CN113886567A

CN113886567A - Teaching method and system based on knowledge graph

Info

Publication number: CN113886567A
Application number: CN202111010796.5A
Authority: CN
Inventors: 胡增芳; 吴福全; 刘宝; 刘爽; 许贤丽
Original assignee: Anhui Business College
Current assignee: Anhui Business College
Priority date: 2021-08-31
Filing date: 2021-08-31
Publication date: 2022-01-04

Abstract

The invention discloses a teaching method and a system based on a knowledge graph, which relate to the technical field of teaching methods and comprise the following steps: acquiring big data, wherein the data comprises subject knowledge points, learning resources, attribute values of the knowledge points, relations among the knowledge points and the learning resources; extracting subdata of a required field from the big data based on the big data; converting structured and semi-structured data in the sub data into triples, and extracting the triples from the unstructured data by using a plurality of algorithm models; carrying out knowledge fusion operation of entity alignment and attribute alignment on the triples; establishing an association relation between knowledge points in a knowledge graph of disciplines in a required field and various teaching resources based on knowledge association; and visualizing the incidence relation maps of the knowledge points in the knowledge map and various teaching resources to form a map. According to the method, after the discipline knowledge map is constructed, the knowledge mastering condition of students is accurately depicted, and therefore the teaching quality is improved.

Description

Teaching method and system based on knowledge graph

Technical Field

The invention relates to the technical field of teaching methods, in particular to a teaching method and a teaching system based on a knowledge graph.

Background

Education informatization means that modern information technology is comprehensively and deeply applied in the education field to promote education reform and development. At present, most of education informatization is to paste the technology to education and teaching according to informatization of an original education system. This kind of way, can only improve a part of teaching efficiency usually, and the improvement is not very obvious in the teaching quality.

Disclosure of Invention

In view of the above, the present invention provides a teaching method and system based on knowledge graph, which are used to improve teaching quality.

Based on the above purpose, the teaching method and system based on knowledge graph provided by the invention comprises:

acquiring big data, wherein the data comprises subject knowledge points, learning resources, knowledge point attribute values, relations between the knowledge points and the learning resources;

extracting subdata of a required field from the big data based on the big data;

converting the structured and semi-structured data in the sub-data into triples, and extracting the triples from the unstructured data by using a plurality of algorithm models;

carrying out knowledge fusion operation of entity alignment and attribute alignment on the triples;

establishing an association relation between knowledge points in a knowledge graph of disciplines in a required field and various teaching resources based on knowledge association;

and visualizing the association relationship map of the knowledge points and various teaching resources in the knowledge map to form a map.

Optionally, the extracting sub-data of a required field from the big data includes:

and extracting the subdata of the required field from the big data by using an entity identification and relation extraction method.

Optionally, the entity identification method includes: the sub-data of the required field is extracted by a rule-based method, an unsupervised learning-based method, a feature-based supervised learning method and a deep learning method.

Optionally, the relationship extraction method includes:

firstly, extracting a structured and semi-structured data source;

comprehensively extracting unstructured data by using a rule method, a learning method, a pre-training model and a crowdsourcing method;

and extracting relation data among the knowledge points by using a frequent pattern mining and process mining algorithm.

Optionally, the graph is a central network subgraph, and the central network subgraph is used for analyzing and displaying one node, adjacent nodes and relevant features.

Optionally, the map is a learning condition analysis sub-map, the learning condition analysis sub-map is used for learning condition data of students obtained by a data mining technology before a class, making a teaching strategy and achieving decision datamation, in the class, utilizing the made teaching strategy for targeted teaching, explaining knowledge points, performing group discussion teaching, recommending relevant post-class exercises after the class, recommending the personalized targeted exercise questions according to the learning condition, learning ability and personalized targeted exercise questions of the students, and consolidating wrong questions.

