CN110210787A - The measure of user behavior under online learning environment is measured based on collective's attention - Google Patents
The measure of user behavior under online learning environment is measured based on collective's attention Download PDFInfo
- Publication number
- CN110210787A CN110210787A CN201910516743.7A CN201910516743A CN110210787A CN 110210787 A CN110210787 A CN 110210787A CN 201910516743 A CN201910516743 A CN 201910516743A CN 110210787 A CN110210787 A CN 110210787A
- Authority
- CN
- China
- Prior art keywords
- user
- attention
- network
- collective
- flow
- 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
- 230000006399 behavior Effects 0.000 title claims abstract description 41
- 238000005259 measurement Methods 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 36
- 238000004458 analytical method Methods 0.000 claims description 16
- 238000000034 method Methods 0.000 claims description 14
- 230000007704 transition Effects 0.000 claims description 13
- 238000009825 accumulation Methods 0.000 claims description 9
- 238000003780 insertion Methods 0.000 claims description 7
- 230000037431 insertion Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000003542 behavioural effect Effects 0.000 claims description 3
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 239000012530 fluid Substances 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 3
- 240000000233 Melia azedarach Species 0.000 claims 1
- 238000013461 design Methods 0.000 abstract description 4
- 230000003993 interaction Effects 0.000 abstract description 4
- 239000000523 sample Substances 0.000 abstract description 3
- 238000011156 evaluation Methods 0.000 abstract 1
- 238000011160 research Methods 0.000 description 8
- 238000004422 calculation algorithm Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000001568 sexual effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000002459 sustained effect Effects 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
- 230000007306 turnover Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3438—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment monitoring of user actions
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- 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
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Educational Administration (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Technology (AREA)
- General Engineering & Computer Science (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Computer Hardware Design (AREA)
- Operations Research (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses the measures that user behavior under online learning environment is measured based on collective's attention, based on flow network theory, establish a set of on-line study resource platform user group and measure evaluation scheme, helpdesk designs more efficient teaching system.The clickstream data that the present invention is generated from a large number of users on-line study, construct the opening flow network based on collective's attention, and the indexs such as network flow, stream distance, spatial embedding are utilized respectively, go to probe into collective's Automobile driving rule of different student groups in on-line study.It interaction that the present invention is gone to consider between teaching and teaching with the angle of system and influences each other with student group collective behavior, to more dynamic and the intuitively objective law in announcement on-line study data, so that measurement results are more objective, have more practical guided significance.
Description
Technical field
The present invention relates to network analysis, study analysis and educational assessment field are directed generally to research collective's attention and exist
Assignment problem in each educational resource helps the part learning behavior of rational learning person, seeks different learning type persons
Habit mode, to measure assessment and design more efficient education resource platform.
Background technique
Online Open Course (MOOCs) and campus course mixed mode start, change educator for
Knowledge is how to generate, the view propagated and consumed.We have had been introduced into one by " information abundant creates attention
There is a serious shortage in the supply " epoch.Although the open flexible characteristic of on-line study can let us at any time, any place is learned
It practises, but does not consider that the cost problem of Automobile driving will substantially reduce the study benefit of ours on it.When learner faces
Resource quantity when reach certain scale, they can only distribute limited energy and learn certain amount of series of knowledge, and
It is not to catch all in one draft all teaching resources.Therefore, although this opening and flexible characteristic are usually described as nothing by people
The human attention of constraint, and think that the resource of magnanimity and cheap learning cost can allow education to become more to popularize, but
As frequently reporting, this marketing slogan must receive query and verifying from educational research person.Student's is effective
How study depends on them by Automobile driving to useful resource, as simmons is previously mentioned in connectionism,
In wide information space, content is just as an isolated resource node.Student needs to set the learning objective of oneself, specifies
With adjust oneself learning strategy, " in the complicated mode for understanding information in thread environment and find oneself ", study and knowledge
Essence changed.Study no longer individually occur it is a in vivo, it includes the pathfinding and consciousness of student.Knowledge no longer has solid
Fixed structure, it is the information network structure formed by student based on the self-construction of informational linkage, we more should be recognized that collection
Body notices how force flow circulates in knowledge abundant.Especially with the continuous increasing of open educational resources acquisition capability
Long, this will be the key that form is experienced in the on-line study that futuristic design is more preferable out, more flexible.
