A method and system for auxiliary generation of test questions for computer-based examinations

Abstract

The present application provides a method and system for auxiliary generation of test questions for a computer-based test, comprising: obtaining a list of knowledge points specified in a test syllabus; selecting test questions that match the list of knowledge points from a test question bank based on the list of knowledge points and a preset difficulty level; performing random processing on the test questions to generate initial test questions, wherein the random processing includes: randomly selecting a question stem, scrambling the order of options, and randomizing numerical parameters; performing semantic optimization on the initial test questions using a pre-trained natural language processing model to generate target test questions; adding the target test questions in batches to a computer-based test system, and assigning a corresponding test question identifier to each test question, so that high-quality computer-based test questions can be efficiently created by intelligently generating and optimizing test questions, while ensuring the diversity and fairness of the test questions.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of computer-based examinations, and more particularly to a method and system for auxiliary generation of test questions for computer-based examinations. Background Art

[0002] In the traditional computer-based exam preparation process, teachers are required to manually write a large number of test questions to cover all the knowledge points in the course syllabus and design test questions according to different difficulty levels. This approach is not only time-consuming and labor-intensive, but also difficult to ensure the quality and diversity of the test questions, especially when the test question bank needs to be updated frequently to prevent students from memorizing the answers.

[0003] Existing methods for generating computer-based test questions usually rely on teachers' manual writing or simple random replacement of variables in templates. This method has the following major defects:

[0004] Inefficiency: Manually writing a large number of test questions consumes a lot of teachers' time and energy.

[0005] Unstable quality: Due to the lack of unified standards, manually written test questions may have problems such as unclear expression and mismatched difficulty.

[0006] Lack of personalization: The content and difficulty of test questions cannot be flexibly adjusted according to students' actual learning situation.

[0007] Insufficient security: The question bank is updated slowly and can be easily mastered by students in advance, which reduces the fairness of the exam. Summary of the invention

[0008] The embodiments of the present application provide a method and system for auxiliary generation of test questions for computer-based examinations, so as to solve the problems existing in the above-mentioned background technology.

[0009] In a first aspect, an embodiment of the present application provides a method for assisting in generating test questions for a computer-based test, comprising: obtaining a list of knowledge points specified in the test syllabus; selecting test questions that match the list of knowledge points from a test question bank based on the list of knowledge points and a preset difficulty level; performing random processing on the test questions to generate initial test questions, wherein the random processing includes: randomly selecting question stems, scrambling the order of options, and randomizing numerical parameters; performing semantic optimization on the initial test questions using a pre-trained natural language processing model to generate target test questions; adding the target test questions in batches to the computer-based test system, and assigning a corresponding test question identifier to each test question.

[0010] Furthermore, based on the knowledge point list and the preset difficulty level, test questions matching the knowledge point list are selected from the test question bank, including: calculating a weight value for each knowledge point in the knowledge point list, wherein the weight value is a weight value determined based on the importance and frequency of occurrence of the knowledge point in the examination syllabus; and selecting test questions matching the knowledge point list from the test question bank according to the difficulty level and the weight value of the knowledge point, wherein the difficulty level of the test question matches the preset difficulty level, and the knowledge point involved in the test question matches at least one knowledge point in the knowledge point list.

[0011] Furthermore, the test questions are randomized to generate initial test questions, including: for test questions containing numerical parameters, multimodal distribution fusion processing is performed on the numerical parameters using a preset distribution type to generate new numerical parameters, wherein the preset distribution type includes: normal distribution, uniform distribution or Poisson distribution; using a preset machine learning model and a real-time feedback mechanism, the new numerical parameters are adjusted to obtain target numerical parameters; using preset checking rules and a dynamic adjustment mechanism, the target numerical parameters are checked for validity, wherein the preset checking rules include: basic grammar checking rules, logical consistency checking rules and problem-solving feasibility checking rules; if the check passes, the original numerical parameters in the test questions containing numerical parameters are replaced with the target numerical parameters to generate the initial test questions.

[0012] Furthermore, the initial test questions are semantically optimized using a pre-trained natural language processing model to generate target test questions, including: inputting the initial test questions into the pre-trained natural language processing model so that the pre-trained natural language processing model outputs grammatical errors of the intermediate test questions, wherein the grammatical errors include: correcting spelling errors, punctuation errors and subject-verb inconsistency; using a multimodal natural language processing model and combining the text and context information of the initial test questions to correct the grammatical errors to obtain intermediate test questions; using a deep learning model to perform semantic similarity analysis on the initial test questions and the intermediate test questions to obtain analysis results; using the analysis results to adjust the expression structure of the intermediate test questions to obtain the target test questions.

