Intelligent self-adaptive learning exercise pushing system and method

Abstract

The invention discloses an intelligent self-adaptive learning exercise pushing system and method, and relates to the technical field of computer information processing. A knowledge graph of a subject is generated through mind mapping software, knowledge points involved in the exercise solving process are collected, and the corresponding relation between exercises and nodes in the knowledge graph is established; the problem solving steps submitted by the exerciser correspond to the correct problem solving steps, the wrong problem solving steps in the problem solving content submitted by the exerciser are obtained from the correction record of the exercises, the corresponding nodes of knowledge points involved in the problem solving steps of the exercises in the knowledge graph are calculated, all the nodes in the neighborhood range of the corresponding nodes are collected, and all the nodes in the neighborhood range of the corresponding nodes are collected; the method comprises the steps of obtaining a neighborhood node set, calculating a difference value of a coincidence degree, obtaining knowledge point evaluation values of knowledge points related to the first target step, comparing the knowledge point evaluation values of two question doing records, and pushing exercises related to the knowledge points to related question doing persons.

Description

Technical Field

[0001] The present invention relates to the technical field of computer information processing, and in particular to an intelligent adaptive learning exercise push system and method. Background Art

[0002] With the integration of computer technology and Internet platforms with the traditional education industry, many education platforms can push corresponding exercises to learners based on their learning records. Learners can use the exercise recommendation system to reduce the time spent searching for suitable exercises and avoid doing exercises blindly.

[0003] Most of the existing technologies compare the similarity between the pictures and texts of the exercises, find similar exercises and push them to learners. This method only considers the similarity of the presentation of the exercises, but lacks the similarity of the knowledge points covered by the exercises in the subject. Therefore, the existing technology has the problem that the resource recommendation is not accurate enough, the knowledge points included in the recommended resources are different from the original questions, and the learners' usage experience is not good. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent adaptive learning exercise push system and method to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for pushing intelligent adaptive learning exercises, the method comprising the steps of: Step S100: Collect knowledge points in the subject knowledge content, mark the knowledge points and the connections between them, generate a knowledge map of the subject through mind mapping software, collect the knowledge points involved in the problem-solving process, and establish a corresponding relationship between the problem and the nodes in the knowledge map; Step S200: Divide the problem-solving process of the exercise into several steps, obtain the wrong problem-solving steps submitted by the exerciser, and the correct problem-solving steps corresponding to the wrong problem-solving steps; Step S300: respectively obtain the correct problem-solving steps and the incorrect problem-solving steps, and the corresponding nodes of the knowledge points involved in the previous problem-solving steps in the knowledge graph, and collect all nodes within the neighborhood of the corresponding nodes to obtain a neighborhood node set; Step S400: Calculate the overlap degree of the correct problem-solving steps and the incorrect problem-solving steps with the neighborhood node set of the previous problem-solving steps in the topological graph, calculate the difference value of the overlap degree, and obtain the knowledge review valuation of the knowledge point involved in the first target step; Step S500: Obtain the knowledge points involved in the first target step and the test taker's two test-taking records including the knowledge points, compare the knowledge review valuations of the two test-taking records, and when the knowledge review valuation of the second test-taking record is greater than the knowledge review valuation of the first test-taking record, push exercises involving the knowledge points to the relevant test taker.

[0006] Furthermore, step S100 includes: Step S101: Visualize the knowledge graph, obtain the topological structure of the knowledge graph, and correspond a part of the knowledge content of the subject to several knowledge nodes in the topological structure; Step S102: Collect the correct solution steps of the exercises and the analysis of the answers to the corresponding knowledge points, collect the incorrect solution steps and the analysis of the wrong questions of the corresponding knowledge points, obtain the characteristic content of the analysis of the answers and the wrong questions, the characteristic content includes characteristic words and characteristic images, mark the knowledge points involved in the characteristic content, collect the knowledge points involved in each characteristic content respectively, and establish a knowledge point data set corresponding to each characteristic content; Step S103: Acquire the knowledge points corresponding to the characteristic contents in the answer analysis of the exercise, collect the nodes corresponding to the knowledge points, and obtain the exercise node set of the exercise.

