Personalized test question recommendation method and system, electronic device, and storage medium

Abstract

The present disclosure provides a personalized test question recommendation method, system, electronic device and storage medium to solve the technical problem that the test questions done by learners are not targeted, the method comprising: profiling the learners to obtain the mastery of the knowledge points by the learners; labeling the test questions to be recommended to determine the knowledge points possessed by the test questions and the difficulty levels of the knowledge points; constructing a learner-knowledge point matrix according to the mastery of the knowledge points by the learners; constructing a knowledge point-test question matrix according to the difficulty levels of the knowledge points in the test questions to be recommended; obtaining the latent interaction between the learners and the test questions according to the learner-knowledge point matrix and the knowledge point-test question matrix, and recommending personalized test questions with corresponding difficulty levels to the learners. The present disclosure can recommend personalized test questions with corresponding difficulty levels to the learners, and solve the problem that the test questions done by the learners at the present stage are highly repetitive and not targeted.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, specifically to a personalized test question recommendation method, a personalized test question recommendation system, an electronic device, and a computer-readable storage medium. Background Technology

[0002] Currently, in order to improve students' grades and consolidate their knowledge, schools and various training systems often adopt a "sea of ​​questions" approach, assigning students a large number of test questions. These test questions contain a large number of questions that repeat knowledge points, and each student receives the same test questions. Students have to repeat the questions they have mastered very well, and they also waste time trying to master the very difficult questions. The test questions are universal but not targeted, which increases students' academic burden, wastes a lot of time, and does not significantly improve their abilities. Summary of the Invention

[0003] To at least address the technical problem in existing technologies where learners' test questions are general but lack specificity, this disclosure provides a personalized test question recommendation method, a personalized test question recommendation system, an electronic device, and a computer-readable storage medium. This system can recommend personalized test questions of appropriate difficulty to learners, thus solving the problem of high repetition and lack of specificity in test questions currently provided by learners.

[0004] Firstly, this disclosure provides a personalized test question recommendation method, the method...

[0005] include:

[0006] Create learner profiles to understand their level of mastery of knowledge points;

[0007] Tag the recommended test questions to determine the knowledge points they cover and the level of difficulty of those knowledge points;

[0008] Construct a learner-knowledge point matrix based on the learner's level of mastery of the knowledge points;

[0009] Construct a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions;

[0010] Based on the learner-knowledge point matrix and the knowledge point-question matrix, the latent interaction between learners and questions is obtained, and personalized questions with matching difficulty are recommended to learners.

[0011] Furthermore, the process of creating learner profiles to obtain information about learners' mastery of various knowledge points includes:

[0012] Extract the various characteristic knowledge points needed from the test questions. These characteristic knowledge points include newly learned knowledge points, incorrect knowledge points, difficult knowledge points, weak knowledge points, and unclear knowledge points.

[0013] Determine the learner's level of mastery of the knowledge points;

[0014] A profile is created for each learner, and the profile is tagged with learner information, subject, knowledge point, difficulty level of the knowledge point, and past test questions.

[0015] Furthermore,

[0016] The extracted test questions contain various feature knowledge points, including:

[0017] Extract new knowledge points from the new chapters learned by learners;

[0018] Extracting key knowledge points from historical mistakes;

[0019] In historical test questions, if the actual time spent on a question / the estimated time spent on a question is greater than a preset value, and the preset value is greater than 1, then the knowledge point of that question is extracted as a difficult knowledge point.

[0020] If the answer to a test question in a history test is modified more than the preset number of times, the knowledge points related to that question will be identified as weak knowledge points.

[0021] In historical test questions, for those questions that have been marked with uncertainties, the knowledge points of those questions are extracted as fuzzy knowledge points;

[0022] Determining the learner's level of mastery of the knowledge points includes:

[0023] For newly learned knowledge points, their level of mastery should be rated as low.

[0024] For knowledge points that were answered incorrectly, were difficult, were weak, or were unclear, the learner's level of mastery of the knowledge point was determined based on the difficulty level of each type of knowledge point in historical test questions.

[0025] Furthermore, the method also includes:

[0026] The learners' recommended test questions were deduplicated, including:

[0027] By using learners' information, subjects, and corresponding historical test questions, test questions that learners have already taken are removed during the process of selecting test questions to be recommended.

[0028] The test questions are tagged with the knowledge points and difficulty levels of the knowledge points to remove duplicates of similar questions from the recommended test questions.

[0029] Furthermore,

[0030] The construction of the learner-knowledge point matrix based on the learner's mastery of knowledge points includes:

[0031] The learner's mastery of the knowledge points is expressed as a certain numerical value; the higher the value, the higher the mastery.

[0032] Construct a learner-knowledge point matrix based on the numerical values ​​of each learner's mastery of each knowledge point;

[0033] The process of constructing a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions includes:

[0034] The difficulty level of each knowledge point in the recommended test questions is represented by a certain value. The higher the value, the more difficult it is to master. If there is no such knowledge point in the test questions, the value is recorded as 0.

