WO2021253480A1 - 习题智能推荐方法、装置、计算机设备及存储介质 - Google Patents

习题智能推荐方法、装置、计算机设备及存储介质 Download PDF

Info

Publication number
WO2021253480A1
WO2021253480A1 PCT/CN2020/098828 CN2020098828W WO2021253480A1 WO 2021253480 A1 WO2021253480 A1 WO 2021253480A1 CN 2020098828 W CN2020098828 W CN 2020098828W WO 2021253480 A1 WO2021253480 A1 WO 2021253480A1
Authority
WO
WIPO (PCT)
Prior art keywords
exercises
error rate
exercise
preset
knowledge point
Prior art date
Application number
PCT/CN2020/098828
Other languages
English (en)
French (fr)
Inventor
梁瑾
冯心
盛亮
谢保林
Original Assignee
平安科技(深圳)有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技(深圳)有限公司 filed Critical 平安科技(深圳)有限公司
Publication of WO2021253480A1 publication Critical patent/WO2021253480A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33Querying
    • G06F16/335Filtering based on additional data, e.g. user or group profiles
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/22Matching criteria, e.g. proximity measures
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10Services
    • G06Q50/20Education

Definitions

  • This application relates to the field of artificial intelligence, in particular to a method, device, computer equipment and storage medium for intelligently recommending exercises.
  • the main purpose of this application is to provide a method, device, computer equipment, and storage medium for intelligently recommending exercises to automatically generate personalized recommended exercises for students' own homework and knowledge mastery.
  • this application proposes a method for intelligently recommending exercises, including:
  • this application also proposes an intelligent recommendation device for exercises, including:
  • the obtaining module is used to obtain the wrong question in the answer result, and search the corresponding knowledge point A in the preset knowledge point mapping table according to the wrong question;
  • the first calculation module is used to calculate the similarity between the multiple knowledge points in the exercise library and the knowledge point A according to the first preset formula, and obtain that all the similarities with the knowledge point A are greater than the first similarity Threshold value of the knowledge point B, and sort all the knowledge points B in the order of similarity from high to low to form a knowledge point B set;
  • the second calculation module is configured to obtain the exercise quality of at least one first exercise corresponding to the knowledge point A and the multiple second exercises corresponding to the set of knowledge points B according to the second preset formula, and follow the exercises Sort the first exercises and the second exercises in descending order of quality to form a first exercise set and a second exercise set;
  • the extraction module is used to obtain the single recommended number of recommended exercises, and extract the first exercises from the first exercise set and the second exercises from the second exercise set respectively according to a preset recommendation ratio to form The recommended exercises; wherein, when extracting the first exercises and the second exercises, they are extracted in descending order of the quality of the exercises.
  • this application also proposes a storage medium on which a computer program is stored, and when the computer program is executed by a processor, a method for intelligently recommending exercises is realized.
  • the steps of the method for intelligently recommending exercises include:
  • this application proposes a method, device, computer equipment and storage medium for intelligently recommending exercises. Draw out the knowledge points that each student has mastered in place and the knowledge points that may not be mastered; then through the second preset formula, further calculate the appropriate exercises to recommend to the students, and solve the current recommendation of wrong questions.
  • the wrong question is relatively single , And the quality of the wrong question recommendation is not well-targeted. It can automatically generate personalized recommended practice exercises according to the students’ own homework and knowledge points, so that students can spend their time on the questions that should be practiced, and help students learn efficiently, so as to save time for doing exercises and achieve personalization The purpose of learning to improve learning efficiency.
  • FIG. 2 is a schematic diagram of modules of an intelligent recommendation device for exercises in an embodiment of this application
  • FIG. 3 is a schematic block diagram of modules of a computer device in an embodiment of the application.
  • FIG. 4 is a schematic block diagram of modules of a storage medium in an embodiment of the application.
  • a method for intelligently recommending exercises which involves the field of artificial intelligence, where the method mainly includes the following steps:
  • the homework recommendation system first obtains the student, such as student S, the past answer results, which are obtained by correcting the exercises submitted by the student, and determine the student’s wrong question based on the student’s answer result , In order to know the weakness of the student’s knowledge.
  • the answering result includes the result obtained by the student from answering the exercises assigned by the teacher.
  • the answering result also includes the result obtained by the student from answering the recommended exercise of the system. Then, in order to further determine the students’ weak points of knowledge, according to the obtained wrong questions, look up the corresponding knowledge point A in the knowledge point mapping table. There is at least one knowledge point A, and the problem library is preset in the knowledge point mapping table.
  • the knowledge points corresponding to each exercise in the exercises such as Tang poetry, Song poetry, magnetic field mechanics, thermodynamics, contradiction, cell structure and other knowledge points, that is, the student's unmastered knowledge points. All the wrong questions obtained are divided into categories for follow-up The homework recommendations provide a better data basis.
  • the system calculates the similarity between the multiple knowledge points in the exercise library and the knowledge point A according to the first preset formula, and obtains all the knowledge points B whose similarity to the knowledge point A is greater than the first similarity threshold, and according to Sort all knowledge points B in the order of similarity from high to low to form a set of knowledge points B; in a specific embodiment, the first preset formula is:
  • knowledge point A and knowledge point B are represented by vectors in the exercise library, x A and x B respectively refer to the feature value of knowledge points in the exercise library of knowledge point A and knowledge point B, and each knowledge point has For a corresponding feature value of a knowledge point, the similarity between different knowledge points is negatively correlated with the difference of the feature value of the knowledge point.
  • the feature value of the knowledge point is represented by a multi-digit number. Each number represents the subject, category, chapter, section, point, etc. in the knowledge attribute of the knowledge point.
  • the knowledge point feature value of knowledge point A is 545863, and the knowledge attributes represented by each number in the knowledge point feature value are shown in the table One shown.
  • the last digit in the feature value of the knowledge point is different according to the previous digit.
  • the first digit is The two digit 4 represents organic chemistry in chemistry, but if it is in the knowledge point feature value 445863, since the first digit 5 represents physics, then the second digit 4 represents physics.
  • each digit in the feature value of the knowledge point has a specific corresponding knowledge attribute, and the latter digit does not change according to the change of the previous digit.
  • the knowledge point is described by the feature value of the knowledge point, the similarity between different knowledge points will be negatively correlated with the difference of the feature value of the knowledge point, that is, if the knowledge between two knowledge points is The closer the point eigenvalues are, the smaller the difference is, and the higher the similarity between the two knowledge points is.
  • y A and y B respectively refer to the difficulty of knowledge point A and knowledge point B.
  • the difficulty of knowledge points can be divided into proficiency, memorization, application, understanding, etc., and different values are used to correspond to the difficulty of knowledge points, such as proficient.
  • the difficulty value of knowledge points is 1, the difficulty value of memorizing knowledge points is 2, the difficulty value of applied knowledge points is 3, the difficulty value of understanding knowledge points is 4, etc.
  • the feature value of the knowledge point and the difficulty value of the knowledge point together form a two-dimensional vector, which is used to describe the corresponding knowledge point.
  • the similarity between knowledge point A and knowledge point can be obtained, and the value of cosine similarity is [-1,1], the larger the value is It shows that the similarity between knowledge point A and knowledge point is higher.
  • the result value is 1, it means that knowledge point A is exactly the same as knowledge point B.
  • the result value is -1, it means that knowledge point A and knowledge point B totally different.
  • the first exercise corresponding to knowledge point A and the second corresponding to knowledge point B set should be obtained respectively.
