A smart teaching method and system based on NLP and knowledge graph

Abstract

The application discloses a kind of wisdom teaching method based on NLP and knowledge graph, it is related to wisdom teaching field, and knowledge graph is constructed by NLP natural language processing technology, and the teaching material and evaluation material corresponding to each knowledge point entity in knowledge graph are acquired, the difficulty coefficient of the knowledge point entity corresponding to teaching material and evaluation material of marking evaluation material is obtained, and the target knowledge graph containing teaching material and evaluation material is obtained, the user skill score corresponding to each knowledge point entity in the process of learning teaching material and evaluation based on target learning path is acquired, to master the learning situation of student to each knowledge point entity in real time, and the recommendation frequency corresponding to knowledge point entity is acquired according to the user skill score and target skill score corresponding to knowledge point entity, according to the recommendation frequency corresponding to each knowledge point entity, adjust the current learning path and the evaluation material of knowledge point entity in current learning path, the individualized teaching to different students is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of smart teaching, and in particular to a smart teaching method and system based on NLP and knowledge graph. BACKGROUND

[0002] Domestic online education has a development history of several years, and the development speed is very fast, but online education only moves the offline resources to the online, and the technology still does not change the essence or mode of online education. Excellent teachers usually adjust the teaching content and teaching method according to the overall learning situation of the class, and teach students according to their aptitude. However, general teachers cannot accurately evaluate the learning situation of the students in the class, so they carry out the same teaching without differentiation, which causes the students with poor grades to learn nothing, and the students with good grades to learn nothing, and the overall learning performance of the students in the class cannot be improved.

[0003] Teachers usually need to spend a lot of time preparing lessons and making test questions, and cannot focus on teaching content. If the teacher has less experience, the preparation of test questions becomes the bottleneck of teaching. At present, there are many institutions doing smart education in China, but almost all products are aimed at extracurricular tutoring of students, and cannot be combined with existing traditional school education. Therefore, the main problems of the current online education include:

[0004] 1. Teachers cannot evaluate the learning situation of students;

[0005] 2. Teachers spend a lot of time preparing teaching content;

[0006] 3. The existing smart education products cannot be combined with traditional education. SUMMARY

[0007] In order to make individualized teaching plan for different students and master the learning situation of students in real time, the present application provides a smart teaching method based on NLP and knowledge graph, which comprises:

[0008] Text data of each subject is crawled from the network by a crawler software;

[0009] Knowledge point entities and the leading relationship between the knowledge point entities corresponding to each subject are extracted from the crawled text data by NLP natural language processing technology; the knowledge point entities include knowledge point name and knowledge point definition;

[0010] The knowledge system corresponding to each subject is constructed by the leading relationship between the knowledge point entities, and the knowledge graph is constructed by the association relationship between the knowledge systems corresponding to each subject;

[0011] A knowledge point entity in a knowledge graph is taken as a skill dimension, a skill dimension score corresponding to each knowledge point entity is set as a target skill score, and teaching materials and evaluation materials corresponding to each knowledge point entity in the knowledge graph are obtained, the knowledge point entity corresponding to the teaching materials and the evaluation materials is marked, and the difficulty coefficient of the evaluation materials is obtained to obtain a target knowledge graph containing the teaching materials and the evaluation materials; the teaching materials include teaching videos, and the evaluation materials include evaluation questions and scores of the evaluation questions;

[0012] An initial learning path is generated according to the lead relationships of the knowledge point entities in the target knowledge graph, and a skill score obtained by a user based on the target learning path in the process of learning the teaching materials and the evaluation materials, that is, a user skill score corresponding to each knowledge point entity, is obtained; the evaluation materials with a preset difficulty coefficient are set for each knowledge point entity in the initial learning path;

[0013] A recommendation frequency corresponding to a knowledge point entity is obtained according to the user skill score and the target skill score of the knowledge point entity, and an optimized learning path is obtained by adjusting the current learning path and the evaluation materials corresponding to the knowledge point entities in the current learning path according to the recommendation frequency corresponding to each knowledge point entity; the target learning path is the initial learning path or the optimized learning path; the user skill score corresponding to the knowledge point entity is in a proportional relationship with the difficulty coefficient of the evaluation materials corresponding to the knowledge point entity in the adjustment process.

[0014] Further, the text data of each subject is obtained from the network by using a crawler software, including:

[0015] S1: sending a subject data request to a target learning website through an http library;

[0016] S2: obtaining response content corresponding to the subject data request;

[0017] S3: obtaining text data corresponding to each subject by analyzing the response content through an analysis module; the analysis module includes a regular expression, a third-party analysis library and a json module.

