Test paper generation method and device, electronic equipment and program product

Abstract

The invention discloses a test paper generation method and device. The test paper generation method comprises the steps of obtaining a question bank comprising a plurality of questions; determining a plurality of knowledge point feature vectors corresponding to a plurality of knowledge points, wherein the plurality of questions are associated with the plurality of knowledge points; determining a preset number of target knowledge points at least according to the feature space distances among the plurality of knowledge point vectors; and selecting a question associated with the target knowledge point from the question bank so as to generate the test paper. Through intelligent topic selection of vectorization calculation, uniform distribution of test paper knowledge points is realized, and the problem of knowledge point repetition is avoided.

Description

Technical Field

[0001] The present invention relates to the field of computer application technology, and in particular to a test paper generating method and device, electronic equipment and program products. Background Art

[0002] In the field of exam question setting, the rationality of the knowledge point distribution of the test paper is an important consideration. Existing test papers are often set manually, and there may be problems such as repeated knowledge points and uneven distribution. Specifically, in the same test paper, there may be multiple questions testing the same knowledge point, which not only wastes test question resources, but also fails to fully test students' mastery of different knowledge points.

[0003] At the same time, the traditional manual question setting method is inefficient when dealing with large-scale question banks, and it is difficult to quickly screen and organize test questions.

[0004] Therefore, it is hoped to provide an automated test paper generation solution that can ensure uniform distribution of test paper knowledge points and avoid duplication of knowledge points.

[0005] The content of this background technology description is only for facilitating understanding of the relevant technology in this field and is not regarded as an admission of the prior art. Summary of the invention

[0006] The embodiment of the present invention aims to provide a test paper generation solution, which can automatically generate a test paper with uniform knowledge point distribution and avoid duplication of knowledge points.

[0007] In a first aspect, an embodiment of the present invention provides a test paper generating method, which includes:

[0008] Acquire a question bank including a plurality of questions, wherein the question information includes a knowledge point identifier;

[0009] Determining a plurality of knowledge point feature vectors corresponding to a plurality of knowledge points, wherein the plurality of topics are associated with the plurality of knowledge points;

[0010] Determining a preset number of target knowledge points at least according to the feature space distances between the plurality of knowledge point vectors; and

[0011] Questions associated with the target knowledge points are selected from the question bank to generate a test paper.

[0012] In some embodiments, the method further comprises: determining a plurality of topic feature vectors corresponding to the plurality of topics;

[0013] The step of selecting questions associated with the target knowledge point from the question bank to generate a test paper includes:

[0014] Determining a first vector space similarity between a knowledge point feature vector of the target knowledge point and a topic feature vector of at least part of the topics; and

[0015] A set of questions associated with the target knowledge point is determined according to the first vector space similarity.

[0016] In some embodiments, the method further includes: determining a mastery feature vector that characterizes the mastery of a knowledge point of at least one examinee, wherein each examinee has a corresponding mastery feature vector, and the mastery feature vector has a dimension corresponding to the number of the plurality of knowledge points. In these embodiments, determining a preset number of target knowledge points based at least on the feature space distance between the plurality of knowledge point vectors includes: determining the preset number of target knowledge points based on the feature space distance and the mastery feature vector.

[0017] In some embodiments, determining the preset number of target knowledge points according to the feature space distance and the mastery feature vector includes:

[0018] Determining feature space distances between the plurality of knowledge points;

[0019] Determining feature components of corresponding dimensions of the plurality of knowledge points in the mastery feature vector; and

[0020] According to a weighted combination of the first weight of the feature space distance and the second weight based on the feature component, a preset number of knowledge points are determined as target knowledge points.

[0021] In some embodiments, determining a mastery feature vector representing the mastery degree of at least one examinee's knowledge point includes:

[0022] creating an initial mastery feature vector for the at least one examinee; and

[0023] The corresponding mastery feature vector is dynamically updated according to the question-solving history information of the at least one examinee, wherein at least part of the questions in the question-solving history information are associated with at least one of the knowledge points.

[0024] In some embodiments, dynamically updating the corresponding mastery feature vector according to the question-solving history information includes:

[0025] Determine the second vector space similarity between the test question vector of the question in the test question history information and the plurality of knowledge point feature vectors; and

[0026] According to the problem-solving situation of the question and the similarity in the second vector space, the feature components of the dimensions corresponding to the knowledge points associated with the question in the mastery feature vector are updated.