[0007] Further, step S200 includes: Step S201: setting unit step areas, dividing the answer content into a number of unit step areas, and arranging the unit step areas according to the logical order of answering questions, each unit step area corresponding to a correct step; Step S202: taking a wrong question of a certain student as a first target question, obtaining the answer content of the certain student in answering the first target question, dividing the answer content into a plurality of unit step areas, each unit step area corresponding to a problem-solving step, obtaining all correct steps of the first target question, and according to the order of the correct steps, corresponding the problem-solving steps to the correct steps according to the answer logic order of the first target question; Step S203: Obtain the evaluation results of the teacher on the answer content, take the first wrong unit step in the answer content as the first target step, the step before the first target step as the second target step, and the correct step corresponding to the first target step as the third target step.

[0008] Furthermore, step S300 includes: Step S301: Identify the feature content in the first target step, obtain the knowledge points included in the feature content, obtain the nodes corresponding to the knowledge points in the knowledge graph, collect the nodes to obtain the first knowledge node set A1, and collect all d-order neighbor nodes of the first knowledge node set A1 in the knowledge graph to the first neighborhood node set K1, wherein all d-order neighbor sets include 1st-order neighbor nodes, 2nd-order neighbor nodes, 3rd-order neighbor nodes, ... and the union of d-order neighbor nodes, d is a positive integer; Step S302: Identify the feature content in the second target step, obtain the knowledge points included in the feature content, obtain the nodes corresponding to the knowledge points in the knowledge graph, aggregate the nodes to obtain a second knowledge node set A2, and aggregate all d-order neighbor nodes of the second knowledge node set A2 in the knowledge graph into a second neighborhood node set K2; Step S303: Identify the characteristic content in the third target step, obtain the knowledge points included in the characteristic content, obtain the nodes corresponding to the knowledge points in the knowledge graph, aggregate the nodes to obtain a third knowledge node set A3, and aggregate all d-order neighbor nodes of the third knowledge node set A3 in the knowledge graph to the third neighborhood node set K3.

[0009] Furthermore, step S400 includes: Step S401: Calculate the relationship evaluation index L between the first target step and the second target step 12 , calculate the relationship evaluation index L between the second target step and the third target step 23 and calculate the relationship evaluation index L between the first target step and the third target step 13 ; Step S402: taking the union of the first knowledge node set A1, the second knowledge node set A2 and the third knowledge node set A3 as the first target node set U1; Step S403: Calculate the knowledge review estimate m1 of the target knowledge point set. , associate the knowledge review valuation with the first target step.

[0010] Furthermore, step S400 includes: Step S4-1: taking the union F12 of the first neighborhood node set K1 and the second neighborhood node set K2 as the first reference set, obtaining any two nodes in the first reference set, which are denoted as ei and ej respectively; Step S4-2: Obtain the degree of node ei and the degree of node ej from the knowledge graph. When there is an edge between node ei and node ej, the relationship coefficient between node ei and node ej is: α ij =1; when there is no edge between node ei and node ej, obtain the first-order neighbor nodes of ei in the knowledge graph, merge the first-order neighbor nodes into the set Ni, obtain the first-order neighbor nodes of ej in the knowledge graph, merge the first-order neighbor nodes into the set Nj, and calculate the association coefficient α between node ei and node ej ij , , where the num function represents the number of knowledge points in the calculation set; Step S4-3: Calculate the relationship evaluation index L between the second target step and the first target step 12 , ; Step S4-4: taking the union Q23 of the third neighborhood node set K3 and the second neighborhood node set K2 as the second reference set, obtaining any two nodes in the second reference set, which are respectively denoted as fu and fv; Step S4-5: Repeat the method in step S4-1 to step S4-3 to calculate the relationship evaluation index L between the second target step and the third target step 23 ; Step S4-6: taking the union F13 of the first neighborhood node set K1 and the third neighborhood node set K3 as a third reference set, obtaining any two nodes in the third reference set, which are respectively denoted as px and py; Step S4-7: Obtain the degree of node px and the degree of node py from the knowledge graph respectively. When there is an edge between node px and node py, the relationship coefficient between node px and node py is: β xy =1; when there is no edge between node px and node py, obtain the first-order neighbor nodes of px in the knowledge graph, merge the first-order neighbor nodes into the set Nx, obtain the first-order neighbor nodes of py in the knowledge graph, merge the first-order neighbor nodes into the set Ny, and calculate the correlation coefficient β between node px and node py xy , , where the num function represents the number of knowledge points in the calculation set; Step S4-8: Calculate the relationship evaluation index L between the first target step and the third target step 13 , .