This concept of collective's attention initially proposes that they attempt with a new model by Huberman and Wu Fang
Come portray user online attention dynamic evolution, the popularity and novelty of itself and information (knowledge) are connected.With biography
Attention type in system psychological study is different, since online behavior expression goes out mode similar with human attention, such as
Short-time characteristic, selectivity, circulation and dissipation.The collective that this concept emphasizes that a large amount of individual behaviors generate emerges in large numbers benefit,
To measure the importance of whole attention level, facilitate moving for our its online behavior developed at any time of more preferable simulation
State.
Have benefited from the extensive prevailing of on-line study, it is study analysis that a large amount of learner, which browses and clicks learning behavior data,
Research provides a unprecedented chance.Previous study analysis uses various techniques to understand clickstream data, such as logical
It crosses clustering algorithm to classify to learner, and the analysis of usage log sequence is to predict achievement.These click steam research trends in
It owes generalities (under-conceptualized) and crosses method (over-methodologized).If clickstream data
It is considered only as the mass data pond about human behavior, is inputted as complicated algorithm, to cluster or predict to learner
Future, then its value will have a greatly reduced quality.Since the proximal segment time, ecosystem viewpoint has been acknowledged as the effective of on-line study
The concept of network ecosystem is introduced into appraisement system by means in the present invention, it is intended to open the systematic perspective with balance from it
High turnover rate present in Explanation-based Learning And Its person and sharply unequal participation scale emphasize that describing learning behavior in collective's level moves
The importance of state feature, including the learner of (or failure) of doing very well, and the learner that may be discontinued one's studies.
Summary of the invention
Past study analysis lacks generalities and crosses method, merely from the click of the angle analysis student of data mining
Flow data, not from the angle of system, large scale consider the interaction between imparting knowledge to students and imparting knowledge to students in system and with student group
Body collective behavior influences each other.In view of the above-mentioned problems, present invention combination flow network is theoretical, a set of pair of on-line study is had devised
The measure of user behavior under environment constructs open flow network from the angle of collective's attention, passes through using clickstream data
The index analysis such as stream distance, spatial embedding in flow network theory pay attention to the accumulation of force flow, circulation, dissipation mode, thus more
Dynamic and the intuitive objective law disclosed in on-line study data, so that measurement results are more objective, anticipate with more practical guidance
Justice.
Technical problem
The clickstream data that the present invention is generated from a large number of users on-line study constructs the opening based on collective's attention
Flow network goes to probe into on-line study with correlation theory in Network Science and method, collective's attention of different student groups
Distribution dynamics.The present invention is respectively from network flow, stream distance, spatial embedding etc., by analyzing entire open flows network flow
The visualization insertion of the stream distance of the accumulation and dissipation of amount, each network node to source and sink nodes and two-dimentional studying space with
And the rules such as feature distribution, help online education worker to solve following problems:
Does is the first, which kind of feature and difference presented in the accumulation of collective, different user group attention, circulation, dissipation?
The second, collective, different user group attention in course immerse it is horizontal how?
How is the learning law of third, different user group in different course units?
Technical solution
Compared to traditional study analysis method, rough use to user clickstream data, the present invention is then from system
Angle is set out, and using user clickstream data, is constructed using click steam as side, and course is node based on user collective attention
Open flow network, and with flow, stream distance and the spatial embedding in flow network theory, to complete the analysis of on-line study system
And measurement.The specific method is as follows:
Step 1) user clickstream data prediction
1-4) data source: clickstream data is Tsinghua University in 2013 with online (xuetangX) from school
The first Chinese released is opened on a large scale for online course platform;
1-5) data format: clickstream data is user behaviors log of the user in on-line study website, has recorded user and is working as
Click behavior in preceding course learning space.Wherein each row record includes: user id, timestamp are (when page open and closing
Between), URLs, page title, page residence time, page type etc..
