Question attack intelligent analysis method and device

Abstract

The invention discloses a question attack intelligent analysis method and device. Exercise questions can be reasonably pushed according to the actual situation of a class. The question attack intelligent analysis method comprises the steps of obtaining a plurality of questions and score information of overall students in each question; obtaining a score corresponding to the mode based on the distribution condition and normal distribution of the score information; calculating a target group index according to the distribution condition of the score students of the score corresponding to the mode; wherein the distribution condition comprises the distribution condition in the whole students and the distribution condition in the students in the class, and the target group index represents the index that the students in the class have strong or weak mastering of the questions in the research range; according to the target group index, the question difficulty coefficient, the score segment position of the score corresponding to the mode and the score coefficient, obtaining a question attack sequence value; and according to the question attack sequence value, the question similarity and the difficulty level, extracting exercises from a question bank to perform question attack.

Description

Technical Field

[0001] This application relates to the field of information analysis technology, and particularly to an intelligent analysis method and device for tackling problems. Background Art

[0002] In an exam, due to the different learning abilities and learning situations of each student in the class, the degree of mastery of questions varies greatly. At present, the personal subjective factors in the problem-solving assignment strategy are relatively large, and the characteristics of the questions themselves and the overall mastery of the class are not considered comprehensively enough. The result of independent assignment is not easy to achieve an ideal effect for rapid score improvement, and it fails to achieve precise positioning and tackling. As a result, the class fails to master more questions within the effective time, improve more scores, wastes time on learning questions that have already been mastered, or spends a large amount of energy on learning questions with low scores and slow mastery. Therefore, it is impossible to formulate a comprehensive and reasonable problem-solving plan for each class. Summary of the Invention

[0003] To solve the above technical problems, this application is proposed. Embodiments of this application provide an intelligent analysis method and device for tackling problems, which can reasonably push practice questions according to the actual situation of the class.

[0004] According to one aspect of this application, an intelligent analysis method for tackling problems is provided, including: obtaining a plurality of questions and the score information of all students in each of the questions; obtaining the score corresponding to the mode based on the distribution of the score information and the normal distribution; calculating a target group index according to the distribution of the students who scored the score corresponding to the mode; where the distribution includes the distribution among all students and the distribution among the class students, and the target group index represents the index of the strength or weakness of the class students' mastery of the questions within the research scope; obtaining a problem-solving sequence value according to the target group index, the question difficulty coefficient, the score segment corresponding to the mode score, and the score coefficient; extracting practice questions from the question bank according to the problem-solving sequence value, question similarity, and difficulty level for problem-solving.

[0005] In one embodiment, obtaining a problem-solving sequence value according to the target group index, the question difficulty coefficient, the score segment corresponding to the mode score, and the score coefficient includes: Y = (difficulty coefficient * (TGI / 100 + a) * score segment) / score coefficient; where Y represents the problem-solving sequence value, the problem-solving sequence value is used for sorting the problem-solving sequence, the problem-solving sequence value is inversely proportional to the tackling priority, TGI represents the target group index, the difficulty coefficient is extracted from the attribute settings in the question, the score segment represents the position of the mode score in the score range, the score range represents [the lowest score of the question among all students, the highest score of the question among all students], the score coefficient represents the total score of the question, and a represents a correction value.

[0006] In one embodiment, calculating a target group index according to the distribution of students scoring the score corresponding to the mode includes: obtaining a first number of students in the class who score greater than or equal to the score corresponding to the mode; calculating a class achievement rate of the score corresponding to the mode according to the first number of students and the total number of students in the class; and obtaining a second number of students in the overall population who score greater than or equal to the score corresponding to the mode; calculating an overall achievement rate of the score corresponding to the mode according to the second number of students and the total number of students in the overall population; and calculating the target group index according to the class achievement rate and the overall achievement rate.

[0007] In one embodiment, calculating the target group index according to the class achievement rate and the overall achievement rate includes: (the class achievement rate / the overall achievement rate) * standard number 100 = the target group index.

[0008] In one embodiment, the intelligent analysis method for problem tackling further includes: performing label setting and attribute setting on the problem according to the score information of the overall students in the problem; wherein, the labels include knowledge points, question types, and examination points, and the attributes include difficulty coefficients.

[0009] In one embodiment, the intelligent analysis method for problem tackling further includes: calculating the problem similarity between the current problem and the problems in the question bank according to a similarity formula; wherein, the similarity formula is: In the formula, item_tags(ti) stores the annotation vector of question i, item_tags(tj) stores the annotation vector of question j, question i is the current question or a question in the question bank, question j is a question in the question bank or the current question, the annotation vector includes knowledge points, question types, and examination points, the similarity value ranges from [0, 1], and the similarity value is proportional to the problem similarity.

[0010] In one embodiment, the intelligent analysis method for problem tackling further includes: performing difficulty level division on multiple problems and the problems in the question bank according to the multiple problems and the score information of the overall students in each problem; determining a difficulty level matching priority based on the score segment, the difficulty coefficient, and the difficulty level division; the score segment indicates the position of the score corresponding to the mode in the score interval.