[0013] Furthermore, the calculation formula of the weight value of the knowledge point is:

[0014] ;

[0015] in, For the The weight of each knowledge point, , , , , , and is the adjustment coefficient, For the The importance function of each knowledge point is For the The frequency function of knowledge points, For the The average score rate function of knowledge points, For the The difficulty function of each knowledge point is For the The correlation function of knowledge points, For the The context relevance function of each knowledge point, For the The time sensitivity function of each knowledge point.

[0016] Furthermore, the method further includes: determining the difficulty level of the knowledge point based on the weight value of the knowledge point and a preset difficulty level function; adjusting the test question based on the difficulty level; the preset difficulty level function is:

[0017] ;

[0018] in, Difficulty level The thresholds for difficulty levels, The highest difficulty level.

[0019] Furthermore, after generating the target test question, the target test question and related information are stored in a database, wherein the related information includes at least: the type of test question, the difficulty level, the knowledge points involved and the test question identifier; and a mapping relationship is established between the test question identifier and the test question information.

[0020] In the second aspect, an embodiment of the present application provides an auxiliary test question generation system for a computer-based test, comprising: an acquisition unit for acquiring a list of knowledge points specified in an examination syllabus; a matching unit for selecting test questions that match the list of knowledge points from a test question bank based on the list of knowledge points and a preset difficulty level; a randomization processing unit for performing randomization processing on the test questions to generate initial test questions, wherein the randomization processing includes: randomly selecting a question stem, scrambling the order of options, and randomizing numerical parameters; a semantic optimization unit for performing semantic optimization on the initial test questions using a pre-trained natural language processing model to generate target test questions; an adding unit for adding the target test questions in batches to the computer-based test system, and assigning a corresponding test question identifier to each test question.

[0021] In a third aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the method described in the first aspect above.

[0022] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, which, when executed by a computer, implements the method described in the first aspect above.

[0023] In an embodiment of the present invention, a method and system for auxiliary generation of test questions for a computer-based test are provided, including: obtaining a list of knowledge points specified in an examination syllabus; selecting test questions that match the list of knowledge points from a test question bank based on the list of knowledge points and a preset difficulty level; performing random processing on the test questions to generate initial test questions, wherein the random processing includes: randomly selecting a question stem, scrambling the order of options, and randomizing numerical parameters; performing semantic optimization on the initial test questions using a pre-trained natural language processing model to generate target test questions; adding the target test questions in batches to a computer-based test system, and assigning a corresponding test question identifier to each test question, so that high-quality computer-based test questions can be efficiently created by intelligently generating and optimizing test questions, while ensuring the diversity and fairness of the test questions.

[0024] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to 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 application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0026] Figure 1 A flowchart of a method for assisting in generating test questions for a computer-based test provided in an embodiment of the present application;

[0027] Figure 2 A schematic diagram of a system for assisting in generating test questions for a computer-based test provided in an embodiment of the present application;

[0028] Figure 3 A schematic diagram of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. 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.

[0030] Embodiment 1:

[0031] According to an embodiment of the present invention, an embodiment of a method for assisting in generating test questions for a computer-based test is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0032] Figure 1 is a flow chart of a method for auxiliary generation of test questions for a computer-based test according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0033] Step S102, obtaining a list of knowledge points specified in the examination syllabus;

[0034] Step S104, based on the knowledge point list and the preset difficulty level, selecting a test question matching the knowledge point list from a test question bank;

[0035] Step S106, performing random processing on the test questions to generate initial test questions, wherein the random processing includes: randomly selecting the question stem, disrupting the order of the options, and randomizing the numerical parameters;

[0036] Step S108, using a pre-trained natural language processing model to perform semantic optimization on the initial test question to generate a target test question;

[0037] Step S110, adding the target test questions in batches to the computer-based test system, and assigning a corresponding test question identifier to each test question.

[0038] The above steps are explained in detail below, taking the auxiliary generation of test questions for "Advanced Mathematics" as an example.