[0011] Furthermore, step S500 includes: Step S501: obtaining a first target node set U1, taking all exercise node sets including the exercises of the first target node set U1 as candidate exercises, and recording the rth wrong exercise of the student after the first target exercise as the second target exercise; Step S502: Obtain the first target step, the second target step and the third target step of the second target exercise, respectively corresponding to obtaining the first knowledge node set B1, the second knowledge node set B2 and the third knowledge node set B3 of the second target exercise, and use the union of the first knowledge node set B1, the second knowledge node set B2 and the third knowledge node set B3 as the second target node set U2; Step S503: When U2 meets the condition , calculate the knowledge review value m2 associated with the first target step of the second target exercise, and when m2≥m1, push the alternative exercise to the student. When m2<m1, the second target wrong question is used as the new first target wrong question, and the second target node set U2 is used as the new first target node set U1, and steps 200 to S500 are repeated until the incorrect problem-solving steps do not include the knowledge points corresponding to the nodes in the first target node set U1.

[0012] In order to better implement the above method, an intelligent adaptive learning exercise push system is also proposed, the system includes: The subject knowledge management module, the exercise management module, the neighborhood knowledge point management module, the knowledge review valuation calculation module and the exercise push module. The subject knowledge management module is used to manage the knowledge graph of the subject and the connection relationship between each knowledge point in the knowledge graph. The exercise management module is used to obtain the correction records of the exercises and obtain the wrong problem-solving steps and the corresponding correct problem-solving steps. The neighborhood knowledge point management module is used to obtain the neighborhood knowledge points of the knowledge points of the problem-solving steps and manage the knowledge points included in each problem-solving step separately. The knowledge review valuation calculation module is used to calculate the relationship evaluation index of the problem-solving steps according to the overlap degree of the neighborhood knowledge points in the knowledge graph. The exercise push module is used to evaluate the wrong problem-solving steps and push exercises including the same knowledge points as the problem-solving steps to the relevant question-takers. Further, the neighborhood knowledge point management module includes: a first neighborhood knowledge point set management unit, a second neighborhood knowledge point set management unit and a third neighborhood knowledge point set management unit, wherein the first neighborhood knowledge point set management unit is used to manage the neighborhood knowledge point set of the first target step, the second neighborhood knowledge point set management unit is used to manage the neighborhood knowledge point set of the second target step, and the third neighborhood knowledge point set management unit is used to manage the neighborhood knowledge point set of the third target step; Further, the knowledge review valuation calculation module includes: a first relationship evaluation index calculation unit, a second relationship evaluation index calculation unit, a third relationship evaluation index calculation unit and a knowledge review valuation calculation unit, wherein the first relationship evaluation index calculation unit is used to calculate the relationship evaluation index between the first target step and the second target step, the second relationship evaluation index calculation unit is used to calculate the relationship evaluation index between the second target step and the third target step, the third relationship evaluation index calculation unit is used to calculate the relationship evaluation index between the first target step and the third target step, and the knowledge review valuation calculation unit is used to calculate the knowledge review valuation of the target knowledge point set; Furthermore, the exercise push module includes: a knowledge comment valuation comparison unit, a loop control unit and an exercise push unit, wherein the knowledge comment valuation comparison unit is used to compare knowledge comment valuations, the loop control unit is used to determine whether to enter a loop based on a judgment condition, and to adjust the first target wrong question, and the exercise push unit is used to push exercises to the exercise taker.

[0013] Compared with the prior art, the beneficial effects of the present invention are: by analyzing the connection between the steps of the exercises through the knowledge graph, analyzing the difference between the wrong steps and the correct steps, so as to record the knowledge content mastered by the learners and the knowledge content not mastered by the learners, and accurately characterize the learning situation of the learners. Further determine the next learning direction of the learners. Push relevant learning content step by step according to the learning situation of the learners, and dynamically adjust the pushed learning content to meet the learning requirements of the learners. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a structural schematic diagram of an intelligent adaptive learning exercise push system of the present invention; Figure 2 The present invention is a flowchart of an intelligent adaptive learning exercise push method. DETAILED DESCRIPTION

[0015] 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.

