estimated Expiration
2044-03-20

AI Technical Summary

Technical Problem

Traditional question-setting methods in education platforms fail to specifically assess areas where students have not yet mastered, as they randomly select questions without considering individual student proficiency.

Method used

An intelligent question-setting method using a deep learning model, specifically a decision tree model, that determines test questions based on student identification information and preset feature values, including test question topics, accuracy, and difficulty coefficients, to generate tailored test papers.

Benefits of technology

Enables automatic question setting that accurately targets students' knowledge gaps, improving the coverage of weak areas to 96.72% through personalized test paper generation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118279101B_ABST
    Figure CN118279101B_ABST
Patent Text Reader

Abstract

The present invention provides an intelligent question-setting method and device based on a deep learning model, which relates to the field of neural network technology. The method comprises: determining the identification information of a student and a preset feature value range; inputting the identification information of the student and the preset feature value range into a decision tree model, determining the test question features and their values ​​corresponding to the identification information of the student, and determining at least one target test question identifier according to the values ​​of the test question features and the preset feature value range; wherein the decision tree model comprises a plurality of nodes, each node represents the features of each test question of a single student, and the features include the test question identifier, test question topic, correct rate, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each branch includes feature values, including test question topic values, difficulty coefficient values, correct rate values, and question type values; and generating a corresponding test paper according to at least one target test question identifier.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of neural network technology, and in particular to an intelligent question-setting method and device based on a deep learning model. Background Art

[0002] With the vigorous development of Internet technology and artificial intelligence technology, artificial intelligence technology has begun to show its powerful capabilities in the field of education and has become a popular application field.

[0003] In traditional training platforms, test questions are generally randomly selected from the question bank, or different test questions are selected from the question bank in a fixed ratio according to the focus of the test content. This kind of test based on the traditional way of setting questions and assembling papers often does not specifically assess the areas that students have not yet mastered in their studies. Summary of the invention

[0004] The present invention provides an intelligent question-setting method and device based on a deep learning model, which are used to solve the technical defect that the question-setting method in the prior art cannot specifically test different students.

[0005] The present invention provides an intelligent question-setting method based on a deep learning model, comprising:

[0006] Determine the identification information of the students and the preset characteristic value range;

[0007] Inputting the identification information of the student and the preset feature value range into the decision tree model, determining the test question features and their values ​​corresponding to the identification information of the student, and determining at least one target test question identification according to the values ​​of the test question features and the preset feature value range;

[0008] The decision tree model includes a plurality of nodes, each node represents the characteristics of each test question of a single student, and the characteristics include the test question topic, correctness, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each branch includes the value of the characteristic, including the test question topic value, difficulty coefficient value, correctness value, and question type value;

[0009] Generate a corresponding test paper according to at least one target test question identifier.

[0010] According to an intelligent question setting method based on a deep learning model provided by the present invention, at least one target question identifier to be selected is determined according to the value of the question feature and the preset feature value range, including:

[0011] Determine a weighted value corresponding to each test question according to the value of the test question feature and the preset feature value range; wherein the higher the weighted value, the greater the probability of being selected as the target test question mark;

[0012] At least one target test question identifier is determined according to the weighted value corresponding to each test question.

[0013] According to an intelligent question-setting method based on a deep learning model provided by the present invention, the training method of the decision tree model includes:

[0014] Collecting students' original learning data, wherein the original learning data includes: students' personal information, number of exams taken, previous exam scores, and test question classification;

[0015] Extracting original features and original feature values ​​of the original learning data; wherein the original features include original test question topics, original accuracy rates, question types corresponding to the original test questions, and original test question difficulty coefficients; and the original feature values ​​include original test question topic values, original difficulty coefficient values, original accuracy rate values, and original question type values;

[0016] The extracted original features are used as nodes, and the original feature values ​​are added to the decision tree model as branches corresponding to the nodes, and the parameters of the decision tree model are adjusted until the fit of the decision tree model for the features is higher than the set threshold.

[0017] According to an intelligent question-setting method based on a deep learning model provided by the present invention, after collecting the original learning data of the students, the method further includes:

[0018] Perform data cleaning, missing value processing and outlier processing on the original learning data to obtain preprocessed learning data;

[0019] The preprocessed learning data is input into a decision tree model to extract features.