[0039] The exam outline for Advanced Mathematics contains the following list of knowledge points: "Limits and Continuity (Basics), Derivatives and Their Applications (Advanced), Integrals and Their Applications (Advanced)". These knowledge points will be recorded in a structured file, such as an Excel spreadsheet or a database table. It should be noted that the knowledge points are all the knowledge points listed in the exam outline that need to be tested. These knowledge points are usually organized by chapter or topic.

[0040] According to the above list of knowledge points, we can set the number of questions at the "basic" difficulty level to 10 and the number of questions at the "advanced" difficulty level to 20. Then select questions from the question bank that match the list of knowledge points, for example, select 5 basic difficulty questions from the "limit and continuity" knowledge point, select 10 advanced difficulty questions from the "derivative and its application" knowledge point, etc. The difficulty level is set at different levels to distinguish the abilities of students at different levels, such as elementary, intermediate, advanced, etc.

[0041] For a test question about "finding the limit", the original question might be "find the function exist The limit of the function can be generated by randomization. exist The initial test question was obtained by changing the function form, parameter value and the point where the limit is sought.

[0042] Next, by using a pre-trained natural language processing (NLP) model, such as BERT (Bidirectional Encoder Representations from Transformers), the initial test questions can be optimized into a more understandable form to obtain the target test questions, for example, "Find the function when Close to the limit of 1 o'clock".

[0043] Finally, the target test questions obtained after semantic optimization are imported into the computer-based examination system. Each test question will be assigned a unique test question identifier, such as "MATH123456789", so that the computer-based examination system can effectively manage and track each test question.

[0044] In an embodiment of the present invention, based on the knowledge point list and the preset difficulty level, selecting a test question matching the knowledge point list from a test question bank includes:

[0045] Calculating a weight value for each knowledge point in the knowledge point list, wherein the weight value is a weight value determined based on the importance and occurrence frequency of the knowledge point in the examination syllabus;

[0046] According to the difficulty level and the weight value of the knowledge point, a test question matching the knowledge point list is selected from a test question bank, wherein the difficulty level of the test question matches the preset difficulty level, and the knowledge point involved in the test question matches at least one knowledge point in the knowledge point list.

[0047] In the embodiment of the present invention, the examination outline of "Advanced Mathematics" lists the following knowledge points: limit and continuity (basic), derivative and its application (advanced) and integral and its application (advanced).

[0048] First, the weight value corresponding to each knowledge point is calculated by the calculation formula of the weight value of the knowledge point. The preferred calculation formula of the weight value of the knowledge point is:

[0049] ;

[0050] in, For the The weight of each knowledge point, , , , , , and is the adjustment coefficient, For the The importance function of each knowledge point is For the The frequency function of knowledge points, For the The average score rate function of knowledge points, For the The difficulty function of each knowledge point is For the The correlation function of knowledge points, For the The context relevance function of each knowledge point, For the The time sensitivity function of each knowledge point.

[0051] The following is a detailed description of each of the above functions:

[0052] ;

[0053] Experts on knowledge points Professional level rating, It is the frequency of appearance of the knowledge point in the examination syllabus.

[0054] ;

[0055] Representing knowledge points The number of co-occurrences with other knowledge points, is the total number of occurrences of all knowledge points.

[0056] ;

[0057] For Students The weight of can reflect the student's learning ability or performance in the class. For Students In the knowledge point The score on the topic, is the total number of students,

[0058] ;

[0059] is the adjustment coefficient, which is used to adjust the impact of completion time and accuracy. Students complete the knowledge points The average time for the questions, It takes the longest time to complete among all the knowledge points. For knowledge points The correctness of the questions, The highest accuracy rate among all questions corresponding to all knowledge points.

[0060] ;

[0061] Representing knowledge points And knowledge points The number of times it appears at the same time, Representing knowledge points The total number of occurrences, and Knowledge points And knowledge points Last updated on and is the adjustment coefficient.

[0062] ;

[0063] Indicated in context Knowledge Points The number of occurrences, It is a collection of all knowledge points.

[0064] ;

[0065] is the current time, For knowledge points Last updated time, is the decay constant, It is the average update interval of knowledge points.

