Student ability detection platform for education and teaching

Abstract

The invention discloses a student ability detection platform for education and teaching, belongs to the technical field of low-pressure casting machines, and aims to solve the problem that the existing student ability test cannot perform personalized accurate test according to the actual mastering condition of each chapter unit of a student. Through the test question bank storage module, by means of subject unit classification, knowledge point labeling association and question type difficulty management and control, a rich question type and difficulty system is combined, a teacher can conveniently and accurately select questions and form test papers, the purpose of providing scientific and reasonable test content for teaching is achieved, the historical data storage module keeps comprehensive score data and performs deep analysis, and the test efficiency is improved. The test question extraction module defines a state space, an action space and a reward function based on a reinforcement learning algorithm and according to historical answer data of students, and continuously optimizes a test question selection strategy, so that test questions closely meet actual demands of the students, and the effect of accurately detecting knowledge mastering conditions of each chapter unit of the students is achieved. And education and teaching quality improvement is effectively promoted.

Description

Technical Field

[0001] The invention relates to the technical field of low-pressure casting machines, and in particular to a student ability testing platform for education and teaching. Background Art

[0002] In the field of modern education and teaching, the student ability testing system is a key link in measuring students' learning outcomes and guiding teaching practice. However, the existing student ability testing system faces many difficulties, especially in using reinforcement learning technology to achieve personalized and accurate testing.

[0003] Most traditional testing systems rely on fixed testing modes, and lack flexibility in test content and difficulty settings. They fail to fully tap the individual differences of students in learning each chapter and unit, and cannot customize test content according to students' specific weaknesses. For example, in a class, different students have different levels of knowledge mastery in different chapters and units, but existing systems often provide unified test papers and cannot effectively detect each student's unique weaknesses. This makes it difficult for students to accurately identify their own learning problems, affecting their learning efficiency and self-confidence. For teachers, due to the lack of accurate grasp of students' weaknesses in each chapter and unit, it is difficult to implement personalized teaching strategies, which can easily lead to a waste of teaching resources and fail to achieve teaching according to students' aptitude. Summary of the invention

[0004] The purpose of the present invention is to provide a student ability testing platform for education and teaching, which solves the problem that the existing student ability test in the background technology cannot conduct personalized and accurate testing based on the students' actual mastery of each chapter unit.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a student ability testing platform for education and teaching, characterized in that it includes: a user management module, a question bank storage module, a historical data storage module, a test question extraction module, an answering module, a scoring module and a feedback report module.

[0006] Furthermore, the user management module includes a student information management submodule and a teacher information management submodule. The student information management submodule records in detail the student's name, gender, date of birth, student ID number and grade and class information. The teacher information management submodule comprehensively stores the teacher's name, contact information, subjects taught, teaching experience, professional qualification certificate number, teaching expertise and details of past teaching achievement evaluations.

[0007] Furthermore, the question bank storage module is used to store test questions for each subject and corresponding chapter unit, and the question bank storage module includes a subject unit classification submodule, a knowledge point annotation association submodule and a question type difficulty control submodule; the subject unit classification submodule accurately classifies the test questions into each subject and corresponding chapter unit according to the education syllabus and textbook system; the knowledge point annotation association submodule deeply annotates the core and secondary knowledge points and complex association graphs for each test question through knowledge graph technology; the question type difficulty control submodule includes multiple-choice questions, fill-in-the-blank questions, short-answer questions, essay questions, experimental operation questions, case analysis questions and creative design questions, and divides the questions into a 1-5 level difficulty system, and is equipped with detailed analysis documents to facilitate teachers to accurately select questions and assemble papers according to teaching objectives and student levels.

[0008] Furthermore, the historical data storage module is used to retain the comprehensive score data of students in previous tests, including the total score, the score of each question type and the score details of each knowledge point; the historical data storage module is also used to generate a visual score trend line chart and a score fluctuation chart, and to perform in-depth data analysis. The in-depth data analysis includes calculating the standard deviation and variance of the score to evaluate the stability of the score. The in-depth data analysis also includes calculating the coefficient of variation of the score. The formula is σ is the standard deviation and μ is the mean.