[0018] Further, the skill score obtained by the user based on the target learning path in the process of learning the teaching materials and the evaluation materials, that is, the user skill score corresponding to each knowledge point entity, is obtained, and specifically includes:

[0019] Obtaining a user learning time and a user evaluation score of the teaching materials corresponding to each knowledge point entity in the process of learning the teaching materials and the evaluation materials;

[0020] The user skill score corresponding to the knowledge point entity is obtained by using a preset skill score calculation formula according to the user learning time and the user evaluation score of the teaching materials corresponding to the knowledge point entity;

[0021] The preset skill score calculation formula is:

[0022] Score = (T 学 / T 总 * 20% + E 测 / E 总 * 80%) * Memory amount

[0023] In the formula, the memory amount is obtained according to the Ebbinghaus memory forgetting curve; Score represents a user skill score, T 学 represents a user learning duration of teaching data corresponding to a knowledge point entity, T 总 represents a total duration of teaching data corresponding to the knowledge point entity, E 测 represents a user evaluation score corresponding to the knowledge point entity, E 总 represents an evaluation total score of evaluation data corresponding to the knowledge point entity.

[0024] Further, the acquisition formula for obtaining the recommendation frequency of the knowledge point entity is:

[0025]

[0026] In the formula, i = 1, 2,..., n; n represents a total number of knowledge point entities included in a target learning path; wherein Q A represents a recommendation frequency of an A knowledge point entity, T A is a target skill score of the A knowledge point entity, C A is a user skill score corresponding to the A knowledge point entity, T i represents a target skill score of an i-th knowledge point entity, C i represents a user skill score corresponding to the i-th knowledge point entity.

[0027] The application further provides a wisdom teaching system based on NLP and a knowledge graph, comprising:

[0028] A subject data acquisition module is configured to crawl text data of various subjects from a network through a crawler software;

[0029] A knowledge graph construction module is configured to extract knowledge point entities and leading relationships between the knowledge point entities corresponding to various subjects from the crawled text data through NLP natural language processing technology; the knowledge point entities include knowledge point names and definitions of the knowledge points; a knowledge system corresponding to each subject is constructed through the leading relationships between the knowledge point entities, and a knowledge graph is constructed through the association relationships between the knowledge systems corresponding to various subjects;

[0030] The target knowledge graph acquisition module is configured to set a skill dimension for each knowledge point entity in the knowledge graph, set a skill dimension score corresponding to each knowledge point entity as a target skill score, and acquire teaching materials and evaluation materials corresponding to each knowledge point entity in the knowledge graph, identify the knowledge point entities corresponding to the teaching materials and the evaluation materials, and identify the difficulty coefficients of the evaluation materials to obtain a target knowledge graph containing the teaching materials and the evaluation materials; the teaching materials include teaching videos, and the evaluation materials include evaluation questions and scores of the evaluation questions;

[0031] The user skill score acquisition module is configured to generate an initial learning path according to the lead relationships of the knowledge point entities in the target knowledge graph, and acquire skill scores obtained by a user in a process of learning the teaching materials and the evaluation materials based on the target learning path, i.e., user skill scores corresponding to each knowledge point entity; each knowledge point entity in the initial learning path is provided with evaluation materials with a preset difficulty coefficient;

[0032] The path optimization module is configured to acquire a recommended frequency corresponding to each knowledge point entity according to the user skill scores and the target skill scores of the knowledge point entities, adjust a current learning path and evaluation materials corresponding to the knowledge point entities in the current learning path according to the recommended frequencies corresponding to the knowledge point entities, and obtain an optimized learning path; the target learning path is the initial learning path or the optimized learning path; the user skill score of the knowledge point entity is in a proportional relationship with a difficulty coefficient of the evaluation materials corresponding to the knowledge point entity in the adjustment process.

[0033] Further, the subject data acquisition module includes:

[0034] The request initiation unit is configured to send a subject data request to a target learning website through an http library.

[0035] The response acquisition unit is configured to acquire response content corresponding to the subject data request.

[0036] The analysis unit is configured to acquire text data corresponding to each subject by analyzing the response content through an analysis module; the analysis module includes a regular expression, a third-party analysis library, and a json module.

[0037] Further, the user skill scores obtained by the user in a process of learning the teaching materials and the evaluation materials based on the target learning path, i.e., the user skill scores corresponding to each knowledge point entity, are as follows:

[0038] The user learning time and the user evaluation scores of the teaching materials corresponding to each knowledge point entity in a process of learning the teaching materials and the evaluation materials are acquired.

