Teaching Resource Dynamic Allocation System Based on Knowledge Encoding and LFNN Model

Abstract

The present invention discloses a dynamic teaching resource allocation system based on knowledge encoding and LFNN model. This system is based on a cybernetic model and includes a controller model formed by professional teachers and an intelligent control module, as well as a controlled object model formed by three subsystems: a knowledge encoding system, a learning situation management system, and a teaching plan system. By introducing knowledge encoding into teaching, the present invention transforms teaching content into quantifiable data. Based on the Ebbinghaus forgetting curve, an online and offline intelligent teaching system based on the cybernetic model is established, which can accurately analyze the process learning situation of students as a whole and individuals, and adjust the teaching plan and personalized autonomous learning plan based on scientific memory methods in real time. It drives teaching with data intelligence, promotes individualized instruction, and improves teaching efficiency.

Description

Technical Field

[0001] The present invention relates to the fields of intelligent teaching and education, and particularly relates to an online and offline intelligent teaching system based on knowledge coding and the LFNN model. Background Art

[0002] Intelligent teaching, namely educational informatization, refers to the process of comprehensively and deeply applying modern information technology in the field of education to promote educational reform and development, which is expected to significantly improve teaching efficiency and achieve large-scale individualized teaching. Among them, educational informatization and its application innovation are the technical keys to promoting intelligent education. At present, the main problems in promoting intelligent education are as follows: 1. The application mode of educational informatization in teaching; 2. The repositioning of teachers in intelligent teaching; 3. The systematic design of intelligent teaching involving intelligent teaching, learning, management, and evaluation.

[0003] To comprehensively improve the ability to cultivate talents, with the development of students as the center and promote the cultivation of students' autonomous learning ability, online and offline blended teaching integrates the advantages of online teaching and traditional face-to-face offline teaching while avoiding their disadvantages, and has become a hot spot for the reform and innovation of teaching models. Blended teaching has the characteristics of flexible teaching content, form, time, and space, and has significant advantages in improving students' autonomous learning ability. At present, the core problem of blended teaching is how to avoid formal blending and achieve deep integration. The main reasons for the formalization of blended teaching are as follows: 1. The unclear blending goal; 2. The unclear basis for class hour allocation; 3. The lack of an organic connection between online and offline; 4. The great difficulty in realizing students' independent autonomous learning and the limited completion degree. Summary of the Invention

[0004] To solve the above problems, the present invention provides an online and offline intelligent teaching system based on knowledge coding and the Ebbinghaus forgetting curve with systematic design, aiming to deeply integrate online and offline education, accurately design an efficient teaching process and teaching plan based on knowledge coding and the Ebbinghaus forgetting curve, accurately analyze the process learning conditions of students as a whole and individuals, establish a control theory-based model, and use professional teachers and intelligent control modules to adjust the teaching plan based on the scientific memory method and the personalized autonomous learning plan in real time, drive teaching with data intelligence, take students as the center, and take teachers as regulators and guides to promote large-scale individualized teaching, improve teaching efficiency, and cultivate students' autonomous learning ability.

[0005] The present invention provides the following technical solutions:

[0006] The present invention is based on a cybernetic model, including a controller model composed of professional teachers and intelligent control modules, and a controlled object model composed of three subsystems: a knowledge coding system, a learning situation management system, and a teaching plan system. The teachers and intelligent control modules are responsible for receiving feedback information and exerting control effects on the states and parameters of the three subsystems. The three subsystems interact with each other and are closely related.

[0007] The knowledge coding system is responsible for the microscopic decomposition and hierarchical quantification of teaching knowledge, encodes knowledge using a centralized strategy, and is the basis for establishing intelligent mapping relationships with modules such as learning situation analysis, teaching process, collective teaching plan, and autonomous learning plan, and is regulated by teachers and intelligent control modules.

[0008] The teaching plan system is responsible for planning teaching content, teaching forms, and teaching processes, intelligently generates teaching plans based on the Ebbinghaus forgetting curve and the knowledge coding system, and adjusts the teaching plans in real time according to the feedback from the learning situation management system by professional teachers and intelligent control modules.

[0009] The learning situation management system is responsible for accurately outputting learning situation analysis in units of microscopic knowledge elements, statistically analyzing the learning situation of students from both subjective and objective aspects based on the teaching plan system, accurately analyzing the learning situation down to microscopic knowledge points based on the knowledge coding system, and feeding back to professional teachers and intelligent control modules.