[0016] Example: Figure 1 and Figure 2 As shown, the present invention provides a technical solution, a method for pushing intelligent adaptive learning exercises: Step S100: Collect knowledge points in the subject knowledge content, mark the knowledge points and the connections between them, generate a knowledge map of the subject through mind mapping software, collect the knowledge points involved in the problem-solving process, and establish a corresponding relationship between the problem and the nodes in the knowledge map; Wherein, step S100 includes: Step S101: Visualize the knowledge graph, obtain the topological structure of the knowledge graph, and correspond a part of the knowledge content of the subject to several knowledge nodes in the topological structure; Step S102: Collect the correct solution steps of the exercises and the analysis of the answers to the corresponding knowledge points, collect the incorrect solution steps and the analysis of the wrong questions of the corresponding knowledge points, obtain the characteristic content of the analysis of the answers and the wrong questions, the characteristic content includes characteristic words and characteristic images, mark the knowledge points involved in the characteristic content, collect the knowledge points involved in each characteristic content respectively, and establish a knowledge point data set corresponding to each characteristic content; Step S103: Acquire the knowledge points corresponding to the characteristic contents in the answer analysis of the exercise, collect the nodes corresponding to the knowledge points, and obtain the exercise node set of the exercise; In an embodiment, textbooks, academic papers, online resources, etc. of relevant subjects may be retrieved and obtained, and basic concepts, definitions, theorems, and formulas of the subject may be used as feature content. For example, characters, names, and terms in definitions and theorems may be collected as feature words, and image information of the formula may be collected as a feature image based on the formula expression of the theorem.

[0017] Step S200: Divide the problem-solving process of the exercise into several steps, obtain the wrong problem-solving steps submitted by the exerciser, and the correct problem-solving steps corresponding to the wrong problem-solving steps; Wherein, step S200 includes: Step S201: setting unit step areas, dividing the answer content into a number of unit step areas, and arranging the unit step areas according to the logical order of answering questions, each unit step area corresponding to a correct step; Step S202: taking a wrong question of a certain student as a first target question, obtaining the answer content of the certain student in answering the first target question, dividing the answer content into a plurality of unit step areas, each unit step area corresponding to a problem-solving step, obtaining all correct steps of the first target question, and according to the order of the correct steps, corresponding the problem-solving steps to the correct steps according to the answer logic order of the first target question; Step S203: Obtain the evaluation result of the teacher on the answer content, take the first wrong unit step in the answer content as the first target step, take the step before the first target step as the second target step, and take the correct step corresponding to the first target step as the third target step; The second target step is the correct problem-solving step. In the embodiment, if it is a multiple-choice question or a fill-in-the-blank question, the content of the question stem is used as the first target step for solving the problem; For example, in English exercises, multiple-choice questions usually have four options: "A", "B", "C" and "D". Each option represents the test taker's use of different knowledge points. By comparing the differences in knowledge points contained in the wrong options and the correct options, the knowledge content that the test taker has overlooked can be obtained.

[0018] Step S300: respectively obtaining the correct problem-solving steps and the incorrect problem-solving steps, and the corresponding nodes of the knowledge points involved in the previous problem-solving steps in the knowledge graph, and collecting all nodes within the neighborhood of the corresponding nodes to obtain a neighborhood node set; Wherein, step S300 includes: Step S301: Identify the feature content in the first target step, obtain the knowledge points included in the feature content, obtain the nodes corresponding to the knowledge points in the knowledge graph, collect the nodes to obtain the first knowledge node set A1, and collect all d-order neighbor nodes of the first knowledge node set A1 in the knowledge graph to the first neighborhood node set K1, wherein all d-order neighbor sets include 1st-order neighbor nodes, 2nd-order neighbor nodes, 3rd-order neighbor nodes, ... and the union of d-order neighbor nodes, d is a positive integer; Step S302: Identify the feature content in the second target step, obtain the knowledge points included in the feature content, obtain the nodes corresponding to the knowledge points in the knowledge graph, aggregate the nodes to obtain a second knowledge node set A2, and aggregate all d-order neighbor nodes of the second knowledge node set A2 in the knowledge graph into a second neighborhood node set K2; Step S303: Identify the characteristic content in the third target step, obtain the knowledge points included in the characteristic content, obtain the nodes corresponding to the knowledge points in the knowledge graph, aggregate the nodes to obtain a third knowledge node set A3, and aggregate all d-order neighbor nodes of the third knowledge node set A3 in the knowledge graph to the third neighborhood node set K3.