[0035] Construct a knowledge point-question matrix based on the difficulty levels of each knowledge point in each question to be recommended;

[0036] The process involves obtaining the latent interaction between learners and test questions based on the learner-knowledge point matrix and the knowledge point-question matrix, and recommending personalized test questions of appropriate difficulty to learners, including:

[0037] The learner-knowledge point matrix multiplied by the knowledge point-question matrix yields the latent interaction matrix between learners and questions.

[0038] Determine the median value in the latent interaction matrix between learners and test items for each learner, and select the test items whose values ​​are close to the median value as the recommended personalized test items based on the number of recommended test items.

[0039] Furthermore, the method also includes:

[0040] Based on the set difficulty level of the personalized test questions, test questions within the corresponding numerical range of the latent interaction matrix between the learner and the test questions are selected as recommended personalized test questions.

[0041] Secondly, this disclosure provides a personalized test question recommendation system, the system comprising:

[0042] The profile module is designed to create a profile of the learner in order to obtain the learner's level of mastery of various knowledge points.

[0043] The tagging module is set to tag the recommended test questions to determine the knowledge points and difficulty level of the knowledge points.

[0044] The module is configured to construct a learner-knowledge point matrix based on the learner's level of mastery of the knowledge points; and,

[0045] Construct a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions;

[0046] The recommendation module is designed to obtain the latent interaction between learners and test questions based on the learner-knowledge point matrix and the knowledge point-test question matrix, and recommend personalized test questions of appropriate difficulty to learners.

[0047] Furthermore,

[0048] The specific configuration of the building module is as follows:

[0049] The learner's mastery of the knowledge points is expressed as a certain numerical value; the higher the value, the higher the mastery.

[0050] Construct a learner-knowledge point matrix based on the numerical values ​​of each learner's mastery of each knowledge point; and,

[0051] The difficulty level of each knowledge point in the recommended test questions is represented by a certain value. The higher the value, the more difficult it is to master. If there is no such knowledge point in the test questions, it is recorded as 0.

[0052] Construct a knowledge point-question matrix based on the difficulty levels of each knowledge point in each question to be recommended;

[0053] The recommendation module is specifically configured as follows:

[0054] The learner-knowledge point matrix multiplied by the knowledge point-question matrix yields the latent interaction matrix between learners and questions.

[0055] Determine the median value in the latent interaction matrix between learners and test items for each learner, and select the test items whose values ​​are close to the median value as the recommended personalized test items based on the number of recommended test items.

[0056] Thirdly, this disclosure provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes a personalized test recommendation method as described in any of the first aspects.

[0057] Fourthly, this disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the personalized test recommendation method described in any of the first aspects above.

[0058] Beneficial effects:

[0059] The personalized test question recommendation method, personalized test question recommendation system, electronic device and storage medium disclosed herein determine the learner's mastery of knowledge points by profiling the learner, tag the test questions according to the knowledge points they contain and the difficulty of those knowledge points, determine the knowledge points and difficulty level of the test questions, extract latent factor knowledge points through low-rank matrix factorization, construct learner-knowledge point matrix and knowledge point-test question matrix, obtain the latent interaction between learners and test questions, and recommend personalized test questions of appropriate difficulty to students. Attached Figure Description

[0060] Figure 1 A flowchart illustrating a personalized test question recommendation method provided in Embodiment 1 of this disclosure;

[0061] Figure 2 This is a flowchart illustrating a personalized test question recommendation method provided in Embodiment 2 of this disclosure;

[0062] Figure 3 This is an architecture diagram of a personalized test question recommendation system provided in Embodiment 3 of this disclosure;

[0063] Figure 4 This is an architectural diagram of an electronic device provided in Embodiment 4 of this disclosure. Detailed Implementation

[0064] To enable those skilled in the art to better understand the technical solutions of this disclosure, the disclosure will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments and drawings described herein are merely for explaining the invention and are not intended to limit the invention.

[0065] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence; furthermore, in the absence of conflict, the embodiments and features in the embodiments of this disclosure can be arbitrarily combined with each other.

[0066] The terminology used in the embodiments of this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of this disclosure. The singular forms “a,” “the,” and “the” as used in the embodiments of this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0067] In the following description, the use of suffixes such as “module,” “part,” or “unit” to denote elements is solely for the purpose of illustrative purposes and has no specific meaning in itself. Therefore, “module,” “part,” or “unit” may be used interchangeably.

[0068] Currently, in order to improve students' academic performance and consolidate their knowledge, schools are implementing a "sea of ​​questions" approach, with repetitive and untargeted questions that increase students' workload, waste a lot of time, and fail to significantly improve their abilities.

[0069] The following detailed embodiments illustrate the technical solutions of this disclosure and how they solve the aforementioned technical problems in the prior art. It is understood that in the embodiments of this application, the executing entity may perform some or all of the steps in the embodiments of this application. These steps or operations are merely examples, and the embodiments of this application may also perform other operations or variations thereof. Furthermore, the steps may be executed in different orders as presented in the embodiments of this application, and it is not necessary to execute all the operations in the embodiments of this application. Moreover, the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0070] Figure 1 This is a flowchart illustrating a personalized test question recommendation method provided in Embodiment 1 of this disclosure, as shown below. Figure 1 As shown, the method includes steps S101 to S105;

[0071] Step S101: Create a profile of the learner to obtain the learner's level of mastery of the knowledge points;

[0072] For learners—including students and various members of society who need to acquire certain knowledge—doing more practice tests is an effective way to improve their mastery of knowledge points. Learners prefer test questions that are neither too easy nor too difficult. Questions that are too easy are unnecessary, while those that are too difficult are incompetent and a waste of time. Therefore, personalized test questions of appropriate difficulty need to be recommended. Since each learner's level of understanding varies, it is necessary to first create a learner profile to determine their mastery of different knowledge points.

[0073] Furthermore, the process of creating learner profiles to obtain information about learners' mastery of various knowledge points includes:

[0074] Extract the various characteristic knowledge points needed from the test questions. These characteristic knowledge points include newly learned knowledge points, incorrect knowledge points, difficult knowledge points, weak knowledge points, and unclear knowledge points.

[0075] Determine the learner's level of mastery of the knowledge points;

[0076] A profile is created for each learner, and the profile is tagged with learner information, subject, knowledge point, difficulty level of the knowledge point, and past test questions.

[0077] To improve the effectiveness of test questions, it is necessary to include various characteristic knowledge points in the test questions to enable learners to have a more comprehensive grasp of the knowledge points. After extracting the characteristic knowledge points that need to be included in students' test questions through various channels, it is necessary to convert the newly learned knowledge points, wrong knowledge points, difficult knowledge points, weak knowledge points, and unclear knowledge points into the students' level of mastery of the knowledge points; and then build a profile for the learners.

[0078] Furthermore,

[0079] The extracted test questions contain various feature knowledge points, including:

[0080] Extract new knowledge points from the new chapters learned by learners;

[0081] Extracting key knowledge points from historical mistakes;

[0082] In historical test questions, if the actual time spent on a question / the estimated time spent on a question is greater than a preset value, and the preset value is greater than 1, then the knowledge point of that question is extracted as a difficult knowledge point.

[0083] If the answer to a test question in a history test is modified more than the preset number of times, the knowledge points related to that question will be identified as weak knowledge points.

[0084] In historical test questions, for those questions that have been marked with uncertainties, the knowledge points of those questions are extracted as fuzzy knowledge points;

[0085] Determining the learner's level of mastery of the knowledge points includes:

[0086] For newly learned knowledge points, their level of mastery should be rated as low.

[0087] For knowledge points that were answered incorrectly, were difficult, were weak, or were unclear, the learner's level of mastery of the knowledge point was determined based on the difficulty level of each type of knowledge point in historical test questions.

[0088] The preset value can be set to various values ​​such as 1.2, 1.5, and 2, and the preset number of times can be set to 3. These can be set according to the actual situation.

[0089] For incorrect, difficult, weak, and unclear knowledge points, since these are acquired after students have completed the test, their level of mastery depends on the difficulty level of the knowledge point in the test. If the knowledge point is easy in the test, the student's mastery level is low; if the knowledge point is medium in difficulty, the student's mastery level is medium; and if the knowledge point is difficult, the student's mastery level is high.

[0090] For example: If a student answers test question 1 incorrectly, and test question 1 contains knowledge point 1 and knowledge point 2, where knowledge point 1 is marked as difficult in the test question and knowledge point 2 is marked as easy in the test question, then student 1 will have the labels shown in Table 1 below.

[0091] Table 1: Student 1's Incorrect Question Tags

[0092] Knowledge Point 1 Knowledge Point 2 History Exam Questions Student 1 Disaster Low Question 1

[0093] By determining the learner's level of mastery of the knowledge points, a learner profile can be created. The learner profile includes several tags, such as learner information (e.g., name, grade, class), subject, knowledge points, difficulty level of the knowledge points, and history test questions.

[0094] Step S102: Tag the recommended test questions to determine the knowledge points and difficulty level of the knowledge points.

[0095] Each test question is labeled with the knowledge point, the difficulty level of the knowledge point, and the estimated time required.

[0096] For example: Question 1 tests knowledge points 1 and 3. Knowledge point 1 is tested in a very simple way in this question, while knowledge point 3 is tested in a relatively difficult way. Therefore, Question 1 has the labels shown in Table 2 below.