[0011] In one embodiment, practice questions are extracted from a question bank according to the question tackling sequence value, question similarity, and difficulty level for question tackling, including: performing a tackling ranking on multiple questions from high to low according to the question tackling sequence value; sequentially performing range matching on each question from high to low according to question similarity to obtain multiple similar questions for each question; obtaining the extraction sequence of the multiple similar questions based on the difficulty level matching priority and similarity value; and extracting practice questions from the question bank based on the extraction sequence for question tackling.

[0012] In one embodiment, obtaining the extraction sequence of the multiple similar questions based on the difficulty level matching priority and similarity value includes: Extraction sequence = similarity weight * difficulty weight; where the similarity weight is proportional to the similarity value, the similarity weight value ranges from 0 to 1, and the difficulty weight is proportional to the difficulty level matching priority.

[0013] According to another aspect of the present application, a question tackling intelligent analysis device is provided, including: an information acquisition module for acquiring multiple questions and the score information of the overall students in each question; a score value acquisition module for obtaining the score value corresponding to the mode based on the distribution of the score information and the normal distribution; an index calculation module for calculating a target group index according to the distribution of the students who obtained the score value corresponding to the mode; where the distribution includes the distribution among the overall students and the distribution among the class students, and the target group index represents the index of whether the class students are strong or weak in mastering the questions within the research scope; an order acquisition module for obtaining a question tackling sequence value according to the target group index, question difficulty coefficient, score segment corresponding to the mode, and score coefficient; and an extraction module for extracting practice questions from the question bank according to the question tackling sequence value, question similarity, and difficulty level for question tackling.

[0014] The question tackling intelligent analysis method and device provided by the present application analyze the mastery of each question by each class in combination with the current class grades from multiple dimensions such as question type, test point, knowledge point, difficulty level, score range (importance), and overall mastery, formulate a reasonable question tackling allocation plan for each class, and intelligently and accurately push practice questions, making the question tackling plan more reasonable, the tackling practice more accurate, fully utilizing limited time and resources, quickly mastering questions, mastering more questions, and improving more scores. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above and other objects, features, and advantages of the present application will become more apparent by describing the embodiments of the present application in more detail with reference to the accompanying drawings. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. They are used to explain the present application together with the embodiments of the present application and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 It is a schematic flowchart of a problem-solving intelligent analysis method provided by an exemplary embodiment of the present application.

[0017] Figure 2 It is a schematic flowchart of a problem-solving intelligent analysis method provided by another exemplary embodiment of the present application.

[0018] Figure 3 It is a schematic flowchart of a problem-solving intelligent analysis method provided by another exemplary embodiment of the present application.

[0019] Figure 4 It is a schematic structural diagram of a problem-solving intelligent analysis device provided by an exemplary embodiment of the present application.

[0020] Figure 5 It is a schematic structural diagram of a problem-solving intelligent analysis device provided by another exemplary embodiment of the present application.

[0021] Figure 6 It is a structural diagram of an electronic device provided by an exemplary embodiment of the present application. Detailed Embodiments

[0022] Next, exemplary embodiments of the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0023] Application Overview

[0024] In an exam, due to the different learning abilities and learning situations of each student in the class, the degree of mastery of the questions varies greatly. Therefore, when pushing unified practice questions to the class, the overall learning situation of the class needs to be considered. The present application analyzes and models according to the question types, test points, knowledge points, difficulty levels, score ranges (importance), and overall mastery situations, combined with the current class scores, to obtain the mastery situation of each class for each question, formulate a reasonable problem-solving assignment plan for each class, and intelligently and accurately push practice questions, making the problem-solving plan more reasonable, the problem-solving practice more accurate, fully utilizing limited time and resources, quickly mastering questions, mastering more questions, and improving more scores.

[0025] Exemplary Method

[0026] Figure 1 FIG. 1 is a flow chart of an intelligent analysis method for solving a problem provided by an exemplary embodiment of the present application. Figure 1 As shown in the figure, the intelligent analysis method for solving the problem includes:

[0027] Step 100: Obtain multiple questions and overall student score information for each question.

[0028] The score details of each question in the unified examination can be extracted from the score management system, including the score details of the entire student and the score details of each class. The entire student can be the students of the entire age in the current school year or the students of the same age in multiple school years. Obtaining multiple questions includes obtaining the knowledge points, question types, test points and difficulty coefficients of multiple questions. The score information indicates the score obtained by the student in the current question.

[0029] Step 200: Based on the distribution of the score information and the normal distribution, obtain the score corresponding to the mode.

[0030] Based on the score information of all students, the corresponding score of the concentration mode is found through normal distribution. The mode refers to the value or set of values ​​that appears most frequently in a set of data. That is, in the current question, the number of students who get this score is greater than the number of students with other scores, and the mode corresponds to this score. The score corresponding to the mode can be used as a reference to reflect the mastery of the question by most students in the class, as well as the average mastery.