[0039] The user skill scores corresponding to each knowledge point entity are acquired by using a preset skill score calculation formula according to the user learning time and the user evaluation scores of the teaching materials corresponding to the knowledge point entity.

[0040] The preset skill score calculation formula is:

[0041] Score=(T 学 / T 总 * 20%+ E 测 / E 总 * 80%)* memory amount;

[0042] In the formula, the memory amount is obtained according to the Ebbinghaus memory forgetting curve; Score represents the user skill score, T 学 represents the user learning time length of the teaching material corresponding to the knowledge point entity, T 总 represents the total time length of the teaching material corresponding to the knowledge point entity, E 测 represents the user evaluation score corresponding to the knowledge point entity, E 总 represents the total score of the evaluation material corresponding to the knowledge point entity.

[0043] Further, the acquisition formula for acquiring the recommendation frequency of the knowledge point entity is:

[0044]

[0045] In the formula, i=1, 2, ……n; n represents the total number of knowledge point entities included in the target learning path; wherein Q A represents the recommendation frequency of the A knowledge point entity, T A is the target skill score of the A knowledge point entity, C A is the user skill score corresponding to the A knowledge point entity, T i represents the target skill score of the i-th knowledge point entity, C i represents the user skill score corresponding to the i-th knowledge point entity.

[0046] Compared with the prior art, the present application has at least the following beneficial effects:

[0047] (1) The present application constructs a knowledge graph through NLP natural language processing technology, and obtains teaching materials and evaluation materials corresponding to each knowledge point entity in the knowledge graph, identifies the knowledge point entities corresponding to the teaching materials and the evaluation materials, and the difficulty coefficient of the evaluation materials to obtain a target knowledge graph containing the teaching materials and the evaluation materials, obtains the user skill score corresponding to each knowledge point entity in the process of learning the teaching materials and the evaluation materials based on the target learning path, thereby real-time mastering the learning situation of the students on each knowledge point entity, and obtaining the recommendation frequency corresponding to the knowledge point entity according to the user skill score and the target skill score of the knowledge point entity, adjusting the current learning path and the evaluation materials corresponding to the knowledge point entity in the current learning path according to the recommendation frequency corresponding to each knowledge point entity, thereby realizing real-time adjustment of the learning path corresponding to the students according to the learning situation of the students in the process of learning, continuously optimizing, realizing teaching of different learning paths for different students, that is, optimizing the current learning path according to the learning situation (user skill score) of the students, so that the students learn in the learning path most suitable for their current level, thereby achieving the optimal learning effect;

[0048] (2) The present application generates a learning path based on the constructed knowledge graph, so as to follow the gradual progress in traditional education, and realizes the combination with traditional education;

[0049] (3) The present application obtains the teaching materials and the evaluation materials corresponding to each knowledge point entity in the knowledge graph, identifies the knowledge point entities corresponding to the teaching materials and the evaluation materials, and the difficulty coefficient of the evaluation materials to obtain a target knowledge graph containing the teaching materials and the evaluation materials, thereby avoiding the problem that the teachers spend more time in preparing teaching contents (including teaching materials and evaluation materials). BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 It is a wisdom teaching method flow chart based on NLP and knowledge graph;

[0051] Figure 2 It is a wisdom teaching system module diagram based on NLP and knowledge graph;

[0052] Figure 3 It is a function related knowledge graph. DETAILED DESCRIPTION

[0053] The following is a specific embodiment of the present application and a further description of the technical scheme of the present application in combination with the drawings, but the present application is not limited to these embodiments.

[0054] Embodiment one

[0055] In order to make individualized teaching scheme for different students, real-time mastering the learning situation of the students, such as Figure 1As shown, the present application proposes a wisdom teaching method based on NLP and knowledge graph, comprising:

[0056] Text data of each subject is crawled from the network through a crawler software;

[0057] The text data of each subject is crawled from the network through a crawler software, comprising:

[0058] S1: sending subject data request Request to the target learning website through the http library; the target learning website includes Baidu, Google;

[0059] S2: obtaining the response content Response corresponding to the subject data request; the Response includes: html, json, picture, video, etc.;

[0060] S3: obtaining the text data corresponding to each subject by parsing the response content through the parsing module; the parsing module includes regular expression, third-party parsing library and json module.

[0061] Specifically, the html data is parsed by regular expression and third-party parsing library; the json data is parsed by json module.