[0066] For example, through the above formula, the weights of Limits and Continuity (Basics), Derivatives and Their Applications (Advanced), and Integrals and Their Applications (Advanced) are calculated to be 20, 45, and 30 respectively, and the difficulty level of this exam is set to "Intermediate", and 50 questions need to be prepared. According to the weights calculated above, questions can be selected in the following ways:

[0067] The proportion of questions on limit and continuity knowledge points = 20 / (20+45+30)≈0.133, that is, about 6 questions are selected (50 * 0.133 ≈ 6).

[0068] The proportion of questions about derivatives and their application knowledge points = 45 / (20+45+30)≈0.400, that is, about 20 questions are selected (50 * 0.400 ≈ 20).

[0069] The proportion of questions about integral calculus and its application knowledge points = 30 / (20+45+30)≈0.467, that is, approximately 24 questions are selected (50 * 0.467 ≈ 24).

[0070] In the question bank, select the corresponding questions according to the difficulty level and the weight value of the knowledge point. For example, select 20 questions of intermediate difficulty in the knowledge point of "derivatives and their applications".

[0071] In the embodiment of the present invention, the test questions are randomized to generate initial test questions, including:

[0072] For test questions containing numerical parameters, multimodal distribution fusion processing is performed on the numerical parameters using a preset distribution type to generate new numerical parameters, wherein the preset distribution type includes: normal distribution, uniform distribution or Poisson distribution;

[0073] Using a preset machine learning model and a real-time feedback mechanism, the new numerical parameters are adjusted to obtain target numerical parameters;

[0074] Using preset checking rules and dynamic adjustment mechanism, the validity of the target numerical parameter is checked, wherein the preset checking rules include: basic grammar checking rules, logical consistency checking rules and problem-solving feasibility checking rules;

[0075] If the check passes, the original numerical parameters in the test question containing the numerical parameters are replaced with the target numerical parameters to generate the initial test question.

[0076] Define preset distribution types for numerical parameters, including normal distribution, uniform distribution, or Poisson distribution, based on the difficulty level and knowledge points of the test questions.

[0077] For example, coefficient a uses a normal distribution, coefficient b uses a uniform distribution, and coefficient c uses a Poisson distribution.

[0078] Set the parameters of the multimodal distribution fusion model, including the weight and parameter range of each distribution type.

[0079] For example, the normal distribution has a mean of 3 and a standard deviation of 1; the uniform distribution is in the range [2, 4]; and the Poisson distribution has a lambda of 2.

[0080] A multimodal distribution fusion model is used to generate multiple candidate numerical parameters according to the preset distribution type.

[0081] For example, the candidate numerical parameters generated are a = 3.5, b = 2.2, c = -1.8.

[0082] From the generated candidate numerical parameters, the best numerical parameter is selected as the new numerical parameter.

[0083] For example, choose a = 3.5, b = 2.2, c = -1.8.

[0084] Build a preset machine learning model to automatically adjust the numerical parameters according to the difficulty level and knowledge points of the test questions, and dynamically adjust the generation strategy according to the actual effects of the generated numerical parameters through a real-time feedback mechanism.

[0085] Using the preset machine learning model and real-time feedback mechanism, the generated numerical parameters are adjusted to obtain the target numerical parameters, ensuring that the generated target numerical parameters are reasonable and meet the test requirements.

[0086] Set up multi-level preset checking rules, including basic grammar checking rules, logical consistency checking rules and problem-solving feasibility checking rules.

[0087] For example, basic grammar checking rules ensure that the format of numerical parameters is correct; logical consistency checking rules ensure that the logical relationship of numerical parameters is correct; and problem-solving feasibility checking rules ensure that the test questions generated by numerical parameters have solutions.

[0088] According to the actual effect of the generated numerical parameters, the inspection rules are dynamically adjusted to ensure that the inspection rules are always effective. If it is found that some inspection rules are not strict enough, the inspection rules are adjusted to improve the strictness.

[0089] Use preset check rules and dynamic adjustment mechanism to check the validity of target numerical parameters and check whether the generated numerical parameters meet all preset check rules.

[0090] Finally, replace the numerical parameters in the original test question with the target numerical parameters.

[0091] By generating new numerical parameters through multimodal distribution fusion, the diversity of test questions can be significantly increased, reducing the possibility of students coping with exams by memorizing question types and answers.

[0092] Use preset machine learning models and real-time feedback mechanisms to adjust numerical parameters to ensure that the generated numerical parameters are reasonable and meet the test requirements, and avoid generating unreasonable or unsolvable numerical parameters.