[0027] In some embodiments, the problem solving status includes the problem solving result. In these embodiments, updating the feature component of the dimension corresponding to the knowledge point associated with the problem in the mastery feature vector according to the problem solving status and the second vector space similarity includes:

[0028] When the result of the test is wrong, reducing the feature component of the dimension corresponding to the associated knowledge point in the mastery feature vector according to the second vector space similarity; and

[0029] When the result of the test is correct, the feature component of the dimension corresponding to the associated knowledge point in the mastery feature vector is increased according to the second vector space similarity.

[0030] In some embodiments, the method further comprises:

[0031] Detecting whether there are repeated questions in the question set;

[0032] When there are repeated questions, redundant repeated questions are removed, and replacement questions are selected from the question bank based on the first vector space similarity until the preset question quantity requirement of the test paper is met.

[0033] In a second aspect, an embodiment of the present invention provides a test paper generating device, comprising:

[0034] An acquisition unit configured to acquire a question bank including a plurality of questions, wherein the question information includes a knowledge point identifier;

[0035] A first determining unit is configured to determine a plurality of knowledge point feature vectors corresponding to a plurality of knowledge points, wherein the plurality of topics are associated with the plurality of knowledge points;

[0036] A second determining unit is configured to determine a preset number of target knowledge points at least according to the feature space distances between the plurality of knowledge point vectors; and

[0037] The generating unit is configured to select questions associated with the target knowledge point from the question bank, thereby generating a test paper.

[0038] In a third aspect, an embodiment of the present invention provides an electronic device, comprising: a processor and a memory storing a computer program, wherein the processor is configured to implement the method as described in the first aspect when executing the computer program.

[0039] In a fourth aspect, an embodiment of the present invention provides a program product, including a computer program, wherein when the computer program is executed by a processor, the method as described in the first aspect is implemented.

[0040] In a fifth aspect, an embodiment of the present invention provides a storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the method described in the first aspect.

[0041] The embodiment of the present invention provides a test paper generation method based on vectorized calculation, by obtaining a question bank containing multiple questions, determining multiple knowledge point feature vectors corresponding to multiple knowledge points, determining a preset number of target knowledge points at least according to the feature space distance between the multiple knowledge point vectors, and selecting questions associated with the target knowledge points from the question bank to generate a test paper. Thus, by vectorizing the knowledge points and using the vector space distance for intelligent question selection, the problem of repeated and uneven distribution of knowledge points in the test paper is effectively solved. Therefore, the embodiment of the present invention achieves uniform coverage of test paper knowledge points, improves question setting efficiency and test paper quality.

[0042] In a further embodiment of the present invention, the test paper generation method also considers the examinee's mastery of knowledge points, dynamically adjusts the topic selection strategy by introducing a mastery feature vector, and dynamically updates the student's mastery feature vector by obtaining the student's historical information on the questions they answer. Therefore, when selecting questions, in addition to considering the feature space distance between knowledge points, the student's mastery feature vector is also considered, so that the test paper is more suitable for the student's weak links, thereby realizing a test paper generation solution for "a thousand papers for a thousand examinees".

[0043] Other optional features and technical effects of the embodiments of the present invention are partially described below, and partially can be understood by reading this document. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. The elements shown are not limited to the proportions shown in the accompanying drawings. The same or similar reference numerals in the accompanying drawings represent the same or similar elements, wherein:

[0045] Figure 1 is a first flow chart of a test paper generating method according to an embodiment of the present invention;

[0046] Figure 2 is a second flow chart of the test paper generating method according to an embodiment of the present invention;

[0047] Figure 3 is a third flow chart of the test paper generating method according to an embodiment of the present invention;

[0048] Figure 4 is a fourth flow chart of the test paper generating method according to an embodiment of the present invention;

[0049] Figure 5 is a fifth flow chart of the test paper generating method according to an embodiment of the present invention;

[0050] Figure 6 is a sixth flow chart of the test paper generating method according to an embodiment of the present invention;

[0051] Figure 7 is a module structure diagram of a test paper generating device according to an embodiment of the present invention; and

[0052] Figure 8 An exemplary structural diagram of an electronic device capable of implementing the method according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific implementation methods and drawings. Here, the exemplary implementation methods of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0054] In the prior art, test questions are often set manually, which may result in uneven distribution of knowledge points and duplication of knowledge points. Specifically, in the same test paper, multiple questions may test the same knowledge point, which not only wastes test question resources, but also fails to comprehensively test students' mastery of different knowledge points. In addition, the traditional manual question setting method is inefficient when processing large-scale question banks, and it is difficult to quickly screen and organize test questions.