[0019] Step S400: Calculate the overlap degree of the correct problem-solving steps and the incorrect problem-solving steps with the neighborhood node set of the previous problem-solving steps in the topological graph, calculate the difference value of the overlap degree, and obtain the knowledge review valuation of the knowledge point involved in the first target step; Wherein, step S400 includes: Step S401: Calculate the relationship evaluation index L between the first target step and the second target step 12 , calculate the relationship evaluation index L between the second target step and the third target step 23 and calculate the relationship evaluation index L between the first target step and the third target step 13 ; Wherein, step S401 includes: Step S4-1: taking the union F12 of the first neighborhood node set K1 and the second neighborhood node set K2 as the first reference set, obtaining any two nodes in the first reference set, which are denoted as ei and ej respectively; Step S4-2: Obtain the degree of node ei and the degree of node ej from the knowledge graph. When there is an edge between node ei and node ej, the relationship coefficient between node ei and node ej is: α ij =1; when there is no edge between node ei and node ej, obtain the first-order neighbor nodes of ei in the knowledge graph, merge the first-order neighbor nodes into the set Ni, obtain the first-order neighbor nodes of ej in the knowledge graph, merge the first-order neighbor nodes into the set Nj, and calculate the association coefficient α between node ei and node ejij , , where the num function represents the number of knowledge points in the calculation set; The order of a neighbor node usually refers to the degree of a node, which represents the number of edges directly connected to the node in the knowledge graph; In the Jaccard index, Ni∩Nj represents the intersection set of the neighbor nodes of the two nodes ei and ej, and Ni∪Nj represents the union set of the neighbor nodes of the two nodes ei and ej, so the relationship coefficient of the node takes a value between 0 and 1; Step S4-3: Calculate the relationship evaluation index L between the second target step and the first target step 12 , ; Step S4-4: taking the union Q23 of the third neighborhood node set K3 and the second neighborhood node set K2 as the second reference set, obtaining any two nodes in the second reference set, which are respectively denoted as fu and fv; Step S4-5: Obtain the degree of node fu and the degree of node fv from the knowledge graph respectively. When there is an edge between node fu and node fv, the relationship coefficient between node fu and node fv is: γ uv =1; when there is no edge between node fu and node fv, obtain the first-order neighbor nodes of fu in the knowledge graph, merge the first-order neighbor nodes into the set Nu, obtain the first-order neighbor nodes of fv in the knowledge graph, merge the first-order neighbor nodes into the set Nv, and calculate the association coefficient γ between node fu and node fv uv , , where the num function represents the number of knowledge points in the calculation set; Calculate the relationship evaluation index L between the second target step and the third target step 23 , ; Step S4-6: taking the union F13 of the first neighborhood node set K1 and the third neighborhood node set K3 as a third reference set, obtaining any two nodes in the third reference set, which are respectively denoted as px and py; Step S4-7: Obtain the degree of node px and the degree of node py from the knowledge graph respectively. When there is an edge between node px and node py, the relationship coefficient between node px and node py is: β xy =1; when there is no edge between node px and node py, obtain the first-order neighbor nodes of px in the knowledge graph, merge the first-order neighbor nodes into the set Nx, obtain the first-order neighbor nodes of py in the knowledge graph, merge the first-order neighbor nodes into the set Ny, and calculate the correlation coefficient β between node px and node py xy , , where the num function represents the number of knowledge points in the calculation set; Step S4-8: Calculate the relationship evaluation index L between the first target step and the third target step 13 , ; Step S402: taking the union of the first knowledge node set A1, the second knowledge node set A2 and the third knowledge node set A3 as the first target node set U1; Step S403: Calculate the knowledge review estimate m1 of the target knowledge point set. , associate the knowledge review valuation with the first target step.

[0020] Step S500: obtaining the knowledge points involved in the first target step and the two test-taking records of the test-taker including the knowledge points, comparing the knowledge review valuations of the two test-taking records, and when the knowledge review valuation of the second test-taking record is greater than the knowledge review valuation of the first test-taking record, pushing exercises involving the knowledge points to the relevant test-taker; Wherein, step S500 includes: Step S501: obtaining a first target node set U1, taking all exercise node sets including the exercises of the first target node set U1 as candidate exercises, and recording the rth wrong exercise of the student after the first target exercise as the second target exercise; Step S502: Obtain the first target step, the second target step and the third target step of the second target exercise, respectively corresponding to obtaining the first knowledge node set B1, the second knowledge node set B2 and the third knowledge node set B3 of the second target exercise, and use the union of the first knowledge node set B1, the second knowledge node set B2 and the third knowledge node set B3 as the second target node set U2; Step S503: When U2 meets the condition , calculate the knowledge review value m2 associated with the first target step of the second target exercise, and when m2≥m1, push the alternative exercise to the student. When m2<m1, the second target wrong question is used as the new first target wrong question, and the second target node set U2 is used as the new first target node set U1, and steps 200 to S500 are repeated until the incorrect problem-solving steps do not include the knowledge points corresponding to the nodes in the first target node set U1.