[0020] According to an intelligent question-setting method based on a deep learning model provided by the present invention, after generating a corresponding test paper according to at least one target test question identifier, the method further includes:

[0021] Scores are given according to the students' answers to the test paper, and the features and feature values ​​of the current answer results are extracted and continue to be input into the decision tree model as training data for the decision tree model.

[0022] The present invention also provides an intelligent question-setting device based on a deep learning model, comprising:

[0023] A determination module is used to determine the identification information of the trainee and the preset characteristic value range;

[0024] A model processing module, used to input the identification information of the student and the preset feature value range into the decision tree model, determine the test question features and their values ​​corresponding to the identification information of the student, and determine at least one target test question identification according to the values ​​of the test question features and the preset feature value range;

[0025] Among them, the decision tree model includes multiple nodes, each node represents the characteristics of each test question of a single student, and the characteristics include the test question topic, accuracy, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each of the branches includes the value of the characteristic, including the test question topic value, difficulty coefficient value, accuracy value, and question type value.

[0026] The test paper generating module is used to generate a corresponding test paper according to at least one target test question identifier.

[0027] According to the intelligent question-setting device based on the deep learning model provided by the present invention, the model processing module is specifically used to: determine the weighted value corresponding to each test question according to the value of the test question feature and the preset feature value range; wherein the higher the weighted value, the greater the probability of being selected as the target test question mark;

[0028] At least one target test question identifier is determined according to the weighted value corresponding to each test question.

[0029] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any of the above-described intelligent question-setting methods based on a deep learning model are implemented.

[0030] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the intelligent question-setting method based on the deep learning model as described in any one of the above are implemented.

[0031] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of any of the above-mentioned intelligent question-setting methods based on deep learning models.

[0032] The intelligent question-setting method and device based on a deep learning model provided by the present invention inputs the identification information of the students and the preset feature value range into the decision tree model, determines the test question features and their values ​​corresponding to the identification information of the students, and determines at least one target test question identification according to the values ​​of the test question features and the preset feature value range, so that different target test questions can be selected according to the identification information of different students, thereby realizing automatic question setting for different students. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 is a schematic diagram of a decision tree model provided by an embodiment of the present invention;

[0035] Figure 2 It is a flowchart of a training method for a decision tree model provided by an embodiment of the present invention;

[0036] Figure 3 It is a flowchart of an intelligent question-setting method based on a deep learning model provided by an embodiment of the present invention;

[0037] Figure 4 is a schematic diagram of a specific example of an intelligent question-setting method based on a deep learning model provided by an embodiment of the present invention;

[0038] Figure 5 is a structural schematic diagram of an intelligent question-setting device based on a deep learning model provided by an embodiment of the present invention;

[0039] Figure 6 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Combine the following Figure 1-Figure 5 Describe the intelligent question-setting method and device based on the deep learning model according to the embodiment of the present invention.

[0042] The intelligent question-setting method disclosed in the embodiment of the present invention uses a deep learning model as the main tool for setting questions, so as to achieve the purpose of automatically setting questions for different students.

[0043] See also Figure 1 , Figure 1 A schematic diagram of a decision tree model of this embodiment is shown.

[0044] The deep learning model in this embodiment can be a decision tree model. The decision tree model is a simple and easy-to-use non-parametric classifier. It is a supervised learning algorithm, mainly used for classification problems, but sometimes also used for regression problems. Select discriminative variables from the data, split the data set into two or more subsets, split step by step, and finally form a "tree". Each leaf node of the "tree" represents the final prediction result of the branch. It does not require any prior assumptions about the data, has a fast calculation speed, is easy to interpret, and is robust.

[0045] In this embodiment, the decision tree model includes multiple nodes, each node represents the characteristics of each test question of a single student, and the characteristics include the test question topic, accuracy, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each branch includes the value of the characteristic, including the test question topic value, difficulty coefficient value, accuracy value, and question type value.

[0046] See also Figure 2 , the training method of the decision tree model in this embodiment includes:

[0047] 201. Collect the original learning data of the students, wherein the original learning data includes: the student's personal information, the number of exams taken, the test scores, and the classification of test questions.