  • the exercise quality of the exercises, and the first exercises and the second exercises are sorted according to the order of the quality of the exercises, and the first exercise set and the second exercise set are formed;
  • the first preset formula it is possible to calculate the knowledge points that each student has mastered and the knowledge points that may not be mastered according to the situation of each student's work; through the second preset formula, the appropriate recommendation to The student’s exercises solve the problem of the single wrong question in the current recommendation of wrong questions and the poor quality of the recommendation of wrong questions.
  • the method further includes:
  • the error rate of each knowledge point A in the answer results is calculated.
  • the corresponding error is determined according to the wrong question
  • knowledge point A of knowledge point A count the total number of wrong questions corresponding to each knowledge point A and the total number of exercises corresponding to knowledge point A in the answer results, and then calculate the error rate of exercises for knowledge point A based on the total number of wrong questions and the total number of exercises .
  • the recommendation ratio refers to the proportion of the number of exercises corresponding to knowledge point A within a certain range of exercise error rate in the total number of recommended exercises.
  • the recommendation ratio contained in the corresponding recommendation plan is 35%
  • the knowledge point A under the error rate of exercises of 80%-100% corresponds to The proportion of the number of exercises in the recommended exercises is 35%.
  • the sum of the recommended ratios in all recommended schemes is less than or equal to 1, in case the sum of the recommended ratios is greater than 1 when the recommended exercises are subsequently generated, causing the problem of incorrect generation of recommended exercises.
  • the recommendation ratio in the recommendation scheme includes the first exercise corresponding to knowledge point A and the second exercise corresponding to the set of knowledge point B, that is, the proportion of the first exercise and the second exercise in the recommended exercises The ratio is the recommended ratio.
  • the recommended scheme found in the preset recommendation table select the exercises corresponding to knowledge point A from the exercise library and the rest of the exercises in the exercise library, and generate recommended exercises and push them to the students.
  • the sum of the recommended proportions according to the recommended schemes found in the preset recommendation table is equal to 1, that is, the student’s knowledge point A has a wider coverage of the exercise error rate.
  • the recommended exercises The selected ones are all exercises corresponding to knowledge point A, and the recommendation ratio of other exercises in the exercise database is 0, that is, the remaining exercises are not recommended.
  • the selection from the remaining exercises in the exercise database will account for 20 of the total number of exercises. % To make up the number. At this time, there are both exercises that the student has not mastered and the exercises that the student has mastered in the recommended exercises. Mixing the rest of the exercises in the exercise library into the recommended exercises has two advantages: one is that it allows students to perform appropriate repeated consolidation exercises to deepen their mastery; the other is that the recommended exercises are not all exercises and questions that students have not mastered. The degree of difficulty is appropriate to increase students' interest in answering questions, thereby optimizing the effect of the questions.
  • the aforementioned second preset formula is:
  • Q z refers to the problem quality of exercise z
  • P z refers to the average score of exercise z
  • P zi refers to the score of exercise z by student i
  • C refers to the degree of collection of exercises z.
  • C max refers to the highest number of collections among all exercises
  • C zf refers to the number of collections of exercise z
  • n refers to the number of students i.
  • the aforementioned exercise quality is calculated according to the second preset formula.
  • the system uses a 1-10 grading system to provide students or teachers to grade exercises.
  • the number of collections can reflect the quality of exercises in addition to the score. The more the number of collections of the exercises, the better the quality of the exercises.
  • C refers to the degree of collection of exercises z
  • C max refers to all exercises.
  • the highest number of collections, C zf refers to the number of collections of exercise z.
  • the number of collections of exercise z After dividing by the highest number of collections of exercises, multiply it by the average score of exercises z, and combine the number of collections with the average score of exercises to reduce errors.
  • the average score of the exercises is consistent with the statistical weight of the degree of collection of exercises.
  • the step S11 of calculating the error rate of all the knowledge points A in the answer results includes:
  • S112 Calculate the error rate of the exercises corresponding to the various question types of the knowledge point A according to the number of the exercises and the total number of the wrong questions.
  • the number of exercise questions corresponding to the answer results of the various question types of knowledge point A and the total number of wrong questions are counted; in a specific embodiment, because there are many types of exercises, such as multiple-choice questions, fill-in-the-blank questions Different question types have different answering methods. Therefore, even if the same student faces the same knowledge point, if the question types are different, the degree of mastery may be different. For example, some people are better at doing multiple-choice questions, and some people are better at doing solution questions. Therefore, we can specifically count the number of exercises and the total number of wrong questions in the answer results of the various question types of knowledge point A, so that a detailed analysis of knowledge point A can be carried out.
  • the error rate of exercises corresponding to the various question types of knowledge point A is calculated. After calculating the error rate of the exercises corresponding to the various question types of knowledge point A, you can get a clearer understanding of the student's work on the knowledge point A, and you can also recommend different question types in the follow-up exercise recommendation. In a specific embodiment, for example, based on the number of exercises and the total number of wrong questions, it is calculated that the error rate of exercises for the same knowledge point A on multiple-choice questions is 10%, and the error rate of exercises on fill-in-the-blank questions is 40%.
  • the exercise error rate on the question is 70%, and different recommended schemes may be found in the preset recommendation table according to the different exercise error rate, and thus have different recommendation ratios.
  • the recommended ratio of exercises corresponding to knowledge point A in multiple-choice questions is 5%
  • the recommended ratio in fill-in-the-blank questions is 15%
  • the recommended ratio in answer questions is 30%. Therefore, for the same knowledge point A.
  • Different question types make targeted recommendations.
  • the step of counting the error rates of all the knowledge points A in the answer results includes:
  • S113 Calculate the first error rate of each of the knowledge points A in all the answer results, and the second error rate of the answer results submitted in the last preset number of times;
  • the second error rate is set as the exercise error rate to reflect the progress of students in a timely manner to avoid the early error rate from affecting the exercise recommendation plan.
  • the first error rate is less than the second error rate, it means that the error rate of the knowledge point A in all answer results is lower than the error rate in the recent answer results, which means that the student is doing the knowledge point In the exercises of A, the early error rate is low, and the recent error rate is high. In the near future, there may be cases of forgetting knowledge points or carelessness in doing the exercises. Therefore, the first error rate is set as the exercise error rate. Based on the overall situation of the students, we can better grasp the situation of students' problems, and avoid the recent abnormal situation that affects the recommended solutions for exercises.
  • the first error rate of each of the knowledge points A in all the answer results is calculated, and the second error rate of the answer results submitted in the most recent preset number of times is calculated.
  • the error rate steps include:
  • S1131 Calculate the first error rate of each knowledge point A in all the answer results, and determine whether the number of wrong questions corresponding to the knowledge point A reaches a preset number;
  • the second error rate of the answer results submitted by the knowledge point A in the latest preset number of times is counted; because if the first error rate is compared with the second error rate, the student can be judged For the overall mastery of the same knowledge point A and recent mastery, it is necessary to have a sufficient number of samples to distinguish the overall situation from the recent situation. If the number of samples is too small, the second error rate cannot actually reflect The situation of students' recent study problems. In a specific embodiment, if the number of wrong questions corresponding to the knowledge point A does not reach the preset number, the first error rate is directly set as the exercise error rate without calculating the second error rate.