[0062] The knowledge point entities and the leading relationships between the knowledge point entities (also including the dependency relationship, the containing relationship, the synonym relationship, the belonging relationship, etc. between the knowledge point entities) corresponding to each subject are extracted from the crawled text data through NLP natural language processing technology; the knowledge point entity includes knowledge point name and knowledge point definition;

[0063] The knowledge system corresponding to each subject is constructed through the leading relationship between the knowledge point entities, and the knowledge graph is constructed through the association relationship between the knowledge systems corresponding to each subject;

[0064] It needs to be explained that online individualized teaching can also be understood as personalized education, which is similar to personalized recommendation. For constructing a personalized education system, the following aspects of work need to be done, first, a complete knowledge point system needs to be constructed, each field will have its own knowledge points and the topological structure between the knowledge points, such as what to learn first and what to learn later, then these knowledge point systems need to be built in advance. When the knowledge system of each field (i.e. each subject) is built, the knowledge systems between different fields are artificially associated according to the association relationship between the knowledge systems corresponding to each subject, so that a very complete knowledge system, i.e. knowledge graph, can be obtained.

[0065] For example, Figure 3The knowledge graph related to functions in high school mathematics is shown (including function concept, basic properties, several transformations, basic elementary functions, piecewise functions, composite functions, abstract functions, functions and equations, and applications of functions). Each node is a knowledge point entity, and the knowledge point entities are linked by directional arrows, indicating that there is a prior knowledge relationship between two knowledge points, that is, to learn a knowledge point well, one must first learn a previous knowledge point well. For example, to learn basic elementary functions well, one must first learn function transformations and function properties, and to learn function properties well, one must first learn function concepts, etc. Of course, this is only a simple graph, and if the graph is expanded, there will be more nodes, and each knowledge point entity can be further split, such as the function concept, which can be further split into function definition, function three elements and function interval, and the function definition can be further split into list method, analytic method, and image method, etc.

[0066] This embodiment adopts a semi-automatic method to construct the knowledge graph. Semi-automatic means human-machine combination. First, important knowledge point entity concepts and dependencies (prior relationships) between knowledge point entity concepts are extracted from massive domain texts through artificial intelligence (crawler + NLP) to construct a preliminary knowledge graph, and then the knowledge graph is corrected manually. The advantage of this approach is that it can partially liberate labor costs and may find knowledge points that humans have not thought of. And manual correction ensures the accuracy of the knowledge graph. If a fully automated method is used to construct the knowledge graph, the accuracy of the graph cannot be guaranteed, which will affect its use.

[0067] A knowledge point entity in the knowledge graph is taken as a skill dimension, a skill dimension score corresponding to each knowledge point entity is set as a target skill score, teaching materials and evaluation materials corresponding to each knowledge point entity in the knowledge graph are obtained, and the knowledge point entities corresponding to the teaching materials and the evaluation materials and the difficulty coefficients (easy, general, and difficult) of the evaluation materials are marked to obtain a target knowledge graph containing the teaching materials and the evaluation materials; the teaching materials include teaching videos, and the evaluation materials include evaluation questions and scores of the evaluation questions;

[0068] It should be noted that the target knowledge graph containing the teaching materials and the evaluation materials is obtained by marking the knowledge point entities corresponding to the teaching materials and the evaluation materials and the difficulty coefficients of the evaluation materials, because the corresponding teaching materials and evaluation materials must be created according to the knowledge point entities in the knowledge graph, and the underlying knowledge graph must be supported to follow the traditional education and realize the combination with the traditional education.

[0069] An initial learning path is generated based on the prior relationships of each knowledge point entity in the target knowledge graph; the skill scores obtained by the user based on the target learning path during the learning and assessment process are obtained, which are the user skill scores corresponding to each knowledge point entity; each knowledge point entity in the initial learning path is equipped with assessment materials with a preset difficulty coefficient (easy).

[0070] The acquisition of the skill score obtained by the user based on the target learning path during the learning and assessment process of teaching materials, i.e., the user skill score corresponding to each knowledge point entity, is specifically as follows:

[0071] During the learning and assessment process, users can obtain the user's learning time, assessment score, and learning duration for each knowledge point entity corresponding to the teaching materials.

[0072] Based on the user's learning time and the current time, the memory amount corresponding to the knowledge point entity is obtained using the Ebbinghaus forgetting curve. Based on the user's learning time and the user's test score of the teaching materials corresponding to the knowledge point entity, the user's skill score is obtained using the preset skill score calculation formula.