[0048] The original learning data may be the test data of the students in previous exams. In a specific example, in order to facilitate the collection of students' learning data, a learning system may be designed to generate learning content and test papers for students after the students log in with their own account and password. After the students complete the test paper and submit it, the system itself will judge the test paper and automatically give a score.

[0049] Furthermore, after collecting the original learning data of the students, the method also includes: performing data cleaning, missing value processing and outlier processing on the original learning data to obtain preprocessed learning data; and inputting the preprocessed learning data into a decision tree model to extract features.

[0050] For example, if the full score of this exam is 100 points, and the regular exam score is 80 points, but the deviation value of a certain exam score is higher than ±60% of the average score deviation value, it can be considered as abnormal data and will not be counted.

[0051] 202. Extract original features and original feature values ​​of the original learning data; wherein the original features include original test question topics, original accuracy rates, question types corresponding to the original test questions, and original test question difficulty coefficients; and the original feature values ​​include original test question topic values, original difficulty coefficient values, original accuracy rate values, and original question type values.

[0052] The value of the test topic may be SQL injection, cross-site vulnerability (XSS) injection or cross-site request forgery (CSRF).

[0053] The question type can be multiple choice, fill-in-the-blank, essay, etc.

[0054] The difficulty level can be low, medium, or high.

[0055] The accuracy rate can be 50%, 60%, 70%, or 80%.

[0056] 203. The extracted original features are added as nodes and the original feature values ​​are added as branches corresponding to the nodes to the decision tree model, and the parameters of the decision tree model are adjusted until the fit of the decision tree model to the features is higher than a set threshold.

[0057] After obtaining the trained decision tree model, its effect can be further evaluated. For example, its performance in the classification of the same test questions under a specific difficulty coefficient can be tested. Methods such as cross-validation and confusion matrix can be used for evaluation.

[0058] For example, cross-validation means taking out most of the sample data from a given modeling sample to build a model, leaving a small part of the sample data to use the newly built model for prediction, and calculating the prediction error of this small part of the sample data, and recording the sum of their squares.

[0059] Through steps 201 to 203, a decision tree model can be established and used for intelligent question setting in subsequent steps.

[0060] See also Figure 3 The intelligent question-setting method of this embodiment includes:

[0061] 301. Determine the identification information of the student and the preset characteristic value range.

[0062] For example, the characteristics of each test question include: test topic, accuracy, question type and test difficulty coefficient. Among them, the values ​​of test topic and question type are relatively fixed and will not change with the changes of students; but the accuracy and difficulty coefficient will vary according to different students.

[0063] For different students, the correct rate and difficulty coefficient of the same test question may be different from each other; for example, for student Zhang San, the correct rate of test question A is 90%, and the difficulty coefficient is low; but for student Li Si, the correct rate of test question A is 60%, and the difficulty coefficient is high. Therefore, the solution of this embodiment needs to maintain the test question features and feature values ​​corresponding to the test questions of the students for different students, so that the questions can be determined according to the identification information of the students in the subsequent steps.

[0064] The feature value range is used to determine which test questions are selected for this test, for example, test questions with a correct rate of 60% to 80%, or test questions with a medium or high difficulty coefficient, so as to screen the test questions and determine the target test question identification.

[0065] 302. Input the student's identification information and preset feature value range into the decision tree model, determine the test question features and their values ​​corresponding to the student's identification information, and determine at least one target test question identification based on the test question feature values ​​and the preset feature value range.

[0066] The decision tree model includes multiple nodes, each node represents the characteristics of each test question of a single student, and the characteristics include the test question identifier, test question topic, accuracy, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each of the branches includes the values ​​of the characteristics, including the test question topic value, difficulty coefficient value, accuracy value, and question type value.

[0067] For example, if the student identification information input is "Zhang San", or the ID corresponding to "Zhang San" is "00001", and the feature value range is "difficulty coefficient is intermediate, accuracy rate is 60% to 70%", then a match is performed in the decision tree model based on the feature value range and the student identification information to determine at least one target test question identifier.