[0097] Table 2: Test Item Tags

[0098] Question 1 Knowledge Point 1 (Easy) Knowledge Point 3 (Difficult) Estimated time: 5 minutes

[0099] Test questions can be categorized using test question tags, and the knowledge points and their difficulty levels within those questions can be determined.

[0100] Step S103: Construct a learner-knowledge point matrix based on the learner's mastery of the knowledge points;

[0101] Step S104: Construct a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions;

[0102] Step S105: Based on the learner-knowledge point matrix and the knowledge point-question matrix, obtain the latent interaction between learners and questions, and recommend personalized questions with matching difficulty to learners.

[0103] After representing learners' mastery of knowledge points as specific numerical values, a learner-knowledge point matrix is ​​constructed. Similarly, a knowledge point-question matrix is ​​constructed by representing the difficulty level of each knowledge point in a test question as a specific numerical value. Then, based on low-rank matrix factorization, a set of latent (hidden) factors is extracted from the learner and question matrices. These factor vectors describe the student and the test question, and the latent factors are identified as knowledge points. Finally, the learner's mastery level of the knowledge points associated with a particular test question is obtained through operations on the learner-knowledge point matrix and the knowledge point-question matrix; the smaller the value, the higher the mastery level. This allows for the selection of personalized test questions with a difficulty level appropriate for the learner.

[0104] Furthermore,

[0105] The construction of the learner-knowledge point matrix based on the learner's mastery of knowledge points includes:

[0106] The learner's mastery of the knowledge points is expressed as a certain numerical value; the higher the value, the higher the mastery.

[0107] Construct a learner-knowledge point matrix based on the numerical values ​​of each learner's mastery of each knowledge point;

[0108] The process of constructing a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions includes:

[0109] The difficulty level of each knowledge point in the recommended test questions is represented by a certain value. The higher the value, the more difficult it is to master. If there is no such knowledge point in the test questions, the value is recorded as 0.

[0110] Construct a knowledge point-question matrix based on the difficulty levels of each knowledge point in each question to be recommended;

[0111] The process involves obtaining the latent interaction between learners and test questions based on the learner-knowledge point matrix and the knowledge point-question matrix, and recommending personalized test questions of appropriate difficulty to learners, including:

[0112] The learner-knowledge point matrix multiplied by the knowledge point-question matrix yields the latent interaction matrix between learners and questions.

[0113] Determine the median value in the latent interaction matrix between learners and test items for each learner, and select the test items whose values ​​are close to the median value as the recommended personalized test items based on the number of recommended test items.

[0114] After obtaining the learner's level of mastery of the knowledge points, the following table (Table 3) can be generated.

[0115] Table 3: Learners' Mastery of Knowledge Points

[0116] Knowledge Point 1 Knowledge Point 2 Knowledge Point 3 ... Knowledge Point n Student 1 Low middle Low ... Student 2 Low middle ... Low Student 3 Low ... high ... ... ... ... ... ... Student m middle Low high ...

[0117] After representing the level of mastery as a certain numerical value, this value can be set according to the actual situation. For example, if a student's mastery of a knowledge point is low, it is recorded as 1; if the student's mastery of a knowledge point is medium, it is recorded as 5; if the student's mastery of a knowledge point is high, it is recorded as 10; if the student has not yet learned the knowledge point, it is recorded as 0. This yields the learner-knowledge point matrix:

[0118]

[0119] Similarly, after marking the difficulty level of each knowledge point in the test questions, we can obtain the following table:

[0120] Table 4: Difficulty Level of Knowledge Points in the Test Questions

[0121]

[0122]

[0123] Then, the difficulty level is expressed as a certain value, such as easy representing 10 points, medium representing 5 points, difficult representing 1 point, and blank representing 0 points if the knowledge point is not in the question.

[0124] After processing the table, we can obtain the following knowledge point-question matrix:

[0125]

[0126] Multiplying the (learner-knowledge point matrix) by the (knowledge point-question matrix) yields the latent interaction matrix between learners and questions, representing the latent interaction between students and questions:

[0127]

[0128] Based on the latent interaction matrix between students and test questions, smaller values ​​indicate a high level of student mastery of the knowledge point, while the difficulty level of the test question containing that knowledge point is low; such test questions are unnecessary for this student. Larger values ​​indicate a low level of student mastery of the knowledge point, while the difficulty level of the test question containing that knowledge point is high; such test questions are also unnecessary for this student. When selecting test questions for recommendation, first determine the median value in the latent interaction matrix. The median value is obtained by dividing the matrix by (maximum + minimum). Based on the number of recommended test questions, select the test questions with values ​​close to the median value as personalized recommendations. Alternatively, determine the range of values ​​in the latent interaction matrix based on the desired number of test questions, and then select the test questions within this range as recommendations. For example, for the aforementioned latent interaction matrix, for student m, the median value is (110 + 5) / 2 = 55. If two recommended test questions are desired, test questions 6 and y are selected; if three recommended test questions are desired, test questions 5, 6, and y are selected. In practical applications, personalized test questions can be recommended to learners based on median values ​​and a certain range of values, as shown in Table 5 below:

[0129] Table 5: List of recommended test questions for students

[0130] Question 1 Question 2 Question 3 Question 4 Question 5 Question 6 ... Test Question y Student 1 recommend ... Student 2 recommend ... recommend Student 3 recommend recommend ... ... ... ... ... ... ... ... ... ... Student m recommend ... recommend

[0131] Furthermore, the method also includes:

[0132] Based on the set difficulty level of the personalized test questions, test questions within the corresponding numerical range of the latent interaction matrix between the learner and the test questions are selected as recommended personalized test questions.

[0133] When recommending test questions, specific difficulty levels can be set for learners. The students' mastery of the knowledge points in the questions is represented by the latent interaction matrix values, and questions within a certain numerical range are used as personalized test recommendations. This provides learners with test sets of varying difficulty levels.

[0134] Furthermore, the method also includes:

[0135] The learners' recommended test questions were deduplicated, including:

[0136] By using learners' information, subjects, and corresponding historical test questions, test questions that learners have already taken are removed during the process of selecting test questions to be recommended.

[0137] The test questions are tagged with the knowledge points and difficulty levels of the knowledge points to remove duplicates of similar questions from the recommended test questions.

[0138] It removes duplicate questions that learners have already done, as well as questions that contain similar knowledge points and are of similar difficulty, so that learners do not waste time doing duplicate questions.

[0139] This embodiment of the disclosure creates a learner profile by including learner information, subject, newly learned knowledge points, incorrect knowledge points, difficult knowledge points, weak knowledge points, and unclear knowledge points. It then tags test questions based on the knowledge points they contain, their difficulty level, and the estimated time required. Using the learner profile and test question tags, it removes duplicate test questions that students have already answered and those containing similar knowledge points or of similar difficulty. Through low-rank matrix factorization, it extracts latent factor knowledge points, constructs a student-knowledge point matrix and a knowledge point-test question matrix, and derives the latent interaction between students and test questions. This allows for the recommendation of personalized test questions with matching difficulty levels. Questions that students have mastered well do not need to be repeated, and overly difficult questions are avoided, thus preventing wasted time. Furthermore, different difficulty levels can be set for learners, making the test questions more targeted.

[0140] Embodiment 2 of this disclosure also provides a personalized test question recommendation method. This method involves creating learner profiles, tagging test questions, removing duplicate questions that students have already answered, and removing questions containing similar knowledge points or of similar difficulty. Using low-rank matrix factorization, latent factor knowledge points are extracted, and a student-knowledge point matrix and a knowledge point-test question matrix are constructed. This yields the latent interaction between students and test questions, allowing for the recommendation of personalized test questions of appropriate difficulty for each student. Figure 2 As shown, the method includes:

[0141] S100: Create learner profiles and test item tags, including:

[0142] 1) Learner profile modeling

[0143] The test questions should include newly learned knowledge points, incorrect knowledge points, difficult knowledge points, weak knowledge points, and unclear knowledge points. The following channels can be used to extract the knowledge points needed from the students' test questions.

[0144] Extract key knowledge points from the new chapters students are learning;

[0145] Extracting knowledge points from historical mistakes to identify the knowledge points for those mistakes;

[0146] If the actual time spent on a problem / the estimated time spent is greater than 1.5, then the knowledge point of that problem is identified as a difficult knowledge point.

[0147] If the answer to a test question is revised more than 3 times, the knowledge points related to that question will be identified as weak knowledge points.

[0148] Mark uncertain test questions and extract the knowledge points of those questions as fuzzy knowledge points;

[0149] The newly learned knowledge points, the knowledge points with mistakes, the knowledge points that are difficult, the knowledge points that are weak, and the knowledge points that are unclear are converted into the students' level of mastery of the knowledge points.

[0150] For newly learned knowledge points, students have just finished learning them and haven't done any tests yet, so their mastery of these knowledge points is low.

[0151] For incorrect, difficult, weak, and unclear knowledge points, since these are acquired after students have completed the test, their level of mastery depends on the difficulty level of the knowledge point in the test. If the knowledge point is easy in the test, the student's mastery level is low; if the knowledge point is medium in difficulty, the student's mastery level is medium; and if the knowledge point is difficult, the student's mastery level is high.

[0152] For example: If a student answers test question 1 incorrectly, and test question 1 contains knowledge point 1 and knowledge point 2, where knowledge point 1 is marked as difficult in the test question and knowledge point 2 is marked as easy in the test question, then student 1 will have the labels shown in Table 1 below.