[0021] The system includes: subject knowledge management module, exercise management module, neighborhood knowledge point management module, knowledge review valuation calculation module and exercise push module; Among them, the subject knowledge management module is used to manage the knowledge graph of the subject and the connection relationship between each knowledge point in the knowledge graph. The exercise management module is used to obtain the correction records of exercises, obtain the wrong problem-solving steps and the corresponding correct problem-solving steps, Wherein, the neighborhood knowledge point management module is used to obtain the neighborhood knowledge points of the knowledge points of the problem-solving steps, and manage the knowledge points included in each problem-solving step separately, wherein the neighborhood knowledge point management module includes: a first neighborhood knowledge point set management unit, a second neighborhood knowledge point set management unit and a third neighborhood knowledge point set management unit, wherein the first neighborhood knowledge point set management unit is used to manage the neighborhood knowledge point set of the first target step, the second neighborhood knowledge point set management unit is used to manage the neighborhood knowledge point set of the second target step, and the third neighborhood knowledge point set management unit is used to manage the neighborhood knowledge point set of the third target step; Wherein, the knowledge comment valuation calculation module is used to calculate the relationship evaluation index of the problem-solving steps according to the degree of overlap of the neighborhood knowledge points in the knowledge graph, wherein the knowledge comment valuation calculation module includes: a first relationship evaluation index calculation unit, a second relationship evaluation index calculation unit, a third relationship evaluation index calculation unit and a knowledge comment valuation calculation unit, wherein the first relationship evaluation index calculation unit is used to calculate the relationship evaluation index between the first target step and the second target step, the second relationship evaluation index calculation unit is used to calculate the relationship evaluation index between the second target step and the third target step, the third relationship evaluation index calculation unit is used to calculate the relationship evaluation index between the first target step and the third target step, and the knowledge comment valuation calculation unit is used to calculate the knowledge comment valuation of the target knowledge point set; Among them, the exercise push module is used to evaluate incorrect problem-solving steps and push exercises that include the same knowledge points as the problem-solving steps to relevant exercise takers. Among them, the exercise push module includes: a knowledge comment valuation comparison unit, a loop control unit and an exercise push unit. Among them, the knowledge comment valuation comparison unit is used to compare knowledge comment valuations, the loop control unit is used to determine whether to enter the loop based on the judgment conditions, and adjust the first target wrong questions. The exercise push unit is used to push exercises to exercise takers.

[0022] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. Connections between knowledge points A method for pushing intelligent adaptive learning exercises, characterized by: Methods include: Step S100: Collect knowledge points in the subject knowledge content, mark the knowledge points and the connections between them, generate a knowledge map of the subject through mind mapping software, collect the knowledge points involved in the problem-solving process, and establish a corresponding relationship between the problem and the nodes in the knowledge map; Step S200: Divide the problem-solving process of the exercise into several steps, obtain the wrong problem-solving steps submitted by the exerciser, and the correct problem-solving steps corresponding to the wrong problem-solving steps; Step S300: respectively obtaining the correct problem-solving steps and the incorrect problem-solving steps, and the corresponding nodes of the knowledge points involved in the previous problem-solving steps in the knowledge graph, and collecting all nodes within the neighborhood of the corresponding nodes to obtain a neighborhood node set; Step S400: Calculate the overlap degree of the correct problem-solving steps and the incorrect problem-solving steps with the neighborhood node set of the previous problem-solving steps in the topological graph, calculate the difference value of the overlap degree, and obtain the knowledge review valuation of the knowledge point involved in the first target step; Step S500: Obtain the knowledge points involved in the first target step and the test taker's two test-taking records including the knowledge points, compare the knowledge review valuations of the two test-taking records, and when the knowledge review valuation of the second test-taking record is greater than the knowledge review valuation of the first test-taking record, push exercises involving the knowledge points to the relevant test taker.

2. The method for pushing intelligent adaptive learning exercises according to claim 1, characterized in that: Step S100 includes: Step S101: Visualize the knowledge graph, obtain the topological structure of the knowledge graph, and correspond a part of the knowledge content of the subject to several knowledge nodes in the topological structure; Step S102: Collect the correct solution steps of the exercises and the analysis of the answers to the corresponding knowledge points, collect the incorrect solution steps and the analysis of the wrong questions of the corresponding knowledge points, obtain the characteristic content of the analysis of the answers and the wrong questions, the characteristic content includes characteristic words and characteristic images, mark the knowledge points involved in the characteristic content, collect the knowledge points involved in each characteristic content respectively, and establish a knowledge point data set corresponding to each characteristic content; Step S103: Acquire the knowledge points corresponding to the characteristic contents in the answer analysis of the exercise, collect the nodes corresponding to the knowledge points, and obtain the exercise node set of the exercise.