[0093] Through multi-level preset checking rules and dynamic adjustment mechanisms, the generated numerical parameters are ensured to be valid and reasonable, avoiding the generation of invalid or unreasonable test questions.

[0094] Through the above steps, initial test questions that are grammatically correct, semantically coherent and meet the test requirements can be generated, thereby improving the quality and fairness of the test questions.

[0095] Through the above steps, more complex and diverse test questions can be generated, effectively preventing students from answering questions by memorizing question types and option sequences, while ensuring that the generated numerical parameters are reasonable and effective and meet the requirements of the test questions.

[0096] In an embodiment of the present invention, semantic optimization is performed on the initial test question using a pre-trained natural language processing model to generate a target test question, including:

[0097] Inputting the initial test question into the pre-trained natural language processing model so that the pre-trained natural language processing model outputs grammatical errors of the intermediate test question, wherein the grammatical errors include: correcting spelling errors, punctuation errors and subject-verb inconsistency;

[0098] Using a multimodal natural language processing model and combining the text and context information of the initial test question, the grammatical error is corrected to obtain an intermediate test question;

[0099] Using a deep learning model to perform semantic similarity analysis on the initial test questions and the intermediate test questions to obtain analysis results;

[0100] The analysis result is used to adjust the expression structure of the intermediate test question to obtain the target test question.

[0101] In the embodiment of the present invention, an initial test question is: "Find the function exist The derivative at , options are: A. 10 B. 12 C. 14 D. 16”.

[0102] The initial test questions are input into the pre-trained natural language processing model, and the pre-trained natural language processing model outputs grammatical errors in the initial test questions.

[0103] By using a multimodal natural language processing model and combining the text and context information of the initial test questions, grammatical errors are corrected to obtain intermediate test questions.

[0104] The deep learning model is used to perform semantic similarity analysis on the initial test questions and the intermediate test questions to obtain the analysis results.

[0105] For example, the semantic similarity between the initial test question and the intermediate test question is calculated, and the similarity score is 0.95.

[0106] By using the analysis results, the statement structure of the intermediate test questions is adjusted to obtain the target test questions.

[0107] According to the results of semantic similarity analysis, the statement structure of the intermediate test questions is adjusted to generate the target test questions.

[0108] Detect and correct grammatical errors through pre-trained natural language processing models to ensure the grammatical correctness of test questions.

[0109] The coherence and logical consistency of test questions are ensured by combining text and context information through multimodal natural language processing models.

[0110] Use deep learning models to perform semantic similarity analysis to ensure that intermediate test questions are semantically consistent with the initial test questions.

[0111] According to the results of semantic similarity analysis, the statement structure of the intermediate test questions is adjusted to generate clearer and easier-to-understand target test questions.

[0112] Through the above steps, test questions with correct grammar, coherent semantics and clear expression can be generated, which effectively solves the problems of grammatical errors and semantic inconsistency in test question generation in the prior art, and improves the quality of test questions and students' understanding.

[0113] In an embodiment of the present invention, the method further includes:

[0114] Determining the difficulty level of the knowledge point based on the weight value of the knowledge point and a preset difficulty level function;

[0115] adjusting the test questions based on the difficulty level;

[0116] The preset difficulty level function is:

[0117] ;

[0118] in, Difficulty level The thresholds for difficulty levels, The highest difficulty level.

[0119] In the embodiment of the present invention, the original test question "Find the derivative of the function f(x)=2x³+4x²-1 in the interval [0, 5]." The weight value of the knowledge point "Find the derivative" in this test question is 8, and the weight value is set to = 4, the preset maximum difficulty level is = 5, the calculated difficulty level is 2.

[0120] The obtained value and After comparison, the smaller value is taken as the final difficulty level. Therefore, the final difficulty level of the knowledge point "derivative" is 2. Considering that the difficulty level of the knowledge point "derivative" is 2, the original test question can be kept unchanged because it already meets this difficulty level.

[0121] In an embodiment of the present invention, after the target test question is generated, the target test question and related information are stored in a database, wherein the related information at least includes: the type of the test question, the difficulty level, the knowledge points involved, and the test question identifier;

[0122] A mapping relationship between a test question identifier and the test question information is established.

[0123] In an embodiment of the present invention, the target test question is: "Find the derivative of the function f(x)=2x³+4x²-1 at x=1." This test question and related information need to be stored in a database.

