Test question recommendation method and device, equipment and storage medium

Abstract

The invention provides a test question recommendation method and device, equipment and a storage medium, which are applied to the technical field of computers, and the method comprises the following steps: determining a test question text included in a test question image according to the test question image input by a target user; performing semantic analysis and knowledge point recognition processing on the test question text by adopting a large model, and determining a target knowledge point corresponding to the test question in the test question text; according to each target knowledge point and a knowledge graph corresponding to the target user, determining target graph information corresponding to each target knowledge point in the knowledge graph; the target map information comprises the mastery degree of a target user on a corresponding target knowledge point, and the knowledge map comprises map information of the knowledge point and an association relationship; and classifying the test questions according to the mastery degree of the target user on each target knowledge point to obtain a test question set of a plurality of recommendation categories, and recommending the test questions according to the priority of the recommendation categories of each test question combination. By adopting the technical scheme of the invention, the review efficiency and review quality of students can be improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a test question recommendation method, device, equipment and storage medium. Background Art

[0002] At present, most students not only have to do the homework assigned by the school after class, but also need to select additional exercises (or test questions) from the teaching materials to consolidate the knowledge points.

[0003] In the related art, when students do homework and exercises in supplementary teaching materials after class, they usually finish all the homework and exercises in supplementary teaching materials to review relevant knowledge points.

[0004] However, as students have limited after-school time, it is generally difficult for them to complete all the exercises, which makes it difficult to ensure their review efficiency and quality. Summary of the invention

[0005] The present invention provides a test question recommendation method, device, equipment and storage medium, which are used to solve the defect in the prior art that it is difficult for general students to complete all exercises when their after-school time is relatively limited, which makes it difficult to ensure the students' review efficiency and review quality. The test questions are classified according to the students' mastery of knowledge points in the test questions to filter out some mastered test questions, thereby reducing the number of test questions for students to review and improving the review efficiency and review quality.

[0006] The present invention provides a test question recommendation method, comprising: Determine, according to the test image input by the target user, a test text included in the test image; the test text includes a plurality of test questions; Use a large model to perform semantic analysis and knowledge point recognition on the test text to determine the target knowledge point corresponding to each test question in the test text; According to the knowledge graphs corresponding to the target knowledge points and the target users, target graph information corresponding to the target knowledge points is determined in the knowledge graph; the target graph information includes the mastery of the target user on the corresponding target knowledge point, and the knowledge graph includes graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points; Multiple test questions are classified according to the target user's mastery of each target knowledge point to obtain a plurality of test question sets of recommended categories, and test questions are recommended according to the priority of the recommended categories of each test question combination; each test question combination of the recommended category includes at least one test question.

[0007] According to a test question recommendation method provided by the present invention, the above-mentioned classification of multiple test questions according to the mastery of each target knowledge point by the target user includes: Obtain the test frequency of each knowledge point within a preset time in the target user's area, and determine the frequency category corresponding to each knowledge point according to the test frequency of each knowledge point; Determine a target frequency category corresponding to each target knowledge point in the frequency category of each knowledge point; According to the mastery of each target knowledge point by the target user and the target frequency label corresponding to each target knowledge point, multiple test questions are classified.

[0008] According to a test question recommendation method provided by the present invention, the multiple test questions are classified according to the mastery of each target knowledge point by the target user and the target frequency label corresponding to each target knowledge point, including: Determine the target fallibility corresponding to each target knowledge point; the target fallibility is used to characterize the fallibility of any user and / or target user for the target knowledge point; According to the mastery of each target knowledge point by the target user, the target frequency label corresponding to each target knowledge point, and the target error proneness corresponding to each target knowledge point, multiple test questions are classified.

[0009] According to a test question recommendation method provided by the present invention, the above-mentioned determination of the target fallibility corresponding to each target knowledge point includes: Obtaining the big data of the examination situation in the area where the target user is located, determining the first fallibility corresponding to each knowledge point according to the big data of the examination situation, and determining the first fallibility corresponding to each target knowledge point from the first fallibility of each knowledge point; Determine the second fallibility of the target user for each target knowledge point based on the target user's practice records of each knowledge point in a preset learning system at a historical time; the preset learning system includes a knowledge graph; The target fallibility of each target knowledge point is determined based on the first fallibility of each target knowledge point and the second fallibility of each target knowledge point.

[0010] According to a test question recommendation method provided by the present invention, the above-mentioned obtaining a set of test questions of multiple recommended categories includes: Determine the recommended category for each test question based on the target user's mastery of each target knowledge point, the target frequency label corresponding to each target knowledge point, and the target error proneness corresponding to each target knowledge point; The test questions with the same recommended category in each test question are grouped into the same test question set to obtain a test question set with multiple recommended categories; the above multiple recommended categories include questions not to be done, optional questions and required questions.

[0011] According to a test question recommendation method provided by the present invention, the test question recommendation is performed according to the priority of the recommendation category of each test question combination, including: Sorting the test question sets of multiple recommended categories from high to low according to the priority of the recommended categories, and recommending test questions according to the sorting results; The priorities of the above-mentioned multiple recommendation categories from high to low are required questions, optional questions, and no questions.

[0012] According to a test question recommendation method provided by the present invention, the above-mentioned classification of multiple test questions according to the mastery of each target knowledge point by the target user includes: Determine the target user's personal learning profile based on the target user's practice records of each knowledge point in the preset learning system at a historical time; the personal learning profile includes the target user's learning ability; Multiple test questions are classified according to the target users' mastery of each target knowledge point and the target users' learning ability.

[0013] According to a test question recommendation method provided by the present invention, the above-mentioned classification of multiple test questions according to the mastery of each target knowledge point by the target user includes: Determine the difficulty level of each knowledge point based on the test situation big data of the target user's area, and determine the difficulty level of each target knowledge point from the difficulty levels of each knowledge point; Multiple test questions are classified according to the target users' mastery of each target knowledge point and the corresponding difficulty level of each target knowledge point.

[0014] The present invention also provides a test question recommendation device, comprising the following modules: A test question text determination module is used to determine the test question text included in the test question image according to the test question image input by the target user; the test question text includes multiple test questions; The knowledge point recognition module is used to perform semantic analysis and knowledge point recognition on the test text using a large model to determine the target knowledge point corresponding to each test question in the test text; A graph information determination module is used to determine target graph information corresponding to each target knowledge point in the knowledge graph according to the knowledge graph corresponding to each target knowledge point and the target user; the target graph information includes the mastery of the corresponding target knowledge point by the target user, and the knowledge graph includes graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points; The classification and recommendation module is used to classify multiple test questions according to the target user's mastery of each target knowledge point, obtain a set of test questions in multiple recommended categories, and recommend test questions according to the priority of the recommended category of each test question combination; each test question combination in the recommended category includes at least one test question.