3. The method for pushing intelligent adaptive learning exercises according to claim 2, characterized in that: Step S200 includes: Step S201: setting unit step areas, dividing the answer content into a number of unit step areas, and arranging the unit step areas according to the logical order of answering questions, each unit step area corresponding to a correct step; Step S202: taking a wrong question of a certain student as a first target question, obtaining the answer content of the certain student in answering the first target question, dividing the answer content into a plurality of unit step areas, each unit step area corresponding to a problem-solving step, obtaining all correct steps of the first target question, and according to the order of the correct steps, corresponding the problem-solving steps to the correct steps according to the answer logic order of the first target question; Step S203: Obtain the evaluation results of the teacher on the answer content, take the first wrong unit step in the answer content as the first target step, the step before the first target step as the second target step, and the correct step corresponding to the first target step as the third target step.

4. The method for pushing intelligent adaptive learning exercises according to claim 3, characterized in that: Step S300 includes: Step S301: Identify the feature content in the first target step, obtain the knowledge points included in the feature content, obtain the nodes corresponding to the knowledge points in the knowledge graph, collect the nodes to obtain the first knowledge node set A1, and collect all d-order neighbor nodes of the first knowledge node set A1 in the knowledge graph to the first neighborhood node set K1, wherein all d-order neighbor sets include 1st-order neighbor nodes, 2nd-order neighbor nodes, 3rd-order neighbor nodes, ... and the union of d-order neighbor nodes, d is a positive integer; Step S302: Identify the feature content in the second target step, obtain the knowledge points included in the feature content, obtain the nodes corresponding to the knowledge points in the knowledge graph, aggregate the nodes to obtain a second knowledge node set A2, and aggregate all d-order neighbor nodes of the second knowledge node set A2 in the knowledge graph into a second neighborhood node set K2; Step S303: Identify the characteristic content in the third target step, obtain the knowledge points included in the characteristic content, obtain the nodes corresponding to the knowledge points in the knowledge graph, aggregate the nodes to obtain a third knowledge node set A3, and aggregate all d-order neighbor nodes of the third knowledge node set A3 in the knowledge graph to the third neighborhood node set K3.

5. The method for pushing intelligent adaptive learning exercises according to claim 4, characterized in that: Step S400 include Step S401: Calculate the relationship evaluation index L between the first target step and the second target step 12 , calculate the relationship evaluation index L between the second target step and the third target step 23 and calculate the relationship evaluation index L between the first target step and the third target step 13 ; Step S402: taking the union of the first knowledge node set A1, the second knowledge node set A2 and the third knowledge node set A3 as the first target node set U1; Step S403: Calculate the knowledge review estimate m1 of the target knowledge point set. , associate the knowledge review valuation with the first target step.

6. The method for pushing intelligent adaptive learning exercises according to claim 5, characterized in that: The method for calculating the relationship evaluation index between the first target step and the second target step in step S400 includes: Step S4-1: taking the union F12 of the first neighborhood node set K1 and the second neighborhood node set K2 as the first reference set, obtaining any two nodes in the first reference set, which are denoted as ei and ej respectively; Step S4-2: Obtain the degree of node ei and the degree of node ej from the knowledge graph. When there is an edge between node ei and node ej, the relationship coefficient between node ei and node ej is: α ij =1; when there is no edge between node ei and node ej, obtain the first-order neighbor node of ei in the knowledge graph, merge the first-order neighbor node into the set Ni, obtain the first-order neighbor node of ej in the knowledge graph, merge the first-order neighbor node into the set Nj, and calculate the association coefficient α between node ei and node ej ij , , where the num function represents the number of nodes in the calculation set; Step S4-3: Calculate the relationship evaluation index L between the second target step and the first target step 12 , .

7. The method for pushing intelligent adaptive learning exercises according to claim 6, characterized in that: The method for calculating the relationship evaluation index between the second target step and the third target step in step S400 includes: Step S4-4: taking the union Q23 of the third neighborhood node set K3 and the second neighborhood node set K2 as the second reference set, obtaining any two nodes in the second reference set, which are respectively denoted as fu and fv; Step S4-5: Repeat the method in step S4-1 to step S4-3 to calculate the relationship evaluation index L between the second target step and the third target step 23 .