[0153] Table 1:

[0154] Knowledge Point 1 Knowledge Point 2 History Exam Questions Student 1 Disaster Low Question 1

[0155] The learner profile includes tags such as name, grade, subject, knowledge point, difficulty level of the knowledge point, and history test questions.

[0156] 2) Tagging test questions

[0157] Each test question is labeled with the knowledge point, the difficulty level of the knowledge point, and the estimated time required.

[0158] For example: Question 1 tests knowledge points 1 and 3. Knowledge point 1 is tested in a very simple way in this question, while knowledge point 3 is tested in a more difficult way. Therefore, Question 1 has the following Table 2.

[0159] Table 2:

[0160] Question 1 Knowledge Point 1 (Easy) Knowledge Point 3 (Difficult) Estimated time: 5 minutes

[0161] All questions include three labels: the knowledge point, the difficulty level of the knowledge point, and the estimated time required.

[0162] 3) Deduplication of test questions

[0163] By using students' names, grades, subjects, and corresponding history test questions, test questions that students have already taken are removed during the test question selection process.

[0164] The test questions are deduplicated by using tags indicating the knowledge points and difficulty levels of those knowledge points.

[0165] S200: Matrix factorization;

[0166] Based on the low-rank matrix factorization method, a set of latent (hidden) factors are extracted from the student and test question matrices, and the students and test questions are described by these factor vectors. These latent factors are identified as knowledge points.

[0167] Student-Knowledge Point Matrix

[0168] Table 3 below shows the students' level of mastery of the knowledge points.

[0169] Table 3:

[0170] Knowledge Point 1 Knowledge Point 2 Knowledge Point 3 ... Knowledge Point n Student 1 Low middle Low ... Student 2 Low middle ... Low Student 3 Low ... high ... ... ... ... ... ... Student m middle Low high ...

[0171] If a student's mastery of a knowledge point is low, it is recorded as 1; if a student's mastery of a knowledge point is medium, it is recorded as 5; if a student's mastery of a knowledge point is high, it is recorded as 10; if a student has not yet learned the knowledge point, it is recorded as 0.

[0172]

[0173] Knowledge Points - Question Matrix

[0174] The difficulty level of each knowledge point in the test questions is marked in Table 4 below.

[0175] Table 4:

[0176] Question 1 Question 2 Question 3 Question 4 Question 5 Question 6 ... Test Question y Knowledge Point 1 Disaster Disaster middle easy ... Knowledge Point 2 easy middle Disaster ... Knowledge Point 3 easy Disaster ... middle ... ... ... ... ... ... ... ... ... Knowledge Point n easy middle Disaster middle ...

[0177] In this matrix, easy represents 10 points, medium represents 5 points, difficult represents 1 point, and empty represents 0 points if the knowledge point is not covered in the question. The resulting knowledge point-question matrix is ​​as follows:

[0178]

[0179] S300: Personalized test question recommendations;

[0180] The potential interaction between students and test questions is represented by (student-knowledge point matrix) × (knowledge point-test question matrix).

[0181]

[0182] Based on the latent interaction matrix between students and test questions, smaller values ​​indicate that the student has a high level of mastery of the knowledge point, while the test question containing that knowledge point is of low difficulty; such a test question is unnecessary for the student. Larger values ​​indicate that the student has a low level of mastery of the knowledge point, while the test question containing that knowledge point is of high difficulty; such a test question is also unnecessary for the student. Other matches are considered valid test questions.

[0183] The median value is obtained by dividing the result by (maximum value + minimum value) / 2. The range of values ​​can be determined based on the number of questions desired. For example, for student m, the median value is (110 + 5) / 2 = 55. If two recommended questions are desired, then questions 6 and y are chosen. If three recommended questions are desired, then questions 5, 6, and y are chosen.

[0184] The above example would then provide a list of recommended test questions for students as follows:

[0185] Question 1 Question 2 Question 3 Question 4 Question 5 Question 6 ... Test Question y Student 1 recommend ... Student 2 recommend ... recommend Student 3 recommend recommend ... ... ... ... ... ... ... ... ... ... Student m recommend ... recommend

[0186] Figure 3 This is a schematic diagram of the structure of a personalized test question recommendation system provided in Embodiment 3 of this disclosure, as shown below. Figure 3 As shown, the system includes:

[0187] The profile module 11 is set to create a profile of the learner in order to obtain the learner’s mastery of various knowledge points.

[0188] The tag module 12 is set to tag the recommended test questions to determine the knowledge points and the difficulty level of the knowledge points in the test questions.

[0189] Module 13 is configured to construct a learner-knowledge point matrix based on the learner's mastery of the knowledge points; and,

[0190] Construct a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions;

[0191] Recommendation module 14 is configured to obtain the latent interaction between learners and test questions based on the learner-knowledge point matrix and the knowledge point-test question matrix, and recommend personalized test questions with matching difficulty to learners.

