CN112488887A - Learner portrait generation method and device based on knowledge graph - Google Patents
Learner portrait generation method and device based on knowledge graph Download PDFInfo
- Publication number
- CN112488887A CN112488887A CN202011388997.4A CN202011388997A CN112488887A CN 112488887 A CN112488887 A CN 112488887A CN 202011388997 A CN202011388997 A CN 202011388997A CN 112488887 A CN112488887 A CN 112488887A
- Authority
- CN
- China
- Prior art keywords
- learner
- emotion
- knowledge
- learning
- course
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 230000008451 emotion Effects 0.000 claims abstract description 204
- 238000004458 analytical method Methods 0.000 claims abstract description 46
- 230000002452 interceptive effect Effects 0.000 claims abstract description 22
- 230000001149 cognitive effect Effects 0.000 claims abstract description 21
- 238000011156 evaluation Methods 0.000 claims description 41
- 238000010276 construction Methods 0.000 claims description 24
- 230000008569 process Effects 0.000 claims description 21
- 230000002996 emotional effect Effects 0.000 claims description 18
- 230000006399 behavior Effects 0.000 claims description 10
- 238000000605 extraction Methods 0.000 claims description 9
- 235000019580 granularity Nutrition 0.000 description 10
- 230000003993 interaction Effects 0.000 description 7
- 238000012800 visualization Methods 0.000 description 7
- 230000000007 visual effect Effects 0.000 description 5
- 238000009825 accumulation Methods 0.000 description 4
- 239000003086 colorant Substances 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 235000005612 Grewia tenax Nutrition 0.000 description 1
- 244000041633 Grewia tenax Species 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000004326 stimulated echo acquisition mode for imaging Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 230000007306 turnover Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
- G06Q50/205—Education administration or guidance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Educational Technology (AREA)
- Educational Administration (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Animal Behavior & Ethology (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- General Business, Economics & Management (AREA)
- Electrically Operated Instructional Devices (AREA)
Abstract
The invention discloses a learner portrait generation method and device based on a knowledge map, wherein the method comprises the following steps: step S1, collecting learner learning information; step S2, the interactive information content of the learner is disassembled, the emotion analysis task is summarized into a plurality of levels such as words, phrases, attributes, sentences, chapters and the like according to the granularity of the text, the emotion dictionary is matched to determine the emotion polarity, the emotion score is calculated according to the weight of the emotion polarity, and the emotion tendency of the learner is output according to the emotion value; step S3, constructing the learning knowledge structure and learning knowledge preference of the learner according to the learning information collected in step S1 and the emotion analysis result in step S2; and step S4, generating a corresponding learner portrait knowledge map according to the knowledge preference and cognitive structure of the learner constructed in the step S3.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a learner portrait generation method and device based on a knowledge graph.
Background
The development of the information age puts more new requirements on human talent culture. To address this development, the need to cultivate new generations of talents has emerged as a revolution in education. Compared with the traditional teaching mode, the teaching mode supported by the informatization technology, such as turnover classroom, STEAM education (comprehensive education integrating multiple fields of science, technology, engineering, art and mathematics) and the like, has great significance on the requirements of thinking ability and overall quality of students. Meanwhile, how to provide personalized and differentiated educational information services for students and promote the development of the services, and representing the learning condition of learners also becomes a problem to be improved urgently at present. Among many methods, using a personalized learner representation to depict aspects of a learner's information is very effective for targeted education.
Currently, learner portrayal applications have been directed to many aspects of the educational field, such as live classroom, assessment tools, talent user portrayal libraries for high-tech intelligent craftsman, and the like. In the study of the live classroom, the basic elements of the learner portrait are obtained through the learning purpose, learning habit, the live classroom participation condition of the learner, the live classroom effect evaluation and the classroom overall satisfaction degree of the student; the learner picture constructed by the assessment tool assesses the learning ability of the learner; the talent is cultured in a personalized way by constructing the talent user portrait of the high-tech intelligent craftsman, and personalized education and the like are realized.
For each learner, there should be a learning picture matching with its learning process and the user can visually see the constructed learner picture information. However, most of the current applications do not make a breakthrough in the automatic construction function and the visual aspect of the portrait constructed by student information. The learner portrait content applied at present mainly aims to link learner information for background workers, and lacks of an important function of intuitively feeding back the learner portrait content to the learner. Most user models of personalized systems, such as ELM-ART models and AHAM models, mainly consider reasoning about learning styles of learners, neglect tracking learning routes of learners, neglect biased diversity of learners and construct learning portraits.
Disclosure of Invention
In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a knowledge-graph-based learner portrait generation method and apparatus, which automatically constructs and transforms a learner portrait according to user information and displays the constructed portrait to a user in a visual manner, thereby making up the deficiencies of the self-adaptive learning system in the aspect of user portrait visualization, enabling a teacher to fully grasp the information of the student's prior knowledge structure, learning bias, etc., and enabling the student to more intuitively understand the student's prior learning structure.