[0073] The preset skill score calculation formula is as follows:

[0074] Score = (T) 学 / T 总 * 20%+ E 测 / E 总 * 80%* Memory capacity;

[0075] In the formula, the memory capacity is obtained based on the Ebbinghaus forgetting curve; Score represents the user's skill score, and T... 学 T represents the user's learning time for the teaching materials corresponding to the knowledge point entity. 总 E represents the total duration of teaching materials corresponding to knowledge point entities. 测 E represents the user rating score corresponding to the knowledge point entity. 总 This represents the total score of the assessment materials corresponding to the knowledge point entity.

[0076] It should be noted that by accumulating the user's skill scores for each knowledge point entity in the target learning path in real time, the user's real-time total skill score can be obtained. By comparing the score with the preset skill target score, it can be understood how far the user's current skill level is from the set target.

[0077] The table corresponding to the Ebbinghaus forgetting curve is shown in Table 1 below:

[0078] Time interval Memory amount Immediately after learning 100％ 20 minutes later 58.2％ 1 hour later 44.2％ 8-9 hours later 35.8％ 1 day later 33.7％ 2 days later 27.8％ 6 days later 25.4％

[0079] According to the user skill score corresponding to the knowledge point entity and the target skill score, a recommendation frequency corresponding to the knowledge point entity is obtained, and according to the recommendation frequency corresponding to each knowledge point entity, the current learning path and the evaluation material corresponding to the knowledge point entity in the current learning path are adjusted to obtain an optimized learning path; the target learning path is the initial learning path or the optimized learning path; the user skill score corresponding to the knowledge point entity is in a positive proportional relationship with the difficulty coefficient of the evaluation material corresponding to the knowledge point entity in the adjustment process.

[0080] It should be noted that the embodiment can set the number of repeated insertion of the knowledge point entity into the current learning path based on the recommendation frequency corresponding to each knowledge point entity (for example, if the recommendation frequency exceeds the first limit value, the knowledge point entity is inserted once, if the recommendation frequency exceeds the second limit value, the knowledge point entity is inserted twice, etc.), and the position of the insertion is repeatedly inserted according to the dependency relationship between the knowledge point entities in the current learning path, thereby realizing the process of completing a cognition, familiarity, mastery, forgetting and re-mastering of the knowledge point entity by the student, so that the student quickly reaches a best learning state, greatly improving the teaching efficiency and the student learning efficiency.

[0081] The acquisition formula for obtaining the recommendation frequency corresponding to the knowledge point entity is:

[0082]

[0083] In the formula, i=1, 2, ……n; n represents the total number of knowledge point entities included in the target learning path; wherein Q A represents the recommendation frequency of the A knowledge point entity, T A is the target skill score of the A knowledge point entity, C A is the user skill score corresponding to the A knowledge point entity, T i represents the target skill score of the i-th knowledge point entity, C i represents the user skill score corresponding to the i-th knowledge point entity.

[0084] The present application constructs a knowledge graph through NLP natural language processing technology, and obtains teaching materials and evaluation materials corresponding to each knowledge point entity in the knowledge graph, identifies the knowledge point entities corresponding to the teaching materials and the evaluation materials, and the difficulty coefficient of the evaluation materials to obtain a target knowledge graph containing the teaching materials and the evaluation materials, obtains the user skill score corresponding to each knowledge point entity in the process of learning the teaching materials and the evaluation materials based on the target learning path, thereby mastering the learning situation of the students on each knowledge point entity in real time, and obtaining the recommendation frequency corresponding to the knowledge point entity according to the user skill score and the target skill score of the knowledge point entity, adjusting the current learning path and the evaluation materials corresponding to the knowledge point entity in the current learning path according to the recommendation frequency corresponding to each knowledge point entity, thereby adjusting the learning path corresponding to the students according to the learning situation of the students in the process of learning, and continuously optimizing, realizing the teaching of different learning paths for different students, that is, optimizing the current learning path according to the learning situation (user skill score) of the students, so that the students learn in the learning path most suitable for their current level, thereby achieving the optimal learning effect.

[0085] Embodiment two

[0086] As Figure 2 shown, the present application also proposes a smart teaching system based on NLP and knowledge graph, comprising:

[0087] A subject data acquisition module is used for crawling text data of each subject from the network through a crawler software;

[0088] The subject data acquisition module comprises:

[0089] A request initiation unit is used for sending a subject material request to a target learning website through an http library;

[0090] A response acquisition unit is used for acquiring response content corresponding to the subject material request;

[0091] An analysis unit is used for obtaining text data corresponding to each subject by analyzing the response content through an analysis module; the analysis module comprises a regular expression, a third-party analysis library and a json module.