[0068] Specifically, step 302 includes: determining a weighted value corresponding to each test question based on the value of the test question feature and a preset feature value range; wherein, the higher the weighted value, the greater the probability of being selected as the target test question identifier; and determining at least one of the target test question identifiers based on the weighted value corresponding to each test question.

[0069] For each test question, there may be only one feature that meets the feature value range, or there may be multiple features that meet the feature value range, so for each test question, the corresponding weighted value is output through the decision tree model.

[0070] 303. Generate a corresponding test paper according to at least one target test question identifier.

[0071] Specifically, the target test question identifier may be used to search in the question bank, and after the corresponding test question is found, the test paper may be generated by combining the questions.

[0072] The intelligent question-setting method based on a deep learning model provided by an embodiment of the present invention determines the question features and their values ​​corresponding to the student's identification information by inputting the student's identification information and a preset feature value range into a decision tree model, and determines at least one target question identification based on the value of the question features and the preset feature value range, so that different target questions can be selected according to the identification information of different students, thereby realizing automatic question setting for different students.

[0073] After step 303, the method further includes: scoring the students' answers to the test paper, extracting the features and feature values ​​of the current answer results, and continuing to input them into the decision tree model as training data of the decision tree model.

[0074] In order to facilitate understanding of the technical solution of this embodiment, a specific example is listed below to schematically illustrate the method of this embodiment.

[0075] See also Figure 4 , the method of this embodiment includes:

[0076] 401. Collection of students’ personal information

[0077] Table 1

[0078]

[0079] 402. Data Preprocessing

[0080] Since Xu xx's score in this test was 12 points, which was -34 points lower than the average score of 46 points, and 170% higher than the average deviation value, this data was considered abnormal data and was not collected for sampling.

[0081] 403. Extracting features

[0082] Table 2

[0083]

[0084] 404. Model Training

[0085] Through repeated test training, relevant data on the test scores are extracted and combined with the decision tree model to continuously revise the test results.

[0086] Table 3

[0087]

[0088] 405. Model Evaluation

[0089] Based on the continuous evaluation of training results, the test questions can solve the students' real level of knowledge and improve their knowledge mastery. According to the preset test difficulty coefficient, the test questions are automatically generated from the question bank according to weighted calculation.

[0090] Table 4

[0091]

[0092] For example, assuming that the preset difficulty coefficient is intermediate and the test accuracy is 70%-80%, the weighted values ​​corresponding to the determined test questions are shown in Table 4. The weighted values ​​in Table 4 are similar to the probability of the question being selected. The higher the weighted value, the higher the probability of the automatic test paper being selected, thereby completing the automatic test paper function.

[0093] As shown in Table 4, question 1 has a difficult coefficient and a correct rate of 40%, so its weight is 1; question 2 has a medium coefficient and a correct rate of 60%, so its weight is 2; question 3 has an easy coefficient and a correct rate of 86%, so its weight is 2. Therefore, in the process of intelligent question setting, questions 2 and 3 will be selected first.

[0094] The solution of this embodiment can generate different test papers for each student according to their individual needs. After several rounds of simulated test questions, the accuracy of automatic test paper generation has been greatly improved, and the coverage of knowledge areas that students are weak in has been increased to 96.72%.

[0095] The following is a description of the intelligent question-setting device based on a deep learning model provided in an embodiment of the present invention. The intelligent question-setting device based on a deep learning model described below and the intelligent question-setting method based on a deep learning model described above can be referenced to each other.

[0096] The embodiment of the present invention provides an intelligent question-setting device based on a deep learning model. Figure 5 ,include:

[0097] Determination module 501, used to determine the identification information of the trainee and the preset characteristic value range;

[0098] The model processing module 502 is used to input the identification information of the student and the preset feature value range into the decision tree model, determine the test question features and their values ​​corresponding to the identification information of the student, and determine at least one target test question identification according to the values ​​of the test question features and the preset feature value range;

[0099] Among them, the decision tree model includes multiple nodes, each node represents the characteristics of each test question of a single student, and the characteristics include the test question topic, accuracy, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each of the branches includes the value of the characteristic, including the test question topic value, difficulty coefficient value, accuracy value, and question type value.

[0100] The test paper generating module 503 is used to generate a corresponding test paper according to at least one target test question identifier.