1-6) data filtering: clickstream data contains the behavioral data of mobile terminal and web terminal, since page jump is patrolled
The data format collect, stored has differences, and only retains the user data of web terminal.Meanwhile also eliminating user's browsing time length
For 0 second page;
Step 2) constructs open flow network
2-4) collective's attention flow network: flow network node is different chapters class journey in the present invention, and even side is then user
The circulation of collective's attention;
2-5) network struction: for convenience of processing data, in the present invention, rule of thumb, use divide as session within 30 minutes
Cut foundation.In each session, click (or access) sequence of each user is considered as once from a course resources
To the attention displacement behavior of another course resources, this transfer is defined as the company side between the two course nodes by we.
Accordingly, we can construct using course resources as node, and user clicks the click flow network that behavior is even side.
2-6) network balances: according to open flows network theory, whole network needs to keep flow equilibrium.For this purpose, we
Step 2-2) building flow network in add the special node of two classes --- source node and sink nodes respectively represent and pay attention to force flow
Source and the end;
Step 3) flow network stream distance, flow dissipation rule and Education Space insertion
It 3-4) flows distance: based on the open flow network of balance built, obtaining collective and pay attention to force flow matrix F, then obtain
Markov Transition Probabilities matrix M.Element in matrix indicates transition probability of collective's attention between resource, calculates public
Formula is as follows:
In formula, mijWhat is indicated is the transition probability from node i to node j.Based on the Markov Transition Probabilities matrix
M, and then stream distance matrix L is calculated, finally obtain the stream distance l of point-to-point transmission in networkij, calculation formula is as follows:
Wherein, U=M+M2+ ...=(I-M)-1, referred to as fundamental matrix is the inverse matrix of the Laplace operator of M.And I
It is then the unit matrix that size is (N+1) * (N+1);
3-5) fluid accumulation, circulation and the rule that dissipates: all courses can be calculated using the stream distance matrix L in step 3-1)
Provide the mean flow distance of source-to-source and sink nodes.Pass through these mean flow distances of the flow network of observation different user group type building
Rule, it can be realized that accumulation, circulation, dissipation characteristics and the difference of collective, different user group attention;Equally, pass through sight
Different course resources are examined to the stream range distribution rule in source and sink nodes, are studied in different user group, each course resources exist
Ecological niche in entire flow network, i.e. representing user immerses level for different course resources for this;
3-6) Education Space is embedded in: we will be flowed using MDS (multidimensional scaling) method apart from square
Battle array is embedded in, and network node (i.e. course resources) is embedded into two-dimensional space.Different in not group by observing, course provides
The different insertion rule in source, learns different user group in learning law and difference.
Beneficial effect
These are answered by the open flow network of collective's attention is established using the clickstream data from MOOC in text
Problem.In this approach, ideal model of the present invention from the learning path of open system angle map individual flexible learning person.
With the theory about open system and network dynamic, an embodiment of the flow data as sustained attention force flow, Ke Yili will click on
Solve situation of change of the human behavior on dynamic is a wide range of.Pay attention to the metaphor of force flow or indicate as the agency for explaining click steam,
This improves us to the continually changing learning behavior for understanding various learner groups.This meets individual learner for design
The educational resource of demand is of great significance.Specific benefit is as follows:
The first, the present invention is based on open flows network theories, can effectively capture various Sexual pattern of on-line study;
The second, in the present invention, exploitation flow network is using course resources as node, and attention flowing is side between course resources
Building, so the learning behavior of user can adequately be shown.It is not only that analysis on-line study behavior provides such as
The means such as stream distance, and the concept map etc. for effectively learning that certain course is contained can be disclosed for us;
Third, the present invention treat on-line study analysis with the visual angle of open system, and on-line study system is considered as a suction
It receives, the ecological space of circulation, the attention force flow that dissipates, by analyzing ecological niche of the different course resources in this space, we
It can formulate and distribute reasonable learning sequence and resource.
Detailed description of the invention
The online website (xuetangX) example in the school Fig. 1;
Fig. 2 social networks, clickstream data session segmentation and flow network building;
Fig. 3 open flows network diagram.
Specific embodiment
Below with reference to attached drawing and based on collective's attention measure online learning environment under user behavior measurement it is specific
Embodiment, the present invention will be described in detail.