To achieve the above and other objects, the present invention provides a learner portrait generation method based on knowledge-base, comprising the steps of:
step S1, collecting learner learning information;
step S2, the interactive information content of the learner is disassembled, the emotion analysis task is summarized into a plurality of levels such as words, phrases, attributes, sentences, chapters and the like according to the granularity of the text, the emotion polarity is determined by matching with a pre-constructed emotion comparison dictionary, the emotion score is calculated according to the weight of the emotion polarity, and the emotion tendency of the learner is output according to the emotion value;
step S3, constructing the learning knowledge structure and learning knowledge preference of the learner according to the learning information collected in step S1 and the emotion analysis result in step S2;
and step S4, generating a learner portrait knowledge map by using the knowledge map according to the knowledge preference and the cognitive structure of the learner constructed in the step S3.
Preferably, the learning information includes, but is not limited to, interactive behaviors of a learner on click through a course on the web, progress of a passed course, evaluation advice on a course, evaluation and advice on a teacher who is a lesson, total number of courses participated, and the like.
Preferably, before step S2, the method further includes the following steps:
and step S2-1, constructing the emotion comparison dictionary according to the Chinese ontology emotion dictionary, setting the emotion polarity of the viewpoint and endowing the set emotion polarity with a weight value.
Preferably, the step S3 further includes:
step S300, constructing a learning knowledge structure of the learner according to the previous learning process information of the learner;
step S301, learning content of the learner is obtained according to the obtained emotional information of the learner, and knowledge preference of the learner is arranged according to the extracted course information of the learning content.
Preferably, step S300 specifically includes:
collecting previous learning process information of the learner;
constructing a user-knowledge point relation knowledge graph;
carrying out entity extraction, relation extraction and semantic recognition;
and obtaining the courses created by the learners and the courses participated by the learners according to the extraction result to form learner course maps and learner cognitive maps.
Preferably, in step S4, according to the knowledge preference and cognitive structure of the learner constructed in step S3, the knowledge graph is classified according to the size of the nodes, and the color classification is performed according to the colors of the nodes to obtain two graph modules, i.e., a learner portrait, of the lesson created by the learner and the lesson participated by the learner.
In order to achieve the above object, the present invention further provides an apparatus for generating a learner profile based on a knowledge-base, comprising:
the learning information collection module is used for collecting learning information of learners;
the emotion analysis module is used for disassembling interactive information content of the learner, generalizing emotion analysis tasks into multiple levels such as words, phrases, attributes, sentences, chapters and the like according to the granularity of the text, matching an emotion dictionary to determine emotion polarity, calculating an emotion value according to the weight of the emotion polarity, and outputting the emotion tendency of the learner according to the emotion value;
the learning knowledge structure and learning knowledge preference construction module is used for constructing the learning knowledge structure and learning knowledge preference of the learner according to the learning information collected by the learning information collection module and the emotion analysis result of the emotion analysis module;
and the learner sketch generation module is used for generating the learner sketch by utilizing the knowledge map according to the learning knowledge structure and the knowledge preference and cognitive structure of the learner, which are constructed in the learning knowledge preference construction module.
The device also comprises an emotion dictionary construction module which is used for constructing the emotion comparison dictionary according to the Chinese ontology emotion dictionary, setting the emotion polarity of the viewpoint and endowing the set emotion polarity with a weight value.
Preferably, the learning knowledge structure and learning knowledge preference building module further comprises:
the learning knowledge structure construction module is used for constructing the cognitive structure of the learner according to the previous learning process information of the learner;
and the learning knowledge preference construction module is used for obtaining the learning content of the learner according to the obtained emotional information of the learner and extracting course information according to the content to arrange the knowledge preference of the learner.
Preferably, the learner learning information includes, but is not limited to, interactive behaviors of a learner on the click amount of the lesson on the web, the progress of the lesson that was taken, evaluation suggestions of the lesson, evaluation and suggestions of a teacher who is in the lesson, the total number of lessons participated, and the like.
Compared with the prior art, the learner portrait generation method and device based on the knowledge graph, provided by the invention, collect the interactive information of the learner, extract the data through methods such as emotion analysis and regular expression and automatically generate the learner portrait in a knowledge graph visualization mode, the generated learner portrait is provided for the teachers to adjust courses according to the learner, the learner is also provided with a learning way for knowing the previous learning condition of the learner, a future suggested learning route is provided, the defects of the self-adaptive learning system in the aspect of user portrait visualization are made up, a teacher can fully master the information of the previous knowledge structure, learning deviation and the like of the student, and the student can also know the previous learning structure of the student more intuitively.
Drawings
FIG. 1 is a flow chart illustrating steps of a method for generating a learner representation based on a knowledge-base map in accordance with the present invention;
FIG. 2 is a flow chart of obtaining a learner's learning knowledge structure according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating the cognitive structure of a learner according to an embodiment of the present invention; a
FIG. 4 is a diagram illustrating knowledge preference of a learner according to an embodiment of the present invention;
FIG. 5a is a lesson knowledge-graph participated in by a learner in an embodiment of the present invention;
FIG. 5b is a learner-created lesson knowledge-graph in accordance with an embodiment of the present invention;
FIG. 6 is a system architecture diagram of an apparatus for learner representation generation based on a knowledge-map according to the present invention;
FIG. 7 is a flow chart of a method for knowledge-graph based learner representation generation in accordance with an embodiment of the present invention;
FIG. 8 is a flowchart illustrating an embodiment of learning bias of a learner.