[0101] Optionally, the model processing module 502 is specifically used to: determine a weighted value corresponding to each test question based on the value of the test question feature and a preset feature value range; wherein, the higher the weighted value, the greater the probability of being selected as the target test question identification; and determine at least one of the target test question identifications based on the weighted value corresponding to each test question.

[0102] Optionally, the device further comprises a training module, for:

[0103] Collecting students' original learning data, wherein the original learning data includes: students' personal information, number of exams taken, previous exam scores, and test question classification;

[0104] Extracting original features and original feature values ​​of the original learning data; wherein the original features include original test question topics, original accuracy rates, question types corresponding to the original test questions, and original test question difficulty coefficients; and the original feature values ​​include original test question topic values, original difficulty coefficient values, original accuracy rate values, and original question type values;

[0105] The extracted original features are used as nodes, and the original feature values ​​are added to the decision tree model as branches corresponding to the nodes, and the parameters of the decision tree model are adjusted until the fit of the decision tree model for the features is higher than the set threshold.

[0106] Optionally, the device also includes: a screening module, which is used to perform data cleaning, missing value processing and outlier processing on the original learning data after collecting the original learning data of the students to obtain pre-processed learning data; and input the pre-processed learning data into a decision tree model to extract features.

[0107] Optionally, the device also includes a subsequent processing module, which is used to: after generating a corresponding test paper according to at least one target test question identifier, score the test paper according to the students' answers to the test paper, and extract the characteristics and characteristic values ​​of the current answer results, and continue to input them into the decision tree model as training data for the decision tree model.

[0108] The intelligent question-generating device based on a deep learning model provided by the present invention determines the test question features and their values ​​corresponding to the trainee's identification information by inputting the trainee's identification information and a preset feature value range into a decision tree model, and determines at least one target test question identifier based on the test question feature values ​​and the preset feature value range, thereby selecting different target test questions according to the identification information of different trainees, thereby realizing automatic question generation for different trainees.

[0109] Figure 6 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 6As shown, the electronic device may include: a processor (processor) 610, a communication interface (Communications Interface) 620, a memory (memory) 630 and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other through the communication bus 640. The processor 610 can call the logic instructions in the memory 630 to execute an intelligent question-setting method based on a deep learning model, which includes: determining the student's identification information and a preset feature value range; inputting the student's identification information and the preset feature value range into a decision tree model, determining the test question features and their values ​​corresponding to the student's identification information, and determining at least one target test question identification based on the values ​​of the test question features and the preset feature value range; wherein the decision tree model includes multiple nodes, each node represents the features of each test question of a single student, and the features include the test question topic, accuracy, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each of the branches includes the values ​​of the features, including the test question topic value, difficulty coefficient value, accuracy value, and question type value; and generating a corresponding test paper based on at least one target test question identification.

[0110] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0111] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the intelligent question-setting method based on the deep learning model provided by the above-mentioned methods, and the method includes: determining the identification information of the student and the preset feature value range; inputting the identification information of the student and the preset feature value range into a decision tree model, determining the test question features and their values ​​corresponding to the identification information of the student, and determining at least one target test question identification according to the values ​​of the test question features and the preset feature value range; wherein the decision tree model includes multiple nodes, each node represents the features of each test question of a single student, and the features include the test question topic, accuracy, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each of the branches includes the values ​​of the features, including the test question topic value, difficulty coefficient value, accuracy value, and question type value; according to at least one target test question identification, a corresponding test paper is generated.

[0112] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the intelligent question-setting method based on a deep learning model provided by the above-mentioned methods, the method comprising: determining the identification information of the student and a preset feature value range; inputting the identification information of the student and the preset feature value range into a decision tree model, determining the test question features and their values ​​corresponding to the identification information of the student, and determining at least one target test question identification based on the values ​​of the test question features and the preset feature value range; wherein the decision tree model comprises a plurality of nodes, each node representing the features of each test question of a single student, the features comprising the test question topic, accuracy, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, each of the branches comprising the values ​​of the features, including the test question topic value, difficulty coefficient value, accuracy value, and question type value; and generating a corresponding test paper based on at least one target test question identification.