[0055] To solve the above technical problems, the embodiment of the present invention provides a test paper generation method based on vectorized calculation. By vectorizing knowledge points and using vector space distance for intelligent question selection, the embodiment of the present invention achieves uniform coverage of test paper knowledge points, effectively improving question setting efficiency and test paper quality.

[0056] like Figure 1 As shown, an embodiment of the present invention provides a test paper generation method, including steps S110 to S140:

[0057] Step S110: Obtain a question bank including multiple questions.

[0058] In an embodiment of the present invention, a question bank containing multiple questions may be obtained. For example, the question bank contains N questions, such as 1000 questions, and each question is associated with a knowledge point.

[0059] In some embodiments, the association between the question and the knowledge point is an explicit association. For example, each question carries at least one knowledge point identifier, which is used to indicate which knowledge points the question is associated with. For example, a question may be associated with two knowledge points, "function extreme value" and "derivative application". The knowledge point identifier can be obtained by manual annotation, or it can be automatically extracted / generated by natural language processing and other technologies. In other embodiments, the association between the question and the knowledge point can be an implicit association, such as by associating the question with the vectorization of the knowledge point, as further described below.

[0060] In some embodiments, the question bank may be stored in a database or in other appropriate data structure formats.

[0061] Step S120: Determine multiple knowledge point feature vectors corresponding to the multiple knowledge points.

[0062] In the embodiment of the present invention, each knowledge point has a corresponding knowledge point feature vector. As mentioned above, multiple questions in the question bank are associated with multiple knowledge points.

[0063] In some embodiments, the knowledge point feature vector corresponding to the knowledge point may be provided in advance. In other embodiments, it may also include vectorizing the knowledge point to obtain the knowledge point feature vector. In some specific embodiments, for example, a pre-trained language model (such as BERT, Word2Vec, etc.) may be used to convert each knowledge point into a feature vector. In an embodiment of the present invention, each knowledge point can be converted into an N-dimensional feature vector, which can characterize the semantic features of the knowledge point. However, it is also conceivable to obtain the knowledge point feature vector in any suitable manner. In some embodiments, the dimensional components of the knowledge point may be interpretable, for example, one or more components map one or more sub-knowledge points; in other embodiments, the dimensional components of the knowledge point may be non-interpretive or low-interpretive, which all fall within the scope of the present invention.

[0064] Step S130: Determine a preset number of target knowledge points based at least on the feature space distances between the plurality of knowledge point vectors.

[0065] Step S140: Select questions associated with the target knowledge point from the question bank to generate a test paper.

[0066] Here, several target knowledge points can be selected for subsequent test paper generation, and the selection of target knowledge points directly affects the uniformity of the distribution of test paper knowledge points, and then affects the coverage of the test paper. In an embodiment of the present invention, in step S130, the semantic discrimination (association relationship) between knowledge points can be established by calculating the feature space distance between knowledge points. The discrimination is conducive to determining the target knowledge points, so as to better cover a wider knowledge space.

[0067] In the embodiment of the present invention, the feature space distance may be determined in a variety of different ways.

[0068] In a specific embodiment, the spatial distances between the feature vectors of all knowledge points can be calculated, for example, to form an M×M distance matrix D (the number of knowledge points is M), where the matrix element D[i][j] represents the feature space distance between knowledge point i and knowledge point j. Then, the knowledge points with farther feature space distances are selected as target knowledge points. For example, the farther knowledge points can be determined by sorting from far to near or by comparing with a predetermined threshold. In an embodiment of the present invention, any suitable spatial distance algorithm can be used to determine the spatial distance, including but not limited to Euclidean distance, Manhattan distance, cosine distance, Chebyshev distance, Mahalanobis distance, and Minkowski distance.