  • step of calculating the second error rate in the answer results submitted by the knowledge point A in the most recent preset number of times if the preset number is reached further include:
  • the preset difference range is within 5%, then This shows that the student’s error rate for this knowledge point A in the early and recent periods is similar, and has not been able to fully grasp the knowledge point A. It is very likely that the knowledge point is not well understood, resulting in a constant error rate in the question. If there is no improvement, then the knowledge point A will be sent to the preset management interface, such as the teacher's attention management page, so that the teacher can timely understand the students' knowledge weaknesses, so as to provide targeted guidance for the knowledge blind area and improve the students Problem-solving efficiency.
  • the preset management interface such as the teacher's attention management page
  • the step of querying a preset recommendation table according to the error rate of the exercises to determine the recommendation scheme corresponding to each knowledge point A at the error rate of the exercises includes :
  • different recommended solutions correspond to different ranges of exercise error rates, and different recommended solutions contain different recommendation ratios.
  • the recommended ratio refers to the knowledge points within a certain range of exercise error rates. The proportion of exercises corresponding to A in the recommended exercises. If the exercise error rate is greater than or equal to the first preset threshold, such as 80%, then the exercise error rate is higher.
  • the exercise error rate is greater than 80%, the exercise corresponding to knowledge point A is directly recommended to the student, one is because the student does not If you don’t have a good grasp of the knowledge point A, and there is no good solution to the exercise, continuing to work on the problem is like a blind person touching the elephant, and you cannot get a good method of doing the problem, and the error rate of doing the problem is not very good. Improvement will only remain high; second, if students are asked to repeat the exercises with a high error rate, it will easily dampen their confidence in solving problems and reduce their interest and efficiency in answering questions.
  • the knowledge point A whose exercise error rate is within this range is better It is also the part that students need to focus on.
  • the first preset recommendation ratio such as 70%, the exercises corresponding to knowledge point A are recommended, and because the error rate of the exercises is still high, the knowledge point A Push to the teacher's preset management interface, so that the teacher has a clear understanding of the students' answering situation.
  • the error rate of exercises is less than the second preset threshold, for example, less than 10%, the error rate of exercises is low at this time, indicating that the student may have done the exercises carelessly, or just temporarily forgot some knowledge and caused the mistakes.
  • the knowledge point A is recommended according to the second preset recommendation ratio, for example, 10%.
  • the preset threshold value is not limited to the first preset threshold value and the second preset threshold value.
  • One or more different preset threshold values can be set according to actual operating conditions, so as to deal with different problem error rates.
  • Knowledge point A, implementation of different recommended schemes, and similar embodiments are all within the protection scope of this application.
  • the first exercises are extracted from the first exercise set and the second exercises are extracted from the second exercise set respectively according to a preset recommendation ratio to form the recommended exercises
  • the steps include:
  • S41 Obtain the most recent recommendation time of each knowledge point A, and determine whether the time interval between the most recent recommendation time and the current time is greater than or equal to a preset recommendation interval time;
  • the recommended program also includes the recommended interval time.
  • the recommended interval time is preset.
  • the preset recommended interval time is set according to the Ebbinghaus forgetting curve, such as one day and seven days. , 15 days, that is, the system selects exercises corresponding to knowledge point A from the exercise library on the second, seventh, and fifteenth days to recommend.
  • the preset recommended time can also be set by itself.
  • the most recent recommendation time of each knowledge point A that is, the time when the knowledge point A appeared in the most recent recommended exercise, and determine whether the time interval between the most recent recommendation time and the current time is greater than or equal to the preset recommendation interval Time; because students may have more knowledge points A, and the number of exercises for a single recommended exercise is limited, so a single recommended exercise may not cover all knowledge points A, and when obtaining recommended exercises, both It can be automatically recommended by the system, or students or teachers can actively request the system to make recommendations.
  • the generation time of the recommendation request does not correspond to the preset recommendation time one-to-one. Therefore, the most recent recommendation time of each knowledge point A needs to be statistically obtained , So as to avoid the time interval of the same knowledge point A appearing in the recommended exercises is too long, which can not play the effect of making students review in time and reduce the efficiency of students.
  • the time interval is greater than or equal to the preset recommendation interval time, it means that the interval time of the knowledge point A in the recommended exercises is too long.
  • This recommendation will give priority to recommending the knowledge point A, with the first preset weight, for example, 0.7 Weight, select the exercises corresponding to knowledge point A from the exercise library;
  • the time interval is less than the preset recommendation interval time, it means that the interval time of the knowledge point A in the recommended exercises is within the control range, and the student may still have an impression of the knowledge point A.
  • This recommendation can give priority to other knowledge points A, use the second preset weight, such as a weight of 0.3, to select the exercises corresponding to knowledge point A from the exercise library.
  • different preset recommendation intervals different knowledge points A are selected in a targeted manner. Under the condition of a limited number of single recommended exercises, a large number of knowledge points A are arranged in a reasonable order to prevent students from forgetting. Review in time to improve the efficiency of doing questions.
  • this application also proposes an intelligent recommendation device for exercises, which mainly includes:
  • the obtaining module 10 is used to obtain the wrong question in the answer result, and search the corresponding knowledge point A in the preset knowledge point mapping table according to the wrong question;
  • the first calculation module 20 is configured to calculate the similarities between the multiple knowledge points in the exercise library and the knowledge point A according to the first preset formula, and obtain that all the similarities with the knowledge point A are greater than the first similarity
  • the knowledge point B with a degree threshold, and sort all the knowledge points B in the order of similarity from high to low to form a set of knowledge points B;
  • the second calculation module 30 is configured to obtain the exercise quality of at least the first exercise corresponding to the knowledge point A and the plurality of second exercises corresponding to the set of knowledge point B according to the second preset formula, and according to the exercise quality Sort the first exercises and the second exercises in descending order of quality to form a first exercise set and a second exercise set;
  • the extraction module 40 is configured to obtain a single recommended number of recommended exercises, and extract the first exercises from the first exercise set and the second exercises from the second exercise set according to a preset recommendation ratio, respectively,
  • the recommended exercises are composed; wherein, when extracting the first exercises and the second exercises, they are extracted in descending order of the quality of the exercises.
  • this application also proposes a computer device, including a memory 1003 and a processor 1002, the memory 1003 stores a computer program 1004, the processor 1002 executes the computer program 1004 to implement a method for intelligently recommending exercises, the exercises
  • the steps of the smart recommendation method include:
  • this application also proposes a computer storage medium 2001.
  • the computer-readable storage medium may be non-volatile or volatile.
  • a computer program 2002 is stored thereon, and the computer program 2002 is processed.
  • a method for intelligently recommending exercises is implemented when the device is executed, and the steps of the method for intelligently recommending exercises include:

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Tourism & Hospitality (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computational Linguistics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Educational Administration (AREA)
  • Educational Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Electrically Operated Instructional Devices (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一种习题智能推荐方法、装置、计算机设备及存储介质,涉及到人工智能领域。该方法先通过第一预设公式,根据每个学生的做题情况,计算出每个学生掌握到位的知识点以及可能掌握不到位的知识点;然后通过第二预设公式,进一步地计算出合适推荐给学生的习题,解决了目前的错题推荐中错题题目较为单一,以及错题推荐的质量针对性不强的问题。能够针对学生自身的作业情况和知识点掌握情况,自动生成个性化的推荐练习习题,使学生将时间花在最该练习的题目上,帮助学生高效学习,从而能够节约做题时间,达到个性化学习,提升学习效率的目的。

Description

习题智能推荐方法、装置、计算机设备及存储介质
本申请要求于2020年6月19日提交中国专利局、申请号为202010568262.3,发明名称为“习题智能推荐方法、装置、计算机设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及到人工智能领域,特别是涉及到一种习题智能推荐方法、装置、计算机设备及存储介质。
背景技术
发明人发现,在现有的错题推荐中,实际上都是错题回顾,一种是以一定的间隔频率将学生做错的题目不断的推送给学生进行重新解题,另一种就是根据相同的知识点进行随机推荐,题目变化性不够灵活、较为单一,且习题质量没有保证,针对性不强。还存在一个缺点就是,只根据做错的知识点进行重复推荐,没有根据相似知识点进行题目推荐,题目整体性不足。
技术问题
本申请的主要目的为提供一种习题智能推荐方法、装置、计算机设备及存储介质,针对学生自身的作业情况和知识点掌握情况,自动生成个性化的推荐练习习题。
技术解决方案
第一方面,本申请提出一种习题智能推荐方法,包括:
获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
第二方面,本申请还提出了一种习题智能推荐装置,包括:
获取模块,用于获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
第一计算模块,用于根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
第二计算模块,用于根据第二预设公式分别获取所述知识点A对应的至少一 个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
抽取模块,用于获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
第三方面,本申请还提出了一种计算机设备,包括存储器和处理器,存储器存储有计算机程序,处理器执行计算机程序时实现一种习题智能推荐方法,所述习题智能推荐方法的步骤包括:
获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
第四方面,本申请还提出了一种存储介质,其上存储有计算机程序,计算机程序被处理器执行时实现一种习题智能推荐方法,所述习题智能推荐方法的步骤包括:
获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
有益效果
本申请与现有技术相比,有益效果是:本申请提出了一种习题智能推荐方法、装置、计算机设备及存储介质,先通过第一预设公式,根据每个学生的做题情况,计算出每个学生掌握到位的知识点以及可能掌握不到位的知识点;然后通过第二预设公式,进一步地计算出合适推荐给学生的习题,解决了目前的错题推荐中错题题目较为单一,以及错题推荐的质量针对性不强的问题。能够针对学生自身的作业情况和知识点掌握情况,自动生成个性化的推荐 练习习题,使学生将时间花在最该练习的题目上,帮助学生高效学习,从而能够节约做题时间,达到个性化学习,提升学习效率的目的。
附图说明
图1为本申请一实施例中习题智能推荐方法的步骤示意图;
图2为本申请一实施例中习题智能推荐装置的模块示意图;
图3为本申请一实施例中计算机设备的模块示意框图;
图4为本申请一实施例中存储介质的模块示意框图。
本发明的最佳实施方式
参照图1,本申请在一实施例中提出了一种习题智能推荐方法,涉及人工智能领域,其中,方法主要包括以下步骤:
S1,获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
S2,根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
S3,根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
S4,获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
在上述步骤实施时,作业推荐系统首先获取学生,例如学生S,以往的答题结果,该答题结果由对学生提交的习题进行批改而得到的,并根据学生的答题结果确定该学生做错的题目,以便知晓该学生的知识薄弱点。在一个具体的实施例中,答题结果包括学生解答教师布置的习题而得到的结果。在另一个具体的实施例中,答题结果还包括学生解答系统的推荐习题而得到的结果。然后,为了进一步地的确定学生的知识薄弱点,根据得到的错题到知识点映射表中查找对应的知识点A,该知识点A至少有一个,在知识点映射表中预设了习题库中每一道习题对应的知识点,例如,唐诗、宋词、磁场力学、热力学、反证法、细胞结构等知识点,也即是该学生的未掌握知识点,将得到的所有错题进行分门别类,为后续的作业推荐提供更好的数据基础。
然后,系统根据第一预设公式计算习题库中的多个知识点分别与知识点A的相似度,得出所有与知识点A的相似度大于第一相似度阈值的知识点B,并按照相似度由高到低的顺序对所有知识点B进行排序,形成知识点B集;在一个具体的实施例中,第一预设公式为:
Figure PCTCN2020098828-appb-000001
其中,知识点A与知识点B在习题库中均用向量进行表示,x A和x B分别指的是知识点A与知识点B在习题库中的知识点特征值,每个知识点具有一个对应的知识点特征值,不同知识点之间的相似度与知识点特征值的差值呈负相关关系, 在一个具体的实施例中,该知识点特征值表现为一个多位数字,每个数字都代表该知识点的知识属性中的学科、类目、章、节、点等,例如知识点A的知识点特征值为545863,其中知识点特征值中各个数字代表的知识属性如表一所示。
表一
Figure PCTCN2020098828-appb-000002
在一个具体的实施例中,知识点特征值中的后一位数字根据前一位数字的不同而不同,例如在知识点特征值545863中,由于第一位数字5代表为化学学科,因此第二位数字4代表的是化学学科中的有机化学,但如果是在知识点特征值445863中,由于第一位数字5代表的是物理学科,则此时第二位数字4代表的是物理学科中的磁力学。在另一个具体的实施例中,知识点特征值中的每一位数字都具有一个特定对应的知识属性,后一位的数字并不根据前一位数字的变化而变化。在习题库中采用知识点特征值对知识点进行描述,不同知识点之间的相似度就会与知识点特征值的差值呈负相关关系,即是说若两个知识点之间的知识点特征值越接近,差值越小,则说明两个知识点之间的相似度就越高。
然后,y A和y B分别指的是知识点A与知识点B的知识点难度,y A和y B值越大,则表示知识点难度越高;在要求学生掌握知识点时,不同知识点之间的知识点难度是存在差别的,也就具有不同的要求,例如知识点难度可以分为熟练、识记、应用、理解等,分别采用不同的值来对应知识点难度,例如熟练的知识点难度值为1、识记的知识点难度值为2、应用的知识点难度值为3、理解的知识点难度值为4等。知识点的知识点特征值与知识点难度值共同构成一个二维向量,用于描述对应的知识点。通过计算知识点A与知识点B之间的余弦相似度,即可得出知识点A与知识点之间的相似度,其中余弦相似度的值为[-1,1],其中值越大则表明知识点A与知识点之间的相似度越高,当结果值为1时,说明知识点A与知识点B完全一样,当结果值为-1时,说明知识点A与知识点B完全不一样。通过余弦相似度,将与知识点A的相似度大于第一相似度阈值的知识点B全部筛选出来,并按照相似度由高到低的顺序对所有知识点B进行排序,形成知识点B集,便可得出习题库中与知识点A相似度较高的所有知识点B。
然后,由于习题库中的习题众多,因此在推荐习题给学生时还需要考虑优先推荐习题质量高的习题,因此还要分别获取知识点A对应的第一习题、知识点B集对应的第二习题的习题质量,并按照习题质量由高到低的顺序分别对第一习题和第二习题进行排序,形成第一习题集和第二习题集;
最后,系统获取推荐习题的单次推荐数量,该单次推荐数量可以为预设的固定数量,例如50题或者100题,也可以由学生或者教师进行自行设定,以契合每个学生的做题水平。然后根据预设推荐比例分别从第一习题集中抽取第一习题以及从第二习题集中抽取第二习题,组成推荐习题推荐给学生S;其中,抽取第一习题和第二习题时,按照习题质量由高到低的顺序进行抽取。在一个具体的实施例中,在推荐习题中,第一习题所占的推荐比例为60%。第二习题所占的推荐比例为40%。通过第一预设公式,能够根据每个学生的做题情况,计算出每个学生掌握到位的知识点以及可能掌握不到位的知识点;通过第二预设公式,进一步地计算出合适推荐给学生的习题,解决了目前的错题推荐中错题题目较为单一,以及错题推荐的质量针对性不强的问题。