Detailed Description
Other advantages and capabilities of the present invention will be readily apparent to those skilled in the art from the present disclosure by describing the embodiments of the present invention with specific embodiments thereof in conjunction with the accompanying drawings. The invention is capable of other and different embodiments and its several details are capable of modification in various other respects, all without departing from the spirit and scope of the present invention.
FIG. 1 is a flow chart illustrating steps of a method for generating a learner representation based on a knowledge-base. As shown in FIG. 1, the present invention relates to a learner sketch generation method based on knowledge-graph, which comprises the following steps:
and step S1, collecting the learning information of the learner.
The learning information includes, but is not limited to, the learner 'S click rate on the course on the internet, the progress of the last course, the evaluation suggestion on the course, the evaluation and suggestion on the teacher who is in the lesson, the total number of the participated courses, and the like, that is, in step S1, when the user takes a lesson using the apparatus to which the present invention is applied, the learner' S click rate on the course on the internet, the progress of the last course, the evaluation suggestion on the course, the evaluation and suggestion on the teacher who is in the lesson, the total number of the participated courses, and the like are first collected.
And step S2, disassembling the interactive information content of the learner, summarizing the emotion analysis task into multiple levels of words, phrases, attributes, sentences, chapters and the like according to the granularity of the text, matching a pre-constructed emotion dictionary to determine the emotion polarity, calculating the emotion score according to the weight of the emotion polarity, and outputting the emotion tendency of the learner according to the emotion value.
Specifically, in step S2, the type and degree of preference of the learner for the current lesson or teacher are obtained mainly through the emotion analysis task. In the present invention, the emotion preferences and emotion values generated by the emotion analysis task are one of the data generated by the learner representation, so through the emotion analysis task, the task attempts to acquire the type and hierarchy of preferences about the learner's preference for the current lesson or teacher. In the invention, an emotion analysis task mainly acquires interaction times of all language comment words and learners under different emotion preference buttons (referring to praise and poor comment buttons in the learning process) and superposes emotion types and emotion values represented by each word or each interaction, so that the emotion analysis task firstly disassembles interaction data of the learners, summarizes the task into multiple levels of words, phrases, sentences, chapters and the like according to the granularity of text languages and matches a pre-constructed emotion comparison dictionary; then, determining the emotion polarity of the word through the emotion comparison dictionary, and calculating the emotion score of the word; because phrases, sentences and chapters are the most basic word components, the emotional tendency and the emotional value of the whole text can be calculated only by further accumulation and taken as the emotional tendency of the learner. The step of disassembling the interactive data of the learner refers to disassembling the number of times that the learner clicks on the course, the learning progress of the course (the learning progress of a student of a certain course is converted into the learning percentage of the student on the course, the high percentage indicates that the student is more interested in the course), the evaluation of the course, the evaluation of a teacher of the course, and the approval of the course (the emotional marks which can be replaced by the approval of the course are manually set).
Specifically, before step S2, the method further includes the following steps:
step S2-1, referring to the emotion dictionary construction method mentioned in the emotion dictionary automatic construction method overview (Wangke, summer Rui. Automation, 2016,42(4):495-511.doi:10.16383/j.aas.2016.c150585) in the prior art, the constructed dictionary is a reference of the prior Chinese ontology emotion dictionaries which can be openly obtained, and the emotion comparison dictionary is constructed. After the emotion comparison dictionary provides the emotion types and emotion values corresponding to more than 7000 emotion vocabularies, the emotion analysis tasks are summarized into multiple levels such as words, phrases, sentences, chapters and the like according to the granularity of the text, the emotion dictionary can accurately judge the emotion polarity of the words/phrases, and the sentence/chapter level emotion analysis tasks can be effectively assisted in an accumulation mode. Considering that some ambiguity exists in the individual evaluation words, a combined evaluation unit (such as less-good and equivalent-good) is also considered in the construction of the emotion dictionary, if the combined evaluation unit is less than the combined evaluation unit, the emotion value is multiplied by the value in the interval of 0.1-1, and if the combined evaluation unit is equivalent, the combined evaluation unit is multiplied by the value between 1.1-10.