[0031] Normal distribution density function formula: f(x) = exp{-(x-μ)2 / 2σ2} / [√(2π)σ];

[0032] μ is the score corresponding to the mode, that is, the score corresponding to the largest number of people (for example, the score range is 0 to 10 points, 30 people get 6 points, and the number of people with other scores is less than 30, so μ is 6); σ takes the smallest span in the score range (such as the interval is 0 to 10 points, the span is 0.5, the possible scores are 0, 0.5, 1, 1.5..., σ takes 0.5), π represents pi (approximately equal to 3.14159), and e represents the base of the natural logarithm (approximately equal to 2.71828).

[0033] Step 300: Calculate the target group index based on the distribution of students with scores corresponding to the mode.

[0034] Among them, the distribution includes the distribution among the overall students and the distribution among the students in the class. The target group index indicates the index of whether the students in the class have a strong or weak grasp of the topic within the research scope.

[0035] The target group index is the TGI index. The TGI (Target Group Index) index is an index that reflects the strength or weakness of the target group within a specific research scope (such as geographic area, demographic field, media audience, product consumer). In this application, the TGI index is used to reflect the strength or weakness of the students in the class in mastering the topic within the research scope. The research scope can be understood as the level within the industry or the level within the school. Taking the score corresponding to the mode of the overall concentration as a reference, the number of people in the class and the whole who are not lower than the score can be used to calculate the pass rate and the TGI index. Taking into account that the class concentration may have a relatively large distribution range, the number of people who are not lower than the corresponding score of the mode of concentration is used as a parameter in the TGI index calculation formula. The higher the target group index, the higher the students' mastery of the topic, and the lower the target group index, the lower the students' mastery of the topic. Based on this, combined with the influence of other factors, the order of tackling the current topic among multiple topics can be defined. The target group index is for a topic, that is, each topic has its own target group index.

[0036] Step 400: Obtain the question attack sequence value according to the target group index, the question difficulty coefficient, the score level corresponding to the mode score, and the score coefficient.

[0037] The class score is at a low level (the corresponding score of the concentration mode is in the middle and low position in the score range), the difficulty coefficient of the questions is low, the target group index is low, it is easy to master quickly, and the score improvement range is large, so it can be prioritized for tackling. The difficulty coefficient of the questions is high, and the overall and class scores are at a low level, occupying few points, and it takes too much time and energy to overcome. The priority can be lowered. By combining multiple factors and the priority of each factor, the order of tackling multiple questions can be sorted, and a question tackling plan can be established to prioritize tackling questions with the greatest benefits and the highest returns, thereby improving learning efficiency.

[0038] Step 500: Extracting exercises from a question bank according to the question attack sequence value, question similarity and difficulty level to attack the questions.

[0039] After arranging the order of solving the questions, that is, after sorting the multiple questions, extract the exercises from the question bank according to the similarity and difficulty level of each question in order to formulate a complete solution strategy. The difficulty level of the questions can be divided according to the actual test situation, and the similarity of the questions can be judged by setting labels for each question and judging the matching degree between the labels. For the extraction principle of the exercises, the principle of high similarity of the questions, matching difficulty coefficient, and priority extraction can be followed, and the similarity can be expanded from high to low, and the difficulty coefficient can be from the same to the next, second, ..., N levels.

[0040] That is, in the finally obtained research plan, multiple questions in the unified examination are sorted for research, and then the exercise questions extracted for each question are sorted to form an ordered research plan, reasonably allocate time and resources, improve scores efficiently, break through key points and difficulties, making the question research plan more reasonable and the research exercises more accurate.

[0041] In one embodiment, step 400 may include: Y = (difficulty coefficient * (TGI / 100 + a) * score segment) / score value coefficient; where Y represents the question research order value, which is used to sort the question research order. The question research order value is inversely proportional to the research priority. TGI represents the target group index. The difficulty coefficient is extracted from the attribute settings in the question. The score segment represents the position of the mode corresponding score in the score range. The score range represents [the lowest score of the question among all students, the highest score of the question among all students]. The score value coefficient represents the total score of the question, and a represents the correction value.

[0042] Each Y value corresponds to a question, that is, each question has its own Y value. The research order of the question is related to the difficulty coefficient, target group index, score segment, and score system. Judging the research order of the question through multiple dimensions is more valuable for reference. For example, the higher the target group index, the higher the degree of mastery of the question by students. The lower the target group index, the lower the degree of mastery of the question by students. Questions with a low degree of mastery can be given higher priority. The difficulty system represents the difficulty coefficient of the question, which is the attribute setting of the question and can be directly extracted from the question. The score segment is the position of the mode corresponding score in the score range in the class concentration. For example, the score range is [0, 3], 0 points: 1 segment (value 1 / 4), 1 point: 2 segments (value 2 / 4), 2 points: 3 segments (value 3 / 4), 3 points: 4 segments (value 4 / 4). The segment value is normalized to reduce the deviation caused by different intervals. By normalizing the values, problems such as some questions that are not difficult but are mostly ignored when students study difficult questions, resulting in low scores, can be solved. Such questions can be quickly improved. However, for questions where the score concentration is already high but individual questions pull down the TGI index, the priority of such questions is lowered. The score value coefficient is the score set for the question. The higher the question score, the more the score can be improved after mastery. Therefore, questions with a higher score can be given higher priority. The correction value a is used for correction in special scenarios (TGI = 0, Y value = 0, and the coefficient has no effect on the Y value). According to the characteristics of the linear model, take a = 1, which does not affect the calculation result (plays a correction role).