[0010] The users targeted by this system include two types: teachers and students. The functions on the teacher side include login and file creation, knowledge base management, teaching calendar management, class management, intelligent question bank management, and teaching resource library management. The functions on the student side include login and file creation, browsing the knowledge base, browsing the teaching calendar, browsing the teaching resource library, online self-testing, learning situation analysis, and learning situation analysis.

[0011] The knowledge base management function on the teacher side is based on the knowledge coding system. After professional teachers log in and create files, they can edit and modify knowledge grading, importance level, difficulty rating, delete or add knowledge elements.

[0012] The teaching calendar management function on the teacher side is based on the teaching plan system. Professional teachers input the start and end times, cycle, total class hours, and special class hour nodes of offline teaching, and the system automatically generates an online and offline teaching calendar plan based on the Ebbinghaus forgetting curve, displaying the microscopic decomposed knowledge elements and corresponding teaching forms.

[0013] The class management function on the teacher side includes a list of class students, an overall learning situation evaluation of the class, and an individual learning situation evaluation of students.

[0014] The intelligent question bank management function on the teacher side includes setting the quantity and difficulty of questions in the student online self-testing module, setting the quantity, difficulty, and chapter ratio of the intelligent test paper function, as well as uploading new questions, checking and associating knowledge point information, and officially adding them to the intelligent question bank after being reviewed by experts on the system side.

[0015] The teaching resource library management function on the teacher side includes downloading teaching resources, uploading new teaching resources, which will be officially added to the teaching resource library after being reviewed by experts on the system side, and setting the permission opening conditions for teaching resources.

[0016] The login and file - creation function on the student side realizes one file per person. According to the learning situation of individual students, the system adaptively and intelligently adjusts the autonomous learning plan. Among them, in the online self - test of individual students, if the correct rate of questions corresponding to a single knowledge point is above 90%, it is defined as good mastery; if the correct rate of questions corresponding to a single knowledge point is below 60%, it is defined as poor mastery; and students with an average self - test score ranking in the top 15% of the class are defined as having good learning conditions.

[0017] The function of browsing the knowledge base on the student side allows students to master the overall knowledge distribution framework of the course, understand the composition, importance, and difficulty rating of each knowledge point.

[0018] The function of browsing the teaching calendar on the student side allows students to understand the learning process and tasks, and conduct collective learning and autonomous learning under the guidance of professional teachers and the system.

[0019] The function of browsing the teaching resource library on the student side includes tasks such as video viewing, case learning, and previewing and reviewing by drawing knowledge networks.

[0020] The online self - test function on the student side is based on the intelligent question bank module of the teaching plan system. The self - test results are fed back to the objective learning situation module of the learning situation management system. In the learning situation analysis function for students and the class management function for teachers, this objective learning situation analysis can be seen.

[0021] The learning situation analysis function on the student side includes subjective learning situation statistics and browsing the learning situation analysis function. On the one hand, after offline teaching, the knowledge points taught on that day are shown in the subjective learning situation statistics. Students can select the knowledge elements that are difficult to master, and the self - evaluation results are fed back to the subjective learning situation module of the learning situation management system. Combined with the objective learning situation module, it is used to intelligently adjust the teaching plan. On the other hand, students can browse the process - based individual learning situation and the overall class learning situation, so that students can understand their own learning situation, promote students to learn purposefully, and stimulate students' learning interest.