Specifically, in step S2, after collecting the learner 'S evaluation of the lesson, the teacher' S evaluation, and the information of the lesson click amount, the emotion value is obtained by matching with the words in the emotion comparison dictionary, the statistics is performed by the total of the button operations of the positive and negative comments, the statistical process is mainly based on the emotion scores accumulated for the positive and negative emotion categories, so as to know the emotion tendency of the learner for the lesson (the strength value of the emotion is represented by the score and the positive and negative emotions are determined according to the positive and negative scores), the word emotion score in the target field is output after analysis, and the emotion scores of all the emotion words are obtained after multi-word iteration. For example, a learner participates in a computer network course and issues a plurality of comments aiming at the course, i.e., "i like the teacher" and "the content of the lesson is rich", etc., the system executes a dictionary matching task of emotion analysis on all the comments, if words like and rich exist in the dictionary, the emotion category and the emotion value corresponding to the words are returned, and then all the comments are superposed to be used as the emotion category and the emotion value of the learner. And in another case, the learner makes statistics on whether the final emotion category of the learner is positive or negative according to the principle that the learner makes positive comments and negative comments on the interactive interface of the course for a plurality of times and a small number of negative comments on other data such as knowledge points or exercises and the like in the computer network course, and the statistical value is also used as the emotion value. And finally adding the emotion types and the emotion values of the two cases to serve as the emotional bias and the emotion value of the learner for the course.
Step S3, according to the learning information collected in step S1 and the emotion analysis result (the degree of the learner 'S tendency to the lesson) in step S2, a learning knowledge structure (the lesson the learner has participated in) and a learning knowledge preference (the degree of the learner' S emotional tendency to the learned lesson has been constructed) of the learner are constructed, the lesson the learner has participated in or clicked on the relevant work of the lesson, such as teaching assistance, audition, and exercise, and is a trial or preferred operation of the learner, the lesson the learner has to complete the whole content of the learning process and learning.
Specifically, step S3 further includes:
step S300, a learning knowledge structure of the learner is constructed according to the previous learning process information of the learner.
As shown in fig. 2, step S300 specifically includes:
collecting previous learning process information of the learner:
collecting knowledge points learned by the learner in the course and new knowledge points created by the learner from the courses participated by the learner (such as editing, discussing and learned course information of the existing courses) and the learning progress in the courses (such as adding contents of the existing courses by the user in the new courses or adding contents of the existing courses); if the learner supplements the corresponding knowledge points or the knowledge experience of the learner under a course, the learner is considered to participate in the resource construction of the course. In addition, the learner can also be used as a teacher to set up a course system of the learner for recording the experience of the learner or the resources of the course.
Constructing a user-knowledge point relation knowledge graph:
according to the courses collected by the learners in the last step for participation, the created courses and the related knowledge points, the linkage between the points of the user-knowledge points is carried out, thus forming a relation knowledge graph of the user-knowledge points;
carrying out entity extraction, relation extraction and semantic recognition;
the knowledge-graph obtained in the last step integrates the courses created by the learner and the courses participated by the learner, and then the two parts of the learner are separated. Namely, two sub knowledge maps are separated from the knowledge map of the previous step: user-created course knowledge-graph and user-participating course knowledge-graph;
and obtaining the courses created by the learners and the courses participated by the learners according to the extraction result to form a learner course map, and forming the learner cognitive map according to knowledge points of the courses.
Step S301, learning content belonging to the preference of the learner is obtained according to the obtained emotion score of the learning content of the learner, and course information is extracted according to the learning content and the knowledge preference of the learner is arranged.
For example, in step S1, the learner' S previous course information is collected: software engineering, PHP language, etc., and then collect the knowledge elements of these courses: and (4) performing summary design, requirement analysis, basic sentences and the like, and describing according to the corresponding course knowledge elements to form the cognitive structure of the learner. Secondly, calculating the click times of the knowledge elements according to the click rates of different sections of the course participated by the user, and changing the sizes of different knowledge elements according to the occurrence times, as shown in FIG. 3; then, the emotional score is calculated by matching the emotional dictionary through the learner's evaluation of the last lesson, the interactive behavior of like the text information corresponding to like behavior (e.g. "like" means "affirming to lesson") and is drawn into the learner's knowledge preference according to the score, as shown in FIG. 4.
In step S4, a learner drawing (including the lesson map in which the learner participated and the lesson map created by the learner) is generated according to the knowledge preference of the learner constructed in step S3 and the learning knowledge structure of the learner constructed in step S3.
Specifically, in step S4, according to the cognitive structure (fig. 3) and knowledge preference (fig. 4) of the learner constructed in step S3, the learner is classified according to the size of the nodes as the criteria, and the nodes are classified by colors, so that the lesson and the knowledge element are distinguished by two colors, thereby having a certain content structure, and obtaining two graphs as shown in fig. 5a and 5b, which are the learner portrait, and the learner portrait includes the lesson in which the learner participates and the two graph modules of the lesson created by the learner. With reference to fig. 5, fig. 5a is a knowledge graph of lessons participated by a learner (the node size represents the learner's preference relative to other nodes), fig. 5b is a knowledge graph created by a learner in the process of participating in the creation of knowledge and learning resources, the learning bias of the learner is shown to be in the category of "research method" and "software engineering" according to the node size, color and created lessons in fig. 5a and 5b, and finally, the two types of learner figures are generated by calling the existing data of the system and by means of corresponding visualization plug-ins, such as Echart and the like.