[0043] Figure 2 is a schematic flowchart of the intelligent analysis method for question research provided by another exemplary embodiment of the present application. As Figure 2 shown, step 300 may include:

[0044] Step 310: Obtain the number of the first students in the class who have scores greater than or equal to the score corresponding to the mode.

[0045] Take the score corresponding to the mode as the reference score for reaching the passing line. The number of students in the class who reach the passing line = (for each question) the number of students in the class who have scores not lower than the score corresponding to the mode. The number of the first students is the number of students in the class who reach the passing line.

[0046] Step 320: Calculate the class achievement rate of the score corresponding to the mode according to the number of the first students and the total number of students in the class.

[0047] The class achievement rate = the number of students in the class who reach the passing line / the total number of students in the class. The class achievement rate can reflect the mastery level of the current class for this question. The number of the first students is the number of students in the class who reach the passing line.

[0048] Step 330: Obtain the number of the second students in the overall who have scores greater than or equal to the score corresponding to the mode.

[0049] Take the score corresponding to the mode as the reference score for reaching the passing line. The number of students in the overall who reach the passing line = (for each question) the number of students in the overall who have scores not lower than the score corresponding to the mode. The number of the second students is the number of students in the overall who reach the passing line.

[0050] Step 340: Calculate the overall achievement rate of the score corresponding to the mode according to the number of the second students and the total number of students in the overall.

[0051] The overall achievement rate = the number of students in the overall who reach the passing line / the total number of students in the overall. The overall achievement rate can reflect the mastery level of the current overall for this question. The number of the second students is the number of students in the overall who reach the passing line.

[0052] Step 350: Calculate the target group index according to the class achievement rate and the overall achievement rate.

[0053] The target group index = [class achievement rate / overall achievement rate] * standard number 100. This target group index can reflect whether the mastery level of the current class for this question is strong or weak in the overall. TGI index > 100: The mastery level of the question is higher than the industry level, and the larger the value, the higher the mastery level; TGI index = 100: The mastery level of the question is equal to the industry level; TGI index < 100: The mastery level of the question is lower than the industry level, and the smaller the value, the lower the mastery level.

[0054] In one embodiment, the above step 350 may include: (class achievement rate / overall achievement rate) * standard number 100 = target group index.

[0055] Target Group Index = [Class Achievement Rate / Overall Achievement Rate] * Standard Number 100. This Target Group Index can reflect whether the mastery level of the current class for this question is strong or weak in the overall context. TGI Index > 100: The mastery level of the question is higher than the industry level, and the larger the value, the higher the mastery level; TGI Index = 100: The mastery level of the question is equal to the industry level; TGI Index < 100: The mastery level of the question is lower than the industry level, and the smaller the value, the lower the mastery level.

[0056] In one embodiment, the above intelligent analysis method for question tackling may further include: setting labels and attributes for the question according to the score information of all students in the question; wherein, the labels include knowledge points, question types, and examination points, and the attributes include difficulty coefficients.

[0057] Based on the examination situation of this question, label the question (knowledge points, question types, examination points) and set attributes (difficulty coefficients). For example, a question may contain multiple knowledge points. In structured disciplines, for multiple-choice questions: generally, there are four options. Therefore, when the four options are completely unrelated, the question can be labeled with a maximum of 4 knowledge point labels. If the four options revolve around one knowledge point or scenario, then the maximum number of knowledge point labels is two; for fill-in-the-blank questions: if a question has N blanks, theoretically, a maximum of N knowledge point labels can be marked. If the knowledge points examined in multiple blanks are the same, the number of labels will be reduced accordingly; for solution questions: usually, it is divided into a question stem and N sub-questions. At most three knowledge points can be marked in the question stem, and at the same time, according to the number and examination content of the sub-questions, each sub-question can be marked with a maximum of one knowledge point label. Question type: Similar examination points, such as examination points being understanding, expression, induction, basic application, appreciation and evaluation, analysis and synthesis, exploration, and expression application. And the examination points can be further divided into multiple sub-points for fine classification of the question, improving the accuracy of subsequent similarity retrieval and closely fitting the examination points for extracting practice questions. The difficulty coefficient of the attribute is an inherent attribute of the question and always accompanies the question.