[0009] Furthermore, the test question extraction module uses an adaptive test question generation algorithm based on reinforcement learning to intelligently extract test questions, specifically including:

[0010] S1. Define the state space, action space and reward function of the reinforcement learning algorithm by analyzing students' historical answer data. The state space S is based on the accuracy of students' historical answer in the current chapter unit. histas the state; action space A: select the next test question from the test question bank, the action is determined by the chapter unit CU to which the question belongs, expressed as a = CU; reward function R: if the student answers the selected question correctly, the reward R = 1, if the answer is wrong, R = -1; a simple action value estimation algorithm is used to determine the test question selection strategy, and the action value estimation value Q(CU) = 0 is initialized for each chapter unit. For each state, that is, the answer accuracy rate of the current chapter unit, the action value estimation value is updated according to the following formula: Q(CU)←Q(CU)+α×(RQ(CU)), where α is the learning rate, and the value range is is 0-1. Each time a test question is selected, a chapter unit question is randomly selected with a certain probability ε, and the chapter unit question with the highest action value estimate Q(CU) is selected with probability 1-ε. When the student completes a certain number of questions, the test question is selected according to the learned action value estimate, that is, the chapter unit question with the highest Q(CU) value is selected as the next test question. In this way, the action value estimate is continuously updated according to the student's answering situation, so that the subsequently generated test questions can better target the student's actual mastery of each chapter unit, and adapt to the dynamic changes in the student's answering process.

[0011] S2. Scientifically set the initial standard weight vector of each unit according to the teaching syllabus W0 = [w 01 ,w 02 ,…w 0n ], calculate the adjustment factor Then through the formula Accurately calculate the adjusted unit weight, is the average of the difficulty-weighted scores of all students in the class, β is the adjustment intensity coefficient, ranging from 0 to 1, and after adjustment, For recurrent neural network models;

[0012] S3.Accurately determine the number of questions for each unit based on the adjusted unit weights Q is the total number of questions in the test paper. According to the principles of knowledge point coverage priority, weak knowledge point reinforcement principle and difficulty adaptation, questions are carefully selected from the question bank to form the test paper. The knowledge point coverage priority principle is to give priority to the questions corresponding to the key knowledge points of each unit; the weak knowledge point reinforcement principle lists the knowledge points with a score rate lower than the threshold θ1 as the key reinforcement area, with an extraction ratio of p1, and the score rate between θ1 and θ2 is the consolidation and improvement area, with an extraction ratio of p2, and the score rate higher than θ2 is the expansion and extension area, with an extraction ratio of p3=1-p1-p2, θ1 takes the value of 0.6, p1 takes the value of 0.7, θ2 takes the value of 0.8, and p2 takes the value of 0.2; the difficulty adaptation principle is based on the average score rate SR of students' history answers on low, medium and high difficulty questions L , SR M , SR H Determine the proportion of low, medium, and high difficulty questions in each unit.L ≥θ3, and SR M ≥θ4, θ3 is 0.8, θ4 is 0.6, then the proportion of low, medium and high difficulty questions is r L :r M :r H =0.3:0.5:0.2, if not satisfied, adjust to r L :r M :r H =0.4:0.4:0.2.

[0013] Furthermore, the answering module is used to provide an answering interactive interface on computers, tablets and mobile terminal devices. The interactive interface of the answering module displays the answering progress and intelligent reminders. The answering progress includes the number of questions answered, the number of remaining questions, the answering time progress bar and the estimated remaining time.

[0014] Furthermore, the scoring module includes an intelligent scoring submodule and a score statistics report generation submodule. The intelligent scoring submodule is used to score and summarize students' answers. The score statistics report generation submodule comprehensively summarizes the scoring data to generate a detailed transcript. The intelligent scoring submodule uses an answer comparison algorithm for objective questions and uses natural language processing technology combined with expert experience rules to score subjective questions, specifically including precise keyword extraction and matching, deep semantic logic analysis, and grammatical error detection and deduction, and a deep learning model to improve scoring accuracy; the score statistics report generation submodule comprehensively summarizes the scoring data to generate a detailed transcript, including the total score, score for each part, score rate, ranking percentile, and comparison with the previous test score, and generates a visual analysis report.

[0015] Compared with the prior art, the present invention has the following beneficial effects:

[0016] The present invention provides a student ability testing platform for education and teaching. Through the test question bank storage module, with the help of subject unit classification, knowledge point annotation association and question type difficulty control, combined with a rich question type and difficulty system, it is convenient for teachers to accurately select questions and compose papers, and the purpose of providing scientific and reasonable test content for teaching is achieved. The historical data storage module retains comprehensive performance data and conducts in-depth analysis to achieve precise teaching intervention. The test question extraction module is based on a reinforcement learning algorithm, and defines the state space, action space and reward function according to the student's historical answer data. By continuously optimizing the test question selection strategy, the test questions are closely aligned with the students' actual needs, so as to achieve the effect of accurately detecting the students' knowledge mastery of each chapter unit, and effectively promote the improvement of education and teaching quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the modules of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] In order to solve the technical problem that the existing student ability test cannot conduct personalized and accurate tests based on the students' actual mastery of each chapter unit, such as Figure 1 As shown, the following preferred technical solutions are provided:

[0020] A student ability testing platform for education and teaching includes: a user management module, a test question bank storage module, a historical data storage module, a test question extraction module, a question answering module, a scoring module and a feedback report module.