The invention uses knowledge graph form to construct visual learner portrait, the richness of nodes in the knowledge graph represents the knowledge density and the association degree of the nodes, the link between the nodes shows the connectivity between the nodes, the connection between the nodes is expressed on the connection line by characters, thus realizing the network representation based on the relationship, revealing the connectivity between the nodes such as knowledge, users and courses, realizing the visual knowledge graph to display the learner learning portrait, and being convenient and intuitive to display the learner information to the users. In addition, in the process of student doing while learning, the method can realize automatic generation of the image of the learner and can also enable the learner to obtain the suggestion of the future learning path in the learning process.
FIG. 6 is a system architecture diagram of an apparatus for learner representation generation based on knowledge-maps in accordance with the present invention. As shown in FIG. 6, the present invention provides an apparatus for generating a learner profile based on a knowledge-map, comprising:
the learning information collecting module 601 is used for collecting learning information of learners.
The learning information includes, but is not limited to, interactive utterance behaviors of a learner on the web for a course, a progress of a taken course, an evaluation suggestion for a course, an evaluation and suggestion for a teacher who is in the course, a total number of courses participated, and the like, that is, when a user takes a course using the apparatus to which the present invention is applied, the learning information collecting module 601 collects interactive utterance behaviors of a click amount for a course, a progress of a taken course, an evaluation suggestion for a course, an evaluation and suggestion for a teacher who is in the course, a total number of courses participated, and the like of a learner on the web.
The emotion analysis module 602 is configured to disassemble the interactive information of the learner, summarize an emotion analysis task into multiple levels, such as words, phrases, attributes, sentences, chapters, and the like, according to the granularity of the text, determine emotion polarities by matching with a pre-constructed emotion dictionary, calculate emotion scores according to weights of the emotion polarities, and output emotion tendencies of the learner according to the emotion values.
The emotion analysis module 602 mainly obtains the preference type and preference degree of the learner for the current course or teacher through an emotion analysis task. In the present invention, the emotion preferences and emotion values generated by the emotion analysis task are one of the data generated by the learner representation, so through the emotion analysis task, the task attempts to acquire the type and hierarchy of preferences about the learner's preference for the current lesson or teacher. In the invention, an emotion analysis task mainly acquires interaction times of all language comment words and learners under different emotion preference buttons (referring to praise and poor comment buttons in the learning process) and superposes emotion types and emotion values represented by each word or each interaction, so that the emotion analysis task firstly disassembles interaction data of the learners, summarizes the task into multiple levels of words, phrases, sentences, chapters and the like according to the granularity of text languages and matches a pre-constructed emotion comparison dictionary; then, determining the emotion polarity of the word through the emotion comparison dictionary, and calculating the emotion score of the word; because phrases, sentences and chapters are the most basic word components, the emotional tendency and the emotional value of the whole text can be calculated only by further accumulation and taken as the emotional tendency of the learner. The step of disassembling the interactive data of the learner refers to disassembling the number of times that the learner clicks on the course, the learning progress of the course (the learning progress of a student of a certain course is converted into the learning percentage of the student on the course, the high percentage indicates that the student is more interested in the course), the evaluation of the course, the evaluation of a teacher of the course, and the approval of the course (the emotional marks which can be replaced by the approval of the course are manually set).
Preferably, the apparatus for generating a learner representation based on a knowledge-map of the present invention further comprises:
the emotion dictionary construction module constructs an emotion comparison dictionary according to the existing Chinese ontology emotion dictionaries which can be obtained openly by referring to an emotion dictionary construction method mentioned in 'an emotion dictionary automatic construction method overview' (Wangke, Chari. automated chemical article, 2016,42(4):495-511.doi: 10.16383/j.aas.2016.2016.c150585) in the prior art. After the emotion comparison dictionary provides the emotion types and emotion values corresponding to more than 7000 emotion vocabularies, the emotion analysis tasks are summarized into multiple levels such as words, phrases, sentences, chapters and the like according to the granularity of the text, the emotion dictionary can accurately judge the emotion polarity of the words/phrases, and the sentence/chapter level emotion analysis tasks can be effectively assisted in an accumulation mode. Considering that some ambiguity exists in the individual evaluation words, a combined evaluation unit (such as less-good and equivalent-good) is also considered in the construction of the emotion dictionary, if the combined evaluation unit is less than the combined evaluation unit, the emotion value is multiplied by the value in the interval of 0.1-1, and if the combined evaluation unit is equivalent, the combined evaluation unit is multiplied by the value between 1.1-10.
Specifically, the emotion analysis module 602 collects information about learner evaluation on a course, teacher evaluation, and course click-and-click amount, and matches the collected information with words in an emotion comparison dictionary to obtain emotion values, and counts the total of click-and-click button operations, and the statistical process mainly includes obtaining emotion tendency of the learner on the course according to the accumulated emotion scores of positive and negative emotion categories (the strength value of the emotion is represented according to the score and the positive or negative emotion is determined according to the score), analyzing the output emotion scores of words in the target field, and performing multi-word iteration to obtain emotion scores of all emotion words. For example, a learner participates in a computer network course and issues a plurality of comments aiming at the course, i.e., "i like the teacher" and "the content of the lesson is rich", etc., the system executes a dictionary matching task of emotion analysis on all the comments, if words like and rich exist in the dictionary, the emotion category and the emotion value corresponding to the words are returned, and then all the comments are superposed to be used as the emotion category and the emotion value of the learner. And in another case, the learner makes statistics on whether the final emotion category of the learner is positive or negative according to the principle that the learner makes positive comments and negative comments on the interactive interface of the course for a plurality of times and a small number of negative comments on other data such as knowledge points or exercises and the like in the computer network course, and the statistical value is also used as the emotion value. And finally adding the emotion types and the emotion values of the two cases to serve as the emotional bias and the emotion value of the learner for the course.