[0069] In another specific embodiment, the feature space distance can be determined based on the seed knowledge point. For example, at least one knowledge point can be set or randomly selected as the seed knowledge point. Subsequently, the spatial distance between other knowledge points and the seed knowledge point can be calculated, and the knowledge point with a longer distance can be selected as the target knowledge point. In a specific example, the knowledge point farthest from the seed knowledge point can be selected as the target knowledge point, and the knowledge point farthest from the seed knowledge point and the selected target knowledge point can be selected as the next target knowledge point. Through such continuous iteration, each new knowledge point selected requires the maximum average distance from the selected knowledge point set until a predetermined number, such as 20 target knowledge points, is selected. In a preferred embodiment, the set seed knowledge point can be an important knowledge point, a basic knowledge point and / or a required knowledge point. In some embodiments, the selection can be made in combination with the importance of the knowledge point. For example, for a certain type of exam, key knowledge points with a high degree of matching with the type of exam can be preferentially selected. In some embodiments, certain required knowledge points can also be pre-set according to the syllabus or exam requirements, and these knowledge points can be directly included in the target knowledge point set and used as seed knowledge points, and then other target knowledge points can be selected from the remaining knowledge points.

[0070] It should be noted that the number of target knowledge points can be adjusted according to actual needs. The selection strategy can also be optimized according to specific application scenarios, for example, factors such as the difficulty distribution of knowledge points and chapter distribution can be considered.

[0071] In the embodiment of the present invention, in step S140, after the target knowledge point is selected, appropriate test questions / examination questions can be selected from the question bank according to the relevance to the target knowledge point. In the embodiment of the present invention, the "question" is interpreted broadly and can include test questions / examination questions used for examinations, and can also include "optional" (test) questions / (examination) questions that are not selected for the examination, and can also include questions related to non-examinations (such as exercises).

[0072] In some embodiments, as described above, the association between the topic and the knowledge point is an explicit association, for example, each topic carries at least one knowledge point identifier. In these embodiments, the associated topic can be directly selected using the knowledge point identifier information of the topic.

[0073] In other embodiments, for example, when the aforementioned questions and knowledge points are implicitly associated, the vector space similarity between the questions and the knowledge points may also be used to select test questions.

[0074] In these embodiments, the test paper generating method may further include step S120' (not shown): determining a plurality of question feature vectors corresponding to the plurality of questions.

[0075] In some embodiments, the topic feature vector has the same dimension as the knowledge point vector.

[0076] In these embodiments, for determining other features of the question feature vector, reference may be made to step S120, which will not be described in detail herein.

[0077] In these embodiments, Figure 2 As shown, step S140 may include:

[0078] Step S210: Determine the first vector space similarity between the knowledge point feature vector of the target knowledge point and the question feature vector of at least part of the questions.

[0079] In the embodiment of the present invention, the similarity between the feature vector of each target knowledge point and the question feature vector of the question in the question bank can be calculated, and the question feature vector can be determined based on step S120'. In the embodiment of the present invention, the vector space similarity can be determined by any suitable algorithm, such as but not limited to cosine similarity; people can also think of using a spatial distance algorithm (such as the distance algorithm described above) that has been appropriately adapted to determine it.

[0080] In a specific example, the first vector space similarity determined based on cosine similarity=(v1·v2) / (|v1|·|v2|), where v1 is a knowledge point feature vector, and v2 is a question feature vector.

[0081] Step S220: Determine a set of questions associated with the target knowledge point based on the first vector space similarity.

[0082] In an embodiment of the present invention, a corresponding set of questions can be established for each target knowledge point based on the determined first vector space similarity. In a specific example, when the similarity between a question and a knowledge point exceeds a preset threshold (such as 0.8), the question will be included in the question set of the corresponding knowledge point. This method can discover potential knowledge point associations and is suitable for large-scale expansion of the question bank. As an explanation, a question may have a high similarity with multiple knowledge points, so that the question is associated with multiple knowledge points and may therefore appear in the question sets of multiple target knowledge points at the same time. In some embodiments, a de-duplication mechanism can also be provided to avoid duplicate test questions, as described below.

[0083] In other embodiments, the characteristics of the degree of mastery of knowledge points by examinees are also introduced to realize the generation of personalized test papers for "one thousand papers for one thousand people".

[0084] In these embodiments, the test paper generating method may further include step S120" (not shown): determining a mastery feature vector representing the mastery degree of at least one examinee's knowledge point.

[0085] In these embodiments, each examinee has a corresponding mastery feature vector.