Optionally, the map is a deep mining sub-map, the deep mining sub-map identifies and connects entities of the electronic publication by using an entity link technology, displays current knowledge information in a form of a knowledge card, and can also be associated with other knowledge related to the current knowledge information, and recommends the related knowledge to help a user connect knowledge in series.

Based on the same invention creation, the invention also provides a teaching system based on the knowledge graph, which comprises:

a data acquisition module: the system is used for acquiring big data, wherein the data comprises subject knowledge points, learning resources, attribute values of the knowledge points, relations among the knowledge points and the learning resources;

a data extraction module: the device comprises a data processing module, a data storage module and a data processing module, wherein the data processing module is used for extracting sub data of a required field from the big data based on the big data;

a data processing module; the data processing device is used for converting the structured and semi-structured data in the sub-data into triples, and extracting the triples from the unstructured data by using a plurality of algorithm models;

a visualization module: and the system is used for visualizing the association relationship map of the knowledge points and various teaching resources in the knowledge map to form a map.

After the discipline knowledge graph is constructed, the association can be constructed with teaching resources (teaching materials, test questions, lectures, teaching videos, test papers and the like), and then the association between knowledge points and users is established through user information and learning records. Through the knowledge map, the knowledge mastering condition of students is accurately described, and resources are accurately described. Accurate study and judgment, study path planning and study resource personalized recommendation for the user are realized.

Meanwhile, teachers are helped to better understand the learning conditions of students, the teaching method is optimized, and the teaching strategy is adjusted. Can be through being associated with the teaching and research data, through initiatively recommending the teaching and research come for mr teaching and research prepare lesson promotion efficiency and quality to supplementary teaching answer system of answering a question with knowledge map question-answering as core technology can effectually alleviate the burden that simple repeated problem brought for mr, also can to a great extent satisfy student's answer demand.

Drawings

In order to more clearly illustrate one or more embodiments or prior art solutions of the present specification, the drawings that are needed in the description of the embodiments or prior art will be briefly described below, and it is obvious that the drawings in the following description are only one or more embodiments of the present specification, and that other drawings may be obtained by those skilled in the art without inventive effort from these drawings.

FIG. 1 is a flow chart of a knowledge-graph-based teaching method according to an embodiment of the present invention.

Detailed Description

To make the objects, technical solutions and advantages of the present disclosure more apparent, the present disclosure is further described in detail below with reference to specific embodiments.

It is to be noted that unless otherwise defined, technical or scientific terms used in one or more embodiments of the present specification should have the ordinary meaning as understood by those of ordinary skill in the art to which this disclosure belongs. The use of "first," "second," and similar terms in one or more embodiments of the specification is not intended to indicate any order, quantity, or importance, but rather is used to distinguish one element from another. The word "comprising" or "comprises", and the like, means that the element or item listed before the word covers the element or item listed after the word and its equivalents, but does not exclude other elements or items. The terms "connected" or "coupled" and the like are not restricted to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "upper", "lower", "left", "right", and the like are used merely to indicate relative positional relationships, and when the absolute position of the object being described is changed, the relative positional relationships may also be changed accordingly.

The embodiment of the invention provides a teaching method based on a knowledge graph. As shown in fig. 1, a knowledge-graph-based teaching method includes:

s101: acquiring big data, wherein the data comprises subject knowledge points, learning resources, knowledge point attribute values, relations between the knowledge points and the learning resources;

the learning resources can be dynamically changed, the attribute values of the knowledge points comprise knowledge point classification tags and corresponding tag attributes, the knowledge map is educated, the requirement on multi-mode knowledge is strong, the knowledge points can be represented in various forms of pictures, videos and voices, the knowledge points also accord with the habit of people for recognizing things, and the knowledge points also need to be perceived except for character description.