A learning knowledge structure and learning knowledge preference constructing module 603, configured to construct learning knowledge structure and learning knowledge preference of the learner according to the learning information collected by the learning information collecting module 601 and the emotion analysis result of the emotion analyzing module 602.
Specifically, the learning knowledge structure and learning knowledge preference building module 603 further comprises:
and the learning knowledge structure construction module is used for constructing the cognitive structure of the learner according to the previous learning process information of the learner.
And the learning knowledge preference construction module is used for obtaining the learning course and the knowledge element preferred by the learner according to the obtained emotional information of the learner.
For example, the learning information collection module 601 first collects the course information of the learner: software engineering, PHP language, etc., and then collect the knowledge elements of these courses: summary design, requirement analysis, basic sentences and the like, wherein the learning knowledge structure building module outputs the learning knowledge structure according to the corresponding course knowledge elements to form a cognitive structure of the learner, as shown in FIG. 3; then, the emotion analysis module 602 calculates an emotion value through the learner's evaluation of the last lesson, the interactive behavior of like praise, and the learning knowledge preference construction module marks the lesson with positive tendency and draws the lesson into the knowledge preference of the learner, as shown in fig. 4.
The learner representation generation module 604 is configured to generate a learner representation using a knowledge map according to the knowledge preferences and cognitive structures of the learner, which are constructed in the learning knowledge structure and learning knowledge preference construction module 603.
Specifically, the learner profile generation module 604 classifies importance levels according to the learning knowledge structure and the learning knowledge structure constructed by the learning knowledge structure and learning knowledge preference construction module 603 by using a knowledge graph in terms of node size, performs color classification by using node colors, and has a certain content structure, so as to obtain two graphs as shown in fig. 5a and 5b, which are the learner profiles. Learner representation provides two map modules of learner created course and learner participated course, combining with figure 4, upper course map created for learner person, lower cognitive map formed for learner person in course of participating creating knowledge and learning resource. From fig. 5a, fig. 5b, it can be seen that the learning bias of the learner is in the category of "research methods" and "software engineering". Finally, the two learner representations are generated by calling the system existing data and by means of corresponding visualization plug-ins, such as Echart.
Examples
As shown in fig. 7, in the present embodiment, a method for generating a learner representation based on a knowledge-map includes the following steps:
step one, collecting learner learning information: collecting interactive speaking behaviors of learners on the online, such as the click quantity of the lessons, the progress of the passed lessons, evaluation suggestions of the lessons, evaluation and suggestions of teachers in any lesson, the total number of the participated lessons and the like;
step two, emotion analysis: and constructing the emotion dictionary by referring to the method of Wangke and Xiurei for automatically constructing the emotion dictionary, and performing emotion analysis under different granularities according to the emotion tendencies of words provided by the emotion dictionary. And constructing an emotion dictionary, setting the emotion polarity of the viewpoint and endowing the set emotion polarity with a weight value. And (3) decomposing the interactive information content of the learner to form a text, summarizing the emotion analysis task into a plurality of levels of words, phrases, attributes, sentences, chapters and the like according to the granularity of the text, and matching the emotion dictionary to process the emotion polarity. And calculating an emotion value (emotion score) according to the weight of the emotion polarity, and outputting preferred learning content of the learner according to the emotion value. Generally, a larger emotion value indicates a higher preference degree of the learner for the course or the knowledge element, and a smaller emotion value indicates a lower preference degree of the learner;
step three, constructing a learning knowledge structure and learning knowledge preference of the learner: learning content with the learner emotion value being a positive tendency can be obtained according to the obtained learner emotion information, and knowledge preference of the learner is arranged according to the extracted course information of the learning content, as shown in fig. 8; secondly, the cognitive structure of the learner is constructed according to the previous learning process information of the learner
Step four, the learner draws a portrait: according to the two maps constructed in the third step, the curriculum map created by the learner and the cognitive map formed by the learner in the process of participating in creating knowledge and learning resources can directly obtain the learner portrait and visually display the learner portrait in the form of the knowledge map.