[0092] A knowledge graph construction module is used for extracting knowledge point entities corresponding to each subject and the leading relationship between the knowledge point entities from the crawled text data through NLP natural language processing technology; the knowledge point entities comprise knowledge point names and definitions of knowledge points; a knowledge system corresponding to each subject is constructed through the leading relationship between the knowledge point entities, and a knowledge graph is constructed through the association relationship between the knowledge systems corresponding to each subject;

[0093] The target knowledge graph acquisition module is configured to set a skill dimension for each knowledge point entity in the knowledge graph, set a skill dimension score corresponding to each knowledge point entity as a target skill score, and acquire teaching materials and evaluation materials corresponding to each knowledge point entity in the knowledge graph, identify the knowledge point entities corresponding to the teaching materials and the evaluation materials, and identify the difficulty coefficient of the evaluation materials to obtain a target knowledge graph containing the teaching materials and the evaluation materials; the teaching materials include teaching videos, and the evaluation materials include evaluation questions and scores of the evaluation questions;

[0094] The user skill score acquisition module is configured to generate an initial learning path according to the lead relationship of each knowledge point entity in the target knowledge graph, and acquire a skill score obtained by a user in the process of learning the teaching materials and the evaluation materials based on the target learning path, i.e., a user skill score corresponding to each knowledge point entity; each knowledge point entity in the initial learning path is provided with evaluation materials with a preset difficulty coefficient;

[0095] The user skill score acquisition module is configured to generate an initial learning path according to the lead relationship of each knowledge point entity in the target knowledge graph, and acquire a skill score obtained by a user in the process of learning the teaching materials and the evaluation materials based on the target learning path, i.e., a user skill score corresponding to each knowledge point entity; each knowledge point entity in the initial learning path is provided with evaluation materials with a preset difficulty coefficient;

[0096] The user skill score acquisition module is configured to generate an initial learning path according to the lead relationship of each knowledge point entity in the target knowledge graph, and acquire a skill score obtained by a user in the process of learning the teaching materials and the evaluation materials based on the target learning path, i.e., a user skill score corresponding to each knowledge point entity; each knowledge point entity in the initial learning path is provided with evaluation materials with a preset difficulty coefficient;

[0097] The user skill score acquisition module is configured to generate an initial learning path according to the lead relationship of each knowledge point entity in the target knowledge graph, and acquire a skill score obtained by a user in the process of learning the teaching materials and the evaluation materials based on the target learning path, i.e., a user skill score corresponding to each knowledge point entity; each knowledge point entity in the initial learning path is provided with evaluation materials with a preset difficulty coefficient;

[0098] The preset skill score calculation formula is:

[0099] Score = (T 学 / T 总 * 20% + E 测 / E 总 * 80%) * memory amount;

[0100] In the formula, the memory amount is obtained according to the Ebbinghaus memory forgetting curve; Score represents the user skill score, T 学 represents the user learning time of the teaching materials corresponding to the knowledge point entity, T 总 represents the total time of the teaching materials corresponding to the knowledge point entity, E 测 represents the user evaluation score of the knowledge point entity, and E 总 represents the total score of the evaluation materials corresponding to the knowledge point entity.

[0101] The path optimization module is configured to obtain a recommended frequency corresponding to each knowledge point entity according to a user skill score corresponding to the knowledge point entity and a target skill score, and adjust a current learning path and evaluation materials corresponding to the knowledge point entity in the current learning path according to the recommended frequency corresponding to each knowledge point entity to obtain an optimized learning path; the target learning path is the initial learning path or the optimized learning path; the user skill score corresponding to the knowledge point entity is in a proportional relationship with a difficulty coefficient of the evaluation materials corresponding to the knowledge point entity in the adjustment process.

[0102] The formula for obtaining the recommended frequency corresponding to each knowledge point entity is as follows:

[0103]

[0104] In the formula, i = 1, 2,..., n; n represents the total number of knowledge point entities included in the target learning path; Q A represents the recommended frequency of the A knowledge point entity, T A is the target skill score of the A knowledge point entity, C A is the user skill score corresponding to the A knowledge point entity, T i represents the target skill score of the i-th knowledge point entity, C i represents the user skill score corresponding to the i-th knowledge point entity.

[0105] The learning path is generated based on the constructed knowledge graph, so that the combination with the traditional education is realized by following the step-by-step progression in the traditional education; in addition, the target knowledge graph including the teaching materials and the evaluation materials is obtained by obtaining the teaching materials and the evaluation materials corresponding to each knowledge point entity in the knowledge graph, marking the knowledge point entities corresponding to the teaching materials and the evaluation materials, and marking the difficulty coefficient of the evaluation materials, so that the problem of time-consuming in preparing the teaching content (including the teaching materials and the evaluation materials) by the teachers is avoided.