[0022] The beneficial effects of the present invention

[0023] In this system, three subsystems, namely the knowledge coding system, the learning situation management system, and the teaching plan system, interact with each other and are closely linked. The knowledge coding system microscopically decomposes and quantifies knowledge hierarchically, and conducts knowledge coding with a centralized strategy. The control methods include feedback control and feedforward control. Feedback control dynamically controls and adjusts the system based on the output feedback of the system, making the value of the controlled object approach the expected value of the system. According to the teaching feedback of students, the teaching plan is adjusted afterwards. Feedforward control directly adjusts the controlled object according to the coding feature size and historical correction according to specific control rules, which belongs to pre-control. The teaching plan system is based on the knowledge coding system. Based on the Ebbinghaus forgetting curve and according to the attribute characteristics of knowledge micro-elements, it outputs the teaching process arrangement, the teaching content and methods of collective teaching and autonomous learning. The learning situation management system is based on the knowledge coding system and accurately outputs the learning situation analysis in units of knowledge micro-elements. The output result of the controlled object model is fed back to the controller model through the reverse transmission channel, and professional teachers and the intelligent control module of the system adjust the relevant parameters of the three subsystems to obtain higher teaching efficiency. By introducing knowledge coding into teaching, this invention transforms teaching content into quantifiable data, and based on the Ebbinghaus forgetting curve, establishes an online and offline intelligent teaching system based on the cybernetic model, which can accurately analyze the process learning situation of students as a whole and individuals, and adjust the teaching plan based on the scientific memory method and the personalized autonomous learning plan in real time, driving teaching with data intelligence, promoting individualized teaching, and improving teaching efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is the system model diagram of the present invention;

[0025] Figure 2 is the schematic diagram of the case of multiple consolidation plans for a single knowledge point based on the Ebbinghaus forgetting curve of the present invention;

[0026] Figure 3 is the functional structure diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] The present invention will be described in conjunction with the accompanying drawings and embodiments.

[0028] As Figure 1 shown, a teaching resource dynamic allocation system based on knowledge coding and the LFNN (Learning and Forgetting Neural Network) model. This system is based on the cybernetic model and includes a controller model formed by professional teachers and an intelligent control module, and a controlled object model formed by three subsystems, namely the knowledge coding system, the learning situation management system, and the teaching plan system. The three subsystems interact with each other and are closely linked. The system process is mainly as follows:

[0029] S1: Professional teachers decompose teaching knowledge according to teaching experience, referring to materials such as textbooks, lesson plans, and teaching syllabuses, to form multi-level knowledge such as chapters, sections, major knowledge points, and minor knowledge points.

[0030] S2: Using difficulty and importance as grading indicators, quantitatively grade the smallest knowledge elements. The difficulty is divided into levels 1-3. The criteria for judging difficulty are mainly based on the cognitive depth and comprehensiveness of the knowledge elements. Knowledge elements with a difficulty level of 1 are mainly for memorization, with little need for understanding / easy to understand, and the amount of memorization content is average, and students can fully study by themselves. Knowledge elements with a difficulty level of 2 are mainly for understanding or have a large amount of content to be memorized. Knowledge elements with a difficulty level of 3 involve comprehensive understanding, calculation, and memorization of multiple knowledge points. The importance is also divided into levels 1-3. The grading criteria for judging importance are mainly based on the supporting relationship between the curriculum objectives in the teaching syllabus and the graduation requirement index points. Finally, generate the weight value of the knowledge element according to the difficulty and importance ratings.

[0031] S3: Implement a centralized coding strategy for the quantitatively graded teaching knowledge, encoding the teaching knowledge in the form of highly concise numbers and letters. The encoding contains information such as the level, subordination, importance, and difficulty of the knowledge. For example, if the importance of the first knowledge element of the first major knowledge point of the first section of the first chapter is level 3 and the difficulty is level 2, it can be encoded as A1B1C1D1Z3N2. The knowledge coding system is regulated by teachers and the intelligent control module. According to the feedback of the learning situation management system, professional teachers and the intelligent control module intelligently and collaboratively adjust the difficulty rating of the knowledge according to the set rules, and professional teachers regulate the importance rating of the knowledge according to the teaching syllabus.

[0032] S4: As Figure 2 shown, it is a schematic diagram of the multiple consolidation plan case of a single knowledge point based on the Ebbinghaus forgetting curve in the present invention. Based on the Ebbinghaus forgetting curve, using 12 hours, 1 day, 4 days, 7 days, 30 days, 45 days, 60 days, and 90 days as the initial consolidation periods, arrange the online self-study hours and offline collective teaching hours alternately to form an online and offline learning hour schedule. Form the original network function y = f(x).

[0033] S5: Connect the weight value of the knowledge element of the teaching plan system and the knowledge coding system. Teachers select the teaching content of the courses in this semester according to the knowledge coding system, establish the mapping relationship between the weight value of the knowledge element and the time weight value, and allocate it proportionally to the offline learning hours as the content of the first class according to the time weight value and the offline learning hour schedule. Then establish the mapping relationship between the weight value of the knowledge element, the consolidation period, the number of times, and the teaching form, and intelligently arrange it into the online and offline learning hour schedule, and finally fine-tune and confirm by the teacher. The fine-tuning behavior is recorded and learned by the intelligent control module.