在一个较优的实施例中,在所述获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A的步骤之后,还包括:
S11,分别统计每个所述知识点A在所述答题结果中对应的习题错误率;
S12,根据所述习题错误率在预设推荐表中进行查询,以确定在所述习题错误率下每个所述知识点A对应的推荐方案;其中,在所述预设推荐表中,不同的所述推荐方案对应不同范围的所述习题错误率,不同所述推荐方案所包含的推荐比例不同,所述推荐比例指的是位于一定范围所述习题错误率内的所述知识点A对应的习题在推荐习题中所占的比例;在所述预设推荐表中,所有所述推荐方案中的所述推荐比例之和小于等于1。
上述步骤实施时,在根据预设的知识点映射表确定了所有的知识点A之后,统计各个知识点A在答题结果中的习题错误率,在一个具体的实施例中,根据错题确定对应的知识点A之后,统计每个知识点A对应的错题总数以及在知识点A在答题结果中对应的习题总数,然后根据错题总数和习题总数即可计算出知识点A的习题错误率。
然后,根据习题错误率在预设推荐表中进行查询,以确定在该习题错误率下知识点A对应的推荐方案;其中,不同的推荐方案对应不同范围的习题错误率,不同推荐方案所包含的推荐比例不同,推荐比例指的是位于一定范围习题错误率内的知识点A对应的习题数目在推荐习题的总数目中所占的比例。在一个具体的实施例中,例如习题错误率位于80%-100%时对应的推荐方案所包含的推荐比例为35%,则在80%-100%的习题错误率下的知识点A所对应的习题数在推荐习题中所占的比例就为35%。值得注意的是,在预设推荐表中,所有推荐方案中的推荐比例之和小于等于1,以防出现在后续生成推荐习题时出现推荐比例之和大于1,导致推荐习题生成错误的问题出现。在一个具体的实施例中,推荐方案中的推荐比例包括了知识点A对应的第一习题以及知识点B集对应的第二习题,即第一习题和第二习题在推荐习题中所占的比例就为该推荐比例。
最后,根据在预设推荐表中查找到的推荐方案从习题库中挑选对应知识点A的习题以及习题库中的其余习题,并生成推荐习题推送给学生。在一个具体的实施例中,若根据在预设推荐表中查找到的推荐方案的推荐比例之和等于1,即该学生的知识点A的习题错误率覆盖范围较广,此时在推荐习题中挑选的则全部都是知识点A所对应的习题,在习题库中的其他习题的推荐比例为0,即不推荐其余习题。在另一个具体的实施例中,若根据在预设推荐表中查找到的推荐方案的推荐比例之和小于1,例如为80%,则从习题库中的其余习题中进行挑选占习题总数20%的数目进行补足,此时在推荐习题中既存在学生未掌握的习题,也存在学生已经掌握的习题。在推荐习题中混入习题库的其余习题具有两个好处:一是可以让学生进行适当的反复巩固练习,加深其掌握情况;二是在推荐习题中不至于全部都是学生未掌握的习题,题目难度适当,提高学生答题兴趣,从而优化做题效果。
在一个较优的实施例中,前述第二预设公式为:
Figure PCTCN2020098828-appb-000003
其中,Q z指的是习题z的习题质量,P z指的是习题z的习题平均评分,P zi指的是学生i对习题z的习题评分,C指的是习题z的习题收藏度,C max指的是 在所有习题中最高的收藏次数,C zf指的是习题z的收藏次数,n指的是学生i的个数。
前述的习题质量是根据第二预设公式计算得出,在一个具体的实施例中,系统采用1-10的评分制提供给学生或教师对习题进行评分,习题的评分越高,则说明习题的质量越好;若习题是新加入习题库中的,则默认给该习题一个中等评分,例如5分或者6分。另一方面,能够反映习题质量的除了评分还有收藏次数,习题的收藏次数越多则说明习题质量可能越好,C指的即是习题z的习题收藏度,C max指的是在所有习题中最高的收藏次数,C zf指的是习题z的收藏次数,为了消除由于特定习题的收藏次数过多或者过少而引起的习题质量评分偏高或偏低的现象发生,习题z的收藏次数除以习题最高的收藏次数之后,再乘以习题z的习题平均评分,将收藏次数与习题平均评分结合,以减少误差。在另一个具体的实施例中,习题平均评分与习题收藏度的统计权重一致。
在一个较优的实施例中,所述统计所有所述知识点A在所述答题结果中的习题错误率的步骤S11,包括:
S111,统计所述知识点A的各种题型在所述答题结果中对应的习题数目以及错题总数;
S112,根据所述习题数目以及所述错题总数,计算出所述知识点A的各种题型对应的所述习题错误率。
在上述步骤实施时,统计知识点A的各种题型在答题结果中对应的习题数目以及错题总数;在一个具体的实施例中,由于习题的题型较多,例如选择题、填空题、判断题、解答题等,不同的题型其应对的解答方法也就不太一样,因此即便是同一个学生在面对同样的知识点时,如果题型不同,其掌握程度也可能会不同,例如有人比较擅长做选择题,有人比较擅长做解答题等。因此具体统计出知识点A的各种题型在答题结果中对应的习题数目以及错题总数,以便后续对知识点A进行一个详细的分析。
最后,根据习题数目以及错题总数,计算出知识点A各种题型对应的习题错误率。统计出知识点A各种题型对应的习题错误率之后,可以更加清晰的了解到该学生对该知识点A的做题情况,也可以在后续的习题推荐中针对不同题型进行推荐。在一个具体的实施例中,例如根据习题数目以及错题总数,计算出同一个知识点A在选择题上的习题错误率为10%,在填空题上的习题错误率为40%,在解答题上的习题错误率为70%,则根据习题错误率不同在预设推荐表中可能查找到不同的推荐方案,从而具有不同的推荐比例。例如该知识点A对应的习题在选择题中的推荐比例为5%,在填空题中的推荐比例为15%,在解答题中的推荐比例为30%,从而针对同一个知识点A,在不同的题型中进行针对性推荐,习题错误率越高的题型则推荐比例也就越高,使学生能够针对自身的薄弱点进行更好的改进,避免对已掌握的题型进行过多重复。
在一个较优的实施例中,所述统计所有所述知识点A在所述答题结果中的习题错误率的步骤,包括:
S113,统计出各个所述知识点A在所有所述答题结果中的第一错误率,以及在最近预设次数提交的所述答题结果中的第二错误率;
S114,对所述第一错误率和所述第二错误率进行比较;
S115,若所述第一错误率大于等于所述第二错误率,则将所述第二错误率设定为所述习题错误率;
S116,若所述第一错误率小于所述第二错误率,则将所述第一错误率设定为 所述习题错误率。
在上述步骤实施时,首先统计出各个知识点A在所有答题结果中的第一错误率,以及在最近预设次数提交的答题结果中的第二错误率;即是统计出各个知识点A在所有的答题结果中的错误率,以及在近期提交的,例如最近五次的答题结果中的错误率,以进行比较。
然后,对第一错误率和第二错误率进行比较;若第一错误率大于等于第二错误率,则说明该知识点A在所有的答题结果中的错误率要高于在近期的答题结果中的错误率,也即是说明该学生在做该知识点A的习题时,前期错误率较高,近期错误率较低,对知识点A可能已经较好的掌握了,能够降低答题的错误率,因此将第二错误率设定为习题错误率,及时反映学生的进步情况,避免前期错误率影响习题推荐方案。若第一错误率小于第二错误率,则说明该知识点A在所有的答题结果中的错误率要低于在近期的答题结果中的错误率,也即是说明该学生在做该知识点A的习题时,前期错误率较低,近期错误率较高,近期有可能出现遗忘知识点或者做题时粗心大意的情况,因此将第一错误率设定为习题错误率,从对该学生的整体情况出发,更好的把握学生做题情况,避免近期出现的异常情况影响到习题推荐方案。
在一个较优的实施例中,所述统计出各个所述知识点A在所有所述答题结果中的第一错误率,以及统计出在最近预设次数提交的所述答题结果中的第二错误率的步骤,包括:
S1131,统计出各个所述知识点A在所有所述答题结果中的第一错误率,并判断所述知识点A对应的所述错题的数量是否达到预设数量;
S1132,若达到所述预设数量,则统计出所述知识点A在最近预设次数提交的所述答题结果中的第二错误率;
S1133,若未达到所述预设数量,则将所述第一错误率设定为所述习题错误率。
在上述步骤实施时,首先,统计出各个知识点A在所有答题结果中的第一错误率,并判断知识点A对应的错题的数量是否达到预设数量;
然后,若达到预设数量,则统计出知识点A在最近预设次数提交的答题结果中的第二错误率;因为若要将第一错误率与第二错误率进行比较,从而判断出学生针对同一个知识点A的整体掌握情况以及近期掌握情况,就需要有足够多的样本数量将整体情况与近期情况进行区分,若样本数量太少,则第二错误率实际上并不能够反映出学生在近期的学习做题情况。在一个具体的实施例中,若知识点A对应的错题的数量未达到预设数量,则直接将第一错误率设定为习题错误率,而不用计算第二错误率。
在一个较优的实施例中,在所述若达到所述预设数量,则统计出所述知识点A在最近预设次数提交的所述答题结果中的第二错误率的步骤之后,还包括:
S1134,若所述第一错误率与所述第二错误率的差值位于预设差值范围内,则将所述知识点A发送到预设管理界面中。
在上述步骤实施时,若根据计算,获知第一错误率与第二错误率的差值位于预设差值范围内,在一个具体的实施例中,预设差值范围在5%以内,则说明该学生在前期与近期针对该知识点A的做题错误率都差不多,一直没能完全掌握该知识点A,很有可能是对该知识点理解不到位,导致在做题错误率上一直没有改善,则将知识点A发送到预设管理界面中,例如教师的关注管理页面中,使得老师可以及时了解到学生一直以来存在的知识薄弱点,从而针对知识盲区进行针对 性辅导,改善学生做题效率。
在一个较优的实施例中,所述根据所述习题错误率在预设推荐表中进行查询,以确定在所述习题错误率下每个所述知识点A对应的推荐方案的步骤,包括:
S121,若所述习题错误率大于等于第一预设阈值,则停止推荐对应的所述知识点A,并将所述知识点A推送至预设管理界面中;