[0106] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative positional relationship, movement condition, etc. between components in a certain posture (as shown in the drawings), and if the certain posture changes, the directional indications also change accordingly.

[0107] In addition, the description such as "first", "second", "one" and the like in the present application is only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first" and "second" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "a plurality of" is at least two, such as two, three, etc., unless otherwise specifically limited.

[0108] In the present application, unless otherwise explicitly specified and limited, the terms "connection", "fixation" and the like should be understood in a broad sense, for example, "fixation" can be fixed connection, or detachable connection, or integral; can be mechanical connection, or electrical connection; can be directly connected, or indirectly connected through intermediate medium, can be internal communication of two elements or interaction relationship of two elements, unless otherwise explicitly limited. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0109] In addition, the technical solutions among various embodiments of the present application can be combined with each other, but it must be based on the fact that a person skilled in the art can realize it, when the combination of technical solutions appears contradictory or unachievable, it should be considered that the combination of technical solutions does not exist, nor is it within the protection scope required by the present application.

Claims

1. A smart teaching method based on NLP and knowledge graph, characterized in that, The method comprises the steps of: obtaining text data of various disciplines from the network through a crawler software; extracting knowledge point entities and the leading relationships between the knowledge point entities corresponding to various disciplines from the obtained text data through NLP natural language processing technology; the knowledge point entities include knowledge point names and definitions of the knowledge points; constructing a knowledge graph by constructing a knowledge system corresponding to each discipline through the leading relationships between the knowledge point entities and constructing an association relationship between the knowledge systems corresponding to various disciplines; setting a skill dimension score corresponding to each knowledge point entity in the knowledge graph as a target skill score, obtaining teaching materials and evaluation materials corresponding to each knowledge point entity in the knowledge graph, and identifying the knowledge point entities corresponding to the teaching materials and the evaluation materials and a difficulty coefficient of the evaluation materials to obtain a target knowledge graph containing the teaching materials and the evaluation materials; the teaching materials include teaching videos, and the evaluation materials include evaluation questions and scores of the evaluation questions; generating an initial learning path according to the leading relationships of the knowledge point entities in the target knowledge graph, and obtaining skill scores of the user corresponding to each knowledge point entity in the learning process of the user based on the target learning path; each knowledge point entity in the initial learning path is provided with evaluation materials with a preset difficulty coefficient; obtaining a recommendation frequency corresponding to each knowledge point entity according to the user skill score and the target skill score of the knowledge point entity, adjusting the current learning path and the evaluation materials corresponding to the knowledge point entities in the current learning path according to the recommendation frequency corresponding to each knowledge point entity to obtain an optimized learning path; the target learning path is the initial learning path or the optimized learning path; the user skill score corresponding to the knowledge point entity and the difficulty coefficient of the evaluation materials corresponding to the knowledge point entity in the adjustment process are in a proportional relationship. 2.The wisdom teaching method based on NLP and knowledge graph of claim 1, wherein, The method of obtaining text data of various disciplines from the network through a crawler software comprises the steps of: S1: sending a discipline data request to a target learning website through an http library; S2: obtaining response content corresponding to the discipline data request; S3: obtaining text data corresponding to various disciplines by analyzing the response content through an analysis module; the analysis module includes a regular expression, a third-party analysis library and a json module. 3.The wisdom teaching method based on NLP and knowledge graph of claim 2, characterized in that, The method of obtaining skill scores of the user corresponding to each knowledge point entity in the learning process of the user based on the target learning path comprises the steps of: obtaining user learning time and user evaluation scores of each knowledge point entity corresponding to the teaching materials in the learning process of the user; obtaining the user skill score corresponding to the knowledge point entity by using a preset skill score calculation formula according to the user learning time and the user evaluation scores of the knowledge point entity corresponding to the teaching materials; the preset skill score calculation formula is: Score = (T 学 / T 总 *20% + E 测 / E 总 *80%)* memory amount; In the formula, the memory amount is obtained according to the Ebbinghaus memory forgetting curve; Score represents the user skill score, T 学 represents the user learning duration of the teaching material corresponding to the knowledge point entity, T 总 represents the total duration of the teaching material corresponding to the knowledge point entity, E 测 represents the user evaluation score corresponding to the knowledge point entity, E 总 represents the total evaluation score of the evaluation material corresponding to the knowledge point entity. 4.The wisdom teaching method based on NLP and knowledge graph of claim 3, characterized in that, the formula for obtaining the recommendation frequency corresponding to the knowledge point entity is: In the formula, i = 1, 2, … n; n represents the total number of knowledge point entities included in the target learning path; wherein Q A represents the recommendation frequency of the A knowledge point entity, T A is the target skill score of the A knowledge point entity, C A is the user skill score corresponding to the A knowledge point entity, T i represents the target skill score of the i th knowledge point entity, C i represents the user skill score corresponding to the i th knowledge point entity.