[0086] In these embodiments, the mastery feature vector has dimensions corresponding to the number of knowledge points. For example, for N knowledge points, the mastery feature vector may be N-dimensional, each dimension corresponding to one knowledge point, and representing the examinee's mastery of the knowledge point, and the mastery of the knowledge point may be adjusted according to the examinee's question-solving history, which is not limited to the test-solving history, but may also include the practice question-solving history, as further described below.

[0087] In these embodiments, for determining other features of the mastery feature vector, reference may be made to steps S120 and S120', which will not be described in detail herein.

[0088] In these embodiments, the test paper generating method, more specifically step S130 , may include step S130 ′ (not shown): determining the preset number of target knowledge points according to the feature space distance and the mastery feature vector.

[0089] In a further embodiment, Figure 3 As shown, step S130' may include:

[0090] Step S310: determining feature space distances between multiple knowledge points;

[0091] In these embodiments, the features of step S310 may refer to the relevant features of determining the feature space distance in step S130, which will not be described in detail here.

[0092] Step S320: Determine the feature components of the corresponding dimensions of the plurality of knowledge points in the mastery feature vector.

[0093] In an embodiment of the present invention, the mastery (feature component) corresponding to each knowledge point can be extracted from the examinee's mastery feature vector. For example, for knowledge point i, the value s[i] of the i-th component in the mastery feature vector v can be obtained, which reflects the examinee's mastery of this knowledge point. For example, if s[i] = 0.8, it means that the examinee has a high degree of mastery of knowledge point i; if s[i] = 0.3, it means that the mastery is low. However, the present invention is not limited to this. For example, the mastery feature vector (component) can represent the degree of non-mastery, which falls within the scope of the present invention.

[0094] In some embodiments, the characteristic components of the mastery feature vector can be normalized, for example, limited to the interval [0,1], for example: 0 means that the knowledge point is not mastered at all; 0.3 and below means poor mastery; 0.3 to 0.7 means average mastery; 0.7 and above means good mastery; 1 means complete mastery of the knowledge point.

[0095] As an example, a candidate's mastery feature vector may be as follows:

[0096] For the knowledge point of "Extremum of Function": s[i] = 0.85, indicating that the candidate has a good grasp of this knowledge point i;

[0097] For the knowledge point "Definition of derivative": s[j] = 0.45, indicating that the candidate has a general grasp of this knowledge point j;

[0098] For the knowledge point of "Sequence Limit": s[k]=0.25, it shows that the candidate has a poor grasp of this knowledge point k.

[0099] In these embodiments, by combining the characteristic spatial distance between knowledge points and the examinee's mastery level, the comprehensiveness of the knowledge points tested can be better guaranteed while paying appropriate attention to the examinee's weak links.

[0100] Step S330: Determine the target knowledge point based on a weighted combination of the first weight of the feature space distance and the second weight based on the feature component.

[0101] In an embodiment of the present invention, a weighted combination method can be used to comprehensively evaluate the priority of each knowledge point to be selected as a target knowledge point. This evaluation method takes into account two key factors: the uniformity of distribution between knowledge points and the knowledge mastery of the examinee. In some embodiments, the weights of the two factors can be adjusted, thereby flexibly controlling the pertinence of the test paper.

[0102] Thus, in these embodiments, for each candidate knowledge point, a weighted combined comprehensive score can be determined. The higher the score, the more suitable the knowledge point is to be selected as a target knowledge point.

[0103] In a specific example, for knowledge point i, the formula for determining its comprehensive score can be:

[0104] Score_i=w1×D_i+w2×(1-s[i])

[0105] Among them, D_i is the feature space distance between knowledge point i and the selected target knowledge point (or other knowledge points), including the average feature space distance or a single average feature space distance; w1 is the first weight; w2 is the second weight; Score_i is the comprehensive score of the weighted combination.

[0106] In some embodiments, the score of each candidate knowledge point can be calculated in parallel or iteratively, and the knowledge point with the highest score is selected to be added to the target knowledge point set until a preset target number of knowledge points (eg, 20) is reached.

[0107] In some preferred embodiments, the feature space distance can be dynamically updated during the selection process.

[0108] In some embodiments, the above-mentioned mastery feature vector can be dynamically updated based on the test-taking history of the examinee.

[0109] In these embodiments, Figure 4 As shown, step S320 may include:

[0110] Step S410: Creating an initial mastery feature vector for at least one examinee.