S122,若所述习题错误率大于等于第二预设阈值,且小于所述第一预设阈值,则按照第一预设推荐比例推荐所述知识点A,并将所述知识点A推送至所述预设管理界面中;
S123,若所述习题错误率小于所述第二预设阈值,则按照第二预设推荐比例推荐所述知识点A。
在上述步骤实施时,在预设推荐表中,不同的推荐方案对应不同范围的习题错误率,不同推荐方案所包含的推荐比例不同,推荐比例指的是位于一定范围习题错误率内的知识点A对应的习题在推荐习题中所占的比例。若习题错误率大于等于第一预设阈值,例如80%,此时习题错误率较高,若直接将习题错误率大于80%的知识点A对应的习题推荐给学生,一是由于该学生并没有很好的掌握该知识点A,对于该习题没有很好地解决办法,继续接着做题就犹如盲人摸象,也并不能得到很好的做题方法,对于做题错误率并不能得到很好地改善,只会一直高居不下;二是如果让学生重复做错误率过高的习题,容易打击学生的解题自信,降低学生答题兴趣和效率;因此对于习题错误率过高的知识点A,则停止推荐对应的知识点A,并将知识点A推送至教师的预设管理界面中;然后待教师在后续进行针对性辅导以及布置习题之后,若系统根据最新的答题结果计算出学生的重复错误率低于第一预设阈值,说明该学生对于该知识点A的掌握程度有所加深,此时可以进行习题推荐,以便进一步地进行针对性改进。
若习题错误率大于等于第二预设阈值,且小于第一预设阈值,例如大于等于25%,小于80%,在实际的答题过程中,习题错误率位于这一范围内的知识点A较多,也是学生需要重点进行攻坚的的部分,此时则按照第一预设推荐比例,例如70%,推荐知识点A对应的习题,并由于该习题错误率仍旧较高,因此将知识点A推送至教师的预设管理界面中,以便教师对学生的答题情况有一个清晰的了解。
若习题错误率小于第二预设阈值,例如小于10%,此时习题错误率较低,说明该学生很可能只是粗心大意将该习题做错,或者只是暂时性的有些遗忘知识而导致做错,则按照第二预设推荐比例,例如10%,推荐知识点A。针对习题错误率较低的知识点A,以一个较低的推荐比例进行推荐,尽量增大推荐习题中习题错误率较高的知识点A,降低偶发错误对推荐习题产生的影响。在具体的实施过程中,预设阈值并不局限于第一预设阈值和第二预设阈值,可以根据实际操作情况设定一个或多个不同的预设阈值,从而针对习题错误率不同的知识点A,执行不同的推荐方案,类似的实施例都处于本申请的保护范围之内。
在一个较优的实施例中,所述根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题的步骤,包括:
S41,获取各个所述知识点A的最近一次推荐时间,并判断所述最近一次推荐时间与当前时间之间的时间间隔是否大于等于预设推荐间隔时间;
S42,若大于等于所述预设推荐间隔时间,则以第一预设权重从所述习题库中挑选所述知识点A对应的习题;
S43,若小于所述预设推荐间隔时间,则以第二预设权重从所述习题库中挑选所述知识点A对应的习题。
在上述步骤实施时,在推荐方案还包括了推荐间隔时间,推荐间隔时间为预设,在一个具体的实施例中,预设推荐间隔时间根据艾宾浩斯遗忘曲线进行设置,如一天、七天、十五天,即系统在第二日、第七日和第十五日从习题库中选取知识点A对应的习题进行推荐。在另一个具体的实施例中,该预设推荐时间也可以自行设置。
然后,获取各个知识点A的最近一次推荐时间,即该知识点A出现在最近一次的推荐习题中的时间,并判断最近一次推荐时间与当前时间之间的时间间隔是否大于等于预设推荐间隔时间;由于学生的知识点A可能较多,而单次的推荐习题的习题数量有限,所以在单次的推荐习题中可能不会覆盖到所有的知识点A,且在获取推荐习题时,既可以由系统进行自动推荐,还可以由学生或者老师主动请求系统进行推荐,推荐请求的生成时间与预设推荐时间并不一一对应,因此需要对各个知识点A的最近一次推荐时间进行统计获取,以免在推荐习题中同一知识点A出现的间隔时间过长,无法起到使学生及时复习的效果,降低了学生的做题效率。
若时间间隔大于等于预设推荐间隔时间,则说明该知识点A在推荐习题中出现的间隔时间过长了,此次推荐要优先推荐该知识点A,以第一预设权重,例如0.7的权重,从习题库中挑选知识点A对应的习题;
若时间间隔小于预设推荐间隔时间,则说明该知识点A在推荐习题中出现的间隔时间在控制范围之内,学生可能还对该知识点A有印象,此次推荐可以优先推荐其他知识点A,则以第二预设权重,例如0.3的权重,从习题库中挑选知识点A对应的习题。通过不同的预设推荐间隔时间,对不同的知识点A进行针对性挑选,在单次推荐习题数量有限的条件下,对数量众多的知识点A进行合理的推荐顺序安排,从而防止学生遗忘,及时复习,提高做题效率。
参照图2,本申请还提出了一种习题智能推荐装置,主要包括:
获取模块10,用于获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
第一计算模块20,用于根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
第二计算模块30,用于根据第二预设公式分别获取所述知识点A对应的至少第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
抽取模块40,用于获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
其中上述模块10-40分别用于执行的操作与前述实施方式的习题智能推荐方法的步骤一一对应,在此不再赘述。
进一步地,对应前述实施方式的习题智能推荐方法的细分步骤,上述模块10-40相应的包含了子模块、单元或子单元,用于执行前述习题智能推荐方法的 细分步骤,在此也不再赘述。
参照图3,本申请还提出了一种计算机设备,包括存储器1003和处理器1002,存储器1003存储有计算机程序1004,,处理器1002执行计算机程序1004时实现一种习题智能推荐方法,所述习题智能推荐方法的步骤包括:
获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
参照图4,本申请还提出了一种计算机存储介质2001,所述计算机可读存储介质可以是非易失性,也可以是易失性,其上存储有计算机程序2002,,计算机程序2002被处理器执行时实现一种习题智能推荐方法,所述习题智能推荐方法的步骤包括:
获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
以上仅为本申请的优选实施例,并非因此限制本申请的专利范围,凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换,或直接或间接运用在其他相关的技术领域,均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种习题智能推荐方法,其中,包括:
    获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
    根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
    根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
    获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
  2. 根据权利要求1所述的习题智能推荐方法,其中,在所述获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A的步骤之后,还包括:
    分别统计每个所述知识点A在所述答题结果中对应的习题错误率;
    根据所述习题错误率在预设推荐表中进行查询,以确定在所述习题错误率下每个所述知识点A对应的推荐方案;其中,在所述预设推荐表中,不同的所述推荐方案对应不同范围的所述习题错误率,不同所述推荐方案所包含的推荐比例不同,所述推荐比例指的是位于一定范围所述习题错误率内的所述知识点A对应的习题在推荐习题中所占的比例;在所述预设推荐表中,所有所述推荐方案中的所述推荐比例之和小于等于1。
  3. 根据权利要求1所述的习题智能推荐方法,其中,所述第二预设公式为:
    Figure PCTCN2020098828-appb-100001
    其中,所述Q z指的是习题z的所述习题质量,所述P z指的是习题z的习题平均评分,所述P zi指的是学生i对习题z的习题评分,所述C指的是习题z的习题收藏度,所述C max指的是在所有习题中最高的收藏次数,所述C zf指的是习题z的所述收藏次数,所述n指的是学生i的个数。
  4. 根据权利要求2所述的习题智能推荐方法,其中,所述统计所有所述知识点A在所述答题结果中的习题错误率的步骤,包括:
    统计出各个所述知识点A在所有所述答题结果中的第一错误率,以及在最近预设次数提交的所述答题结果中的第二错误率;
    对所述第一错误率和所述第二错误率进行比较;
    若所述第一错误率大于等于所述第二错误率,则将所述第二错误率设定为所述习题错误率;
    若所述第一错误率小于所述第二错误率,则将所述第一错误率设定为所述习题错误率。
  5. 根据权利要求4所述的习题智能推荐方法,其中,所述统计出各个所述 知识点A在所有所述答题结果中的第一错误率,以及统计出在最近预设次数提交的所述答题结果中的第二错误率的步骤,包括:
    统计出各个所述知识点A在所有所述答题结果中的第一错误率,并判断所述知识点A对应的所述错题的数量是否达到预设数量;
    若达到所述预设数量,则统计出所述知识点A在最近预设次数提交的所述答题结果中的第二错误率;
    若未达到所述预设数量,则将所述第一错误率设定为所述习题错误率。
  6. 根据权利要求2所述的习题智能推荐方法,其中,所述根据所述习题错误率在预设推荐表中进行查询,以确定在所述习题错误率下每个所述知识点A对应的推荐方案的步骤,包括:
    若所述习题错误率大于等于第一预设阈值,则停止推荐对应的所述知识点A,并将所述知识点A推送至预设管理界面中;
    若所述习题错误率大于等于第二预设阈值,且小于所述第一预设阈值,则按照第一预设推荐比例推荐所述知识点A,并将所述知识点A推送至所述预设管理界面中;
    若所述习题错误率小于所述第二预设阈值,则按照第二预设推荐比例推荐所述知识点A。
  7. 根据权利要求1所述的习题智能推荐方法,其中,所述根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题的步骤,包括:
    获取各个所述知识点A的最近一次推荐时间,并判断所述最近一次推荐时间与当前时间之间的时间间隔是否大于等于预设推荐间隔时间;
    若大于等于所述预设推荐间隔时间,则以第一预设权重从所述习题库中挑选所述知识点A对应的习题;
    若小于所述预设推荐间隔时间,则以第二预设权重从所述习题库中挑选所述知识点A对应的习题。
  8. 一种习题智能推荐装置,其中,包括:
    获取模块,用于获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
    第一计算模块,用于根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
    第二计算模块,用于根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
    抽取模块,用于获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
  9. 一种计算机设备,包括存储器和处理器,所述存储器存储有计算机程序,其中,所述处理器执行所述计算机程序时实现一种习题智能推荐方法,所述习题智能推荐方法的步骤包括:
    获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
    根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
    根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
    获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
  10. 根据权利要求9所述的计算机设备,其中,在所述获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A的步骤之后,还包括:
    分别统计每个所述知识点A在所述答题结果中对应的习题错误率;
    根据所述习题错误率在预设推荐表中进行查询,以确定在所述习题错误率下每个所述知识点A对应的推荐方案;其中,在所述预设推荐表中,不同的所述推荐方案对应不同范围的所述习题错误率,不同所述推荐方案所包含的推荐比例不同,所述推荐比例指的是位于一定范围所述习题错误率内的所述知识点A对应的习题在推荐习题中所占的比例;在所述预设推荐表中,所有所述推荐方案中的所述推荐比例之和小于等于1。
  11. 根据权利要求9所述的计算机设备,其中,所述第二预设公式为:
    Figure PCTCN2020098828-appb-100002
    其中,所述Q z指的是习题z的所述习题质量,所述P z指的是习题z的习题平均评分,所述P zi指的是学生i对习题z的习题评分,所述C指的是习题z的习题收藏度,所述C max指的是在所有习题中最高的收藏次数,所述C zf指的是习题z的所述收藏次数,所述n指的是学生i的个数。
  12. 根据权利要求10所述的计算机设备,其中,所述统计所有所述知识点A在所述答题结果中的习题错误率的步骤,包括:
    统计出各个所述知识点A在所有所述答题结果中的第一错误率,以及在最近预设次数提交的所述答题结果中的第二错误率;
    对所述第一错误率和所述第二错误率进行比较;
    若所述第一错误率大于等于所述第二错误率,则将所述第二错误率设定为所述习题错误率;
    若所述第一错误率小于所述第二错误率,则将所述第一错误率设定为所述习题错误率。
  13. 根据权利要求12所述的计算机设备,其中,所述统计出各个所述知识点A在所有所述答题结果中的第一错误率,以及统计出在最近预设次数提交的所述答题结果中的第二错误率的步骤,包括:
    统计出各个所述知识点A在所有所述答题结果中的第一错误率,并判断所述 知识点A对应的所述错题的数量是否达到预设数量;
    若达到所述预设数量,则统计出所述知识点A在最近预设次数提交的所述答题结果中的第二错误率;
    若未达到所述预设数量,则将所述第一错误率设定为所述习题错误率。
  14. 根据权利要求10所述的计算机设备,其中,所述根据所述习题错误率在预设推荐表中进行查询,以确定在所述习题错误率下每个所述知识点A对应的推荐方案的步骤,包括:
    若所述习题错误率大于等于第一预设阈值,则停止推荐对应的所述知识点A,并将所述知识点A推送至预设管理界面中;
    若所述习题错误率大于等于第二预设阈值,且小于所述第一预设阈值,则按照第一预设推荐比例推荐所述知识点A,并将所述知识点A推送至所述预设管理界面中;
    若所述习题错误率小于所述第二预设阈值,则按照第二预设推荐比例推荐所述知识点A。
  15. 一种存储介质,其上存储有计算机程序,其中,所述计算机程序被处理器执行时实现一种习题智能推荐方法,所述习题智能推荐方法的步骤包括:
    获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A;
    根据第一预设公式计算习题库中的多个知识点分别与所述知识点A的相似度,得出所有与所述知识点A的相似度大于第一相似度阈值的所述知识点B,并按照相似度由高到低的顺序对所有所述知识点B进行排序,形成知识点B集;
    根据第二预设公式分别获取所述知识点A对应的至少一个第一习题、所述知识点B集对应的多个第二习题的习题质量,并按照所述习题质量由高到低的顺序分别对所述第一习题和所述第二习题进行排序,形成第一习题集和第二习题集;
    获取推荐习题的单次推荐数量,并根据预设推荐比例分别从所述第一习题集中抽取所述第一习题以及从所述第二习题集中抽取所述第二习题,组成所述推荐习题;其中,抽取所述第一习题和所述第二习题时,按照所述习题质量由高到低的顺序进行抽取。
  16. 根据权利要求15所述的存储介质,其中,在所述获取答题结果中的错题,并根据所述错题在预设的知识点映射表中查找对应的知识点A的步骤之后,还包括:
    分别统计每个所述知识点A在所述答题结果中对应的习题错误率;
    根据所述习题错误率在预设推荐表中进行查询,以确定在所述习题错误率下每个所述知识点A对应的推荐方案;其中,在所述预设推荐表中,不同的所述推荐方案对应不同范围的所述习题错误率,不同所述推荐方案所包含的推荐比例不同,所述推荐比例指的是位于一定范围所述习题错误率内的所述知识点A对应的习题在推荐习题中所占的比例;在所述预设推荐表中,所有所述推荐方案中的所述推荐比例之和小于等于1。
  17. 根据权利要求15所述的存储介质,其中,所述第二预设公式为:
    Figure PCTCN2020098828-appb-100003
    其中,所述Q z指的是习题z的所述习题质量,所述P z指的是习题z的习题 平均评分,所述P zi指的是学生i对习题z的习题评分,所述C指的是习题z的习题收藏度,所述C max指的是在所有习题中最高的收藏次数,所述C zf指的是习题z的所述收藏次数,所述n指的是学生i的个数。
  18. 根据权利要求16所述的存储介质,其中,所述统计所有所述知识点A在所述答题结果中的习题错误率的步骤,包括:
    统计出各个所述知识点A在所有所述答题结果中的第一错误率,以及在最近预设次数提交的所述答题结果中的第二错误率;
    对所述第一错误率和所述第二错误率进行比较;
    若所述第一错误率大于等于所述第二错误率,则将所述第二错误率设定为所述习题错误率;
    若所述第一错误率小于所述第二错误率,则将所述第一错误率设定为所述习题错误率。
  19. 根据权利要求18所述的存储介质,其中,所述统计出各个所述知识点A在所有所述答题结果中的第一错误率,以及统计出在最近预设次数提交的所述答题结果中的第二错误率的步骤,包括:
    统计出各个所述知识点A在所有所述答题结果中的第一错误率,并判断所述知识点A对应的所述错题的数量是否达到预设数量;
    若达到所述预设数量,则统计出所述知识点A在最近预设次数提交的所述答题结果中的第二错误率;
    若未达到所述预设数量,则将所述第一错误率设定为所述习题错误率。
  20. 根据权利要求16所述的存储介质,其中,所述根据所述习题错误率在预设推荐表中进行查询,以确定在所述习题错误率下每个所述知识点A对应的推荐方案的步骤,包括:
    若所述习题错误率大于等于第一预设阈值,则停止推荐对应的所述知识点A,并将所述知识点A推送至预设管理界面中;
    若所述习题错误率大于等于第二预设阈值,且小于所述第一预设阈值,则按照第一预设推荐比例推荐所述知识点A,并将所述知识点A推送至所述预设管理界面中;
    若所述习题错误率小于所述第二预设阈值,则按照第二预设推荐比例推荐所述知识点A。
PCT/CN2020/098828 2020-06-19 2020-06-29 习题智能推荐方法、装置、计算机设备及存储介质 WO2021253480A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010568262.3A CN111723193A (zh) 2020-06-19 2020-06-19 习题智能推荐方法、装置、计算机设备及存储介质
CN202010568262.3 2020-06-19