[0058] Setting labels can not only perform similarity retrieval on questions from multiple dimensions but also perform similarity retrieval on questions from a single label, and allocate question tackling according to a certain single specific attribute to achieve key breakthroughs.

[0059] In one embodiment, the above intelligent analysis method for question tackling may further include: calculating the question similarity between the current question and the questions in the question bank according to the similarity formula; wherein, the similarity formula is: In the formula, item_tags(ti) stores the annotation vector of question i, item_tags(tj) stores the annotation vector of question j. Question i is the current question or a question in the question bank, and question j is a question in the question bank or the current question. The annotation vector includes knowledge points, question types, and examination points. The similarity value ranges from [0, 1], and the similarity value is directly proportional to the question similarity.

[0060] After determining the order of attacking the questions, it is necessary to extract practice questions from the question bank to attack the current question. In order to attack the current question, it is best to select questions with highly similar knowledge points, question types, and test points to achieve the attack on the current question. Therefore, the calculation of question similarity considers the tags and attributes of the questions to achieve high-efficiency and high-accuracy retrieval, and push accurate practice questions to achieve question attack. For example, using a similarity formula, calculate considering knowledge points, question types, and test points. The similarity value ranges from [0, 1]. The similarity value is proportional to the question similarity. The larger the value, the higher the similarity. That is, when the similarity value is 1, the similarity is the highest. Based on this, extract from the question bank according to tags (knowledge points, question types, test points) and attributes (difficulty level). The general order can be: 1. High similarity & the difficulty level to be extracted; 2. Medium similarity & the difficulty level to be extracted; 3. Low similarity & the difficulty level to be extracted. Considering that the priority of similarity is higher than the difficulty level to be extracted, so only when the similarities are equal, then extract and sort according to the difficulty level to be extracted. The difficulty level to be extracted can be customized.

[0061] Figure 3 It is a schematic flowchart of the intelligent analysis method for question attack provided by another exemplary embodiment of the present application, as Figure 3 shown, the above intelligent analysis method for question attack may further include:

[0062] Step 600: Divide the difficulty levels of multiple questions and the questions in the question bank according to multiple questions and the score information of the overall students in each question.

[0063] The difficulty level of the question can be determined according to the score segment and the difficulty coefficient, and the division rule of the difficulty level can be customized. For example, the first priority is to match the difficulty level of the current question (the current question is more difficult and the score is in the low segment, reduce the difficulty for the first priority match), the second priority match, the second priority match is to increase one difficulty level on the current basis, the third priority match, the expansion principle is to have one level above and below the existing difficulty level range respectively, the fourth priority match, the expansion principle is to have two levels above and below the existing difficulty level range respectively.

[0064] Step 700: Determine the priority of difficulty level matching based on the score segment, the difficulty coefficient, and the difficulty level division.

[0065] Among them, the score segment represents the position of the mode corresponding score in the score interval.

[0066] For example: According to the difficulty level of the current question, the score segments are grouped and pushed in stages. The difficulty level of the test questions can generally be divided into 5 levels: easy, relatively easy, moderate, relatively difficult, and difficult. The score segments are divided into high, medium, and low: according to the percentage of the score range [0-30] [30-70] [70-100].

[0067] a. For difficulty levels of moderate or below (easy, relatively easy, moderate), groups and tiered push are carried out according to the score level; the current difficulty is matched as the first priority, and the difficulty level up is matched as the second priority for expansion, and the difficulty level up or down is matched as the third priority for expansion.

[0068] In actual matching, the current question difficulty is: relatively difficult. The current group score is at a low level, and the order of difficulty is: first priority is to extract a lower difficulty level (moderate), second priority is to extract the same difficulty level (relatively difficult) to expand the quantity, and third priority is to extract (relatively easy, difficult).

[0069] b. For difficulty levels above moderate (relatively difficult, difficult), they are grouped and pushed in stages according to the score level; if the current score is in a low level: reduce the difficulty by one and extract it as the first priority; if the difficulty level goes up one level, the current difficulty level will be matched as the second priority for expansion; if the difficulty level goes up or down one level in the current difficulty range, it will be matched as the third priority for expansion; if the difficulty level goes up or down two levels in the current difficulty range, it will be matched as the fourth priority, and so on.

[0070] In actual matching, the current question difficulty is: relatively difficult. The current group score segment is in the middle and high level. The order of difficulty extraction is: first, the current difficulty level (relatively difficult), second, the next difficulty level (difficult) for expansion, and third, the difficulty level (moderate).

[0071] Therefore, by combining the scoring range, difficulty coefficient and difficulty level division, we can achieve accurate delivery of questions, consider multi-dimensional factors, formulate reasonable solution plans, and effectively and efficiently improve scores.

[0072] In one embodiment, if Figure 3 As shown, the above step 500 may include:

[0073] Step 510: sorting the multiple questions from high to low according to the question attack sequence values.