[0015] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the test question recommendation method as described above is implemented.

[0016] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the test question recommendation method as described in any one of the above is implemented.

[0017] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the test question recommendation method as described above is implemented.

[0018] The test question recommendation method, device, equipment and storage medium provided by the present invention determine the test question text included in the test question image based on the test question image input by the target user, then use a large model to perform semantic analysis and knowledge point recognition processing on the test question text, determine the target knowledge point corresponding to each test question in the test question text, and determine the target graph information corresponding to each target knowledge point in the knowledge graph based on the knowledge graph corresponding to each target knowledge point and the target user, and then classify multiple test questions according to the target user's mastery of the target knowledge point in the target graph information to obtain a set of test questions of multiple recommended categories, and recommend test questions according to the priority of each recommended category; wherein the test question text includes multiple test questions, the knowledge graph includes graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points, and each recommended category of test question combination includes at least one test question. In this method, since the students' mastery of the knowledge points in the test questions can be obtained through the students' knowledge graph, and the test questions are classified according to the students' mastery of the knowledge points in the test questions to filter out some mastered test questions according to the recommended category priority, this can reduce the number of test questions that students review and the recommended test questions are all in line with the students' own learning situation, thereby improving the students' review efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 This is one of the flow charts of the test question recommendation method provided by the present invention.

[0021] Figure 2 This is the second flow chart of the test question recommendation method provided by the present invention.

[0022] Figure 3 This is the third flow chart of the test question recommendation method provided by the present invention.

[0023] Figure 4 It is a structural schematic diagram of the test question recommendation device provided by the present invention.

[0024] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] At present, most students not only need to complete the homework assigned by the teacher after school, but also choose exercises in the supplementary teaching materials for consolidation. However, the superposition of extracurricular homework and a large number of exercises, especially when students have limited time, makes it difficult for students to complete all the questions, resulting in difficulty in ensuring review efficiency and quality. In the traditional review mode, students with different mastery levels and different learning levels do exactly the same exercises, and students themselves often find it difficult to judge which questions are "must-do questions" that meet their own learning situation, and which are "optional questions" that can be completed selectively. This lack of targeted practice method can easily lead students to invest a lot of time in the content they have mastered, affecting the consolidation and breakthrough of weak knowledge points, and the effect of homework practice or review is greatly reduced. Based on this, the embodiment of the present invention provides a test question recommendation method, device, equipment and storage medium, which can solve this technical problem.

[0027] Combine the following Figure 1-Figure 3 The test question recommendation method according to an embodiment of the present invention is described.

[0028] It should be noted that the execution subject of the embodiment of the present invention can be a test question recommendation device, or an electronic device, or a preset learning system installed in the electronic device, or a background server connected to the electronic device, etc. This embodiment does not make specific limitations on this. The following embodiments will take the electronic device as the execution subject as an example for explanation.

[0029] It is understandable that the electronic device may be a terminal, such as a mobile phone, a tablet computer, a learning machine, a personal computer, a desktop computer, a wearable device, etc. The electronic device may include an information or data input device, and may also include a display interface, etc.

[0030] Figure 1 It is one of the flow charts of the test question recommendation method provided by the present invention, such as Figure 1 As shown, the method comprises the following steps: S102, determining a test question text included in the test question image according to the test question image input by the target user; the test question text includes a plurality of test questions.

[0031] The target user refers to a user who uses an electronic device to review test questions, for example, the target user may be a student who is reviewing test questions.

[0032] Specifically, an image acquisition device (such as a camera) may be used to photograph the paper test papers of the target user to obtain test paper images, and the test paper images may be input or imported into the electronic device; or an image acquisition device on the electronic device may be directly used to photograph the paper test papers of the target user to obtain test paper images, and the test paper images may be input or imported into the electronic device; or other methods may be used, as long as the test paper images input by the target user can be obtained.

[0033] It is understandable that the test question image input each time can be a single image or a video including multiple images. Each test question image can include multiple test questions. The paper test questions of the target user can be the test questions in the target user's homework or the test questions in the target user's extracurricular auxiliary materials.

[0034] After the electronic device obtains the test question image input by the user, optionally, the question knowledge point parsing module in the electronic device can use an OCR (Optical Character Recognition) algorithm to recognize the characters in the test question image, obtain the characters included in the test question image, and then obtain the test question text through the recognized characters, which generally can include multiple test questions.

[0035] S104, using a large model to perform semantic analysis and knowledge point recognition processing on the test text, and determine the target knowledge point corresponding to each test question in the test text.

[0036] The large model may be a large language model. After obtaining the test text, the test text may be input into the large model. In the large model, semantic analysis may be performed on the content of each test question in the test text to obtain the semantic analysis result of each test question. Then, the knowledge points in each test question may be identified based on the semantic analysis result of each test question to obtain the identified knowledge points in each test question.

[0037] Optionally, each test question may include one or more knowledge points. When the large model identifies the knowledge points in each test question, the weight of the knowledge points identified in each test question may be set in the large model, and finally the core knowledge points of each test question may be identified and output. The core knowledge points here may be, for example, the knowledge points with the largest weight among the knowledge points of each test question. The size of the weight here represents the importance of the knowledge point. In other words, the target knowledge point of each test question finally determined is the core knowledge point of each test question, and the core knowledge point is the most important knowledge point in the test question. Generally, each test question has one core knowledge point. Of course, when the weights of multiple knowledge points are set to be the same in the large model, it can also be determined that there are multiple core knowledge points for a test question.

[0038] Here, by identifying the core knowledge points of each test question as the target knowledge points, it is convenient to recommend accurate test questions to target users in the future.

[0039] S106, according to the knowledge graphs corresponding to the target knowledge points and the target users, determine the target graph information corresponding to the target knowledge points in the knowledge graph; the target graph information includes the mastery of the corresponding target knowledge points by the target users, and the knowledge graph includes the graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points.