The acquisition and integration of the data resources not only depend on the technologies of data crawling, multi-source heterogeneous data governance, distributed data storage and the like, but also depend on strong external data resource cooperation capacity and internal pushing capacity. For example, in the aspect of knowledge graph construction, when extracting entities, relations and attributes, although the difficulties of disambiguation, alignment, fusion and the like are faced as well, the knowledge structure and the data model can be designed more accurately when the knowledge graph is constructed through the participation of professional experts with professional knowledge.

S102: extracting subdata of a required field from the big data based on the big data;

s103: converting the structured and semi-structured data in the sub-data into triples, and extracting the triples from the unstructured data by using a plurality of algorithm models;

to facilitate computer processing and understanding, knowledge maps use a more formal and simplified way to represent knowledge, namely triples. The triple is entity (entity), entity relationship (relationship), entity (entity). Considering an entity as a node and an entity relationship (including attributes, categories, etc.) as an edge, a knowledge base containing a large number of triples becomes a huge knowledge graph.

The multiple algorithm models comprise a domain relation extraction model and an open domain relation extraction model, the text data related to the domain relation extraction model is generally composed of multiple texts under one or more similar topics, such as financial news data, and the text data required to be processed by the open domain relation extraction model is domain-open, is not limited in topic quantity, and is typically encyclopedia data.

S104: carrying out knowledge fusion operation of entity alignment and attribute alignment on the triples;

entity alignment: it is determined whether two or more entities from different sources are pointing to the same object in the real world. If a plurality of entities represent the same object, an alignment relation is constructed among the entities, and meanwhile information contained in the entities is fused and aggregated.

And (3) attribute alignment: and after judging whether two or more attributes can represent the same attribute, performing information fusion on the attributes with the same source or name but same representation, thereby obtaining richer and more accurate information.

S105: establishing an association relation between knowledge points in a knowledge graph of disciplines in a required field and various teaching resources based on knowledge association;

s106: and visualizing the association relationship map of the knowledge points and various teaching resources in the knowledge map to form a map.

The educational knowledge resource construction taking the knowledge map as the core, the knowledge map is utilized to establish the association between the domain knowledge, and the association is established between the knowledge points and various educational resources of teaching materials, teaching aids, lectures, videos, test questions and the like of different versions to form an integral network. The upper layer application is supported by the associated networks.

In some embodiments, the extracting the sub data of the required domain from the big data includes:

The method of rules refers to explicit use of rules, and the set of rules is specified by a person and does not change when used. The complex task rules are difficult to describe and cover complex input conditions, so that the method is only suitable for simple and easily-described tasks.

The method for unsupervised learning comprises the following steps: the target variables in similar classifications and regression do not exist in advance.

The method with supervised learning comprises the following steps: the machine can deduce the possible result of the specified target variable from the input sample set, and the machine only needs to predict a suitable model from the input data and calculate the result of the target variable from the predicted model.

The deep learning method is to learn the intrinsic rules and the expression levels of sample data, and the information obtained in the learning process is very helpful to the interpretation of data such as characters, images and sounds. The final aim of the method is to enable the machine to have the analysis and learning capability like a human, and to recognize data such as characters, images and sounds.

The scarcity of annotation data in the field determines that under the condition of cost constraint, a mature Named Entity Recognition (NER) model cannot be directly used in the open field in many cases. But enterprises in the field can accumulate word banks in the field. The thesaurus can be used together with some engineering methods and strategies to solve the problem, such as rules application, unsupervised learning and the like.

Optionally, the relationship extraction method includes:

s201: firstly, extracting a structured and semi-structured data source;

s202: comprehensively extracting unstructured data by using a rule method, a learning method, a pre-training model and a crowdsourcing method;

the learned method rule set is automatically generated by the model. Accordingly, the learning method can be divided into an explicit rule set and an implicit rule set.

The pre-training model is a stage of enabling natural language processing to enter large-scale reproducible industrial development from the original stage of manually tuning parameters and depending on ML experts. And the pre-training model extends from single language, to multi-language, multi-modal tasks.