[0021] The user management module includes a student information management submodule and a teacher information management submodule. The student information management submodule records in detail the student's name, gender, date of birth, student ID number and grade and class information. The teacher information management submodule comprehensively stores the teacher's name, contact information, subjects taught, teaching years, professional qualification certificate number, teaching expertise and details of past teaching achievement evaluations.

[0022] The question bank storage module is used to store the questions of each subject and the corresponding chapter unit. The question bank storage module includes a subject unit classification submodule, a knowledge point annotation association submodule and a question type difficulty control submodule; the subject unit classification submodule accurately classifies the questions into each subject and the corresponding chapter unit according to the education syllabus and textbook system; the knowledge point annotation association submodule deeply annotates the core and secondary knowledge points and complex association maps for each question through knowledge graph technology; the question type difficulty control submodule includes multiple-choice questions, fill-in-the-blank questions, short-answer questions, essay questions, experimental operation questions, case analysis questions and creative design questions, and divides the questions into a 1-5 level difficulty system, and is equipped with detailed analysis documents, so that teachers can accurately select questions and assemble papers according to teaching objectives and student levels.

[0023] The historical data storage module is used to retain the comprehensive score data of students in previous tests, including the total score, the score of each question type and the score details of each knowledge point; the historical data storage module is also used to generate a visual score trend line chart and a score fluctuation chart, and to perform in-depth data analysis. The in-depth data analysis includes calculating the standard deviation and variance of the score to evaluate the stability of the score. The in-depth data analysis also includes calculating the coefficient of variation of the score. The formula is σ is the standard deviation and μ is the mean. The coefficient of variation is used to compare the degree of dispersion of scores at different stages and more accurately evaluate the trend of changes in student score stability. For example, an increase in the coefficient of variation indicates that score fluctuations have intensified, which may indicate learning problems or external interference factors, providing a quantitative basis for timely intervention.

[0024] The test question extraction module uses an adaptive test question generation algorithm based on reinforcement learning to intelligently extract test questions, specifically including:

[0025] S1. Define the state space, action space and reward function of the reinforcement learning algorithm by analyzing students' historical answer data. The state space S is based on the accuracy of students' historical answer in the current chapter unit. hist as the state; action space A: select the next test question from the test question bank, the action is determined by the chapter unit CU to which the question belongs, expressed as a = CU; reward function R: if the student answers the selected question correctly, the reward R = 1, if the answer is wrong, R = -1; a simple action value estimation algorithm is used to determine the test question selection strategy, and the action value estimation value Q(CU) = 0 is initialized for each chapter unit. For each state, that is, the answer accuracy rate of the current chapter unit, the action value estimation value is updated according to the following formula: Q(CU)←Q(CU)+α×(RQ(CU)), where α is the learning rate, and the value range is is 0-1. Each time a test question is selected, a chapter unit question is randomly selected with a certain probability ε, and the chapter unit question with the highest action value estimate Q(CU) is selected with probability 1-ε. When the student completes a certain number of questions, the test question is selected according to the learned action value estimate, that is, the chapter unit question with the highest Q(CU) value is selected as the next test question. In this way, the action value estimate is continuously updated according to the student's answering situation, so that the subsequently generated test questions can better target the student's actual mastery of each chapter unit, and adapt to the dynamic changes in the student's answering process.

[0026] S2. Scientifically set the initial standard weight vector of each unit according to the teaching syllabus W0 = [w 01 ,w 02 ,…w 0n ], calculate the adjustment factor Then through the formula Accurately calculate the adjusted unit weights to customize the unit test weights according to individual students and the overall situation of the class, ensuring that the test questions closely match the actual needs of students. is the average of the difficulty-weighted scores of all students in the class, β is the adjustment intensity coefficient, ranging from 0 to 1, and after adjustment, For recurrent neural network models;