[0040] In this step, the knowledge graph of the target user can be pre-constructed through the knowledge graph construction module in the electronic device. The specific construction method may include: importing the textbook content of any subject (such as mathematics) of the target user into the electronic device, and extracting the knowledge points in each chapter of the textbook, and then determining the graph information of each knowledge point, taking each knowledge point and its graph information as a node, and connecting each node in sequence according to the previous and next level association relationship (or dependency relationship) between each knowledge point to obtain a constructed knowledge graph.

[0041] For the graph information of each of the above-mentioned knowledge points, the target user's learning data may be included, and the learning data may include, for example, the target user's mastery of the corresponding knowledge point, and of course, other information or data may also be included, such as the textbook chapter information to which the knowledge point belongs. Among them, the target user's mastery of each knowledge point indicates the target user's mastery of the knowledge point, and the mastery may be expressed in the form of a numerical percentage, or may be divided into multiple categories according to actual conditions, such as mastered, partially mastered, not mastered, unclear mastery, etc. In addition, for the mastery of different categories of each knowledge point, it can be quickly distinguished by marking / displaying different colors in the nodes of the corresponding knowledge points in the knowledge graph, such as when the corresponding node of the knowledge point is marked in green, it means that the target user has mastered the knowledge point, when the corresponding node of the knowledge point is marked in yellow, it means that the target user has partially mastered the knowledge point, when the corresponding node of the knowledge point is marked in red, it means that the target user has not mastered the knowledge point, and when the corresponding node of the knowledge point is marked in gray, it means that the target user's mastery of the knowledge point is unclear.

[0042] Furthermore, the target user's mastery of each knowledge point can be determined through the learning situation analysis module in the electronic device based on the target user's practice records of each knowledge point test in historical time. After each practice or a set number of practices by the target user, the electronic device or a preset learning system (a learning system specifically used by students to practice test questions, such as a precision learning system) can automatically analyze the target user's mastery of each knowledge point, and then update the graph information of the corresponding knowledge point in the target user's knowledge graph.

[0043] It is understandable that the knowledge graph corresponding to each user is constructed based on the user's own learning situation, so the knowledge graph corresponding to each user is different. The knowledge graph of the target user constructed here can provide structured data support for subsequent personalized test question recommendations.

[0044] Specifically, after the target knowledge points of each test question are identified through the large model, the target knowledge points of each test question can be matched in the knowledge graph of the target user constructed above, and the graph information corresponding to the knowledge points matched by each target knowledge point in the knowledge graph is obtained, which is recorded as the target graph information. Optionally, the target graph information of each target knowledge point may include the target user's mastery of the target knowledge point, or may also include other information, such as the target user's fallibility for the target knowledge point, the frequency of examinations for the target knowledge point in the target user's area, and other information. Here, by mapping the identified test question knowledge points to the corresponding knowledge points and their graph information in the target user's knowledge graph, it can be ensured that the knowledge points corresponding to the test questions can be effectively associated with the target user's academic data, so as to facilitate the subsequent personalized test question recommendations for the target user.

[0045] S108, classifying multiple test questions according to the target user's mastery of each target knowledge point, obtaining multiple test question sets of recommended categories, and recommending test questions according to the priority of the recommended categories of each test question combination; each test question combination of the recommended category includes at least one test question.

[0046] In this step, after obtaining the mastery of each target knowledge point of each test question by the target user, the personalized homework recommendation module in the electronic device can be used to directly classify each test question according to the mastery of each target knowledge point of each test question by the target user, such as classifying the test questions corresponding to the mastered target knowledge points into the first category, classifying the test questions corresponding to the partially mastered target knowledge points into the second category, classifying the test questions corresponding to the unmastered target knowledge points into the third category, and classifying the test questions corresponding to the target knowledge points whose mastery is unclear into the fourth category. Of course, each test question can also be classified by mastery combined with other data. The number of specific categories is not specifically limited here.

[0047] In addition, if there are multiple target knowledge points corresponding to each test question and the mastery of the multiple target knowledge points is different, the mastery of all target knowledge points in the test question can be determined by weighted average or other methods, and the test questions can be classified accordingly. Alternatively, classification can be adopted, which is not specifically limited here.

[0048] It should be noted that, after dividing the multiple test questions into multiple categories, the multiple test questions can be divided into multiple groups of test question sets, and then each group of test question sets can be presented to the target user in a list classification. At the same time, each group of test question sets corresponds to a category (which can be recorded as a recommended category), and each group of test question sets corresponds to a different recommended category, and the recommendation priority of different recommended categories is also different. Then as an optional embodiment, the test question sets of multiple recommended categories can be sorted from high to low according to the priority of the recommended category, and test questions can be recommended according to the sorting results; the priorities of the above multiple recommended categories are from high to low, namely, required questions, optional questions, and no-questions.

[0049] In other words, for example, the recommendation category of the unmastered test questions is generally classified as must-do questions, which have the highest recommendation priority, and the recommendation category of the mastered test questions is generally classified as do not-do questions, which have the lowest recommendation priority. This allows the target users to pay more attention to the unmastered test questions and ignore the mastered test questions, thereby reducing the review time for the mastered test questions to improve the review efficiency, and at the same time pay more attention to the unmastered test questions, so as to effectively and specifically improve the target users' review quality.

[0050] In addition, when recommending test questions based on the ranking results of the recommendation categories from high to low priority, the labels of the recommendation categories can be displayed in sequence in the display interface according to the ranking results. When the target user clicks on a label of a recommendation category in the display interface, the display interface can display the corresponding test question set under the label of the recommendation category. Among them, the label of the recommendation category with a higher priority is displayed more forward, so that the user can pay more attention to the test questions under the recommendation category with a higher priority for review.

[0051] As an optional embodiment, when recommending test questions, the test question set with the lowest priority can also be filtered out, and only one or more test question sets with higher priority can be recommended. This can specifically reduce the number of test questions that students review, provide personalized test question recommendations for users, and improve users' review efficiency.

[0052] In this embodiment, the test question text included in the test question image is determined based on the test question image input by the target user, and then the large model is used to perform semantic analysis and knowledge point recognition processing on the test question text to determine the target knowledge point corresponding to each test question in the test question text, and based on the knowledge graph corresponding to each target knowledge point and the target user, the target graph information corresponding to each target knowledge point is determined in the knowledge graph, and then multiple test questions are classified according to the target user's mastery of the target knowledge point in the target graph information to obtain a set of test questions of multiple recommended categories, and test questions are recommended according to the priority of each recommended category; wherein the test question text includes multiple test questions, the knowledge graph includes graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points, and each recommended category of test question combination includes at least one test question. In this method, since the students' mastery of the knowledge points in the test questions can be obtained through the students' knowledge graph, and the test questions are classified according to the students' mastery of the knowledge points in the test questions to filter out some mastered test questions according to the recommended category priority, this can reduce the number of test questions that students review and the recommended test questions are all in line with the students' own learning situation, thereby improving the students' review efficiency and quality.