S203: and extracting relation data among the knowledge points by using a frequent pattern mining and process mining algorithm.

Frequent pattern mining refers to a class of patterns that occur frequently in datasets. A frequent item set refers to a set consisting of elements that contain a certain class of frequent patterns. Thus, frequent pattern mining is also often called frequent itemset mining.

The process mining algorithm refers to reconstructing a process model from an event log, and the mining model can be used as an objective feedback mechanism to check the specification implementation of an information system from initial deployment to application development.

By the method, knowledge point learning paths are mined from data such as catalog arrangement, lecture chapter arrangement, test paper knowledge point distribution, user behavior logs and the like of teaching materials of different versions.

In some embodiments, the graph is a central network subgraph used to analytically show one node, neighboring nodes and related features.

And drawing a knowledge graph of one degree, two degrees, three degrees and the like by taking one node to be analyzed as a center. A network subgraph centered on a single student may also be referred to as a client representation of the student.

To perform comprehensive quality evaluation of students in the current complex data environment, complex data results need to be presented in a knowledge graph form from a real customer portrait by using a customer portrait method. The school can further collect and analyze the main information of the students/teachers, such as learning behaviors, knowledge abilities, living habits and the like, through the visual data presentation method, and then completely describe the characteristics of the education target group.

In some embodiments, the atlas is a learning condition analysis subgraph, the learning condition analysis subgraph is used for learning condition data of students obtained by a data mining technology before a class, a teaching strategy is formulated to realize decision-making datamation, the formulated teaching strategy is used for targeted teaching, knowledge points are explained, teaching is discussed in groups, relevant after-class exercises are recommended after the class, and personalized targeted exercise recommendation and wrong exercise consolidation are performed according to the learning condition, the learning capacity and the personalized targeted exercise recommendation of the students.

Dynamic data analysis and dynamic learning condition diagnosis run through the whole teaching process, the purpose of teaching according to the material is realized, and the teaching decision is digitalized and intelligentized. Through the accurate analysis to the condition of learning, the system carries out relevant consolidation exercise recommendation, and the pertinence of formulating the teaching strategy promotes the teaching pertinence, carries out accurate teaching.

In some embodiments, the map is a deep mining sub-map, the deep mining sub-map performs entity identification and connection on the electronic publication by using an entity link technology, displays current knowledge information in a form of a knowledge card, and can also be associated with other knowledge related to the current knowledge information, and performs recommendation of the related knowledge to help a user to connect knowledge in series.

In order to better implement the invention creation, the invention also provides a teaching system based on knowledge graph, comprising:

a data processing module: the data processing device is used for converting the structured and semi-structured data in the sub-data into triples, and extracting the triples from the unstructured data by using a plurality of algorithm models;

The system has the following effects:

1. accurate portrait

The user portrait based on the knowledge map can enhance user portrait data and is more comprehensive and accurate in user portrait.

2. Analysis of learning emotion

And the knowledge graph can be used for more accurate study and situation analysis. The method is based on knowledge map, big data analysis and the like, the objective learning process of the learner is mined, analysis is carried out from multiple dimensions, the data can be mined from multiple dimensions, the data is not limited to knowledge mastering conditions mined from behavior tracks such as test results, wrong exercise books, learning records and the like, and the dominant characteristics of weak knowledge can be mined, and some deep invisible characteristics such as learning speed, learning preference, cognitive level and the like can also be mined. The analysis result is more personalized and objective.

3. Quality of teaching

The knowledge graph assists teachers to complete the work of lesson preparation, teaching and research, question setting and test question analysis in the auxiliary teaching application. The system can recommend similar related data for teachers in a recommendation mode to improve teaching efficiency of teachers, and the required content can be returned more accurately through map-based searching.

In some embodiments, the system further comprises an educational robot module.