[0027] S3.Accurately determine the number of questions for each unit based on the adjusted unit weights Q is the total number of questions in the test paper. According to the principles of knowledge point coverage priority, weak knowledge point reinforcement principle and difficulty adaptation, questions are carefully selected from the question bank to form the test paper, ensuring that the test paper is of moderate difficulty, prominent and targeted; the principle of knowledge point coverage priority is to give priority to the selection of questions corresponding to the key knowledge points of each unit to ensure a comprehensive examination of the knowledge system; the principle of weak knowledge point reinforcement lists the knowledge points with a score rate lower than the threshold θ1 as the key reinforcement area, with an extraction ratio of p1, and the score rate between θ1 and θ2 is the consolidation and improvement area, with an extraction ratio of p2, and the score rate higher than θ2 is the expansion and extension area, with an extraction ratio of p3=1-p1-p2, θ1 takes the value of 0.6, p1 takes the value of 0.7, θ2 takes the value of 0.8, and p2 takes the value of 0.2; the principle of difficulty adaptation is based on the average score rate SR of students' history answers on low, medium and high difficulty questions L , SR M , SR H Determine the proportion of low, medium, and high difficulty questions in each unit. L ≥θ3, and SR M ≥θ4, θ3 is 0.8, θ4 is 0.6, then the proportion of low, medium and high difficulty questions is r L :r M :r H =0.3:0.5:0.2, if not satisfied, adjust to r L :r M :r H =0.4:0.4:0.2.

[0028] The answering module is used to provide an answering interactive interface on computers, tablets and mobile terminal devices. The interactive interface of the answering module displays the answering progress and intelligent reminders. The answering progress includes the number of questions answered, the number of remaining questions, the answering time progress bar and the estimated remaining time.

[0029] The scoring module includes an intelligent scoring submodule and a score statistics report generation submodule. The intelligent scoring submodule is used to score and summarize students' answers. The score statistics report generation submodule comprehensively summarizes the scoring data to generate a detailed transcript. The intelligent scoring submodule uses an answer comparison algorithm for objective questions and uses natural language processing technology combined with expert experience rules to score subjective questions, specifically including accurate keyword extraction and matching, in-depth analysis of semantic logic and grammatical error detection and deduction, and uses a deep learning model to improve scoring accuracy; the score statistics report generation submodule comprehensively summarizes the scoring data to generate a detailed transcript, including the total score, score of each part, score rate, ranking percentile, and comparison with the previous test score, and generates a visual analysis report.

[0030] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0031] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A student ability testing platform for education and teaching, characterized in that: include: User management module, test question bank storage module, historical data storage module, test question extraction module, question answering module and scoring module; The user management module includes a student information management submodule and a teacher information management submodule; The test question bank storage module is used to store test questions of various subjects and corresponding chapter units; The historical data storage module is used to retain the comprehensive score data of students in previous tests, including the total score, the score of each question type and the score details of each knowledge point; The test question extraction module uses an adaptive test question generation algorithm based on reinforcement learning to intelligently extract test questions, specifically including: S1. Define the state space, action space and reward function of the reinforcement learning algorithm by analyzing students' historical answer data. The state space S is based on the accuracy of students' historical answer in the current chapter unit. hist as the state; action space A: select the next test question from the test question bank, the action is determined by the chapter unit CU to which the question belongs, expressed as a = CU; reward function R: if the student answers the selected question correctly, the reward R = 1, if the answer is wrong, R = -1; a simple action value estimation algorithm is used to determine the test question selection strategy, and the action value estimation value Q(CU) = 0 is initialized for each chapter unit. For each state, that is, the answer accuracy rate of the current chapter unit, the action value estimation value is updated according to the following formula: Q(CU)←Q(CU)+α×(RQ(CU)), where α is the learning rate, and the value range is is 0-1. Each time a test question is selected, a chapter unit question is randomly selected with a certain probability ε, and the chapter unit question with the highest action value estimate Q(CU) is selected with probability 1-ε. When the student completes a certain number of questions, the test question is selected according to the learned action value estimate, that is, the chapter unit question with the highest Q(CU) value is selected as the next test question. In this way, the action value estimate is continuously updated according to the student's answering situation, so that the subsequently generated test questions can better target the student's actual mastery of each chapter unit, and adapt to the dynamic changes in the student's answering process. S2. Scientifically set the initial standard weight vector of each unit according to the teaching syllabus W0 = [w 01 ,w 02 ,…w 0n ], calculate the adjustment factor It is the average of the difficulty-weighted score rates of all students in the class. i WSR is the student's weighted score rate for the unit. i =SR i ×(1-D i ), D i is the average difficulty coefficient of the history test questions in this unit, and then through the formula Accurately calculate the adjusted unit weight, is the average of the difficulty-weighted scores of all students in the class, β is the adjustment intensity coefficient, ranging from 0 to 1, and after adjustment, S3.Accurately determine the number of questions for each unit based on the adjusted unit weights Q is the total number of questions in the test paper. The test paper is composed of carefully selected questions from the question bank according to the principles of knowledge point coverage priority, weak knowledge point reinforcement principle and difficulty adaptation principle; The answering module is used to provide an answering interactive interface on computers, tablets and mobile terminal devices; The scoring module includes an intelligent scoring submodule and a score statistics report generation submodule. The intelligent scoring submodule is used to summarize the scores of students' answers, and the score statistics report generation submodule comprehensively summarizes the score data to generate a detailed transcript.