[0192] Furthermore, the portrait module 11 is specifically configured as follows:

[0193] Extract the various characteristic knowledge points needed from the test questions. These characteristic knowledge points include newly learned knowledge points, incorrect knowledge points, difficult knowledge points, weak knowledge points, and unclear knowledge points.

[0194] Determine the learner's level of mastery of the knowledge points;

[0195] A profile is created for each learner, and the profile is tagged with learner information, subject, knowledge point, difficulty level of the knowledge point, and past test questions.

[0196] Furthermore,

[0197] The extracted test questions contain various feature knowledge points, including:

[0198] Extract new knowledge points from the new chapters learned by learners;

[0199] Extracting key knowledge points from historical mistakes;

[0200] In historical test questions, if the actual time spent on a question / the estimated time spent on a question is greater than a preset value, and the preset value is greater than 1, then the knowledge point of that question is extracted as a difficult knowledge point.

[0201] If the answer to a test question in a history test is modified more than the preset number of times, the knowledge points related to that question will be identified as weak knowledge points.

[0202] In historical test questions, for those questions that have been marked with uncertainties, the knowledge points of those questions are extracted as fuzzy knowledge points;

[0203] Determining the learner's level of mastery of the knowledge points includes:

[0204] For newly learned knowledge points, their level of mastery should be rated as low.

[0205] For knowledge points that were answered incorrectly, were difficult, were weak, or were unclear, the learner's level of mastery of the knowledge point was determined based on the difficulty level of each type of knowledge point in historical test questions.

[0206] Furthermore, the system also includes a deduplication module 15:

[0207] The deduplication module 15 is configured to deduplicat the learner's recommended test questions, including:

[0208] By using learners' information, subjects, and corresponding historical test questions, test questions that learners have already taken are removed during the process of selecting test questions to be recommended.

[0209] The test questions are tagged with the knowledge points and difficulty levels of the knowledge points to remove duplicates of similar questions from the recommended test questions.

[0210] Furthermore,

[0211] The construction module 13 is specifically configured as follows:

[0212] The learner's mastery of the knowledge points is expressed as a certain numerical value; the higher the value, the higher the mastery.

[0213] Construct a learner-knowledge point matrix based on the numerical values ​​of each learner's mastery of each knowledge point; and,

[0214] The difficulty level of each knowledge point in the recommended test questions is represented by a certain value. The higher the value, the more difficult it is to master. If there is no such knowledge point in the test questions, it is recorded as 0.

[0215] Construct a knowledge point-question matrix based on the difficulty levels of each knowledge point in each question to be recommended;

[0216] The recommendation module 14 is specifically configured as follows:

[0217] The learner-knowledge point matrix multiplied by the knowledge point-question matrix yields the latent interaction matrix between learners and questions.

[0218] Determine the median value in the latent interaction matrix between learners and test items for each learner, and select the test items whose values ​​are close to the median value as the recommended personalized test items based on the number of recommended test items.

[0219] Furthermore, the recommendation module 14 is also configured as follows:

[0220] Based on the set difficulty level of the personalized test questions, test questions within the corresponding numerical range of the latent interaction matrix between the learner and the test questions are selected as recommended personalized test questions.

[0221] The personalized test question recommendation system of this disclosure is used to implement the personalized test question recommendation method in Method Embodiment 1 and Method Embodiment 2, so the description is relatively simple. For details, please refer to the relevant descriptions in Method Embodiment 1 and Method Embodiment 2 above, which will not be repeated here.

[0222] In addition, such as Figure 4 As shown, Embodiment 4 of this disclosure also provides an electronic device, including a memory 100 and a processor 200. The memory 100 stores a computer program. When the processor 200 runs the computer program stored in the memory 100, the processor 200 executes the various possible methods described above.

[0223] The memory 100 is connected to the processor 200. The memory 100 can be a flash memory, a read-only memory, or another type of memory. The processor 200 can be a central processing unit or a microcontroller.

[0224] Furthermore, embodiments of this disclosure also provide a computer-readable storage medium storing a computer program, which is executed by a processor using the various possible methods described above.

[0225] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program modules or other data). Computer-readable storage media includes, but is not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), Digital Video Disc (DVD) or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.

[0226] It is understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of this disclosure, and this disclosure is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of this disclosure, and these modifications and improvements are also considered to be within the scope of protection of this disclosure.