In summary, the method and apparatus for generating a learner portrait based on a knowledge-graph according to the present invention collect interaction information of a learner, extract data by methods such as emotion analysis and regular expression, and automatically generate a learner portrait in a form of a knowledge-graph visualization, so that the generated learner portrait is provided to a lecturer for adjusting a course according to the learner, and the learner can also know a previous learning situation of the learner. The invention makes up the defects of the self-adaptive learning system in the aspect of user portrait visualization, can enable a teacher to fully master the information of the prior knowledge structure, learning deviation and the like of students, and can enable the students to more intuitively know the prior learning structure and have deeper visual feeling on the preference of the students.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Modifications and variations can be made to the above-described embodiments by those skilled in the art without departing from the spirit and scope of the present invention. Therefore, the scope of the invention should be determined from the following claims.
Claims (10)
1. A learner portrait generation method based on knowledge-graph includes the following steps:
step S1, collecting learner learning information;
step S2, the interactive information content of the learner is disassembled, the emotion analysis task is summarized into a plurality of levels such as words, phrases, attributes, sentences, chapters and the like according to the granularity of the text, the emotion polarity is determined by matching with a pre-constructed emotion comparison dictionary, the emotion score is calculated according to the weight of the emotion polarity, and the emotion tendency of the learner is output according to the emotion value;
step S3, constructing the learning knowledge structure and learning knowledge preference of the learner according to the learning information collected in step S1 and the emotion analysis result in step S2;
and step S4, generating a learner portrait knowledge map by using the knowledge map according to the knowledge preference and the cognitive structure of the learner constructed in the step S3.
2. A method of generating a learner representation based on a knowledge-map as claimed in claim 1, wherein: the learning information includes, but is not limited to, interactive behaviors of a learner on the online click through a course, the progress of a last course, evaluation suggestions for the course, evaluation and suggestions for teachers who are lessons, the total number of courses participated, and the like.
3. The method of claim 2, wherein before step S2, the method further comprises the steps of:
and step S2-1, constructing the emotion comparison dictionary according to the Chinese ontology emotion dictionary, setting the emotion polarity of the viewpoint and endowing the set emotion polarity with a weight value.
4. The method of claim 3, wherein the step S3 further comprises:
step S300, constructing a learning knowledge structure of the learner according to the previous learning process information of the learner;
step S301, learning content of the learner is obtained according to the obtained emotional information of the learner, and knowledge preference of the learner is arranged according to the extracted course information of the learning content.
5. The method as claimed in claim 4, wherein the step S300 further comprises:
collecting previous learning process information of the learner;
constructing a user-knowledge point relation knowledge graph;
carrying out entity extraction, relation extraction and semantic recognition;
and obtaining the courses created by the learners and the courses participated by the learners according to the extraction result to form learner course maps and learner cognitive maps.
6. The method as claimed in claim 5, wherein the step S4 is performed to classify the learner' S knowledge preferences and cognitive structure by node size according to the knowledge graph constructed in the step S3, and the color classification by node color has a certain content structure, so as to obtain two graph modules, i.e. learner-created course and learner-participated course.
7. An apparatus for generating a learner profile based on a knowledge-graph, comprising:
the learning information collection module is used for collecting learning information of learners;
the emotion analysis module is used for disassembling interactive information content of the learner, generalizing emotion analysis tasks into multiple levels such as words, phrases, attributes, sentences, chapters and the like according to the granularity of the text, matching an emotion dictionary to determine emotion polarity, calculating an emotion value according to the weight of the emotion polarity, and outputting the emotion tendency of the learner according to the emotion value;
the learning knowledge structure and learning knowledge preference construction module is used for constructing the learning knowledge structure and learning knowledge preference of the learner according to the learning information collected by the learning information collection module and the emotion analysis result of the emotion analysis module;
and the learner sketch generation module is used for generating the learner sketch by utilizing the knowledge map according to the learning knowledge structure and the knowledge preference and cognitive structure of the learner, which are constructed in the learning knowledge preference construction module.
8. The apparatus of claim 7, further comprising an emotion dictionary construction module for constructing the emotion comparison dictionary according to the Chinese ontology emotion dictionary, setting emotion polarities of viewpoints and weighting the emotion polarities set.
9. The apparatus of claim 8, wherein the learning knowledge structure and learning knowledge preference building module further comprises:
the learning knowledge structure construction module is used for constructing the cognitive structure of the learner according to the previous learning process information of the learner;
and the learning knowledge preference construction module is used for obtaining the learning content of the learner according to the obtained emotional information of the learner and extracting course information according to the content to arrange the knowledge preference of the learner.