2. A student ability testing platform for education and teaching as claimed in claim 1, characterized in that: The student information management submodule records in detail the student's name, gender, date of birth, student ID number and grade and class information, and the teacher information management submodule comprehensively stores the teacher's name, contact information, subjects taught, teaching experience, professional qualification certificate number, teaching expertise and details of past teaching achievement evaluations.

3. A student ability testing platform for education and teaching as claimed in claim 1, characterized in that: The test question bank storage module includes a subject unit classification submodule, a knowledge point annotation association submodule and a question type difficulty control submodule; the subject unit classification submodule accurately classifies the test questions into various subjects and corresponding chapter units according to the education syllabus and teaching material system; The knowledge point marking and association sub-module uses knowledge graph technology to deeply mark the core and secondary knowledge points and complex association graphs for each test question; the question type difficulty control sub-module includes multiple-choice questions, fill-in-the-blank questions, short-answer questions, essay questions, experimental operation questions, case analysis questions and creative design questions. The questions are divided into a 1-5 level difficulty system and are equipped with detailed analysis documents to facilitate teachers to accurately select questions and compile papers according to teaching objectives and student levels.

4. A student ability testing platform for education and teaching as claimed in claim 1, characterized in that: The historical data storage module is also used to generate a visual performance trend line chart and performance fluctuation chart, and to perform in-depth data analysis, which includes calculating the performance standard deviation and variance for evaluating performance stability.

5. A student ability testing platform for education and teaching as claimed in claim 1, characterized in that: The principle of priority in knowledge point coverage in step S3 is to give priority to selecting questions corresponding to the key knowledge points of each unit.

6. A student ability testing platform for education and teaching as claimed in claim 1, characterized in that: The weak knowledge point strengthening principle in step S3 lists the knowledge points with a score rate lower than the threshold θ1 as the key strengthening area, with an extraction ratio of p1. The score rate between θ1 and θ2 is the consolidation and improvement area, with an extraction ratio of p2. The score rate higher than θ2 is the expansion and extension area, with an extraction ratio of p3=1-p1-p2. θ1 takes a value of 0.6, p1 takes a value of 0.7, θ2 takes a value of 0.8, and p2 takes a value of 0.

2.

7. A student ability testing platform for education and teaching as claimed in claim 1, characterized in that: The difficulty adaptation principle in step S3 is based on the average score rate SR of students' history answers on low, medium and high difficulty questions. L , SR M , SR H Determine the proportion of low, medium, and high difficulty questions in each unit. L ≥θ3, and SR M ≥θ4, θ3 is 0.8, θ4 is 0.6, then the proportion of low, medium and high difficulty questions is r L :r M :r H =0.3:0.5:0.2, if not satisfied, adjust to r L :r M :r H =0.4:0.4:0.

2.

8. A student ability testing platform for education and teaching as claimed in claim 1, characterized in that: The interactive interface of the answering module displays the answering progress and intelligent reminders. The answering progress includes the number of questions answered, the number of remaining questions, the answering time progress bar and the estimated remaining time.

9. A student ability testing platform for education and teaching as claimed in claim 1, characterized in that: The intelligent scoring submodule adopts an answer comparison algorithm for objective questions and uses natural language processing technology combined with expert experience rules to score subjective questions, including precise keyword extraction and matching, deep analysis of semantic logic and grammatical error detection and deduction, and a deep learning model to improve scoring accuracy; the score statistical report generation submodule comprehensively summarizes the score data to generate a detailed transcript, including the total score, score of each part, score rate, ranking percentile, and comparison with the previous test score, and generates a visual analysis report.

10. A student ability testing platform for education and teaching as claimed in claim 4, characterized in that: The in-depth data analysis also includes calculating the coefficient of variation of the scores, the formula is σ is the standard deviation and μ is the mean.

Images