[0034] S6: The collective teaching module includes regular teaching forms (lecture, question-and-answer, in-class exercise and explanation, heuristic case teaching) and special teaching forms (test, project-based case discussion, mid-term exam, final exam). The regular teaching forms are automatically generated by connecting with the knowledge element weight values and consolidation cycles in the knowledge coding system. The class hours corresponding to the special teaching forms are selected by professional teachers. After selection, the regular teaching process is adaptively adjusted by the intelligent control module. The autonomous learning module involves teaching forms such as watching videos, online self-testing, independently drawing knowledge networks, and online project-based case discussions, including two sub-modules: the intelligent question bank and the teaching resource library. The intelligent question bank is connected to the knowledge coding system. Professional teachers determine the knowledge elements included in each question, associate the question coding with the coding of the included knowledge elements, and the weight values of the difficulty and importance of the questions are determined by the number of included knowledge elements and the comprehensive weight values of the knowledge elements. In the online self-test session, according to the knowledge points that need to be consolidated in the online self-test on the current day in the teaching schedule, the initial quantity and difficulty of the online self-test questions set by the teacher, self-test questions are automatically generated. The self-test results are fed back to the learning situation management system. After statistical analysis, they are fed back to professional teachers and the intelligent control module. The intelligent control module is responsible for adjusting the learning plan of the autonomous learning module for individual students, and provides methods for adjusting the difficulty rating of knowledge elements for the whole class of students and the teaching plan of the collective teaching module, which are finally determined by professional teachers. For mid-term and final exams, random test questions are automatically generated according to the test paper difficulty, chapter proportion, question types and quantity set by teachers. After the exam results are imported into the system, they are fed back to the learning situation management system to form an analysis of the exam situation for the whole class and individual students based on knowledge elements, and are fed back to teachers and individual students respectively. The teaching resource library is connected to the knowledge coding system and includes teaching videos, knowledge graphs, application cases, and ideological and political cases. Relevant teaching resources are opened to students according to the resource type and associated knowledge coding.

[0035] S7: The learning situation management system includes four modules: subjective learning situation, objective learning situation, overall class learning situation, and individual student learning situation. For the subjective learning situation, students independently check the knowledge points that are difficult to master in the previous offline teaching content, and the limit on the number of checked items is set by professional teachers. The objective learning situation is obtained through the statistical analysis of the results of online self-tests and mid-term tests in the teaching plan system. The overall class learning situation is obtained through the statistical analysis of the subjective and objective learning situations of all students in the class. The individual student learning situation is obtained through the statistical analysis of the subjective and objective learning situations of individual students, with one file for each person. The learning situation management system is associated with the knowledge coding system and the teaching plan system. The overall class learning situation is associated with the difficulty rating of knowledge elements in the knowledge coding system and the regulation of the consolidation times in the teaching plan system. For knowledge elements that most students have difficulty mastering, the difficulty rating will be increased overall, and the consolidation times will be increased. The individual student learning situation is associated with the independent learning module in the teaching plan system. For knowledge elements with poor mastery, the consolidation times in the independent learning plan will be increased, and for knowledge elements with good mastery, the consolidation times will be reduced accordingly. For individual students with good learning performance, the intelligent control module will intelligently reduce the independent learning time and increase the independent learning difficulty.

[0036] According to the original network function y = f(x) in S4, the state transition model is defined as a multi-stage decision problem model. The scheduling process can be divided into multiple interconnected stages. The strategy of the current stage depends on the previous stage's strategy and affects the overall effect in the future. After the strategy for each stage is selected, a decision sequence is formed. The purpose of the LKNN model is to find the decision sequence with the best benefit.

[0037] Define the following assumptions:

[0038] The total number of stages (class hours) is n, v is the number of course days, w is the number of class hours per course day, and n = v * w.

[0039] The serial number of each stage is k, where k = 1, 2, 3... n.

[0040] The starting state of the k-th stage is s k , assuming that knowledge points A1 and A2 are scheduled before the k-th class hour, denoted as s k = {A1, A2}.

[0041] The decision for the k-th stage is u k (s k ), assuming that knowledge points B1 and B2 are scheduled in the k-th class hour, denoted as u k (s k ) = {B1, B2}.

[0042] The final strategy sequence {u1(s1), u2(s2)…u n (s n )}.

[0043] The sub-strategy of the k-th stage is denoted as pk,n (s k ),

[0044] The total number of knowledge points in the stage is m,

[0045] The serial number of each knowledge point is j, where j = 1, 2, 3... m,

[0046] The benefit of the knowledge point is S,

[0047] The importance of the knowledge point is I,

[0048] The difficulty of the knowledge point is D,

[0049] The weight of each knowledge point is W, and the initial weight is W = 1.

[0050] Define S[j] = W[j] * D[j] * I[j],

[0051] There exists the Ebbinghaus memory curve function g(k), and the benefit of the knowledge point changes with the stage, then:

[0052] S[j,k] = W[j] * D[j] * I[j] * g(k),

[0053] Let f[j,k] represent the maximum benefit obtained by putting the first i knowledge points into the first k class hours

[0054] Then p k,n (s k )'s state transition equation is f[j,k] = max{f[j - 1,k], f[j - 1][k - 1] + S[j,k]}

[0055] By recursion, the overall model evaluation function S = S[n,m] can be obtained. For this evaluation function, the student samples are divided into a training set and a prediction set. The original rating system from 1 to 9 is normalized to the [0,1] interval. The Euclidean distance between the network prediction value and the student prediction set samples is calculated to obtain the correlation coefficient. The correlation coefficients of each class are arranged as a one-dimensional vector to obtain the vector k, and the length of k is calculated, which is the evaluation function of LFNN.

[0056] LFNN designs an automatic learning situation feedback algorithm based on neural network through optimizing the mathematical model of curriculum arrangement for the curriculum arrangement decision-making process. This algorithm can judge the course difficulty and importance, and automatically make customized adjustments according to the individual situation of students. In addition, in curriculum arrangement, the Ebbinghaus memory curve method is also considered. Based on the Jaccard coefficient and neural network method, the "computational complexity" and "strengthened memory" problems caused by tradition in the curriculum arrangement control system are solved.

[0057] S8: As Figure 3As shown in the figure, a teaching resource dynamic allocation system based on knowledge encoding and LFNN (Learning and Forgetting Neural Network) model of the present invention faces two types of users, namely teachers and students. The functions of the teacher end include login and file creation, knowledge base management, teaching calendar management, class management, intelligent question bank management, and teaching resource library management. The functions of the student end include login and file creation, browsing the knowledge base, browsing the learning process, browsing learning tasks, subjective learning situation statistics, online self-testing, browsing learning situation analysis, and browsing the teaching resource library. The knowledge base management of the teacher end is based on the knowledge encoding system. After professional teachers log in and create files, they can edit and modify knowledge classification, importance level, difficulty rating, and delete or add knowledge elements.

[0058] S9: The teaching calendar management of the teacher end is based on the teaching plan system. Professional teachers input the start and end times, cycle, total class hours, and special class hour nodes of offline teaching. The system automatically generates an online and offline teaching calendar plan based on the Ebbinghaus forgetting curve, and displays the micro-decomposed knowledge elements and corresponding teaching forms. The class management module of the teacher end includes the class student list, overall class learning situation evaluation, and individual student learning situation evaluation. The intelligent question bank management includes the question quantity and difficulty setting of the student online self-test module, the question quantity, difficulty, and chapter ratio setting of the intelligent test paper function, as well as uploading new questions, checking and associating knowledge point information, and officially adding them to the intelligent question bank after being reviewed by experts on the system side. The teaching resource library management of the teacher end includes downloading teaching resources, uploading new teaching resources, officially adding them to the teaching resource library after being reviewed by experts on the system side, and setting the permission opening conditions of teaching resources.

[0059] S10: In the functions of the student end, after logging in and creating files, one file per person, according to the learning situation of individual students, the system adaptively and intelligently adjusts the independent learning plan. Students can browse the knowledge base, master the overall knowledge distribution framework of the course, and understand the composition, importance level, and difficulty rating of each knowledge point. They can browse the learning process and learning tasks, and carry out collective learning and independent learning under the guidance of professional teachers and the system. Students can browse the teaching resource library and complete preview and review tasks including video viewing, case learning, and drawing knowledge networks. The online self-test of the student end is based on the intelligent question bank module of the teaching plan system, and the self-test results are fed back to the objective learning situation module of the learning situation management system. After offline teaching, through subjective learning situation statistics, self-evaluate the knowledge elements that are difficult to master and feed them back to the subjective learning situation module of the learning situation management system. Combining with the objective learning situation module, it is used to intelligently adjust the teaching plan. Students can browse the process-based individual learning situation and overall class learning situation, so that students can understand their own learning situation, which is conducive to students' purposeful learning and stimulates students' learning interest.

[0060] The above description is only a preferred embodiment of the present invention and a basic explanation of the basic principles and characteristics. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solution formed by the specific combination of the above technical features. Without departing from the principles and spirit of the present invention, various changes, modifications, substitutions, and variations can be made to these embodiments. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A teaching resource dynamic allocation system based on knowledge encoding and LFNN model, characterized in that The system is based on a cybernetic model, abstracted as an LFNN model, including a controller model composed of professional teachers and an intelligent control module, and a controlled object model composed of three subsystems: a knowledge encoding system, a learning situation management system, and a teaching plan system; The teacher and intelligent control module are responsible for receiving feedback information and exerting control effects on the states and parameters of the three subsystems; The knowledge encoding system is responsible for the microscopic decomposition and hierarchical quantification of teaching knowledge, encodes knowledge using a centralized strategy, and is the basis for establishing an intelligent mapping relationship with the learning situation analysis, teaching process, collective teaching plan, and autonomous learning plan modules, and is regulated by the teacher and intelligent control module; The teaching plan system is responsible for planning teaching content, teaching forms, and teaching processes, intelligently generates teaching plans based on the Ebbinghaus forgetting curve and the knowledge encoding system, and adjusts the teaching plans in real time by professional teachers and the intelligent control module according to the feedback from the learning situation management system; The learning situation management system is responsible for accurately outputting a learning situation analysis in units of microscopic knowledge elements, statistically analyzes the learning situation of students from both subjective and objective aspects based on the teaching plan system, accurately analyzes the learning situation down to microscopic knowledge points based on the knowledge encoding system, and feeds back to professional teachers and the intelligent control module; The teaching plan system consists of three parts: a teaching process, a collective teaching module, and an autonomous learning module; The teaching process is described as follows: Based on the Ebbinghaus forgetting curve, with 12 hours, 1 day, 4 days, 7 days, 30 days, 45 days, 60 days, and 90 days as the initial consolidation periods, the online autonomous learning class hours and offline collective teaching class hours are arranged alternately to form an online and offline class hour progress schedule; The specific steps for forming the online and offline class hour progress schedule are as follows: First, form the original network function y = f(x), define the state transition model as a multi-stage decision-making problem model. The scheduling process is divided into multiple interconnected stages. The strategy of the current stage depends on the previous strategy and affects the overall effect in the future. After the strategy for each stage is selected, a decision sequence is formed. The purpose of the LFNN model is to find the decision sequence with the best benefit; Define the following assumptions: The total number of stages n is the number of class hours, v is the number of course days, w is the number of class hours per course day, and n = v * w, The serial number of each stage is k, k = 1, 2, 3... n, The starting state of the k-th stage is s k , assuming that knowledge points A1 and A2 are arranged before the k-th class hour, denoted as s k = {A1, A2} The decision in the k-th stage is u k (s k ), assuming that knowledge points B1 and B2 are arranged in the k-th class hour, denoted as u k (s k ) = {B1, B2} Final policy sequence {u1(s1), u2(s2) … u n (s n )} The sub-strategy of the k-th stage is denoted as p k,n (s k ) The total number of knowledge points in the stage is m, The serial number of each knowledge point is j, j = 1, 2, 3... m, The benefit of the knowledge point is S, The importance of the knowledge point is I, The difficulty of the knowledge point is D, The weight of each knowledge point is W, and the initial weight is W = 1, Define S[j] = W[j] * D[j] * I[j], There is an Ebbinghaus forgetting curve function g(k), and the benefit of the knowledge point changes with the stage, then: S[j,k] = W[j] * D[j] * I[j] * g(k), Let f[j,k] represent the maximum benefit obtained by putting the first i knowledge points into the first k class hours, Then p k,n (s k ) has a state transition equation of f[j,k] = max{f[j-1,k], f[j-1][k-1] + S[j,k]}, The recursive overall model evaluation function is \(S = S[n,m]\). For this evaluation function, the student samples are divided into a training set and a prediction set. The original rating system from 1 to 9 is normalized to the interval [0,1]. The Euclidean distance is calculated between the network prediction value and the student prediction set samples to obtain the correlation coefficient. The correlation coefficients of each class are arranged as a one-dimensional vector to obtain vector \(k\), and the length of \(k\) is calculated, which is the evaluation function of LFNN.

2. The system according to claim 1, characterized in that, The knowledge encoding system consists of three parts: microscopic knowledge decomposition, hierarchical quantization, and encoding. The microscopic decomposition of teaching knowledge is carried out by professional teachers according to teaching experience, referring to teaching materials, lesson plans, and syllabus materials, forming multi-level knowledge including chapters, sections, major knowledge points, and minor knowledge points. Taking the smallest knowledge element as the unit, with difficulty and importance as the grading indicators, according to the supporting relationship between the course objectives in the syllabus and the graduation requirement index points, the importance of each knowledge element is quantitatively rated. The difficulty of knowledge is hierarchically quantified according to the cognitive depth and comprehensive degree of knowledge. The weight value of the knowledge element is generated from the difficulty and importance ratings. For the hierarchically quantified teaching knowledge, a centralized encoding strategy is adopted, and the teaching knowledge is encoded in a highly concise form of numbers and letters, and the encoding contains information such as the level, subordination, importance, and difficulty of the knowledge.

3. The system according to claim 1, characterized in that, The regulation of the knowledge encoding system by the teacher and the intelligent control module is specifically as follows: According to the feedback of the learning situation management system, the professional teacher and the intelligent control module intelligently cooperate to adjust the difficulty rating of the knowledge according to the set rules, and the professional teacher regulates the importance rating of the knowledge according to the syllabus.

4. The system according to claim 1, wherein The teaching plan system is connected to the weight values of the knowledge elements of the knowledge encoding system. The teacher selects the teaching content of the courses in this semester according to the knowledge encoding system. The selected teaching knowledge elements determine the consolidation period, number of times, and teaching form for each knowledge element according to the weight value, and are intelligently arranged into the online and offline class hour schedule, and finally fine-tuned and confirmed by the teacher. The fine-tuning behavior is recorded and learned by the intelligent control module. The collective teaching module includes conventional teaching forms, including lectures, questions, classroom exercises and explanations, heuristic case teaching; and special teaching forms, including tests, project-based case discussions, mid-term exams, and final exams. The conventional teaching forms are automatically generated in connection with the weight values of the knowledge elements and the consolidation period of the knowledge encoding system. The class hours corresponding to the special teaching forms are selected by the professional teacher. After selection, the conventional teaching process is adaptively adjusted by the intelligent control module. The autonomous learning module involves teaching forms such as watching videos, online self-tests, independently drawing knowledge networks, and online project-based case discussions, including two sub-modules: a smart question bank and a teaching resource library. The smart question bank is connected to the knowledge encoding system. The professional teacher determines the knowledge elements included in each question, associates the question encoding with the encoding of the included knowledge elements, and the difficulty and importance weight values of the question are determined by the number of included knowledge elements and the comprehensive weight value of the knowledge elements. For the online self-test described above, according to the knowledge points that need to be consolidated in the online self-test on the teaching schedule for the day, and the initial values of the quantity and difficulty of the online self-test questions set by the teacher, self-test questions are automatically generated; the self-test results are fed back to the learning situation management system, and after statistical analysis, they are fed back to the professional teacher and the intelligent control module. The intelligent control module is responsible for adjusting the learning plan of the autonomous learning module for individual students, and providing methods for adjusting the difficulty rating of knowledge elements for the whole class of students and the teaching plan of the collective teaching module, which are finally determined by the professional teacher; For the mid-term and final exams described above, according to the test paper difficulty, chapter proportion, question types and quantity set by the teacher, random test questions are automatically generated. After the test results are imported into the system, they are fed back to the learning situation management system to form an analysis of the test situation for the whole class and individual students based on knowledge elements, and are respectively fed back to the teacher and individual students; The teaching resource library is connected to the knowledge coding system and includes teaching videos, knowledge graphs, application cases, and ideological and political cases. Relevant teaching resources are opened to students according to the resource type and associated knowledge coding.

5. The system according to claim 1, wherein The learning situation management system described above includes four modules: subjective learning situation, objective learning situation, overall class learning situation, and individual student learning situation; The subjective learning situation is that students independently check the knowledge points that are difficult to master in the previous offline teaching content. The limit on the number of checks is set by the professional teacher; The objective learning situation is obtained through statistical analysis of the feedback results of the online self-test and mid-term tests in the teaching plan system; The overall class learning situation is obtained through statistical analysis of the subjective and objective learning situations of all students in the class; The individual student learning situation is obtained through statistical analysis of the subjective and objective learning situations of individual students, with one file for each person.

6. The system according to claim 5, wherein The learning situation management system is associated with the knowledge coding system and the teaching plan system. The overall class learning situation is associated with the difficulty rating of knowledge elements in the knowledge coding system and the regulation of the consolidation times in the teaching plan system. The difficulty rating of knowledge elements with poor student mastery will be overall increased, and the consolidation times will be increased; among them: the correct rate of questions corresponding to a single knowledge point above 90% is defined as good mastery, and the correct rate of questions corresponding to a single knowledge point below 60% is defined as poor mastery; The individual student learning situation is associated with the autonomous learning module of the teaching plan system. The knowledge elements with poor mastery will increase the consolidation times in the autonomous learning plan, and the knowledge elements with good mastery will correspondingly reduce the consolidation times; individual students with good learning situations will have their autonomous learning time intelligently reduced and the autonomous learning difficulty increased by the intelligent control module; students with an average self-test score ranking in the top 15% of the class are defined as having good learning situations.

7. The system according to claim 1, wherein The users targeted by this system include two types: teachers and students. The functions of the teacher side include login and file creation, knowledge base management, teaching calendar management, class management, intelligent question bank management, and teaching resource library management; the functions of the student side include login and file creation, browsing the knowledge base, browsing the teaching calendar, browsing the teaching resource library, online self-test, and learning situation analysis.

8. The system according to claim 7, wherein The knowledge base management on the teacher side is based on the knowledge coding system. After the professional teacher logs in and creates a file, they have the permission to edit and modify the knowledge classification, importance level, and difficulty rating, and to delete or add knowledge elements; The teaching calendar management on the teacher side is based on the teaching plan system. Professional teachers input the start and end times, cycle, total class hours, and special class hour nodes of offline teaching. The system automatically generates an online and offline teaching calendar plan based on the Ebbinghaus forgetting curve, showing the micro-decomposed knowledge elements and corresponding teaching forms. The class management module on the teacher side includes the class student list, overall class learning situation evaluation, and individual student learning situation evaluation. The intelligent question bank management includes setting the quantity and difficulty of questions in the online self-test module for students, setting the quantity, difficulty, and chapter ratio of the intelligent test paper function, as well as uploading new questions, checking and associating knowledge point information, and officially adding them to the intelligent question bank after being reviewed by experts on the system side. The teaching resource library management on the teacher side includes downloading teaching resources, uploading new teaching resources, officially adding them to the teaching resource library after being reviewed by experts on the system side, and setting the permission opening conditions for teaching resources.

9. The system according to claim 7, wherein In the functions on the student side, after logging in and creating a file, each student has a personalized file. According to the learning situation of each student, the system adaptively and intelligently adjusts the independent learning plan. Students have the permission to browse the knowledge base, master the overall knowledge distribution framework of the course, and understand the composition, importance, and difficulty rating of each knowledge point. Students have the permission to browse the teaching calendar, understand the learning process and tasks, and carry out collective learning and independent learning under the guidance of professional teachers and the system. Students have the permission to browse the teaching resource library and complete preview and review tasks including video viewing, case learning, and drawing knowledge networks. The online self-test on the student side is based on the intelligent question bank module of the teaching plan system, and the self-test results are fed back to the objective learning situation module of the learning situation management system. The learning situation analysis function includes subjective learning situation statistics and browsing the learning situation analysis function. On the one hand, after offline teaching, through subjective learning situation statistics, self-evaluate the knowledge elements that are difficult to master and feed them back to the subjective learning situation module of the learning situation management system, which is combined with the objective learning situation module to be used for intelligent adjustment of the teaching plan. On the other hand, students browse the process-based individual learning situation and the overall class learning situation, enabling students to understand their own learning situation, facilitating targeted learning, and stimulating students' learning interest.

Images