Educational robots have become an important application in the field of education. By using the teaching robot with the question-answering system as the core, a series of teaching works such as course answering, knowledge retrieval recommendation, teaching management and the like can be realized. Not only reduces the burden and pressure of teachers, but also solves the practical problems of students. An excellent and comprehensive teaching robot is a comprehensive body integrating a plurality of system modules such as a task type question answering system, a knowledge type question answering system, a search recommending system and the like, and has multi-turn question answering capability. The knowledge graph plays an important role in question understanding and knowledge-guided language generation.

It is intended that the one or more embodiments of the present specification embrace all such alternatives, modifications and variations as fall within the broad scope of the appended claims. Therefore, any omissions, modifications, substitutions, improvements, and the like that may be made without departing from the spirit and principles of one or more embodiments of the present disclosure are intended to be included within the scope of the present disclosure.

Claims

1. a teaching method based on knowledge graph, is characterized in that, comprises:

Acquiring big data, the data includes subject knowledge points, learning resources, attribute values of knowledge points, the relationship between knowledge points, and the relationship between knowledge points and learning resources;

Based on the big data, extract the sub-data of the required field from the big data;

Converting structured and semi-structured data in the sub-data into triples, and using multiple algorithm models to extract triples in unstructured data;

performing the knowledge fusion operation of entity alignment and attribute alignment on the triplet;

Establish the relationship between knowledge points and various teaching resources in the subject knowledge map of the required field based on knowledge association;

The graph of the association relationship between knowledge points and various teaching resources in the knowledge graph is visualized to form a graph.

2. A kind of teaching method based on knowledge graph according to claim 1, is characterized in that, described extracting the sub-data of required field from described big data comprises:

The sub-data of the required domain is extracted from the big data by the method of entity recognition and relation extraction.

3. a kind of teaching method based on knowledge graph according to claim 2, is characterized in that, the method for described entity identification comprises: the method based on rules, the method based on unsupervised learning, the supervised learning method based on feature, and deep learning methods to extract sub-data in the desired domain.

4. a kind of teaching method based on knowledge graph according to claim 2, is characterized in that, the method for described relation extraction comprises:

First extract structured and semi-structured data sources;

Extract unstructured data comprehensively by using rule method, learning method, pre-training model and crowdsourcing method;

Use the algorithms of frequent pattern mining and process mining to extract relational data between knowledge points.

5 . The teaching method based on a knowledge graph according to claim 1 , wherein the graph is a central network subgraph, and the central network subgraph is used to analyze and display a node, adjacent nodes and related features. 6 .

6. A kind of teaching method based on knowledge graph according to claim 1, is characterized in that, described graph is learning situation analysis sub-graph, and described learning situation analysis sub-graph is used for before class, utilizes data mining technology to obtain data of students’ learning situation, formulate teaching strategies, and make decision-making digitized. In class, use the formulated teaching strategies to conduct targeted teaching, explain knowledge points, and discuss teaching in groups. After class, relevant after-class exercises are recommended. Learning status, learning ability, personalized recommendation of targeted practice questions, and consolidation of wrong questions.

7. a kind of teaching method based on knowledge graph according to claim 1, is characterized in that, described graph is deep mining subgraph, and described deep mining subgraph utilizes entity link technology to carry out entity identification and identification to electronic publication. Connect, display current knowledge information in the form of knowledge cards, and can also associate with other related knowledge, and recommend related knowledge to help users connect knowledge together.

8. A teaching system based on knowledge graph, characterized in that, comprising:

Data acquisition module: used to acquire big data, the data includes subject knowledge points, learning resources, attribute values of knowledge points, the relationship between knowledge points and the relationship between knowledge points and learning resources;

Data extraction module: used for extracting sub-data of required fields from the big data based on the big data;

A data processing module; used to convert structured and semi-structured data in the sub-data into triples, and use multiple algorithm models to extract triples from unstructured data;

Visualization module: used to visualize the relationship graph of the knowledge points in the knowledge graph and various teaching resources to form a graph.