[0124] Target test related information:

[0125] Type of question: Multiple choice

[0126] Difficulty level: 2

[0127] Knowledge points involved: derivative calculation

[0128] Question Identifier: QID_001

[0129] Information stored in the database:

[0130] Create a database table: First, create a table in the database to store test questions and related information. The table structure is designed as follows:

[0131]

[0132] Enter the information of the above target test questions into the database table.

[0133] Question Identifier (QID): QID_001

[0134] Question Text: Find the derivative of the function f(x)=2x³+4x²-1 at x=1

[0135] Question Type: Multiple choice

[0136] Difficulty Level: 2

[0137] Knowledge Point involved: Derivative calculation

[0138] Mapping relationship establishment:

[0139] In the database, by setting the primary key (for example, using QID as the primary key), a mapping relationship between the question identifier and the question information is naturally established. Each record has a unique QID, which ensures this one-to-one correspondence.

[0140] Embodiment 2:

[0141] An embodiment of the present invention also provides a system for assisting the generation of computer-based examination questions. The system for assisting the generation of computer-based examination questions is used to execute the method for assisting the generation of computer-based examination questions provided in the above content of the embodiment of the present invention. The following is a specific introduction to the system for assisting the generation of computer-based examination questions provided in the embodiment of the present invention.

[0142] like Figure 2 As shown, Figure 2 The schematic diagram of the above-mentioned system for assisting the generation of test questions for the computer-based test is as follows:

[0143] An acquisition unit 10 is used to acquire a list of knowledge points specified in the examination syllabus;

[0144] A matching unit 20, configured to select, from a test question bank, a test question that matches the knowledge point list based on the knowledge point list and a preset difficulty level;

[0145] A randomization processing unit 30, configured to perform randomization processing on the test questions to generate initial test questions, wherein the randomization processing includes: randomly selecting the question stem, disrupting the order of the options, and randomizing the numerical parameters;

[0146] A semantic optimization unit 40, configured to perform semantic optimization on the initial test question using a pre-trained natural language processing model to generate a target test question;

[0147] The adding unit 50 is used to add the target test questions in batches to the computer-based test system and assign a corresponding test question identifier to each test question.

[0148] Embodiment three:

[0149] The embodiment of the present invention further provides a computing device for executing a program of the method described in the first embodiment, such as Figure 3As shown, the computing device may include a storage component 41 and a processing component 42;

[0150] The storage component 41 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 42 .

[0151] The processing component 42 may include one or more processors to execute computer instructions to complete all or part of the steps in the method of embodiment 1. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0152] The storage component 41 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0153] The display component 43 may be an electroluminescent (EL) element, a liquid crystal display or a micro display having a similar structure, or a retinal direct display or a similar laser scanning display.

[0154] Of course, a computing device may also include other components, such as input / output interfaces, communication components, etc.

[0155] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0156] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0157] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0158] Embodiment 4:

[0159] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1The method of the illustrated embodiment.

[0160] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0161] 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.

[0162] 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, 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.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application 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 application.

Claims

1. A method for auxiliary generation of test questions for computer-based examinations, characterized in that: include: Get the list of knowledge points specified in the exam syllabus; Based on the knowledge point list and the preset difficulty level, selecting a test question matching the knowledge point list from a test question bank; Performing random processing on the test questions to generate initial test questions, wherein the random processing includes: randomly selecting the question stem, disrupting the order of options, and randomizing numerical parameters; Using a pre-trained natural language processing model to semantically optimize the initial test questions to generate target test questions; Adding the target test questions in batches to the computer-based test system and assigning a corresponding test question identifier to each test question; Wherein, based on the knowledge point list and the preset difficulty level, selecting test questions matching the knowledge point list from the test question bank includes: Calculating a weight value for each knowledge point in the knowledge point list, wherein the weight value is a weight value determined based on the importance and occurrence frequency of the knowledge point in the examination syllabus; According to the difficulty level and the weight value of the knowledge point, selecting a test question matching the knowledge point list from a test question bank, wherein the difficulty level of the test question matches the preset difficulty level, and the knowledge point involved in the test question matches at least one knowledge point in the knowledge point list; The calculation formula of the weight value of the knowledge point is: ； in, For the The weight of each knowledge point, , , , , , and is the adjustment coefficient, For the The importance function of each knowledge point is For the The frequency function of knowledge points, For the The average score rate function of knowledge points, For the The difficulty function of each knowledge point is For the The correlation function of knowledge points, For the The context relevance function of each knowledge point, For the The time sensitivity function of each knowledge point; Determining the difficulty level of the knowledge point based on the weight value of the knowledge point and a preset difficulty level function; adjusting the test questions based on the difficulty level; The preset difficulty level function is: ； in, Difficulty level The thresholds for difficulty levels, The highest difficulty level.