[0113] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0114] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent question-setting method based on a deep learning model, characterized in that: include: Determine the identification information of the students and the preset characteristic value range; Inputting the identification information of the student and the preset feature value range into the decision tree model, determining the test question features and their values ​​corresponding to the identification information of the student, and determining at least one target test question identification according to the values ​​of the test question features and the preset feature value range; The decision tree model includes a plurality of nodes, each node represents the characteristics of each test question of a single student, and the characteristics include the test question identification, test question topic, correct rate, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each branch includes the value of the characteristic, including the test question topic value, difficulty coefficient value, correct rate value, and question type value; Generate a corresponding test paper according to at least one target test question identifier; The training method of the decision tree model includes: collecting original learning data of students, wherein the original learning data includes: personal information of students, number of examinations taken, previous examination scores and classification of test questions; extracting original features and original feature values ​​of the original learning data; wherein the original features include original test question topics, original accuracy rates, question types corresponding to original test questions, and original test question difficulty coefficients, and the original feature values ​​include original test question topic values, original difficulty coefficient values, original accuracy rates, and original question type values; adding the extracted original features as nodes and the original feature values ​​as branches corresponding to the nodes to the decision tree model, and adjusting the parameters of the decision tree model until the fitting degree of the decision tree model to the features is higher than a set threshold; Determining at least one target test question identifier to be selected according to the value of the test question feature and the preset feature value range, including: Determine a weighted value corresponding to each test question according to the value of the test question feature and the preset feature value range; wherein the higher the weighted value, the greater the probability of being selected as the target test question mark; At least one target test question identifier is determined according to the weighted value corresponding to each test question.

2. The intelligent question-setting method based on a deep learning model according to claim 1, characterized in that: After collecting the original learning data of the students, the method further includes: Perform data cleaning, missing value processing and outlier processing on the original learning data to obtain preprocessed learning data; The preprocessed learning data is input into a decision tree model to extract features.

3. The intelligent question-setting method based on a deep learning model according to claim 1, characterized in that: After generating a corresponding test paper according to at least one target test question identifier, the method further includes: Scores are given according to the students' answers to the test paper, and the features and feature values ​​of the current answer results are extracted and continue to be input into the decision tree model as training data for the decision tree model.

4. An intelligent question-setting device based on a deep learning model, characterized in that: include: A determination module is used to determine the identification information of the trainee and the preset characteristic value range; A model processing module, used to input the identification information of the student and the preset feature value range into the decision tree model, determine the test question features and their values ​​corresponding to the identification information of the student, and determine at least one target test question identification according to the values ​​of the test question features and the preset feature value range; The decision tree model includes a plurality of nodes, each node represents the characteristics of each test question of a single student, and the characteristics include the test question topic, correctness, question type, and test question difficulty coefficient corresponding to each test question; each node is connected to at least one branch, and each branch includes the value of the characteristic, including the test question topic value, difficulty coefficient value, correctness value, and question type value; A test paper generating module, used for generating a corresponding test paper according to at least one target test question identifier; The model processing module is specifically used to: determine the weighted value corresponding to each test question according to the value of the test question feature and the preset feature value range; wherein the higher the weighted value, the greater the probability of being selected as the target test question identifier; and determine at least one of the target test question identifiers according to the weighted value corresponding to each test question; The device also includes a training module, which is used to: collect original learning data of students, wherein the original learning data includes: students' personal information, number of exams taken, previous exam scores and test question topic classification; extract original features and original feature values ​​of the original learning data; wherein the original features include original test question topics, original accuracy rates, question types corresponding to original test questions, and original test question difficulty coefficients, and the original feature values ​​include original test question topic values, original difficulty coefficient values, original accuracy rates, and original question type values; add the extracted original features as nodes and the original feature values ​​as branches corresponding to the nodes to a decision tree model, and adjust the parameters of the decision tree model until the fit of the decision tree model to the features is higher than a set threshold.

5. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the intelligent question-setting method based on the deep learning model as described in any one of claims 1 to 3 are implemented.

6. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intelligent question-setting method based on a deep learning model as described in any one of claims 1 to 3 are implemented.

7. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the intelligent question-setting method based on a deep learning model as described in any one of claims 1 to 3 are implemented.