[0074] The question attack sequence value is used to sort the question attack sequence. The question attack sequence value is inversely proportional to the attack priority. Therefore, the larger the question attack sequence value, the smaller the priority and the lower the attack ranking. The smaller the attack sequence value, the higher the priority and the higher the attack ranking.

[0075] Step 520: Perform range matching on each topic in order from high to low according to the topic similarity, and obtain multiple similar topics for each topic.

[0076] After the sorting is completed, for each question, match it in the question one by one. After multiple questions are matched according to the similarity, sort them according to the similarity level. The higher the similarity, the higher the extraction priority, and the lower the similarity, the lower the extraction priority.

[0077] Step 530: Obtain the extraction order of multiple similar questions based on the matching priority of the difficulty level and the similarity value.

[0078] Among the same similarities, sort the questions with the same similarity according to the matching priority of the difficulty level. The specific sorting rules can be customized. For example, the first priority is to match the current question's difficulty level (if the current question has a high difficulty and a low score in the low segment, reduce the difficulty for the first priority match), the second priority is to match, the second priority is to match one difficulty level up on the current basis, the third priority is to match, the expansion principle is to have one level above and below the existing difficulty level range respectively, the fourth priority is to match, the expansion principle is to have two levels above and below the existing difficulty level range respectively, etc. It can be set according to the student's learning situation and the specific research needs for this question.

[0079] Step 540: Extract practice questions from the question bank based on the extraction order to conduct research on the questions.

[0080] After the hierarchical sorting, extract practice questions from the question bank based on the extraction order. The number of extracted questions can be customized. For example, extract the top 20 questions in the extraction ranking. The number of extracted questions can also be set according to the student's learning situation and the specific research needs for this question.

[0081] In an embodiment, the above step 530 may include: extraction order = similarity weight * difficulty weight; where the similarity weight is proportional to the similarity value, the similarity weight value ranges from 0 to 1, and the difficulty weight is proportional to the matching priority of the difficulty level.

[0082] In addition to first relying on the question similarity and then on the matching priority of the difficulty level for the extraction order, an extraction order formula can also be set, that is, extraction order = similarity weight * difficulty weight. The value of the similarity weight is the value of the above similarity formula, and the value range is [0, 1]. The higher the similarity, the greater the similarity weight, and the lower the similarity, the lower the similarity weight; for example, when the value is 1, the similarity is the largest. The difficulty weight can be set according to the matching priority of the difficulty level. The higher the matching priority of the difficulty level, the greater the difficulty weight value, and the lower the matching priority of the difficulty level, the lower the difficulty weight value. The weights of similarity and difficulty are adjusted according to the actual situation. Therefore, the extraction order is also affected by multiple dimensions and is not determined based on a single factor, and finally more accurate practice questions can be pushed to support the research plan.

[0083] The magnitude of the difficulty weight value is obtained through model training and is specifically set according to business requirements. For example, the first matching level = 1, the second matching level = 0.8, the third matching level = 0.5, …… others = 0.01.

[0084] Exemplary Device

[0085] Figure 4 FIG. is a schematic structural diagram of a question-solving intelligent analysis device provided by an exemplary embodiment of the present application. As Figure 4 shown, the question-solving intelligent analysis device 2 includes: an information acquisition module 21, configured to acquire a plurality of questions and the score information of all students in each question; a score value acquisition module 22, configured to acquire the score value corresponding to the mode based on the distribution of the score information and the normal distribution; an index calculation module 23, configured to calculate a target group index according to the distribution of the students who obtain the score value corresponding to the mode, where the distribution includes the distribution among all students and the distribution among class students, and the target group index represents the index of whether the class students are strong or weak in mastering the questions within the research scope; an acquisition sequence module 24, configured to obtain a question-solving sequence value according to the target group index, the question difficulty coefficient, the score segment corresponding to the mode score value, and the score coefficient; an extraction module 25, configured to extract practice questions from the question bank according to the question-solving sequence value, the question similarity, and the difficulty level for question solving.

[0086] The question-solving intelligent analysis device provided by the present application analyzes the mastery of each question by each class based on the question type, test point, knowledge point, difficulty level, score range (importance), overall mastery situation, etc. of the questions, combines multi-dimensional analysis of the current class scores, formulates a reasonable question-solving allocation plan for each class, intelligently and accurately pushes practice questions, makes the question-solving plan more reasonable, the question-solving practice more accurate, fully utilizes limited time and resources, quickly masters questions, masters more questions, and improves more scores.

[0087] Figure 5 FIG. is a schematic structural diagram of a question-solving intelligent analysis device provided by another exemplary embodiment of the present application. As Figure 5 shown, the above-mentioned index calculation module 23 may be configured as: a first acquisition unit 231, configured to acquire the number of first students in the class who are greater than or equal to the score value corresponding to the mode. A first calculation unit 232, configured to calculate the class achievement rate of the score value corresponding to the mode according to the number of first students and the total number of class students. A second acquisition unit 233, configured to acquire the number of second students in the whole who are greater than or equal to the score value corresponding to the mode. A second calculation unit 234, configured to calculate the overall achievement rate of the score value corresponding to the mode according to the number of second students and the total number of all students. A third calculation unit 235, configured to calculate the target group index according to the class achievement rate and the overall achievement rate.