The present invention devises the measurement side of user behavior under a set of pair of online learning environment from open flows network theory
Method, is broadly divided into two parts, and a part is to utilize the open flow network of user clickstream data building collective's attention, another portion
Divide then is to utilize method in open flows network theory --- stream distance, dissipation are restrained and the behavior pattern of spatial embedding analysis user,
Specific operation is as follows:
Step 1) user clickstream data prediction
Data source: clickstream data is online (xuetangX) from school, is that Tsinghua University released in 2013
First Chinese opens online course platform on a large scale.One of MOOCs platform as largest domestic, school accommodates online is more than
8000000 registration user.By the end of in March, 2018, the course that offer is counted on platform shares more than 1500, and course comes from state
Inside and outside well-known colleges and universities, cover including computer, through pipe, art, mathematics, physics, chemistry, social sciences including etc. 13 big subject doors
Class;In course learning space, by taking " psychology introduction " as an example, a variety of different functional modules are provided for student, such as: courseware,
Curriculum information, zone of discussion etc., as shown in Figure 1.In courseware modules, course provides the content of courses of multiple chapters and sections, comprising: view
Frequency courseware, exercise carry out autonomous learning for student and examine learning effect;In the module of zone of discussion, different themes are contained
Interaction model carries out exchange and interdynamic for student.In addition to this, there are also record student's self-study progress curricular advancement module,
Integrally-built program content module of course etc. is described;" psychology introduction " is school online from 2015, opening
One psychology basis course being under the jurisdiction of under philosophy discipline classification.Since 2015 open up, continuously opened by 2017
If 6 rounds.All rounds opened up between 2015 to 2017 to the subject find that autumn in 2015 opens into analysis
If round in register and possess that the number of student of behavior record, the behavior record quantity of student are most, and data volume is also the most
It is abundant, it is shown in Table 2.Therefore the round is chosen as analysis of cases;
Data format: clickstream data is user behaviors log of the user in on-line study website, has recorded user in current class
Click behavior in journey studying space, being shown in Table 1 is user in " psychology introduction " user behaviors log example.Wherein each row records
: user id, timestamp (page open and shut-in time), URLs, page title, page residence time, page type etc..With
Family id is used to carry out user unique identification, and timestamp is used to be ranked up user clickstream data and session cutting, URLs
For carrying out unique identification to the page in studying space, the page residence time is used to be filtered invalid click access,
And page title and type, then it is the information description to current page;For " psychology introduction " this subject, page mark
Topic is broadly divided into 13 classes and 13 chapters and sections correspond, and specific chapters and sections structure will be shown in table 3;
Data filtering: clickstream data contains the behavioral data of mobile terminal and web terminal, due on mobile terminal and web terminal
Page jump logic, the data format of storage have differences, in order to which data uniformly facilitate research, weed out all of mobile terminal
User data only retains the user data of web terminal.Meanwhile in User action log data, there are a part of browsing time is long
The page that degree is 0 second, these access records can not illustrate that user participates in course in study, therefore also eliminate this part
Data;In the selection of final data sample, this research anonymity is obtained in " psychology introduction " autumn round in 2015, practical
116356 user behaviors log data of 7397 students and its this crowd of student of study in web terminal.This batch data have recorded from
In May, 2015, click of the student in course between the different pages jumped behavior to during in December, 2016, was related to 229, band altogether
The page, comprising: the Video Courseware page, the forum postings page, program content etc..Meanwhile this research also accordingly obtains currently
The academic record data of 7397 students.Student group will be hereafter grouped based on academic record data (as it is very good, one
As, failure, exit), remove building collective's attention flow network using the user behaviors log data of student, and based on this Develop Data point
Analysis and research;
Step 2) constructs open flow network
Collective's attention flow network: different traditional using user as core building network, such as social networks, node is to use
Family, even side is then the interaction between user;Collective's attention open flows network node is different chapters class journey, Lian Bian in the present invention
It is then the circulation of user collective attention;
Network struction: rule of thumb, it is considered as one complete that the online behavior in 25.5 minutes occurs on average
Session, be then another session more than this time.For convenience of processing, we in the present invention, use 30 minutes as session
Divide foundation.In each session, the click of each user (or access) sequence is indicated once from a page jump to another
One page, i.e., once from a course resources to the attention displacement behavior of another course resources, we are this transfer
It is defined as the company side between the two course nodes.Accordingly, we can construct using course resources as node, and user clicks behavior and is
The even click flow network on side, such as Fig. 2.