8. The method for pushing intelligent adaptive learning exercises according to claim 5, characterized in that: The method for calculating the relationship evaluation index between the first target step and the third target step in step S400 includes: Step S4-6: taking the union F13 of the first neighborhood node set K1 and the third neighborhood node set K3 as a third reference set, obtaining any two nodes in the third reference set, which are respectively denoted as px and py; Step S4-7: Obtain the degree of node px and the degree of node py from the knowledge graph respectively. When there is an edge between node px and node py, the relationship coefficient between node px and node py is: β xy =1; when there is no edge between node px and node py, obtain the first-order neighbor node of px in the knowledge graph, merge the first-order neighbor node into the set Nx, obtain the first-order neighbor node of py in the knowledge graph, merge the first-order neighbor node into the set Ny, and calculate the correlation coefficient β between node px and node py xy , , where the num function represents the number of nodes in the calculation set; Step S4-8: Calculate the relationship evaluation index L between the first target step and the third target step 13 , .

9. The method for pushing intelligent adaptive learning exercises according to claim 5, characterized in that: Step S500 includes: Step S501: obtaining a first target node set U1, taking all exercise node sets including the exercises of the first target node set U1 as candidate exercises, and recording the rth wrong exercise of the student after the first target exercise as the second target exercise; Step S502: Obtain the first target step, the second target step and the third target step of the second target exercise, respectively corresponding to obtaining the first knowledge node set B1, the second knowledge node set B2 and the third knowledge node set B3 of the second target exercise, and use the union of the first knowledge node set B1, the second knowledge node set B2 and the third knowledge node set B3 as the second target node set U2; Step S503: When U2 meets the condition , calculate the knowledge review value m2 associated with the first target step of the second target exercise, and when m2≥m1, push the alternative exercise to the student. When m2<m1, the second target wrong question is used as the new first target wrong question, and the second target node set U2 is used as the new first target node set U1, and steps 200 to S500 are repeated until the incorrect problem-solving steps do not include the knowledge points corresponding to the nodes in the first target node set U1.

10. An intelligent adaptive learning exercise push system, used to execute an intelligent adaptive learning exercise push method according to any one of claims 1 to 9, characterized in that: The system includes: The subject knowledge management module, the exercise management module, the neighborhood knowledge point management module, the knowledge review valuation calculation module and the exercise push module. The subject knowledge management module is used to manage the knowledge graph of the subject and the connection relationship between each knowledge point in the knowledge graph. The exercise management module is used to obtain the correction records of the exercises and obtain the wrong problem-solving steps and the corresponding correct problem-solving steps. The neighborhood knowledge point management module is used to obtain the neighborhood knowledge points of the knowledge points of the problem-solving steps and manage the knowledge points included in each problem-solving step separately. The knowledge review valuation calculation module is used to calculate the relationship evaluation index of the problem-solving steps according to the overlap degree of the neighborhood knowledge points in the knowledge graph. The exercise push module is used to evaluate the wrong problem-solving steps and push exercises including the same knowledge points as the problem-solving steps to the relevant question-takers. The neighborhood knowledge point management module includes: a first neighborhood knowledge point set management unit, a second neighborhood knowledge point set management unit and a third neighborhood knowledge point set management unit, wherein the first neighborhood knowledge point set management unit is used to manage the neighborhood knowledge point set of the first target step, the second neighborhood knowledge point set management unit is used to manage the neighborhood knowledge point set of the second target step, and the third neighborhood knowledge point set management unit is used to manage the neighborhood knowledge point set of the third target step; The knowledge review valuation calculation module includes: a first relationship evaluation index calculation unit, a second relationship evaluation index calculation unit, a third relationship evaluation index calculation unit and a knowledge review valuation calculation unit, wherein the first relationship evaluation index calculation unit is used to calculate the relationship evaluation index between the first target step and the second target step, the second relationship evaluation index calculation unit is used to calculate the relationship evaluation index between the second target step and the third target step, the third relationship evaluation index calculation unit is used to calculate the relationship evaluation index between the first target step and the third target step, and the knowledge review valuation calculation unit is used to calculate the knowledge review valuation of the target knowledge point set; The exercise push module includes: a knowledge comment valuation comparison unit, a loop control unit and an exercise push unit, wherein the knowledge comment valuation comparison unit is used to compare the knowledge comment valuations, the loop control unit is used to determine whether to enter a loop based on a judgment condition and adjust the first target wrong question, and the exercise push unit is used to push exercises to the exercise taker.

Images