Publications (1)

Publication Number Publication Date
WO2021253480A1 true WO2021253480A1 (zh) 2021-12-23

Family

ID=72568211

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/098828 WO2021253480A1 (zh) 2020-06-19 2020-06-29 习题智能推荐方法、装置、计算机设备及存储介质

Country Status (2)

Country Link
CN (1) CN111723193A (zh)
WO (1) WO2021253480A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114885216A (zh) * 2022-04-26 2022-08-09 深圳市企鹅网络科技有限公司 习题推送方法、系统、电子设备和存储介质
CN115935032A (zh) * 2022-12-29 2023-04-07 北京十六进制科技有限公司 一种基于智能推荐习题发布个性化作业的方法及装置
CN116450801A (zh) * 2023-03-29 2023-07-18 北京思明启创科技有限公司 编程学习方法、装置、设备和存储介质
CN117648934A (zh) * 2024-01-30 2024-03-05 青岛培诺教育科技股份有限公司 基于错误试题的知识点确定方法、装置、设备和介质

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112380429A (zh) * 2020-11-10 2021-02-19 武汉天有科技有限公司 一种习题推荐方法及装置
CN112347366B (zh) * 2020-12-04 2022-07-08 华侨大学 基于学习者画像与习题相似度的预科中文习题推送方法
CN112699298A (zh) * 2020-12-28 2021-04-23 科大讯飞股份有限公司 一种试题推荐方法、电子设备及存储装置
CN113934840B (zh) * 2021-11-01 2022-04-01 东北师范大学 一种结合覆盖启发式的数量感知练习推荐方法
CN114676334A (zh) * 2022-04-19 2022-06-28 江苏迈拓网络科技有限公司 一种个性化的智能习题推荐方法及推荐系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106599054A (zh) * 2016-11-16 2017-04-26 福建天泉教育科技有限公司 一种题目分类及推送的方法及系统
CN107273490A (zh) * 2017-06-14 2017-10-20 北京工业大学 一种基于知识图谱的组合错题推荐方法
CN103870463B (zh) * 2012-12-10 2018-02-23 中国电信股份有限公司 测试题目的选择方法与系统
US10140880B2 (en) * 2015-07-10 2018-11-27 Fujitsu Limited Ranking of segments of learning materials
CN109299380A (zh) * 2018-10-30 2019-02-01 浙江工商大学 在线教育平台中基于多维特征的习题个性化推荐方法
CN109635100A (zh) * 2018-12-24 2019-04-16 上海仁静信息技术有限公司 一种相似题目的推荐方法、装置、电子设备及存储介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103870463B (zh) * 2012-12-10 2018-02-23 中国电信股份有限公司 测试题目的选择方法与系统
US10140880B2 (en) * 2015-07-10 2018-11-27 Fujitsu Limited Ranking of segments of learning materials
CN106599054A (zh) * 2016-11-16 2017-04-26 福建天泉教育科技有限公司 一种题目分类及推送的方法及系统
CN107273490A (zh) * 2017-06-14 2017-10-20 北京工业大学 一种基于知识图谱的组合错题推荐方法
CN109299380A (zh) * 2018-10-30 2019-02-01 浙江工商大学 在线教育平台中基于多维特征的习题个性化推荐方法
CN109635100A (zh) * 2018-12-24 2019-04-16 上海仁静信息技术有限公司 一种相似题目的推荐方法、装置、电子设备及存储介质

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114885216A (zh) * 2022-04-26 2022-08-09 深圳市企鹅网络科技有限公司 习题推送方法、系统、电子设备和存储介质
CN114885216B (zh) * 2022-04-26 2024-03-19 深圳市企鹅网络科技有限公司 习题推送方法、系统、电子设备和存储介质
CN115935032A (zh) * 2022-12-29 2023-04-07 北京十六进制科技有限公司 一种基于智能推荐习题发布个性化作业的方法及装置
CN116450801A (zh) * 2023-03-29 2023-07-18 北京思明启创科技有限公司 编程学习方法、装置、设备和存储介质
CN117648934A (zh) * 2024-01-30 2024-03-05 青岛培诺教育科技股份有限公司 基于错误试题的知识点确定方法、装置、设备和介质
CN117648934B (zh) * 2024-01-30 2024-04-26 青岛培诺教育科技股份有限公司 基于错误试题的知识点确定方法、装置、设备和介质

Also Published As

Publication number Publication date
CN111723193A (zh) 2020-09-29

Similar Documents

Publication Publication Date Title
WO2021253480A1 (zh) 习题智能推荐方法、装置、计算机设备及存储介质
Theobald Students are rarely independent: When, why, and how to use random effects in discipline-based education research
CN110378818B (zh) 基于难度的个性化习题推荐方法、系统及介质
CN107230174B (zh) 一种基于网络的在线互动学习系统和方法
CN109903617A (zh) 个性化练习方法和系统
CN112184500A (zh) 基于深度学习和知识图谱的课外学习辅导系统及实现方法
CN110399558B (zh) 一种试题推荐方法和系统
CN107248019A (zh) 一种云教学平台在线教学评价系统
Wang et al. Leveraging First Response Time into the Knowledge Tracing Model.
CN111858906A (zh) 习题推荐方法、装置、电子设备及计算机可读存储介质
CN112348725A (zh) 基于大数据的知识点难度定级方法
US20170330133A1 (en) Organizing training sequences
Hardegree Standards-based assessment and high stakes testing: Accuracy of standards-based grading
CN112988844B (zh) 一种基于学生练习序列的知识概念表示学习方法
US20170221163A1 (en) Create a heterogeneous learner group
JP2010243662A (ja) リメディアル教育支援システム、リメディアル教育支援方法、およびメディアル教育支援プログラム
Madsen et al. Best practices for administering attitudes and beliefs surveys in physics
Bernard et al. Integrating research into instructional practice: The use and abuse of meta-analysis
CN114004499A (zh) 一种教学作业的时长计算终端
Bendjebar et al. An improvement of a data mining technique for early detection of at-risk learners in distance learning environments
Chamorro-Atalaya et al. Supervised learning through classification learner techniques for the predictive system of personal and social attitudes of engineering students
Nurjanah et al. Homogeneous group formation in collaborative learning using fuzzy C-means
Arafiyah et al. Monitoring online learners’ performance based on learning progress prediction
Feng et al. Applying learning analytics to support instruction
Seltzer et al. Multilevel analysis

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20941498

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20941498

Country of ref document: EP

Kind code of ref document: A1