[0111] In the embodiment of the present invention, an initial mastery feature vector may be created for each examinee.

[0112] In some embodiments, the initial values ​​of the mastery degree feature vectors are the same. In the initial state, all component values ​​of the vectors are set to 0.5, indicating that the mastery degree of all knowledge points is at a medium level.

[0113] In other embodiments, the initial values ​​of the mastery feature vectors are different, and may be determined based on qualitative or quantitative historical performance, for example.

[0114] Step S420: dynamically update the corresponding mastery feature vector according to the test-taking history information of at least one examinee.

[0115] In an embodiment of the present invention, at least part of the questions, preferably all of the questions in the question-solving history information are associated with at least one of the knowledge points, for example, at least part of the questions, preferably all of the questions in the question-solving history information are vectorized or can be vectorized.

[0116] In the embodiment of the present invention, the updating of the mastery feature vector is a dynamic process. Whenever a candidate completes a question that can be represented by vectorization, the vector can be adjusted accordingly according to the test-taking situation. This real-time updating mechanism ensures that the system can timely reflect the changes in the candidate's knowledge mastery status.

[0117] In these embodiments, Figure 5 As shown, step S420 may include:

[0118] S510: Determine the second vector space similarity between the test question vector of the question in the test history information and the feature vectors of multiple knowledge points.

[0119] In the embodiment of the present invention, the characteristics of determining the vector space similarity in step S510 may refer to step S210, which will not be described in detail herein.

[0120] S520: Based on the problem-solving status of the problem and the similarity in the second vector space, the feature components of the dimensions corresponding to the knowledge points associated with the problem in the mastery feature vector are updated.

[0121] In some embodiments, when determining the second vector space similarity (or by pre-screening), the knowledge points associated with the questions in the question-solving history information may be selected. Thus, in step S520, only the feature components of the dimensions corresponding to the selected knowledge points will be updated. However, in other embodiments, the feature components of the dimensions corresponding to any associated knowledge points may be fully updated according to the second vector space similarity.

[0122] In these embodiments, step S520 may include:

[0123] A1: when the result of the test is wrong, reducing the feature component of the dimension corresponding to the associated knowledge point in the mastery feature vector according to the second vector space similarity; and

[0124] A2: When the result of the test is correct, the feature component of the dimension corresponding to the associated knowledge point in the mastery feature vector is increased according to the second vector space similarity.

[0125] In some preferred embodiments, when adjusting the feature components, not only the second vector space similarity is considered, but also multiple other influencing factors are considered, including but not limited to learning rate, question (correct question / incorrect question) influence value and / or forgetting time, etc. In some embodiments, the above other impression factors for correct questions and incorrect questions can be set to be the same or different.

[0126] In some embodiments, the other influencing factors mentioned above may be used as a weight factor and / or a weight function of the second vector space similarity or a part thereof.

[0127] In a specific example, for example, the dynamic update of the feature component s[i] corresponding to the i-th knowledge point can be determined according to the following formula:

[0128] s[i] = s[i] + α*f(question (correct question / incorrect question) impact value, time decay)*a_j

[0129] Among them, s[i] represents the feature component, α is the learning rate, f() is the weight function, and a_j is the similarity of the second vector space.

[0130] Therefore, the embodiment of the present invention can generate a more personalized and targeted test paper content while generating a test paper that covers knowledge points more comprehensively.

[0131] In some embodiments, after completing the (preliminary) selection of test questions, the question set can also be checked for duplication and replaced.

[0132] In these embodiments, Figure 6 As shown, the test paper generating method may further include:

[0133] Step S150: Check whether there are duplicate questions in the question set.

[0134] In some embodiments, the selected set of questions is traversed and compared to identify repeated questions. In these embodiments, when repeated questions are detected, redundant repeated questions can be removed.

[0135] In some embodiments, the determination of duplication includes not only completely identical test questions, but also test questions with only slight differences in numbers, expressions, etc. For example, two function extreme value calculation questions will also be determined as duplicate questions if only the coefficients in the function expressions are different.

[0136] Step S160: When there are repeated questions, remove the redundant repeated questions, and select replacement questions from the question bank according to the first vector space similarity until the preset question quantity requirement of the test paper is met.

[0137] In these embodiments, after the duplicate questions are removed, new questions can be selected from the question bank for replacement. When selecting replacement questions, the system calculates the first vector space similarity between the candidate questions and the target knowledge point, and preferentially selects questions with higher similarity as replacement questions.

[0138] In some embodiments, replacement can be performed after all unnecessary repetitive questions are completed. In other embodiments, replacement can be performed after removing one or more unnecessary repetitive questions, and subsequent repetition determination and removal (if any) are performed after replacement. This all falls within the scope of the present invention.

[0139] In some embodiments of the present invention, Figure 7 As shown, a test paper generating device 700 is provided, which may include: an acquiring unit 710, a first determining unit 720, a second determining unit 730 and a generating unit 740. Among them,

[0140] The acquisition unit 710 is configured to acquire a question bank including a plurality of questions.

[0141] The first determining unit 720 is configured to determine a plurality of knowledge point feature vectors corresponding to a plurality of knowledge points, wherein the plurality of topics are associated with the plurality of knowledge points.

[0142] The second determination unit 730 is configured to determine a preset number of target knowledge points at least according to the feature space distances between the plurality of knowledge point vectors.

[0143] The generating unit 740 is configured to select questions associated with the target knowledge point from the question bank, thereby generating a test paper.

[0144] In embodiments of the present invention, method features may be combined with apparatus to obtain new embodiments, and vice versa.

[0145] In some embodiments of the present invention, an electronic device is provided, which includes a processor and a memory storing a computer program, wherein the processor is configured to implement a method of any embodiment of the present invention when running the computer program.

[0146] Figure 8 A schematic diagram of an electronic device 800 that can be used to implement a method or implement an embodiment of the present invention is shown. In some embodiments, the number of electronic devices may be more or less than the number shown. In some embodiments, a single or multiple electronic devices may be used for implementation. In some embodiments, cloud or distributed electronic devices may also be used for implementation.

[0147] like Figure 8As shown, the electronic device 800 includes a processor 801 and a memory 802. The processor is used to execute programs stored in the memory, and these programs can implement the methods, steps or functions described in the above embodiments when the computer executes them. The processor 801 may include various types of processors, such as a central processing unit (CPU), a graphics processing unit (GPU), a neural network processor (NPU), a digital signal processor (DSP), etc. The processor 801 and the memory 802 are interconnected via a bus 803. The bus 803 can also be connected to an input / output (I / O) interface, etc. The systems, devices, modules or units described in the above embodiments can be implemented by a computer or its associated components. The computer can be, for example, a mobile terminal, a smart phone, a personal computer, a laptop computer, a vehicle-mounted human-computer interaction device, a personal digital assistant, a media player, a navigation device, a game console, a tablet computer, a wearable device, a smart TV, an Internet of Things system, a smart home, an industrial computer, a server or a combination thereof.

[0148] Although not shown, in an embodiment of the present invention, a program product is provided, including a computer program, and when the computer program is executed by a processor, the method of any one of the embodiments of the present invention is implemented.

[0149] Although not shown, in an embodiment of the present invention, a storage medium is provided, the storage medium storing a computer program, the computer program being configured to implement any method of the embodiments of the present invention when executed.

[0150] Storage media in embodiments of the present invention include permanent and non-permanent, removable and non-removable items that can be used to store information by any method or technology. Examples of storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0151] The methods, programs, systems, devices, etc. of the embodiments of the present invention may be executed or implemented in a single or multiple networked computers, or may be practiced in a distributed computing environment. In the embodiments of this specification, in these distributed computing environments, tasks may be performed by remote processing devices connected via a communication network.

[0152] Those skilled in the art should understand that the embodiments of the present specification may be provided as methods, systems or computer program products. Therefore, those skilled in the art may imagine that the functional modules / units or controllers and related method steps described in the above embodiments may be implemented in software, hardware or a combination of software / hardware.

[0153] Unless explicitly stated, the actions or steps of the methods, programs, and embodiments of the present invention do not have to be performed in a specific order and can still achieve the desired results. In some implementations, multitasking and parallel processing are also possible or may be advantageous.

[0154] In this article, multiple embodiments of the present invention are described, but for the sake of brevity, the description of each embodiment is not exhaustive, and the same or similar features or parts between the embodiments may be omitted. In this article, "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" are meant to be applicable to at least one embodiment or example according to the present invention, but not all embodiments. The above terms do not necessarily mean to refer to the same embodiment or example. In the absence of contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples.

[0155] The exemplary systems and methods of the present invention have been specifically shown and described with reference to the above embodiments, which are merely examples of the best modes for implementing the present systems and methods. It will be appreciated by those skilled in the art that various changes may be made to the embodiments of the systems and methods described herein when implementing the present systems and / or methods without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. A test paper generation method, characterized in that: include: Get a question bank including multiple questions; Determining a plurality of knowledge point feature vectors corresponding to a plurality of knowledge points, wherein the plurality of topics are associated with the plurality of knowledge points; Determining a preset number of target knowledge points at least according to the feature space distances between the plurality of knowledge point vectors; as well as Questions associated with the target knowledge points are selected from the question bank to generate a test paper.

2. The method according to claim 1, characterized in that: The method further comprises: determining a plurality of topic feature vectors corresponding to the plurality of topics; The step of selecting questions associated with the target knowledge point from the question bank to generate a test paper includes: Determining a first vector space similarity between a knowledge point feature vector of the target knowledge point and a topic feature vector of at least part of the topics; and A set of questions associated with the target knowledge point is determined according to the first vector space similarity.

3. The method according to claim 2, characterized in that The method further includes: determining a mastery feature vector representing the mastery degree of at least one examinee's knowledge point, wherein each examinee has a corresponding mastery feature vector, and the mastery feature vector has a dimension corresponding to the number of the plurality of knowledge points; The step of determining a preset number of target knowledge points at least based on the feature space distance between the plurality of knowledge point vectors includes: determining the preset number of target knowledge points based on the feature space distance and the mastery feature vector.

4. The method according to claim 3, characterized in that The step of determining the preset number of target knowledge points according to the feature space distance and the mastery feature vector includes: Determining feature space distances between the plurality of knowledge points; Determine the feature components of the corresponding dimensions of the plurality of knowledge points in the mastery feature vector; According to a weighted combination of the first weight of the feature space distance and the second weight based on the feature component, a preset number of knowledge points are determined as target knowledge points.

5. The method according to claim 3, characterized in that: The step of determining a mastery feature vector representing the mastery degree of at least one examinee's knowledge point comprises: creating an initial mastery feature vector for the at least one examinee; and The corresponding mastery feature vector is dynamically updated according to the question-solving history information of the at least one examinee, wherein at least part of the questions in the question-solving history information are associated with at least one of the knowledge points.

6. The method according to claim 5, characterized in that The dynamically updating the corresponding mastery feature vector according to the historical information of the questions includes: Determine the second vector space similarity between the test question vector of the question in the test question history information and the plurality of knowledge point feature vectors; and According to the problem-solving situation of the question and the similarity in the second vector space, the feature components of the dimensions corresponding to the knowledge points associated with the question in the mastery feature vector are updated.

7. The method according to claim 6, characterized in that The answering status of the question includes the answering result; The updating of the feature components of the dimension corresponding to the knowledge point associated with the question in the mastery feature vector according to the problem-solving situation of the question and the second vector space similarity includes: When the result of the test is wrong, reducing the feature component of the dimension corresponding to the associated knowledge point in the mastery feature vector according to the second vector space similarity; and When the result of the test is correct, the feature component of the dimension corresponding to the associated knowledge point in the mastery feature vector is increased according to the second vector space similarity.

8. The method according to any one of claims 2 to 7, characterized in that: The method further comprises: Detecting whether there are repeated questions in the question set; When there are repeated questions, redundant repeated questions are removed, and replacement questions are selected from the question bank based on the first vector space similarity until the preset question quantity requirement of the test paper is met.

9. A test paper generating device, characterized in that: include: An acquisition unit configured to acquire a question bank including a plurality of questions; A first determining unit is configured to determine a plurality of knowledge point feature vectors corresponding to a plurality of knowledge points, wherein the plurality of topics are associated with the plurality of knowledge points; A second determining unit is configured to determine a preset number of target knowledge points based at least on the feature space distances between the plurality of knowledge point vectors; as well as The generating unit is configured to select questions associated with the target knowledge point from the question bank, thereby generating a test paper.

10. An electronic device, characterized in that: include: A processor and a memory storing a computer program, wherein the processor is configured to implement the method according to any one of claims 1 to 8 when executing the computer program.

11. A program product, characterized in that The method comprises a computer program, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images