[0088] In one embodiment, the above-mentioned third calculation unit 235 may be configured as: (class achievement rate / overall achievement rate) * standard number 100 = target group index.

[0089] In one embodiment, the above-mentioned intelligent analysis device 2 for problem-solving can also be configured to: set tags and attributes for the questions according to the score information of all students in the questions; wherein, the tags include knowledge points, question types, and examination points, and the attributes include difficulty coefficients.

[0090] In one embodiment, the above-mentioned intelligent analysis device 2 for problem-solving can also be configured to: calculate the question similarity between the current question and the questions in the question bank according to the similarity formula; wherein, the similarity formula is: In the formula, item_tags(ti) stores the annotation vector of question i, item_tags(tj) stores the annotation vector of question j, question i is the current question or a question in the question bank, question j is a question in the question bank or the current question, the annotation vector includes knowledge points, question types, and examination points, the similarity value ranges from [0, 1], and the similarity value is directly proportional to the question similarity.

[0091] In one embodiment, as Figure 5 shown, the above-mentioned intelligent analysis device 2 for problem-solving can also be configured to: a division module 26, which is used to divide the difficulty levels of multiple questions and the questions in the question bank according to multiple questions and the score information of all students in each question; a determination priority module 27, which is used to determine the difficulty level matching priority based on the score segment, difficulty coefficient, and difficulty level division. Among them, the score segment represents the position of the mode corresponding score in the score interval.

[0092] In one embodiment, as Figure 5 shown, the above-mentioned extraction module 25 can be configured to: a sorting unit 251, which is used to sort multiple questions from high to low according to the problem-solving order value of the questions; a matching unit 252, which is used to perform range matching on each question in turn from high to low according to the question similarity to obtain multiple similar questions for each question; an extraction order unit 253, which is used to obtain the extraction order of multiple similar questions based on the difficulty level matching priority and the similarity value; an extraction practice question unit 254, which is used to extract practice questions from the question bank based on the extraction order for problem-solving.

[0093] In one embodiment, the above-mentioned extraction order unit 253 can be configured as: extraction order = similarity weight * difficulty weight; wherein, the similarity weight is directly proportional to the similarity value, the similarity weight value ranges from 0 to 1, and the difficulty weight is directly proportional to the difficulty level matching priority.

[0094] Exemplary Electronic Device

[0095] An electronic device, the electronic device comprising: a processor; a memory for storing processor-executable instructions; the processor for executing the problem-solving intelligent analysis method described in the embodiments provided in the present application.

[0096] Next, reference is made to Figure 6 to describe the electronic device according to an embodiment of the present application. The electronic device may be either or both of the first device and the second device, or a stand-alone device independent of them, and the stand-alone device may communicate with the first device and the second device to receive the input signals collected therefrom.

[0097] Figure 6 The block diagram of the electronic device according to an embodiment of the present application is illustrated.

[0098] As Figure 6 shown, the electronic device 10 includes one or more processors 11 and a memory 12.

[0099] The processor 11 may be a central processing unit (CPU) or other form of processing unit having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0100] The memory 12 may include one or more computer program products, and the computer program products may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory, etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage media, and the processor 11 may run the program instructions to implement the problem-solving intelligent analysis method of the various embodiments of the present application described above and / or other desired functions. Various contents such as input signals, signal components, noise components, etc. may also be stored in the computer-readable storage media.

[0101] In one example, the electronic device 10 may further include: an input device 13 and an output device 14, and these components are interconnected through a bus system and / or other forms of connection mechanisms (not shown).

[0102] When the electronic device is a stand-alone device, the input device 13 may be a communication network connector for receiving the input signals collected from the first device and the second device.

[0103] In addition, the input device 13 may further include, for example, a keyboard, a mouse, and so on.

[0104] The output device 14 can output various information to the outside, including the determined distance information, direction information, etc. The output device 14 can include, for example, a display, a speaker, a printer, a communication network, and remote output devices connected thereto, and the like.

[0105] Of course, for simplicity, Figure 6 only some of the components related to the present application in the electronic device 10 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application scenarios, the electronic device 10 may further include any other appropriate components.

[0106] The computer program product can be written in any combination of one or more programming languages for executing the program code of the operations of the embodiments of the present application. The programming languages include object-oriented programming languages such as Java, C++, etc., and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, executed as a stand-alone software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0107] A computer-readable storage medium stores a computer program for executing the problem-solving intelligent analysis method described in the embodiments provided by the present application.

[0108] The computer-readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0109] The above description has been given for purposes of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions, and sub-combinations thereof.