5. A smart teaching system based on NLP and knowledge graph, characterized in that, The method comprises the steps of: a discipline data acquisition module is configured to obtain text data of various disciplines from the network through a crawler software; The knowledge graph construction module is configured to extract knowledge point entities and leading relationships between the knowledge point entities corresponding to each discipline from the crawled text data through NLP natural language processing technology; the knowledge point entities include knowledge point names and definitions of the knowledge points; The knowledge graph is constructed by constructing a knowledge system corresponding to each discipline through the leading relationships between the knowledge point entities and by constructing the knowledge graph through the association relationships between the knowledge systems corresponding to each discipline; The target knowledge graph acquisition module is configured to set a skill dimension score corresponding to each knowledge point entity as a target skill score with one knowledge point entity in the knowledge graph as the skill dimension, to acquire teaching materials and evaluation materials corresponding to each knowledge point entity in the knowledge graph, and to identify the knowledge point entities corresponding to the teaching materials and the evaluation materials and a difficulty coefficient of the evaluation materials to obtain a target knowledge graph containing the teaching materials and the evaluation materials; the teaching materials include teaching videos, and the evaluation materials include evaluation questions and scores of the evaluation questions; The user skill score acquisition module is configured to generate an initial learning path according to the leading relationships of the knowledge point entities in the target knowledge graph, to acquire skill scores of the user corresponding to each knowledge point entity in the process of learning the teaching materials and the evaluation materials based on the target learning path, and to acquire the user skill scores of the knowledge point entities. The initial learning path is provided with evaluation materials with a preset difficulty coefficient corresponding to each knowledge point entity. The path optimization module is configured to acquire a recommended frequency of each knowledge point entity according to the user skill score and the target skill score of the knowledge point entity, to adjust a current learning path and evaluation materials corresponding to the knowledge point entities in the current learning path according to the recommended frequency of each knowledge point entity, and to obtain an optimized learning path; the target learning path is the initial learning path or the optimized learning path; the user skill score of the knowledge point entity is in a direct proportional relationship with a difficulty coefficient of the evaluation materials corresponding to the knowledge point entity in the adjustment process. 6.The intelligent teaching system based on NLP and knowledge graph of claim 5, wherein, The discipline data acquisition module includes: A request initiation unit configured to send a discipline data request to a target learning website through an http library; A response acquisition unit configured to acquire response content corresponding to the discipline data request; An analysis unit configured to acquire text data corresponding to each discipline by analyzing the response content through an analysis module; the analysis module includes a regular expression, a third-party analysis library, and a json module. 7.The intelligent teaching system based on NLP and knowledge graph of claim 6, wherein, The user skill score of each knowledge point entity is acquired by: Acquiring a user learning time length and a user evaluation score of each knowledge point entity corresponding to the teaching materials in the process of learning the teaching materials and the evaluation materials; Acquiring the user skill score of the knowledge point entity according to the user learning time length and the user evaluation score of the knowledge point entity corresponding to the teaching materials through a preset skill score calculation formula; The preset skill score calculation formula is: Score = (T 学 / T 总 *20% + E 测 / E 总 *80%)* memory amount; In the formula, the memory amount is obtained according to the Ebbinghaus memory forgetting curve; Score represents the user skill score, T 学 represents the user learning duration of the teaching material corresponding to the knowledge point entity, T 总 represents the total duration of the teaching material corresponding to the knowledge point entity, E 测 represents the user evaluation score corresponding to the knowledge point entity, E 总 represents the total evaluation score of the evaluation material corresponding to the knowledge point entity. 8.The intelligent teaching system based on NLP and knowledge graph of claim 7, wherein, The acquisition formula for acquiring the recommended frequency of the knowledge point entity is: wherein i = 1, 2, … n; n represents the total number of knowledge point entities included in the target learning path; wherein Q A represents the recommendation frequency of the A knowledge point entity, T A is the target skill score of the A knowledge point entity, C A is the user skill score corresponding to the A knowledge point entity, T i represents the target skill score of the i-th knowledge point entity, C i represents the user skill score corresponding to the i-th knowledge point entity.

Images