Network balance: according to open flows network theory, whole network needs to keep flow equilibrium.For this purpose, we are in step
2-2) the special node of two classes is added in the flow network constructed --- source node and sink nodes respectively represent and enter from external environment
Collective in course resources network pays attention to the source of force flow and is dissipated to the end of external learning environment from course resources network,
Such as Fig. 2,3;
Step 3) flow network stream distance, flow dissipation rule and spatial embedding
Flow distance: since flow network is there are source and sink nodes, whole network becomes an open system.Therefore in the past only
The traditional algorithm of close network interior joint distance can be used to calculate and be not suitable for.In order to measure collective's attention flow network interior joint
Between distance, present invention employs a kind of distance metric calculated based on user browsing behavior, referred to as stream distance, represent from
Flow network interior joint A reaches the average first up to distance of node B.Specific calculating process is to be primarily based on the balance built to open
Flow network obtains collective and pays attention to force flow matrix F, then obtains Markov Transition Probabilities matrix M.Element in matrix, indicates
Transition probability of collective's attention between resource, calculation formula are as follows:
In formula, mijWhat is indicated is the transition probability from node i to node j.In the attention flow network of balance, due to
The outflow of attention is not present in sink nodes, therefore the transition probability perseverance of sink nodes to any node is 0.In addition to this, for it
Remaining all node is, the sum of transition probability are constantly equal to 1, i.e.,Illustrate that the collective for being flowed into present node pays attention to
Power finally can all be flowed out in other nodes with different probability distribution, this is also close flow network and open flow network one
A essential difference embodies.Based on Markov method matrix M, and then stream distance matrix L is calculated, finally obtained in network
The stream distance l of point-to-point transmissionij, calculation formula is as follows:
Wherein, U=M+M2+ ...=(I-M)-1, referred to as fundamental matrix is the inverse matrix of the Laplace operator of M.And I
It is then the unit matrix that size is (N+1) * (N+1);
It is calculated based on above matrix conversion, we have just obtained for a pair of of section every in a description attention flow network
Distance matrix is flowed between point.It but is under normal conditions, directive, stream of the node i to node j since the click between the page jumps
Distance and not equal to node j to node i stream distance.MDS algorithm can be used in order to subsequent, the node in distance matrix is embedding by flowing
Enter and visualized into Euclidean space, it would be desirable to obtain a symmetrical stream distance matrix C.It is then simple in the present invention
Matrix L is added with the matrix L .T after its transposition.Stream distance after being added can intuitively be interpreted as the average past of point-to-point transmission
It returns first up to distance.Its calculation formula is as follows: Cij=lij+lji。
Fluid accumulation, circulation and the rule that dissipates: all course resources can be calculated using the stream distance matrix L in step 3-1)
To the mean flow distance in source and sink nodes.Distance to source node shows to obtain course resources after user enters on-line study system
Speed diversity.I.e. it is considered that attention has been run up in these course resources.Stream distance to sink nodes then implies note
A possibility that meaning force flow dissipates in learning process;For different user types, all classes in respective flow network are calculated separately
Journey provides the mean flow distance of source-to-source and sink nodes, accumulation, circulation, the dissipation of collective, available different user group attention
Which kind of feature and difference is presented;Equally, also can use stream distance matrix L calculate different course resources to source and sink nodes stream
Distance, study different user group in, course resources attract and dissipate pay attention to force flow ecological niche, i.e., different user group for
Course resources immerse level;
Education Space insertion: in view of the geometric distance characteristic of stream distance matrix, we use MDS
(multidimensional scaling) insertion, it is in the case where being able to maintain original relative positional relationship, network node is embedding
Enter into two-dimensional space.For different user types, we calculate separately the two dimension insertion of all course resources, study them
The regularity of distribution, thus it can be seen that learning law of the different user group in different course units.
Table 1 " psychology introduction " User action log data format
Number of student statistics in table 2 " psychology introduction " difference round
Table 3 " psychology introduction " course chapters and sections structure
Claims (1)
1. measuring the measure of user behavior under online learning environment based on collective's attention, which is characterized in that difference and biography
The rough use to user clickstream data of system is constructed from the angle of system using user clickstream data to click
Stream is side, and course is the opening flow network based on user collective attention of node, and with the stream distance in flow network theory
And spatial embedding, to complete the analysis and measurement of on-line study system;The specific method is as follows:
Step 1) user clickstream data prediction
1-1) data source: the user that clickstream data records in on-line study system clicks user behaviors log;
1-2) data format: clickstream data has recorded click behavior of the user in course learning space;Wherein each row record
Include: user id, timestamp include page open and shut-in time, URLs, page title, page residence time, classes of pages
Type;
1-3) data filtering: clickstream data contains the behavioral data of mobile terminal and web terminal, due to page jump logic, storage
The data format deposited has differences, and only retains the user data of web terminal, meanwhile, also eliminating user's browsing time length is 0 second
The page;
Step 2) constructs open flow network
2-1) collective's attention flow network: being different chapters class journey in flow network node, Lian Bianzeshi user collective attention
Circulation;
2-2) network struction: 30 minutes are used as session and divides foundation;In each session, the click or visit of each user
Ask that sequence is considered as once from a course resources to the attention displacement behavior of another course resources, this transfer is calmly
Company side of the justice between the two course nodes;Accordingly, using course resources as node, user clicks the click that behavior is even side for building
Flow network;
2-3) network balances: according to open flows network theory, whole network needs to keep flow equilibrium;In step 2-2) building
The special node of two classes is added in flow network --- source node and sink nodes respectively represent the source for paying attention to force flow and the end;
Step 3) flow network stream distance, flow dissipation rule and Education Space insertion
It 3-1) flows distance: based on the open flow network of balance built, obtaining collective and pay attention to force flow matrix F, then obtain Ma Er
Section husband transition probability matrix M;Each element in matrix indicates transition probability of collective's attention between resource, and calculation formula is such as
Shown in lower:
In formula, mijWhat is indicated is the transition probability from node i to node j;Based on Markov Transition Probabilities matrix M, into
And stream distance matrix L is calculated, finally obtain the stream distance l of point-to-point transmission in networkij, calculation formula is as follows:
Wherein, U=M+M2+ ...=(I-M)-1, referred to as fundamental matrix is the inverse matrix of the Laplace operator of M;And I is then
Size is the unit matrix of (N+1) * (N+1);
3-2) fluid accumulation, circulation and the rule that dissipates: all course resources are calculated to source using the stream distance matrix L in step 3-1)
With the mean flow distance of sink nodes;By observing the rule of these mean flow distances of the flow network of different user group type building,
Recognize accumulation, circulation, dissipation characteristics and the difference of collective, different user group attention;Equally, by observing different courses
The stream range distribution rule for providing source-to-source and sink nodes, is studied in different user group, each course resources are in entire flow network
In ecological niche, this represents user and immerses level for different course resources;
3-3) Education Space is embedded in: it is embedded in using MDS, that is, multidimensionalscaling method by distance matrix is flowed,
Network node, that is, course resources are embedded into two-dimensional space;It is different in not group by observing, different embedding of course resources
Enter rule, learns different user group in learning law and difference.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910516743.7A CN110210787A (en) | 2019-06-14 | 2019-06-14 | The measure of user behavior under online learning environment is measured based on collective's attention |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910516743.7A CN110210787A (en) | 2019-06-14 | 2019-06-14 | The measure of user behavior under online learning environment is measured based on collective's attention |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110210787A true CN110210787A (en) | 2019-09-06 |
Family
ID=67792701
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910516743.7A Pending CN110210787A (en) | 2019-06-14 | 2019-06-14 | The measure of user behavior under online learning environment is measured based on collective's attention |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110210787A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112380465A (en) * | 2020-11-13 | 2021-02-19 | 兰州七度数聚技术有限公司 | Quantitative evaluation method of internet site influence based on attention flow network model |
CN112435152A (en) * | 2020-12-04 | 2021-03-02 | 北京师范大学 | Online learning investment dynamic evaluation method and system |
CN112529141A (en) * | 2020-12-11 | 2021-03-19 | 中国海洋大学 | Learning path generation method based on improved immune algorithm |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104484454A (en) * | 2014-12-27 | 2015-04-01 | 西安交通大学 | Knowledge map oriented network learning behavior and efficiency analysis method |
CN108985993A (en) * | 2018-08-21 | 2018-12-11 | 北京航空航天大学 | A kind of the network self-adapting control system and control method of on-line study |
-
2019
- 2019-06-14 CN CN201910516743.7A patent/CN110210787A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104484454A (en) * | 2014-12-27 | 2015-04-01 | 西安交通大学 | Knowledge map oriented network learning behavior and efficiency analysis method |
CN108985993A (en) * | 2018-08-21 | 2018-12-11 | 北京航空航天大学 | A kind of the network self-adapting control system and control method of on-line study |
Non-Patent Citations (3)
Title |
---|
仇建平: "互联网关注力的模型分析", 《智能系统学报》 * |
张洪玲等: "基于在线学习行为数据分析的网络教育教学研究与对策", 《情报科学》 * |
王俊: "有向加权网络节点有效距离模型及其应用研究", 《中国优秀硕士学位论文全文数据库 社会科学Ⅰ辑》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112380465A (en) * | 2020-11-13 | 2021-02-19 | 兰州七度数聚技术有限公司 | Quantitative evaluation method of internet site influence based on attention flow network model |
CN112435152A (en) * | 2020-12-04 | 2021-03-02 | 北京师范大学 | Online learning investment dynamic evaluation method and system |
CN112435152B (en) * | 2020-12-04 | 2023-04-18 | 北京师范大学 | Online learning investment dynamic evaluation method and system |
CN112529141A (en) * | 2020-12-11 | 2021-03-19 | 中国海洋大学 | Learning path generation method based on improved immune algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Felton* et al. | Web-based student evaluations of professors: the relations between perceived quality, easiness and sexiness | |
CN110210787A (en) | The measure of user behavior under online learning environment is measured based on collective's attention | |
Kay | Exploring the use of web-based learning tools in secondary school classrooms | |
Luo et al. | Predicting Student Grade Based on Free-Style Comments Using Word2Vec and ANN by Considering Prediction Results Obtained in Consecutive Lessons. | |
Mahajan et al. | Usage patterns discovery from a web log in an Indian e-learning site: A case study | |
Fatahi et al. | An empirical study of using sequential behavior pattern mining approach to predict learning styles | |
Pirolli et al. | A knowledge-tracing model of learning from a social tagging system | |
Baragash et al. | Sentiment analysis in higher education: a systematic mapping review | |
Watson et al. | What is the role of statistics in integrating STEM education? | |
CN110245133B (en) | Online Learning Course Analysis Method Based on Collective Attention Flow Network | |
Li | Role of machine learning and artificial intelligence algorithms for teaching reform of linguistics | |
Bittermann et al. | The landscape of research on prior knowledge and learning: A bibliometric analysis | |
Sorour et al. | Evaluation of effectiveness of time-series comments by using machine learning techniques | |
Tong et al. | An evaluation model based on procedural behaviors for predicting MOOC learning performance: students’ online learning behavior analytics and algorithms construction | |
Su et al. | Research on the improvement effect of machine learning and neural network algorithms on the prediction of learning achievement | |
Wei et al. | Personalized recommendation of learning resources based on knowledge graph | |
Than et al. | A meta-analysis on factors influencing student satisfaction in higher education | |
Seif et al. | A model for predicting intention to use e-learning based on epistemological beliefs | |
Mu et al. | Research on a personalized learning path recommendation system based on cognitive graph with a cognitive graph | |
Kearney et al. | Examining the impact of critical feedback on learner engagement in secondary Mathematics classrooms: a multi-level analysis | |
ARI et al. | Examining Turkish Students' Views on the Reading Literacy Tasks of PISA 2018 in Terms of Different Variables. | |
Kasper | Codifying figurational theory and mapping common ground in sociology... and beyond | |
Pavlekovic et al. | Comparison of intelligent systems in detecting a child’s mathematical gift | |
Xie et al. | SQKT: A student attention-based and question-aware model for knowledge tracing | |
Kawale et al. | Development of an e-learning system based on artificial intelligence |
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 |