Claims

1. An intelligent analysis method for tackling problems, characterized in that, Including: Obtain multiple questions and the score information of all students in each of the questions; Based on the distribution of the score information and the normal distribution, obtain the score corresponding to the mode; According to the distribution of the students who scored the score corresponding to the mode, calculate the target group index; wherein, the distribution includes the distribution among all students and the distribution among class students, and the target group index represents the index of whether the class students are strong or weak in mastering the questions within the research scope; According to the target group index, the question difficulty coefficient, the score segment corresponding to the mode score, and the score coefficient, obtain the question tackling order value; Extract practice questions from the question bank according to the question tackling order value, the question similarity, and the difficulty level for question tackling.

2. The intelligent analysis method for tackling problems according to claim 1, characterized in that, Obtaining the question tackling order value according to the target group index, the question difficulty coefficient, the score segment corresponding to the mode score, and the score coefficient includes: Y = (difficulty coefficient * (TGI / 100 + a) * score segment) / score coefficient; Wherein, Y represents the question tackling order value, the question tackling order value is used for sorting the question tackling order, the question tackling order value is inversely proportional to the tackling priority, TGI represents the target group index, the difficulty coefficient is extracted from the attribute settings in the question, the score segment represents the position of the score corresponding to the mode in the score interval, the score interval represents [the lowest score of the question among all students, the highest score of the question among all students], the score coefficient represents the total score of the question, and a represents the correction value.

3. The intelligent analysis method for tackling problems according to claim 1, characterized in that, Calculating the target group index according to the distribution of the students who scored the score corresponding to the mode includes: Obtain the number of first students in the class who are greater than or equal to the score corresponding to the mode; According to the number of first students and the total number of class students, calculate the class achievement rate of the score corresponding to the mode; and Obtain the number of second students in the whole who are greater than or equal to the score corresponding to the mode; According to the number of second students and the total number of all students, calculate the overall achievement rate of the score corresponding to the mode; and Calculate the target group index according to the class achievement rate and the overall achievement rate.

4. The intelligent analysis method for tackling problems according to claim 3, characterized in that, Calculating the target group index according to the class achievement rate and the overall achievement rate includes: (class achievement rate / overall achievement rate) * standard number 100 = target group index.

5. The intelligent analysis method for tackling problems according to claim 1, characterized in that, The intelligent analysis method for question tackling further includes: Perform label setting and attribute setting on the question according to the score information of all students in the question; wherein, the labels include knowledge points, question types, test points, and the attributes include difficulty coefficient.

6. The intelligent analysis method for tackling problems according to claim 1, characterized in that, The intelligent analysis method for question tackling further includes: Calculate the question similarity between the current question and the questions in the question bank according to the similarity formula; wherein, the similarity formula is: In the formula, item_tags(ti) stores the annotation vector of question i, and item_tags(tj) stores the annotation vector of question j. Question i is the current question or a question in the question bank, and question j is a question in the question bank or the current question. The annotation vector includes knowledge points, question types, and examination points. The similarity value ranges from [0, 1], and the similarity value is directly proportional to the question similarity.

7. The intelligent analysis method for tackling problems according to claim 6, characterized in that, The intelligent analysis method for question tackling further includes: Based on multiple questions and the score information of the overall students in each of the questions, perform a difficulty level division on the multiple questions and the questions in the question bank; Based on the score segment, difficulty coefficient, and difficulty level division, determine the difficulty level matching priority; the score segment represents the position of the mode corresponding score in the score interval.

8. The intelligent analysis method for tackling problems according to claim 7, characterized in that, Extract practice questions from the question bank according to the question tackling sequence value, question similarity, and difficulty level for question tackling, including: Perform a tackling ranking on multiple questions from high to low according to the question tackling sequence value; Perform a range matching on each question in turn from high to low according to the question similarity to obtain multiple similar questions for each question; Based on the difficulty level matching priority and the similarity value, obtain the extraction sequence of the multiple similar questions; Extract practice questions from the question bank based on the extraction sequence for question tackling.

9. The intelligent analysis method for problem-solving as claimed in claim 8, wherein, The obtaining the extraction sequence of the multiple similar questions based on the difficulty level matching priority and the similarity value includes: Extraction sequence = similarity weight * difficulty weight; where the similarity weight is directly proportional to the similarity value, the similarity weight ranges from 0 to 1, and the difficulty weight is directly proportional to the difficulty level matching priority.

10. An intelligent analysis device for problem-solving, wherein, It includes: An information acquisition module for acquiring multiple questions and the score information of the overall students in each of the questions; A score acquisition module for acquiring the score corresponding to the mode based on the distribution of the score information and the normal distribution; An index calculation module for calculating the target group index according to the distribution of the students who obtained the score corresponding to the mode; where the distribution includes the distribution among the overall students and the distribution among the class students, and the target group index represents the index of whether the class students are strong or weak in mastering the questions within the research scope; An order acquisition module for obtaining the question tackling sequence value according to the target group index, question difficulty coefficient, score segment corresponding to the mode, and score coefficient; An extraction module for extracting practice questions from the question bank according to the question tackling sequence value, question similarity, and difficulty level for question tackling.

Images