[0053] At present, there are differences in the emphasis of examinations in different regions. Some knowledge points are frequently tested in some regions, but relatively rarely in other regions. Under the same homework and exercises, students cannot effectively identify the key questions of the homework. In addition, facing the unchanging homework questions, students' learning enthusiasm during the review process is also affected. Although there are some intelligent learning platforms on the market that can recommend exercises, there is still a lack of accurate analysis and screening of individual students' mastery and test frequency, and it is impossible to make intelligent choices based on students' weak knowledge. Based on this, an embodiment of the present invention provides a solution for recommending test questions in combination with the test frequency of knowledge points in the region.

[0054] Figure 2 FIG. 2 is a flow chart of the test question recommendation method provided by the present invention. Figure 2 As shown, the "classifying multiple test questions according to the mastery of each target knowledge point by the target user" in the above S108 may include the following steps: S202, obtaining the test frequency of each knowledge point within a preset time in the area where the target user is located, and determining the frequency category corresponding to each knowledge point according to the test frequency of each knowledge point.

[0055] The preset time may be a time period set according to actual conditions, such as three years or five years closest to the current time.

[0056] In this step, for a certain subject, all the test questions (including large and small test questions) for the subject in the target user's area within the preset time can be obtained, and the knowledge points in all the test questions can be semantically analyzed and recognized through the large model to identify all the knowledge points in all the test questions; at the same time, the number of examinations for each knowledge point in all the test questions (i.e., the test frequency) can be counted.

[0057] In addition, different frequency categories (also called test frequency categories) can be pre-set here, and each frequency category corresponds to a test frequency range. After obtaining the test frequency of each knowledge point in all the test questions, the test frequency of each knowledge point can be matched with each test frequency range, and the test frequency category corresponding to the matched test frequency range is the frequency category of the knowledge point. The number of frequency categories here can be set according to actual conditions, for example, it can be set to three categories, high-frequency test points, medium-frequency test points and low-frequency test points, where the test frequencies of high-frequency test points, medium-frequency test points and low-frequency test points are reduced in turn, that is, the high-frequency test points correspond to the knowledge points with the highest number of tests / test frequencies. In addition, the above-mentioned test frequency range can include only one test frequency threshold, or it can also include a range consisting of an upper limit value and a lower limit value.

[0058] S204, determining a target frequency category corresponding to each target knowledge point in the frequency category of each knowledge point.

[0059] In this step, for a certain subject, after obtaining the frequency categories of the knowledge points involved in all the test questions in the target user's area, the corresponding knowledge points can be found in the knowledge graph of the target user, and the frequency category information determined by the system can be added to the graph information of the corresponding knowledge points.

[0060] After obtaining the target knowledge points corresponding to the multiple test questions inputted above, the target graph information corresponding to the target knowledge point can be obtained in the knowledge graph, and the corresponding frequency category can be obtained from the target graph information, which is recorded as the target frequency category.

[0061] S206, classifying the plurality of test questions according to the target user's mastery of each target knowledge point and the target frequency label corresponding to each target knowledge point.

[0062] In this step, after obtaining the target user's mastery of each target knowledge point and the target frequency label corresponding to each target knowledge point, each test question can be directly classified based on these two parts of information of each test question, or each test question can be classified in combination with other information. In short, the recommended category corresponding to each test question can be obtained.

[0063] For example, the recommended category of test questions corresponding to target knowledge points that have been mastered and are low-frequency test points is determined as do not do questions, the recommended category of test questions corresponding to target knowledge points that have not been mastered and are high-frequency test points is determined as required questions, and the recommended category of test questions corresponding to target knowledge points that have not been mastered and are low-frequency test points is determined as optional questions.

[0064] In this embodiment, multiple test questions are classified based on the target user's mastery of each target knowledge point and the examination frequency of each target knowledge point in the target user's area. This can help students effectively distinguish between key questions and non-key questions in their area, ensuring that students in the area can review test questions in a targeted manner, further improving review efficiency and quality.

[0065] The following embodiment illustrates the process of further combining the test information big data to classify the test questions.

[0066] Figure 3 FIG. 3 is a flow chart of the test question recommendation method provided by the present invention. Figure 3 As shown, the above S206 may include the following steps: S302, determining the target fallibility corresponding to each target knowledge point; the target fallibility is used to characterize the fallibility of any user and / or target user for the target knowledge point.

[0067] Among them, fallibility refers to the degree of error-proneness in a knowledge point. For example, it can be expressed as the probability of students making mistakes when answering test questions corresponding to the knowledge point.

[0068] In this embodiment, the target fallibility of the target knowledge point may include the overall fallibility of the target knowledge point by all users or any user in the target user's area, or may include the local fallibility of the target user himself to the target knowledge point. As an optional embodiment, the following steps may be used to determine the target fallibility: Obtain exam big data in the area where the target user is located, determine the first fallibility corresponding to each knowledge point based on the exam big data, and determine the first fallibility corresponding to each target knowledge point from the first fallibility of each knowledge point; determine the second fallibility of each target knowledge point by the target user based on the target user's practice records of each knowledge point in a preset learning system at a historical time; the above-mentioned preset learning system includes a knowledge graph; determine the target fallibility of each target knowledge point based on the first fallibility of each target knowledge point and the second fallibility of each target knowledge point.

[0069] Among them, the test big data may include the test results of all users in the area where the target user is located on each knowledge point and the difficulty of each knowledge point, among which the test results may include the error rate, correct rate, number of errors, number of correct answers and other data of all users on each knowledge point.

[0070] After obtaining the big data of the examination situation, for example, the error rate of each knowledge point for all users in the big data of the examination situation can be used to determine the fallibility of each knowledge point. For example, the error rate of each knowledge point can be used as the fallibility of each knowledge point, and all are recorded as the first fallibility. At the same time, the target user can practice the test questions in a preset learning system (such as a precision learning system) including the knowledge graph in the previous / historical time, so that the practice record of the target user in the preset learning system can be obtained. Each test question will have a corresponding knowledge point, and the practice record may include the number of errors or error rates of each knowledge point when the target user practices in the historical time, and then the fallibility of the target user for the corresponding knowledge point can be determined by the number of errors or error rates of each knowledge point, which is recorded as the second fallibility. Afterwards, any one of the first fallibility and the second fallibility of each knowledge point can be used as the target fallibility of the knowledge point, or the first fallibility and the second fallibility of each knowledge point can be used as the target fallibility of the knowledge point.

[0071] S304, classifying the plurality of test questions according to the mastery of each target knowledge point by the target user, the target frequency label corresponding to each target knowledge point, and the target fallibility corresponding to each target knowledge point.

[0072] In this step, after obtaining the target user's mastery of each target knowledge point, the target frequency label corresponding to each target knowledge point, and the target fallibility of each target knowledge point, each test question can be directly classified based on these three parts of information of each test question, or each test question can be classified in combination with other information. In short, the recommended category corresponding to each test question can be obtained.

[0073] Optionally, the specific classification process may include the following steps: Determine the recommended category corresponding to each test question based on the target user's mastery of each target knowledge point, the target frequency label corresponding to each target knowledge point, and the target error-proneness corresponding to each target knowledge point; group the test questions with the same recommended category in each test question into the same test question set to obtain a test question set of multiple recommended categories; the above multiple recommended categories include questions not to be done, optional questions, and required questions.

[0074] For example, the recommended category of test questions corresponding to low-frequency test points in the local area and target knowledge points with a low error rate (i.e., not easy to make mistakes) that have been mastered can be determined as not to do questions. When test questions are recommended, they are automatically filtered to remind target users not to practice such test questions.

[0075] For example, the recommended categories of test questions corresponding to target knowledge points that have not been mastered and are frequently tested in the local area, as well as target knowledge points with a high error rate (i.e., easy to make mistakes, such as making multiple mistakes on this knowledge point in recent weeks) can be identified as required questions. This can help students focus on their weak and frequently tested knowledge points, ensuring that they are practiced first.

[0076] For example, the recommended categories of test questions corresponding to low-frequency test points in the local area that have not been mastered and target knowledge points with a low degree of error can be determined as optional questions. Students can choose whether to practice them according to their own time arrangements. This can improve students' review or practice efficiency.

[0077] Through the above method, each test question in the test question image can be divided into a corresponding recommendation category, and then the test questions of the same recommendation category can be put into the same test question set, so that multiple test question sets of different recommendation categories can be obtained, and then the test question sets can be screened / filtered, displayed, recommended, etc. according to different recommendation categories.

[0078] In this embodiment, multiple test questions are classified based on the target user's mastery of each target knowledge point and the examination frequency of each target knowledge point in the target user's area, and combined with the fallibility of the target knowledge point. This makes it easy to distinguish multiple test questions based on the student's own local learning situation and the overall learning situation of the region. The test questions divided in this way are more suitable for the students themselves, and can therefore further effectively improve the students' review efficiency and quality.

[0079] The following embodiment illustrates the process of further classifying test questions in combination with the user's personal learning ability.

[0080] In some embodiments, the “classifying multiple test questions according to the target user's mastery of each target knowledge point” in S108 may include the following steps: Determine the target user's personal learning profile based on the target user's practice records of each knowledge point in the preset learning system at a historical time; the personal learning profile includes the target user's learning ability; Multiple test questions are classified according to the target users' mastery of each target knowledge point and the target users' learning ability.

[0081] Among them, after the target user practices knowledge point-related test questions in the preset learning system, a practice record of the target user can be generated. The practice record may include data such as the number of errors or error rate, the number of correct answers or correct rate, practice time, mastery of each knowledge point, mastery of the prerequisite knowledge points of each knowledge point, and mastery of the prerequisite knowledge points of each knowledge point when the target user practices test questions in historical time.

[0082] After the target user's practice record is generated, a personal learning profile of the target user can be generated based on the data of multiple dimensions in the practice record. The personal learning profile can include the target user's learning ability, or can also include other data. The target user's learning ability can include the target user's overall learning ability, or can also include the target user's local learning ability for each knowledge point.

[0083] After obtaining the personal learning portrait of the target user, the test questions in the test image can be directly classified based on the target user's mastery of each target knowledge point, or other data such as the target frequency label of each target knowledge point, the target error-proneness of each target knowledge point, etc. can be further combined to jointly classify the test questions in the test image.

[0084] In this embodiment, the test questions in the test question image are classified based on the target user's mastery of each target knowledge point and the target user's personal learning ability. In this way, the classified test questions are more in line with the user's personal learning needs, thereby effectively ensuring the user's review efficiency and review quality.

[0085] The following embodiment illustrates the process of further classifying test questions in combination with the difficulty of knowledge points.

[0086] In some embodiments, the “classifying multiple test questions according to the target user's mastery of each target knowledge point” in S108 may include the following steps: Determine the difficulty level of each knowledge point based on the test situation big data of the target user's area, and determine the difficulty level of each target knowledge point from the difficulty levels of each knowledge point; Multiple test questions are classified according to the target users' mastery of each target knowledge point and the corresponding difficulty level of each target knowledge point.

[0087] Among them, the difficulty of a knowledge point refers to the difficulty probability of the knowledge point. For example, a knowledge point with a smaller difficulty probability can be considered to have a relatively easy difficulty level, and a knowledge point with a larger difficulty probability can be considered to have a relatively difficult difficulty level.

[0088] The above-mentioned big test data may include the degree of difficulty of each knowledge point, so the degree of difficulty of each knowledge point can be directly obtained by analyzing the big test data. Or the big test data includes data such as the error rate and correct rate of each knowledge point for all users, so the degree of difficulty of each knowledge point can be obtained by analyzing the error rate and correct rate of each knowledge point. For example, knowledge points with higher error rates can be determined as test questions with higher difficulty.

[0089] After obtaining the difficulty level of each knowledge point, the test questions in the test image can be directly classified based on the target user's mastery of each target knowledge point, or the test questions in the test image can be further classified based on other data such as the target user's personal learning ability, the target frequency label of each target knowledge point, the target error-proneness of each target knowledge point, etc.

[0090] After classification, in order to make personalized test question recommendations for target users, for multiple test questions in the test question set of each recommended category, the recommended order of practice for each test question can be given comprehensively according to the difficulty of each test question. For example, the test questions in the test question set can be recommended to the target user in order of difficulty from low to high. This will allow students to practice test questions with increasing difficulty, improve their sense of accomplishment in daily practice, and then enhance their enthusiasm for learning, making their review more efficient and effective.

[0091] In this embodiment, the test questions in the test question image are classified based on the target user's mastery of each target knowledge point and the difficulty of each target knowledge point. In this way, each classified test question is more in line with the user's personal learning needs, thereby effectively ensuring the user's review efficiency and review quality.

[0092] The technical solution of the embodiment of the present invention is illustrated below by an application example, and a preset learning system in an electronic device is used as an execution subject for explanation.

[0093] Assume that a seventh-grade student in a certain province uses the preset learning system in this embodiment to analyze and screen questions in mathematics and uploads some exercises in the homework. The student has previously used the preset learning system to learn the knowledge points of the chapter "Polymers and Addition and Subtraction". The preset learning system records all the knowledge points and their associations in the chapter in the student's knowledge map, and the student's mastery of each knowledge point (i.e., mastery) can be obtained through the student's previous adaptive practice records.

[0094] After the student takes a photo and uploads the practice questions, the preset learning system can automatically identify the content of the questions and analyze the knowledge points, and then obtain the knowledge points associated with each question, as well as the local test data and student mastery of the knowledge points in the knowledge graph. Based on this, the following gives several related knowledge point test question classifications and recommended examples.

[0095] 1. Question 1 tests the knowledge point: addition and subtraction operations of polynomials (arrangement in descending powers).

[0096] The preset learning system recognizes that the knowledge point corresponding to question 1 is the addition and subtraction of integers (descending power arrangement). The mastery degree of this knowledge point is displayed as 93% on the knowledge graph (the mastery degree is greater than or equal to 80%, which means it has been mastered). The student has also mastered the prerequisite basic knowledge points and subsequent advanced knowledge points of this knowledge point. This shows that the student has fully learned and mastered the relevant content of this knowledge point in previous studies, and this knowledge point is not frequently tested in local exams. It has only been tested 6 times in the past three years. Taking all dimensions into consideration, the preset learning system divides / marks this question as "Do not do question (also known as ignorable question)" to prevent students from repeating and ineffective practice on such questions.

[0097] 2. Question 2 tests knowledge points: related concepts of polynomials (differentiation).

[0098] The preset learning system identified that the knowledge point corresponding to question 2 is the concept related to polynomials (discrimination). This knowledge point has been tested 260 times in local exams in the past three years, which is a very high-frequency test point. However, this knowledge point is a basic test point with relatively low overall difficulty, and students generally make fewer mistakes. At the same time, it was identified that the student had just answered correctly under this knowledge point in the past week. Combining various dimensions, the preset learning system can classify / mark this question as "do not do question (also known as ignorable question)" to prevent the student from repeating ineffective practice.

[0099] 3. Question 3 tests the knowledge point: definition of like terms (evaluation).

[0100] The preset learning system identified that the knowledge point corresponding to question 3 is the definition (evaluation) of like items. This knowledge point has been tested 704 times in local exams in the past three years. It is an important high-frequency test point, and students have a high error rate. The students' mastery of this knowledge point in the preset learning system is 56% (less than 80%, which is partial mastery), which is relatively weak. In addition, the student has made two mistakes in the past week, which indicates that the student has a big deficiency in this knowledge point and is likely to lose points in the exam. Combining various dimensions, the preset learning system can classify / mark this question as a "must-do question" and strongly recommends that students complete it to ensure their proficiency in high-frequency test points.

[0101] 4. Question 4 tests knowledge points: Simplification and evaluation of polynomials (substituting the whole into the equation and evaluating).

[0102] The preset learning system identified that the knowledge point corresponding to question 4 is the simplification and evaluation of the whole expression (substitution and evaluation). The preset learning system shows that the student has never answered and studied this knowledge point, that is, the current mastery of this knowledge point is unknown. By analyzing the test situation in the student's area, this knowledge point is a high-frequency test point in the exam, and has been tested 165 times in the past three years. Based on various dimensions, the preset learning system can classify / mark this question as "optional question" and recommend that students decide whether to complete it based on their review time and needs.

[0103] It can be seen from the above examples that the preset learning system in the embodiment of the present invention can utilize the knowledge graph and the students' learning data, combined with regionalized local test analysis, to intelligently screen out questions suitable for students' current learning situation, help students reasonably allocate time for homework questions, and avoid spending too much time on "invalid questions", so that students can allocate more time to valuable questions and weak points, thereby improving review efficiency and review quality.

[0104] Based on this, the technical solution of the embodiment of the present invention has the following technical effects: 1. Reduce repetitive exercises and improve review efficiency.

[0105] In the traditional student homework practice mode, students need to complete a large number of homework and supplementary exercises after class. They are usually unable to distinguish between the knowledge points they have mastered and those they have not mastered, and are easily trapped in the dilemma of repeated practice. In the embodiment of the present invention, by performing correlation analysis between the student's homework questions and the learning data under the knowledge graph, the questions corresponding to the knowledge points that the students have mastered are filtered out, thereby avoiding unnecessary repeated practice, allowing students to focus their time on the exercises corresponding to the knowledge points that have not been mastered, and significantly improving the students' review efficiency.

[0106] 2. Help students identify their weaknesses and focus on practicing and mastering them.

[0107] In the traditional student homework practice mode, it is often difficult for students to accurately identify their weak knowledge points, which easily leads to a lack of focus in homework practice and an inability to overcome weak knowledge points in a timely manner. In the embodiment of the present invention, based on the knowledge map and learning data formed by students' previous practice, the knowledge mastery behind the students' homework questions is analyzed, and questions that belong to students' weak points and are high-frequency test points are marked as "must-do questions", ensuring that students practice these homework first to overcome their weaknesses, improve the effect and pertinence of daily homework practice, and save time costs.

[0108] 3. Based on the test data, optimize the review priority.

[0109] In the traditional student homework practice mode, it is difficult for students to distinguish the priority of practice questions and cannot reasonably arrange time to practice key high-frequency test points. In the embodiment of the present invention, by adding the analysis data of regional test frequencies, the knowledge points are divided into multiple frequency categories according to the test frequencies, and students are given priority to practice the exercises of high-frequency test points, thereby effectively solving the problem of no focus between the test situation analysis and the students' actual homework practice, making the students' daily homework practice process more practical and more targeted.

[0110] 4. Personalized recommendations to enhance students’ learning initiative and confidence.

[0111] In the traditional student homework practice mode, students are prone to fatigue and frustration because the questions do not have personalized recommendations. In the embodiment of the present invention, personalized screening and scientific practice classification (for example, divided into multiple recommendation categories such as required questions, optional questions, and non-questions) help students more clearly understand the focus and priority of their next practice. At the same time, the system will also give a recommended order of test questions based on the difficulty of the questions and the current learning situation of the students, so that the difficulty of the test questions practiced by students continues to increase, improving the sense of accomplishment of students' daily practice, and at the same time, it can improve students' learning enthusiasm, making review more efficient and more effective.

[0112] In summary, the embodiments of the present invention provide a more targeted, effective and personalized test question recommendation solution, which significantly solves the problems of students' repeated practice of knowledge points in daily homework exercises and their inability to efficiently screen exercises suitable for themselves, making students' daily homework exercises more scientific and efficient.

[0113] The test question recommendation device provided by the present invention is described below. The test question recommendation device described below and the test question recommendation method described above can be referenced to each other.

[0114] Figure 4 is a schematic diagram of the structure of the test question recommendation device provided by the present invention, see Figure 4 As shown, the device may include: The test question text determination module 410 is used to determine the test question text included in the test question image according to the test question image input by the target user; the test question text includes multiple test questions; The knowledge point identification module 420 is used to perform semantic analysis and knowledge point identification processing on the test text using a large model to determine the target knowledge point corresponding to each test question in the test text; A graph information determination module 430 is used to determine target graph information corresponding to each target knowledge point in the knowledge graph according to the knowledge graph corresponding to each target knowledge point and the target user; the target graph information includes the mastery of the corresponding target knowledge point by the target user, and the knowledge graph includes graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points; The classification and recommendation module 440 is used to classify multiple test questions according to the target user's mastery of each target knowledge point, obtain a set of test questions in multiple recommended categories, and recommend test questions according to the priority of the recommended category of each test question combination; each test question combination in the recommended category includes at least one test question.

[0115] Optionally, the above-mentioned classification and recommendation module 440 is specifically used to sort the question sets of multiple recommended categories from high to low according to the priority of the recommended categories, and recommend questions according to the sorting results; the priorities of the above-mentioned multiple recommended categories from high to low are compulsory questions, optional questions and no-questions.

[0116] In some embodiments, the classification and recommendation module 440 may include: A frequency category determination unit, used to obtain the test frequency of each knowledge point within a preset time in the area where the target user is located, and determine the frequency category corresponding to each knowledge point according to the test frequency of each knowledge point; A target frequency category determination unit, used to determine a target frequency category corresponding to each target knowledge point in the frequency category of each knowledge point; The classification unit is used to classify multiple test questions according to the target user's mastery of each target knowledge point and the target frequency label corresponding to each target knowledge point.

[0117] In some embodiments, the above-mentioned classification unit is specifically used to determine the target fallibility corresponding to each target knowledge point; the above-mentioned target fallibility is used to characterize the fallibility of any user and / or target user for the target knowledge point; and multiple test questions are classified according to the target user's mastery of each target knowledge point, the target frequency label corresponding to each target knowledge point, and the target fallibility corresponding to each target knowledge point.

[0118] Optionally, the above-mentioned classification unit is specifically used to obtain test situation big data in the area where the target user is located, determine the first fallibility corresponding to each knowledge point based on the test situation big data, and determine the first fallibility corresponding to each target knowledge point among the first fallibility of each knowledge point; determine the second fallibility of each target knowledge point by the target user based on the target user's practice records of each knowledge point in a preset learning system at a historical time; the above-mentioned preset learning system includes a knowledge graph; determine the target fallibility of each target knowledge point based on the first fallibility of each target knowledge point and the second fallibility of each target knowledge point.

[0119] In some embodiments, the above-mentioned classification unit is specifically used to determine the recommended category corresponding to each test question based on the target user's mastery of each target knowledge point, the target frequency label corresponding to each target knowledge point, and the target fallibility corresponding to each target knowledge point; the test questions with the same recommended category in each test question are grouped into the same test question set to obtain a test question set of multiple recommended categories; the above-mentioned multiple recommended categories include questions not to be done, optional questions, and compulsory questions.

[0120] In some embodiments, the above-mentioned classification and recommendation module 440 is specifically used to determine the personal learning portrait of the target user based on the target user's practice records of each knowledge point in a preset learning system at a historical time; the above-mentioned personal learning portrait includes the target user's learning ability; and classify multiple test questions according to the target user's mastery of each target knowledge point and the target user's learning ability.

[0121] In some embodiments, the above-mentioned classification and recommendation module 440 is specifically used to determine the difficulty level corresponding to each knowledge point based on the test situation big data of the target user's area, and determine the difficulty level corresponding to each target knowledge point among the difficulty levels of each knowledge point; and classify multiple test questions based on the target user's mastery of each target knowledge point and the difficulty level corresponding to each target knowledge point.

[0122] It should be noted here that the above-mentioned device provided in the embodiment of the present invention can implement all the method steps implemented in the above-mentioned method embodiment, and can achieve the same technical effect. The parts and beneficial effects that are the same as the method embodiment in this embodiment will not be described in detail here.

[0123] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5As shown, the electronic device may include: a processor (processor) 510, a communication interface (Communications Interface) 520, a memory (memory) 530 and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call the logic instructions in the memory 530 to execute the test question recommendation method, which includes: determining the test question text included in the test question image according to the test question image input by the target user; the above-mentioned test question text includes multiple test questions; using a large model to perform semantic analysis and knowledge point recognition processing on the test question text to determine the target knowledge point corresponding to each test question in the test question text; according to the knowledge graph corresponding to each target knowledge point and the target user, determining the target graph information corresponding to each target knowledge point in the knowledge graph; the above-mentioned target graph information includes the target user's mastery of the corresponding target knowledge point, and the above-mentioned knowledge graph includes the graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points; classifying multiple test questions according to the target user's mastery of each target knowledge point, obtaining a set of test questions of multiple recommended categories, and recommending test questions according to the priority of the recommended category of each test question combination; each test question combination of the recommended category includes at least one test question.

[0124] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0125] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the test question recommendation method provided by the above methods, which method includes: determining the test question text included in the test question image according to the test question image input by the target user; the above test question text includes multiple test questions; using a large model to perform semantic analysis and knowledge point recognition processing on the test question text to determine the target knowledge point corresponding to each test question in the test question text; according to the knowledge graph corresponding to each target knowledge point and the target user, determining the target graph information corresponding to each target knowledge point in the knowledge graph; the above target graph information includes the target user's mastery of the corresponding target knowledge point, and the above knowledge graph includes graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points; classifying multiple test questions according to the target user's mastery of each target knowledge point, obtaining a set of test questions of multiple recommended categories, and recommending test questions according to the priority of the recommended category of each test question combination; each test question combination of the recommended category includes at least one test question.

[0126] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented by a processor to execute the test question recommendation method provided by the above-mentioned methods when executed, the method comprising: determining the test question text included in the test question image according to the test question image input by the target user; the above-mentioned test question text includes multiple test questions; using a large model to perform semantic analysis and knowledge point recognition processing on the test question text to determine the target knowledge point corresponding to each test question in the test question text; according to the knowledge graph corresponding to each target knowledge point and the target user, determining the target graph information corresponding to each target knowledge point in the knowledge graph; the above-mentioned target graph information includes the target user's mastery of the corresponding target knowledge point, and the above-mentioned knowledge graph includes graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points; classifying multiple test questions according to the target user's mastery of each target knowledge point, obtaining a set of test questions of multiple recommended categories, and recommending test questions according to the priority of the recommended category of each test question combination; each test question combination of the recommended category includes at least one test question.

[0127] 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.

[0128] 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.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A test question recommendation method, characterized in that: include: Determine, according to the test question image input by the target user, a test question text included in the test question image; the test question text includes a plurality of test questions; Using a large model to perform semantic analysis and knowledge point recognition processing on the test text, and determine the target knowledge point corresponding to each test question in the test text; According to the knowledge graphs corresponding to the target knowledge points and the target users, determining the target graph information corresponding to the target knowledge points in the knowledge graph; The target graph information includes the target user's mastery of the corresponding target knowledge point, and the knowledge graph includes graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points; The multiple test questions are classified according to the target user's mastery of each of the target knowledge points to obtain a plurality of test question sets of recommended categories, and test questions are recommended according to the priority of the recommended categories of each test question combination; each test question combination of the recommended category includes at least one test question.

2. The test question recommendation method according to claim 1, characterized in that: The classifying the plurality of test questions according to the mastery of each target knowledge point by the target user comprises: Obtaining the test frequency of each of the knowledge points within a preset time in the area where the target user is located, and determining the frequency category corresponding to each of the knowledge points according to the test frequency of each of the knowledge points; Determining a target frequency category corresponding to each of the target knowledge points in the frequency categories of each of the knowledge points; The multiple test questions are classified according to the target user's mastery of each target knowledge point and the target frequency label corresponding to each target knowledge point.

3. The test question recommendation method according to claim 2, characterized in that: The classifying the plurality of test questions according to the mastery of each target knowledge point by the target user and the target frequency label corresponding to each target knowledge point includes: Determine a target fallibility corresponding to each of the target knowledge points; the target fallibility is used to characterize the fallibility of any user and / or the target user for the target knowledge point; The multiple test questions are classified according to the target user's mastery of each target knowledge point, the target frequency label corresponding to each target knowledge point, and the target fallibility corresponding to each target knowledge point.

4. The test question recommendation method according to claim 3, characterized in that: The determining of the target fallibility corresponding to each target knowledge point includes: Obtaining the test situation big data of the area where the target user is located, determining the first fallibility corresponding to each of the knowledge points according to the test situation big data, and determining the first fallibility corresponding to each of the target knowledge points from the first fallibility of each of the knowledge points; Determining the second fallibility of the target user for each target knowledge point according to the target user's practice record of each knowledge point in a preset learning system at a historical time; the preset learning system includes the knowledge graph; The target fallibility of each target knowledge point is determined according to the first fallibility of each target knowledge point and the second fallibility of each target knowledge point.

5. The test question recommendation method according to claim 3, characterized in that: The step of obtaining a set of test questions of multiple recommended categories includes: Determine a recommended category corresponding to each of the test questions according to the target user's mastery of each of the target knowledge points, the target frequency label corresponding to each of the target knowledge points, and the target fallibility corresponding to each of the target knowledge points; The test questions with the same recommended category among the test questions are grouped into the same test question set to obtain a test question set of multiple recommended categories; the multiple recommended categories include questions not to be done, optional questions, and required questions.

6. The test question recommendation method according to claim 5, characterized in that: The test question recommendation is performed according to the priority of the recommendation category of each test question combination, including: Sorting the test question sets of the multiple recommended categories from high to low according to the priorities of the recommended categories, and recommending test questions according to the sorting results; The priorities of the multiple recommendation categories from high to low are required questions, optional questions, and no-questions.

7. The test question recommendation method according to claim 1, characterized in that: The classifying the plurality of test questions according to the mastery of each target knowledge point by the target user comprises: Determine a personal learning profile of the target user according to the target user's practice records of each of the knowledge points in the preset learning system at a historical time; the personal learning profile includes the target user's learning ability; The multiple test questions are classified according to the target user's mastery of each target knowledge point and the target user's learning ability.

8. The test question recommendation method according to claim 1, characterized in that: The classifying the plurality of test questions according to the mastery of each target knowledge point by the target user comprises: Determine the degree of difficulty corresponding to each of the knowledge points based on the test situation big data of the target user's area, and determine the degree of difficulty corresponding to each of the target knowledge points from the degree of difficulty of each of the knowledge points; The multiple test questions are classified according to the target user's mastery of each target knowledge point and the difficulty level corresponding to each target knowledge point.

9. A test question recommendation device, characterized in that: include: A test question text determination module, used to determine the test question text included in the test question image according to the test question image input by the target user; the test question text includes a plurality of test questions; A knowledge point recognition module is used to perform semantic analysis and knowledge point recognition processing on the test text using a large model to determine the target knowledge point corresponding to each test question in the test text; A graph information determination module, for determining target graph information corresponding to each target knowledge point in the knowledge graph according to the knowledge graph corresponding to each target knowledge point and the target user; The target graph information includes the target user's mastery of the corresponding target knowledge point, and the knowledge graph includes graph information of multiple knowledge points and the association relationship between each knowledge point and other knowledge points; The classification and recommendation module is used to classify the multiple test questions according to the target user's mastery of each target knowledge point, obtain a set of test questions in multiple recommended categories, and recommend test questions according to the priority of the recommended category of each test question combination; each test question combination in the recommended category includes at least one test question.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the test question recommendation method according to any one of claims 1 to 8 is implemented.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the test question recommendation method according to any one of claims 1 to 8 is implemented.

Images