Claims

1. A personalized test question recommendation method, characterized in that, The method includes: Create learner profiles to understand their level of mastery of knowledge points; Tag the recommended test questions to determine the knowledge points they cover and the level of difficulty of those knowledge points; Construct a learner-knowledge point matrix based on the learner's level of mastery of the knowledge points; Construct a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions; Based on the learner-knowledge point matrix and the knowledge point-question matrix, the potential interaction between learners and questions is obtained, and personalized questions with matching difficulty are recommended to learners. The construction of the learner-knowledge point matrix based on the learner's mastery of the knowledge points includes: The learner's mastery of the knowledge points is expressed as a certain numerical value; the higher the value, the higher the mastery. Construct a learner-knowledge point matrix based on the numerical values ​​of each learner's mastery of each knowledge point; The process of constructing a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions includes: The difficulty level of each knowledge point in the recommended test questions is represented by a certain value. The higher the value, the more difficult it is to master. If there is no such knowledge point in the test questions, the value is recorded as 0. Construct a knowledge point-question matrix based on the difficulty levels of each knowledge point in each question to be recommended; The process involves obtaining the latent interaction between learners and test questions based on the learner-knowledge point matrix and the knowledge point-question matrix, and recommending personalized test questions of appropriate difficulty to learners, including: The learner-knowledge point matrix multiplied by the knowledge point-question matrix yields the latent interaction matrix between learners and questions. Determine the median value in the latent interaction matrix between learners and test items for each learner, and select the test items whose values ​​are close to the median value as the recommended personalized test items based on the number of recommended test items.

2. The method according to claim 1, characterized in that, The process of creating learner profiles to assess learners' mastery of various knowledge points includes: Extract the various characteristic knowledge points needed from the test questions. These characteristic knowledge points include newly learned knowledge points, incorrect knowledge points, difficult knowledge points, weak knowledge points, and unclear knowledge points. Determine the learner's level of mastery of the knowledge points; A profile is created for each learner, and the profile is tagged with learner information, subject, knowledge point, difficulty level of the knowledge point, and past test questions.

3. The method according to claim 2, characterized in that, The extracted test questions contain various feature knowledge points, including: Extract new knowledge points from the new chapters learned by learners; Extracting key knowledge points from historical mistakes; In historical test questions, if the actual time spent on a question / the estimated time spent on a question is greater than a preset value, and the preset value is greater than 1, then the knowledge point of that question is extracted as a difficult knowledge point. If the answer to a test question in a history test is modified more than the preset number of times, the knowledge points related to that question will be identified as weak knowledge points. In historical test questions, for those questions that have been marked with uncertainties, the knowledge points of those questions are extracted as fuzzy knowledge points; Determining the learner's level of mastery of the knowledge points includes: For newly learned knowledge points, their level of mastery should be rated as low. For knowledge points that were answered incorrectly, were difficult, were weak, or were unclear, the learner's level of mastery of the knowledge point was determined based on the difficulty level of each type of knowledge point in historical test questions.

4. The method according to claim 2, characterized in that, The method further includes: The learners' recommended test questions were deduplicated, including: By using learners' information, subjects, and corresponding historical test questions, test questions that learners have already taken are removed during the process of selecting test questions to be recommended. The test questions are tagged with the knowledge points and difficulty levels of the knowledge points to remove duplicates of similar questions from the recommended test questions.

5. The method according to claim 1, characterized in that, The method further includes: Based on the set difficulty level of the personalized test questions, test questions within the corresponding numerical range of the latent interaction matrix between the learner and the test questions are selected as recommended personalized test questions.

6. A personalized test question recommendation system, characterized in that, The system includes: The profile module is designed to create a profile of the learner in order to obtain the learner's level of mastery of various knowledge points. The tagging module is set to tag the recommended test questions to determine the knowledge points and difficulty level of the knowledge points. The module is configured to construct a learner-knowledge point matrix based on the learner's level of mastery of the knowledge points; and, Construct a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions; The recommendation module is designed to obtain the latent interaction between learners and test questions based on the learner-knowledge point matrix and the knowledge point-test question matrix, and recommend personalized test questions of appropriate difficulty to learners. The construction of the learner-knowledge point matrix based on the learner's mastery of the knowledge points includes: The learner's mastery of the knowledge points is expressed as a certain numerical value; the higher the value, the higher the mastery. Construct a learner-knowledge point matrix based on the numerical values ​​of each learner's mastery of each knowledge point; The process of constructing a knowledge point-question matrix based on the difficulty level of each knowledge point in the recommended test questions includes: The difficulty level of each knowledge point in the recommended test questions is represented by a certain value. The higher the value, the more difficult it is to master. If there is no such knowledge point in the test questions, the value is recorded as 0. Construct a knowledge point-question matrix based on the difficulty levels of each knowledge point in each question to be recommended; The process involves obtaining the latent interaction between learners and test questions based on the learner-knowledge point matrix and the knowledge point-question matrix, and recommending personalized test questions of appropriate difficulty to learners, including: The learner-knowledge point matrix multiplied by the knowledge point-question matrix yields the latent interaction matrix between learners and questions. Determine the median value in the latent interaction matrix between learners and test items for each learner, and select the test items whose values ​​are close to the median value as the recommended personalized test items based on the number of recommended test items.

7. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the personalized test item recommendation method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the personalized test question recommendation method as described in any one of claims 1-5.

Images