2. The method according to claim 1, characterized in that The test questions are randomized to generate initial test questions, including: For test questions containing numerical parameters, multimodal distribution fusion processing is performed on the numerical parameters using a preset distribution type to generate new numerical parameters, wherein the preset distribution type includes: normal distribution, uniform distribution or Poisson distribution; Using a preset machine learning model and a real-time feedback mechanism, the new numerical parameters are adjusted to obtain target numerical parameters; Using preset checking rules and dynamic adjustment mechanism, the validity of the target numerical parameter is checked, wherein the preset checking rules include: basic grammar checking rules, logical consistency checking rules and problem-solving feasibility checking rules; If the check passes, the original numerical parameters in the test question containing the numerical parameters are replaced with the target numerical parameters to generate the initial test question.

3. The method according to claim 1, characterized in that: The initial test questions are semantically optimized using a pre-trained natural language processing model to generate target test questions, including: Inputting the initial test question into the pre-trained natural language processing model so that the pre-trained natural language processing model outputs grammatical errors of the initial test question, wherein the grammatical errors include: correcting spelling errors, punctuation errors and subject-verb inconsistency; Using a multimodal natural language processing model and combining the text and context information of the initial test question, the grammatical error is corrected to obtain an intermediate test question; Using a deep learning model to perform semantic similarity analysis on the initial test questions and the intermediate test questions to obtain analysis results; The analysis result is used to adjust the expression structure of the intermediate test question to obtain the target test question.

4. The method according to claim 1, characterized in that: The method further comprises: After generating the target test question, storing the target test question and related information in a database, wherein the related information at least includes: the type of the test question, the difficulty level, the knowledge points involved, and the test question identifier; Establish a mapping relationship between the test question identifier and the test question information.

5. A system for assisting the generation of test questions for a computer-based test, characterized in that: include: An acquisition unit is used to obtain a list of knowledge points specified in the examination syllabus; A matching unit, configured to select, from a test question bank, a test question that matches the knowledge point list based on the knowledge point list and a preset difficulty level; A randomization processing unit, used for performing randomization processing on the test questions to generate initial test questions, wherein the randomization processing includes: randomly selecting the question stem, disrupting the order of options, and randomizing numerical parameters; A semantic optimization unit, configured to perform semantic optimization on the initial test question using a pre-trained natural language processing model to generate a target test question; An adding unit, used to add the target test questions in batches to the computer-based test system, and assign a corresponding test question identifier to each test question; Wherein, based on the knowledge point list and the preset difficulty level, selecting test questions matching the knowledge point list from the test question bank includes: Calculating a weight value for each knowledge point in the knowledge point list, wherein the weight value is a weight value determined based on the importance and occurrence frequency of the knowledge point in the examination syllabus; According to the difficulty level and the weight value of the knowledge point, selecting a test question matching the knowledge point list from a test question bank, wherein the difficulty level of the test question matches the preset difficulty level, and the knowledge point involved in the test question matches at least one knowledge point in the knowledge point list; The calculation formula of the weight value of the knowledge point is: ； in, For the The weight of each knowledge point, , , , , , and is the adjustment coefficient, For the The importance function of each knowledge point is For the The frequency function of knowledge points, For the The average score rate function of knowledge points, For the The difficulty function of each knowledge point is For the The correlation function of knowledge points, For the The context relevance function of each knowledge point, For the The time sensitivity function of each knowledge point; Determining the difficulty level of the knowledge point based on the weight value of the knowledge point and a preset difficulty level function; adjusting the test questions based on the difficulty level; The preset difficulty level function is: ； in, Difficulty level The thresholds for difficulty levels, The highest difficulty level.

6. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the method according to any one of claims 1 to 4.

7. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, the method according to any one of claims 1 to 4 is implemented.

Images