10. The apparatus of claim 7, wherein the knowledge-graph based learner representation generating means is further configured to: the learner learning information includes, but is not limited to, interactive behaviors of a learner on the click amount of the lesson on the internet, the progress of the lesson that was taken, evaluation suggestions of the lesson, evaluation and suggestions of teachers in the lesson, the total number of lessons participated in, and the like.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011388997.4A CN112488887A (en) | 2020-12-02 | 2020-12-02 | Learner portrait generation method and device based on knowledge graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011388997.4A CN112488887A (en) | 2020-12-02 | 2020-12-02 | Learner portrait generation method and device based on knowledge graph |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112488887A true CN112488887A (en) | 2021-03-12 |
Family
ID=74938937
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011388997.4A Pending CN112488887A (en) | 2020-12-02 | 2020-12-02 | Learner portrait generation method and device based on knowledge graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112488887A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113535957A (en) * | 2021-07-27 | 2021-10-22 | 哈尔滨工业大学 | Conversation emotion recognition network model based on dual knowledge interaction and multitask learning, construction method, electronic device and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106205248A (en) * | 2016-08-31 | 2016-12-07 | 北京师范大学 | A kind of representative learning person generates system and method at the on-line study cognitive map of domain-specific knowledge learning and mastering state |
| CN107085803A (en) * | 2017-03-31 | 2017-08-22 | 弘成科技发展有限公司 | The individualized teaching resource recommendation system of knowledge based collection of illustrative plates and capability comparison |
| CN110543572A (en) * | 2019-08-26 | 2019-12-06 | 杭州奇迹在线科技有限公司 | learning cognitive path generation method based on cognitive map |
| CN111914162A (en) * | 2020-06-01 | 2020-11-10 | 西安理工大学 | Method for guiding personalized learning scheme based on knowledge graph |
-
2020
- 2020-12-02 CN CN202011388997.4A patent/CN112488887A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106205248A (en) * | 2016-08-31 | 2016-12-07 | 北京师范大学 | A kind of representative learning person generates system and method at the on-line study cognitive map of domain-specific knowledge learning and mastering state |
| CN107085803A (en) * | 2017-03-31 | 2017-08-22 | 弘成科技发展有限公司 | The individualized teaching resource recommendation system of knowledge based collection of illustrative plates and capability comparison |
| CN110543572A (en) * | 2019-08-26 | 2019-12-06 | 杭州奇迹在线科技有限公司 | learning cognitive path generation method based on cognitive map |
| CN111914162A (en) * | 2020-06-01 | 2020-11-10 | 西安理工大学 | Method for guiding personalized learning scheme based on knowledge graph |
Non-Patent Citations (3)
| Title |
|---|
| 夏雪莹等: "基于学生画像的个性化学习辅导策略研究", 黑龙江生态工程职业学院学报, vol. 33, no. 3, pages 125 - 128 * |
| 田娜;周驿;: "基于MOOC课程评论的话题挖掘与情感分析研究", 软件导刊, no. 08 * |
| 白杨: "融合文本情感分析的个性化学习模式", 现代盐化工, no. 3, pages 149 - 150 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113535957A (en) * | 2021-07-27 | 2021-10-22 | 哈尔滨工业大学 | Conversation emotion recognition network model based on dual knowledge interaction and multitask learning, construction method, electronic device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Martins et al. | Findings on teaching machine learning in high school: A ten-year systematic literature review | |
| Lee et al. | Computational thinking integration patterns along the framework defining computational thinking from a disciplinary perspective | |
| Chen et al. | Detecting latent topics and trends in educational technologies over four decades using structural topic modeling: A retrospective of all volumes of Computers & Education | |
| Sorour et al. | Comment data mining to estimate student performance considering consecutive lessons | |
| Liu et al. | Emotion and associated topic detection for course comments in a MOOC platform | |
| Misiejuk et al. | State of the field report on learning analytics | |
| Bryfczynski | BeSocratic: An intelligent tutoring system for the recognition, evaluation, and analysis of free-form student input | |
| Oliva Córdova et al. | An experience making use of learning analytics techniques in discussion forums to improve the interaction in learning ecosystems | |
| Xie et al. | Virtual reality primary school mathematics teaching system based on GIS data fusion | |
| Aridor et al. | Students’ aggregate reasoning with covariation | |
| Huang et al. | Developing argumentation processing agents for computer-supported collaborative learning | |
| CN113934846A (en) | Online forum topic modeling method combining behavior-emotion-time sequence | |
| CN112488887A (en) | Learner portrait generation method and device based on knowledge graph | |
| CN116226410B (en) | Teaching evaluation and feedback method and system for knowledge element connection learner state | |
| Lakho et al. | Development of an integrated blended learning model and its performance prediction on students’ learning using Bayesian network | |
| Gottipati et al. | Automated discussion analysis-framework for knowledge analysis from class discussions | |
| Jiang | [Retracted] Data Analysis of College English Speculative Reading Teaching Based on Multimodal Teaching Theory in a Complex Cross‐Cultural Environment | |
| Li et al. | Data-driven analytics for student reviews in China’s higher vocational education MOOCs: A quality improvement perspective | |
| Yu et al. | Education data mining: How to mine interactive text in MOOCs using natural language process | |
| Favareto et al. | Integrated Solution for Evaluating Emotional Expressions in Learning Management Systems | |
| Scheihing et al. | Understanding the role of micro-blogging in B-learning activities: Kelluwen experiences in Chilean public schools | |
| El Maazouzi et al. | ChatBot Integrated with sentiment anlaysis: Application to E-Learning | |
| Li et al. | Design of a Big Data-based Evaluation System for Teacher-Student Cooperation in College English Writing | |
| Poitras et al. | Preservice teacher disengagement with computer-based learning environments | |
| Milisavljevic-Syed | Quantification of students’ active learning in design, build, and test engineering modules |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |