CN110399982A - Incorporate the initialization of Bayesian Network Learning style and the correction algorithm of emotional factor - Google Patents
Incorporate the initialization of Bayesian Network Learning style and the correction algorithm of emotional factor Download PDFInfo
- Publication number
- CN110399982A CN110399982A CN201810345695.5A CN201810345695A CN110399982A CN 110399982 A CN110399982 A CN 110399982A CN 201810345695 A CN201810345695 A CN 201810345695A CN 110399982 A CN110399982 A CN 110399982A
- Authority
- CN
- China
- Prior art keywords
- learning
- initialization
- module
- learning style
- probability
- 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
- 230000013016 learning Effects 0.000 title claims abstract description 173
- 230000002996 emotional effect Effects 0.000 title claims abstract description 49
- 238000012937 correction Methods 0.000 title claims abstract description 43
- 230000006399 behavior Effects 0.000 claims abstract description 58
- 238000000034 method Methods 0.000 claims abstract description 23
- 230000008569 process Effects 0.000 claims abstract description 7
- 230000036651 mood Effects 0.000 claims description 15
- 230000008447 perception Effects 0.000 claims description 8
- 230000008451 emotion Effects 0.000 claims description 6
- 238000002474 experimental method Methods 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000001755 vocal effect Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000005034 decoration Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 238000007493 shaping process Methods 0.000 description 1
Landscapes
- Electrically Operated Instructional Devices (AREA)
Abstract
The invention discloses a kind of Bayesian Network Learning style initialization for incorporating emotional factor and correction algorithms, including simple problem initialization module, essential information initialization module, study emotional factor module, learning behavior probabilistic module, conditional probability table initialization and correction module and Bayesian network correction module；The simple problem initialization module initializes learning style using simple problem；The essential information initialization module initializes learning style by essential information；The weight that the study emotional factor module allows learner to select emotional factor determines current emotional；The learning behavior probabilistic module is the process for calculating study behavior probability；The conditional probability table initialization and correction module complete initialization and correction conditions probability tables；The Bayesian network correction module is modified learning style.By the above-mentioned means, the present invention can rapidly and accurately complete the initialization and amendment of learning style, learning path is recommended to do basis for adaptive and learning system.
Description
Technical field
The present invention relates to a kind of initialization of learning style and correction algorithm more particularly to a kind of pattra leaves for incorporating emotional factor
The initialization of this e-learning style and correction algorithm.
Background technique
In the trend that e-learning and long-distance education are modern education developments, adaptive and learning system is for learner
Learning style recommendation personalized education resource and learning path are nets with this come the system for improving the learning efficiency of learner
One pith of network study and long-distance education.
The basis that adaptive and learning system is recommended is learning style, and learning style initialization at present mostly uses Saloman to learn greatly
Practise style scale.This method initializes the learning style of learner using questionnaire form, there are problems that initialize questionnaire in compared with
The shortcomings that causing learner more over-burden, this makes learner that can not get on the ball questionnaire learning style is caused to initialize accuracy
Not high problem.The present invention in view of Saloman learning style scale there are the problem of, for learning style, each dimension is mentioned
44 problem reductions in the scale of Saloman are solved the disadvantage that more problems in the scale of Saloman, subtracted by a problem out at 4
The light burden of learner.Consider that learning style is related with the essential informations such as the gender of each learner and age simultaneously, because
And the present invention proposes the method that second of essential information by learner carries out initialization learning style.
Since the learning behavior of people can change with the variation of environment, especially with mood variation and change, and
The conditional probability table of forefathers be taken as after being obtained by expertise and specific experimental data it is unalterable, cannot be with feelings
The variation of thread and change, this can reduce accuracy to a certain extent.Number for different study crowds, in conditional probability table
According to that should be different, only can just be calculated more accurately according to different study crowds using different conditional probability numerical value
The learning style of study crowd out.
Existing learning style update the system, the mode for mostly using Bayesian network analysis learning behavior to be modified greatly,
The update the system of these learning styles has the following disadvantages: can be in study emotional factor whether do not account for learning style accurately
In embody, emotional factor is not incorporated into Bayesian network and learning behavior and corrects learning style jointly.
Summary of the invention
The purpose of the present invention is being initialized and being corrected learning style, to be accurately obtained the learning style of learner,
Recommend learning path for adaptive and learning system and personalized mode of learning be presented to do basis.
To achieve the above object, the technical scheme adopted by the invention is as follows: provide it is a kind of incorporate emotional factor Bayesian network
The initialization of network learning style and correction algorithm, which is characterized in that initialize mould including simple problem initialization module, essential information
Block, study emotional factor module, learning behavior probabilistic module, conditional probability table initialization and correction module and Bayesian network are repaired
Positive module；
The simple problem initialization module carries out initialization learning style by allowing learner to answer four simple problems；
The essential information initialization module is to carry out initialization study according to the gender of learner, age and academic essential information
Style；
The study emotional factor module is to determine current study feelings by allowing learner to select the weight of study emotional factor
Thread；
The learning behavior probabilistic module is to carry out calculating study behavior probability by certain rule；
Conditional probability table initialization and correction module are to be initialized by the data of a large amount of learners and correction conditions
Probability tables；
The Bayesian network correction module is carried out using study emotional factor, conditional probability table numerical value and learning behavior probability
Correct the process of learning style.
The simple problem initialization module is for Information procession, perception, input in learning style and to understand four dimensions
The different characteristics of degree proposes a problem respectively, while designing option there are two each problems, just by each problem
Each learning style dimension can be initialized, comprehensive four learning style dimensions can complete the learning style of simple problem
Initialization procedure.
The essential information initialization module is the learning style value shape by analyzing the learner of various essential informations
Cheng Guan
In the corresponding relationship of the essential information of the value and learner of learning style, need to initialize learning style is obtained thereafter
Habit person's
Essential information is compared by the essential information with the corresponding relationship of a large amount of learners generally counted, by the learner
Most probable
Initial learning style of the learning style as the learner.
The study emotional factor module is will to learn mood to be divided into glad, surprised and detest three types, passes through restriction
The weight of three kinds of type of emotion is between 0 and 1 and the sum of three is equal to 1, will learn the weight of emotional factor as probability number
Value uses in the Bayesian network correction module for incorporating emotional factor.
Steps are as follows for the learning behavior probabilistic module:
(1) upper limit for defining each learning behavior first in a learning cycle learns number；
(2) number and the upper limit that then comparative learning person is learnt in a learning cycle by the behavior learn the phase of number
To size, when study number is more than that upper limit study number then thinks that learner in the probability of the learning behavior is often 1, if
Probability value of the ratio of number and upper limit study number as learning behavior often will then be learnt less than upper limit value；
(3) finally assume to only have learning behavior often and few two values, subtract the probability of learning behavior often using 1
Obtain the few probability value of study number.
The conditional probability table initialization and correction module are to obtain frequency with expertise numerical value through a large number of experiments
Value is as numerical value in conditional probability table:
(1) the initial learning style of each experimenter is obtained by simple problem or essential information initial method first；
(2) then learnt for learner's unrestricted choice by being supplied to the various modes of learning of experimenter and record every learner
Selected mode of learning and study number；
(3) it finally integrates emotional factor and show that the learning style frequency values under each learning behavior and emotional factor replace generally
Rate is worth the numerical value of final conditional probability table.
Correction module is the utilization again of learning style initial method, wherein except that by the data of experimenter
Be converted to the data of learner.
The Bayesian network correction module is the weight using learning behavior probability and emotional factor, in combination with condition
The method that the numerical value of probability tables calculate study demeanour dimension probability value will be in each dimension after calculating probability value
The corresponding learning style value of the value of maximum probability initializes value as the learning style dimension of dimension.
Compared with prior art, the invention has the following beneficial effects:
(1) present invention is selected by the method for providing two kinds of initialization learning styles for learners' corpora, ensure that learner selects
The independence and science selected.The simple problem initialization module of design simplifies the burden of learner simultaneously, is conducive to improve
Efficiency.Essential information initialization module does not need learner and fills in questionnaires so that the mode for initializing learning style is easier；
(2) Bayesian network correction module integrates learning behavior and emotional factor, improves the modified standard of learning style
Exactness.The method for proposing new calculating study behavior probability simultaneously simplifies calculating process, so that being more easier to calculate.Finally I
The method that proposes initialization condition probability tables and how conditional probability table is modified according to different study crowds, mention
The high accuracy of Bayesian network amendment learning style；
(3) it can obtain learner by whole learning style initialization and later period amendment, the present invention and accurately learn wind
Lattice, to play the effect on basis in individualized learning and recommendation learning path.
Detailed description of the invention
Fig. 1 is the Bayesian Network Learning style initialization for incorporating emotional factor and the overall flow figure of correction algorithm, packet
It includes learning style initialization and learning style corrects two parts.
Fig. 2 is the broad flow diagram that learner initializes learning style, the method for initializing learning style including two kinds.
Fig. 3 is using the flow chart of Bayesian network amendment learning style, including learning behavior probability calculation, conditional probability
The initialization of table numerical value and amendment, the involvement of emotional factor, the probability calculation of learning style dimension and the entirety for obtaining learning style
Process.
Fig. 4 is the corresponding relationship of each learning style dimension and its learning behavior and emotional factor.
Specific embodiment
Below in conjunction with specific attached drawing, the present invention is described in detail.
It is a kind of incorporate emotional factor Bayesian network amendment style initialization and correction algorithm, as shown in Figure 1, include letter
Single problem initialization module, essential information initialization module, study emotional factor module, learning behavior probabilistic module, condition are general
The initialization of rate table and correction module and Bayesian network correction module.
The simple problem initialization module, for initializing the learning style of learner, as shown in Fig. 2, being directed to information
Process dimension, propose problem one: when learner selects A to illustrate the learner for active type, on the contrary is reflective style；It is tieed up for perception
Degree proposes problem two: when learner selects A to illustrate the learner for modality of sensation, on the contrary is Intuition；For input dimension, mention
Go wrong three: when learner select A illustrate the learner for optic type, otherwise be verbal type；For dimension is understood, problem is proposed
Four: when learner selects A to illustrate the learner for comprehensive type, on the contrary is sequence type.
Specific the problem of proposing, is as follows:
Problem one: you like team unity and obtain knowledge and still like quiet thinking to go to obtain knowledge.
A. team unity B. peace and quiet are thought deeply
Problem two: you like the content that particular content still likes abstract.
A. particular content B. abstract content
Problem three: you like the thing for remembeing to see and are still good to obtain knowledge from text and oral expression.
A. it is good at and remembers that thing B. text and oral expression obtain knowledge
Problem four: you start to learn the discrete knowledge that still likes learning after liking the entire block diagram of assurance knowledge.
A. entire block diagram B. is held first learns discrete knowledge
Citing: as learner's problem one selects A, problem two to select, B, problem three select B and problem four is selected as A, then
The initialization learning style of habit person is initially active type, Intuition, verbal type and comprehensive.
The essential information initialization module speculates learner's maximum possible by the gender of learner, age and educational background
Learning style as initial learning style, as shown in Figure 2.
The present invention passes through many experiments first and expertise obtains the learning style ratio of different sexes, all ages and classes
The learning style ratio of learning style ratio and different academic backgrounds, see the table below 1,2,3 respectively:
Table 1: the learning style ratio of different sexes
Table 2: the learning style ratio of all ages and classes
Table 3: the learning style ratio of different academic backgrounds
Then the present invention obtains gender, age and the educational background of learner, and corresponds to different sexes, all ages and classes and not classmate
The learning style ratio table gone through obtains ratio of the learner in three tables and is denoted as respectively、、.By three
Probability value takes probability of its average value as the learner under current learning style dimension.Specific calculate sees below formula
(1):
(1)
The sum of probability of two values is 1 in each dimension herein, when the probability of a certain value of certain dimension is more than 0.5 then
Think that the learner is the value.
If last a certain position learner gender is female, age between 19-27 years old and academic level is undergraduate course, according to
The probability of formula (1) calculating active type=(66.2%+64.0%+61.5%)/3=63.9%, therefore the probability of learner's active type
It is 63.9%, the learning style that initialization information processes dimension is active type；Equally calculate the probability that the learner is perception type=(73.1%+69.7%+77.1%)/3=73.3%, therefore the probability that the learner is perception type is 73.3%, initialization perception dimension
The learning style of degree is perception type；For inputting dimension, the probability of verbal type is calculated=(29.9%+27.5% +29.4%)/3=
28.9%, which is that the probability of verbal type is 28.9%, therefore initializing the learner to input the learning style of dimension is vision
Type；For understanding dimension, comprehensive probability is calculated=(49.3%+46.6%+53.2%)/3=49.7%, therefore the learner is comprehensive
The probability of type is 49.7%, initializes the learner and understands that the learning style of dimension is sequence type.
The study emotional factor module, the present invention are designed for learners' corpora selection type of emotion pair on study interface
The block diagram of weight is answered, as shown in Figure 3.By allowing learner voluntarily to select the current corresponding weight of each type of emotion.Tool
The emotional factor type of body and corresponding weight are shown in Table 4.Wherein have glad, surprised and detests three kinds of type of emotion, it is corresponding
The weight of type is respectively、With, and three kinds of weights sums are 1.
Table 4: emotional factor
Select the weight of the type of emotion of each learning style dimension by learner, what system obtained the learner works as cause
Thread type and corresponding weight.
The learning behavior probabilistic module, wherein learning behavior and learning style dimension corresponding relationship are as shown in figure 4, specific
Calculating method it is as follows:
For some learning behavior, the first number of regulation behavior maximum possible appearance in a learning cycle, it is denoted as；Then the number that behavior appearance is extracted from the learning behavior log of learner, is denoted as；Finally compareWithRelative size: if the former be more than or equal to the latter, then the probability more than behavior frequency of occurrence be 1；If
The former is less than the latter, then the probability more than behavior frequency of occurrence isWithRatio, as formula (2):
(2)
Here we provide that each learning behavior only has number few and often two kinds of values, thus the learning behavior is having been calculated
After probability more than frequency of occurrence, it can use the sum of two probability equal to 1 and the few probability of frequency of occurrence, as formula be calculated
(3):
(3)
The conditional probability table initialization and correction module, as shown in figure 3, way is described below:
(1) a large amount of experimenter is selected, obtains each using simple problem initialization module or essential information initialization module
The initial learning style of experimenter；
(2) education resource including forum, example, video, text, outline and summary etc. is provided, allows each experimenter
Unrestricted choice study is carried out, the selected education resource mode of all experimenters and study number are recorded；
(3) for Information procession, perception, input in learning style and four dimensions are understood, respectively using learning behavior probability
Computation rule calculates the frequency values that the corresponding learning behavior frequency of occurrences of each learning style dimension is big and frequency is small；
(4) learning behavior probability and emotional factor are integrated, the frequency for corresponding to value in each dimension is calculated
Value；
(5) method that frequency of use replaces probability is closed according to each learning style dimension is corresponding with its corresponding learning behavior
System is stored in conditional probability table using the calculated frequency values of previous step as probability value, completes the initialization of conditional probability table.
The conditional probability table initial value of Information procession dimension determines that we are obtained by experiment and expertise:
Access forum is few, exemplary browser number is few and mood is that active type frequency is 0.68 in glad number, corresponding heavy
Frequency shared by think of type is 0.32；
Access forum is few, exemplary browser number is few and mood is that active type frequency is 0.56 in surprised number, corresponding heavy
Frequency shared by think of type is 0.44；
Access forum is few, exemplary browser number is few and mood is that active type frequency is 0.12 in detestation's number, corresponding heavy
Frequency shared by think of type is 0.88；
Forum is few for access, exemplary browser often and mood is that active type frequency is 0.08 in glad number, it is corresponding heavy
Frequency shared by think of type is 0.92；
Access forum is few, exemplary browser often and mood is that active type frequency is 0.17 in surprised number, it is corresponding heavy
Frequency shared by think of type is 0.83；
Access forum is few, exemplary browser often and mood is that active type frequency is 0.45 in detestation's number, it is corresponding heavy
Frequency shared by think of type is 0.55；
Access forum is more, exemplary browser number is few and mood is that active type frequency is 0.96 in glad number, corresponding heavy
Frequency shared by think of type is 0.04；
Access forum is more, exemplary browser number is few and mood is that active type frequency is 0.73 in surprised number, corresponding heavy
Frequency shared by think of type is 0.87；
Access forum is more, exemplary browser number is few and mood is that active type frequency is 0.26 in detestation's number, corresponding heavy
Frequency shared by think of type is 0.74；
Forum is more for access, exemplary browser often and mood is that active type frequency is 0.69 in glad number, accordingly sink
Frequency shared by think of type is 0.31；
Access forum is more, exemplary browser often and mood is that active type frequency is 0.61 in surprised number, accordingly sink
Frequency shared by think of type is 0.39；
Access forum is more, exemplary browser often and mood is that active type frequency is 0.23 in detestation's number, accordingly sink
Frequency shared by think of type is 0.77.
Replacing the mode of probability that said frequencies are converged into table with frequency is shown in table 5:
Table 5: the conditional probability table of Information procession dimension
The amendment of the conditional probability table numerical value, particular content are described below:
(1) the record limit value of study behavior database is definedIf not reaching the numerical value, continue waiting for；Work as data
When reaching the value, start the numerical value of correction conditions probability tables；
(2) learning style of each learner is extracted in learning style database according to student number；
(3) learning behavior of corresponding learner is extracted in learning behavior database according to the student number of learner;
(4) by the calculation method of learning behavior probability calculate each learning behavior frequency of occurrence mostly with few general of number
Rate；
(5) each learning style dimension and its learning behavior are mapped, the method weight of use condition probability tables initialization
Newly calculate the respective frequencies of each learning behavior He each learning style dimension；
(6) previous step calculated result and emotional factor are integrated, obtains frequency values, using the frequency as probability numbers
The corresponding position of filling conditional probability table obtains revised conditional probability table.
For the dynamic corrections of Information procession dimension condition probability tables, system extracts forum's browsing in learning behavior database
Number and exemplary browser number calculate learning behavior probability using formula (2) and (3), integrate emotional factor thereafter, calculate shaping
Active type probability value in part probability tables subtracts each using number 1 simultaneously because the sum of probability of each dimension is 1
The probability of reflective style can be obtained in probability, and result see the table below 6 after arrangement:
Table 6: the conditional probability table of Information procession dimension after amendment
The Bayesian network correction module, the present invention are general using learning behavior probability numbers, the weight of emotional factor and condition
Rate table value revision learning style, especially by such as following formula (4):
(4)
In formula(=1,2,3,4) the is indicatedThe learning style of dimension, there are two value, such as understand dimension have sequence type and
Comprehensive two values.Furthermore(=1,2,3,4) a series of nodes are indicated,Indicate a series of nodes
Joint probability distribution.
WhereinMultiplying for the learning behavior probability of conditional probability table numerical value and corresponding node can be used
For product to calculate, specific formula is following (5):
(5)
In formulaIt indicatesCorresponding parent nodeCorresponding conditional probability, value correspond toConditional probability table numerical value.
The actual conditions of Bayesian network amendment learning style are combined, the corresponding relationship in formula (6) is provided.
(6)
The formula concrete meaning:
Since there are two values for a learning style dimension, therefore always there is probability P >=0.5 in two values, therefore is not in
The case where probability P < 0.5；
As 0.5≤P of probability < 0.6, corresponding weight is 1；
As 0.6≤P of probability < 0.7, corresponding weight is 3；
As 0.7≤P of probability < 0.8, corresponding weight is 5；
As 0.8≤P of probability < 0.9, corresponding weight is 7；
As 0.9≤P of probability≤1.0, corresponding weight is 9.
Current dimension pair is finally obtained according to the corresponding relationship between the corresponding weight of formula (6) and the weight of learning style
The learning style type answered and corresponding weight.
Learning style amendment for Information procession dimension, it is often and few general that we calculate learner's browsing forum first
Rate, navigation example often with few probability.It is defined in a study week, the upper limit number for browsing forum is 20, browsing is real
Example upper limit number is 20.A certain learner's browsing forum's number is 16, navigation example number is 10, and formula (2) and (3) is used to calculate
Obtain the probability of browsing forum often=0.8, the few probability of browsing forum's number=0.2, navigation example is often
Probability=0.50, the few probability of navigation example number=0.50.Thereafter the power that the learner selects happy emoticon is obtained
ValueFor the 0.7, weight of surprised moodFor 0.2, the corresponding weight of detestIt is 0.1.The item of combining information processing dimension
Part probability tables (table 6) are simultaneously calculated using formula (5) and (6):
P (active type)=P (active type | forum is few, and example is few, glad) × P (forum is few) × P (example is few) × P (happiness)
+ P (active type | forum is few, and example is few, surprised) × P (forum is few) × P (example is few) × P (surprised)
+ P (active type | forum is few, and example is few, detests) × P (forum is few) × P (example is few) × P (detest)
+ P (active type | forum is few, and example is more, glad) × P (forum is few) × P (example is more) × P (happiness)
+ P (active type | forum is few, and example is more, surprised) × P (forum is few) × P (example is more) × P (surprised)
+ P (active type | forum is few, and example is more, detests) × P (forum is few) × P (example is more) × P (detest)
+ P (active type | forum is more, and example is few, glad) × P (forum is more) × P (example is few) × P (happiness)
+ P (active type | forum is more, and example is few, surprised) × P (forum is more) × P (example is few) × P (surprised)
+ P (active type | forum is more, and example is few, detests) × P (forum is more) × P (example is few) × P (detest)
+ P (active type | forum is more, and example is more, glad) × P (forum is more) × P (example is more) × P (happiness)
+ P (active type | forum is more, and example is more, surprised) × P (forum is more) × P (example is more) × P (surprised)
+ P (active type | forum is more, and example is more, detests) × P (forum is more) × P (example is more) × P (detest)
=0.6623。
Thereafter by the learner known to formula (6) be active type in the learning style of Information procession dimension and weight is 3,
The Bayesian network for completing to incorporate emotional factor corrects learning style.
The above is only presently preferred embodiments of the present invention and oneself, any type of limitation not is done to the present invention.It is all
Technology and methods essence according to the present invention is to any nonessential modification, equivalent variations made by embodiment of above and repairs
Decorations, in the range of still falling within technology and methods scheme of the invention.
Claims (7)
1. a kind of Bayesian Network Learning style initialization for incorporating emotional factor and correction algorithm, it is characterised in that: including letter
Single problem initialization module, essential information initialization module, study emotional factor module, learning behavior probabilistic module, condition are general
The initialization of rate table and correction module and Bayesian network correction module；
The simple problem initialization module carries out initialization learning style by allowing learner to answer four simple problems；
The essential information initialization module is to carry out initialization learning style according to the gender of learner, age and educational background；
The study emotional factor module is to determine current study feelings by allowing learner to select the weight of study emotional factor
Thread；
The learning behavior probabilistic module is to carry out calculating study behavior probability by certain rule；
Conditional probability table initialization and correction module are to be initialized by the data of a large amount of learners and correction conditions
Probability tables；
The Bayesian network correction module is carried out using study emotional factor, conditional probability table numerical value and learning behavior probability
Correct the process of learning style.
2. the Bayesian Network Learning style initialization according to claim 1 for incorporating emotional factor and correction algorithm,
Be characterized in that: the simple problem initialization module is for Information procession, perception, input in learning style and understands four dimensions
Different characteristics propose a problem respectively, while designing each problem only there are two option, carry out initialization learning style.
3. the Bayesian Network Learning style initialization according to claim 1 for incorporating emotional factor and correction algorithm,
Be characterized in that: the essential information initialization module passes through the various essential informations of analytic learning person, and by current learner's
Essential information and the generally statistics of a large amount of learners are corresponding, show that the most probable learning style of the learner is learned as initial
Practise style.
4. the Bayesian Network Learning style initialization according to claim 1 for incorporating emotional factor and correction algorithm,
Be characterized in that: the study emotional factor module will learn mood and be divided into glad, surprised and detest three types, by limiting three
Study emotional factor is dissolved into Bayesian network correction algorithm by the weight of kind type of emotion.
5. the Bayesian Network Learning style initialization according to claim 1 for incorporating emotional factor and correction algorithm,
Be characterized in that: the learning behavior probabilistic module is to define each learning behavior highest upper limit time in a learning cycle
The relative size of number, the number and upper limit number that are learnt in a learning cycle by the behavior by comparing learner obtains
Learning behavior probability.
6. the Bayesian Network Learning style initialization according to claim 1 for incorporating emotional factor and correction algorithm,
Be characterized in that: the conditional probability table initializes and correction module is to obtain through a large number of experiments and refering to expertise numerical value
Frequency values are as probability numbers.
7. the Bayesian Network Learning style initialization according to claim 1 for incorporating emotional factor and correction algorithm,
Be characterized in that: the Bayesian network correction module is the weight and condition using learning behavior probability, study emotional factor
The numerical value of probability tables calculate the weight of the probability numbers acquisition learning style of different learning style dimensions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810345695.5A CN110399982A (en) | 2018-04-18 | 2018-04-18 | Incorporate the initialization of Bayesian Network Learning style and the correction algorithm of emotional factor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810345695.5A CN110399982A (en) | 2018-04-18 | 2018-04-18 | Incorporate the initialization of Bayesian Network Learning style and the correction algorithm of emotional factor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110399982A true CN110399982A (en) | 2019-11-01 |
Family
ID=68319491
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810345695.5A Pending CN110399982A (en) | 2018-04-18 | 2018-04-18 | Incorporate the initialization of Bayesian Network Learning style and the correction algorithm of emotional factor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110399982A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5987415A (en) * | 1998-03-23 | 1999-11-16 | Microsoft Corporation | Modeling a user's emotion and personality in a computer user interface |
| US20120052476A1 (en) * | 2010-08-27 | 2012-03-01 | Arthur Carl Graesser | Affect-sensitive intelligent tutoring system |
| CN104299036A (en) * | 2013-07-19 | 2015-01-21 | 上海宽带技术及应用工程研究中心 | Q-learning based double-layer Bayes network interference algorithm |
| RU2598050C1 (en) * | 2015-07-29 | 2016-09-20 | Военный Институт (Физической Культуры) Военно-Медицинской Академии Им. С.М. Кирова | Method of video-computer-aided diagnosis of the emotional state of trainees and generating decisions on rational division of educational groups |
| CN106991627A (en) * | 2017-03-28 | 2017-07-28 | 广西师范大学 | The distributed intelligence tutoring system acted on behalf of based on domain body and more |
| CN107291654A (en) * | 2016-03-31 | 2017-10-24 | 深圳光启合众科技有限公司 | The intelligent decision system and method for robot |
| CN107423851A (en) * | 2017-07-20 | 2017-12-01 | 上海理工大学 | Adaptive learning method based on learning style context aware |
| US20190110726A1 (en) * | 2016-06-24 | 2019-04-18 | Tata Consultancy Services Limited | Method and system for detection and analysis of cognitive flow |
-
2018
- 2018-04-18 CN CN201810345695.5A patent/CN110399982A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5987415A (en) * | 1998-03-23 | 1999-11-16 | Microsoft Corporation | Modeling a user's emotion and personality in a computer user interface |
| US20120052476A1 (en) * | 2010-08-27 | 2012-03-01 | Arthur Carl Graesser | Affect-sensitive intelligent tutoring system |
| CN104299036A (en) * | 2013-07-19 | 2015-01-21 | 上海宽带技术及应用工程研究中心 | Q-learning based double-layer Bayes network interference algorithm |
| RU2598050C1 (en) * | 2015-07-29 | 2016-09-20 | Военный Институт (Физической Культуры) Военно-Медицинской Академии Им. С.М. Кирова | Method of video-computer-aided diagnosis of the emotional state of trainees and generating decisions on rational division of educational groups |
| CN107291654A (en) * | 2016-03-31 | 2017-10-24 | 深圳光启合众科技有限公司 | The intelligent decision system and method for robot |
| US20190110726A1 (en) * | 2016-06-24 | 2019-04-18 | Tata Consultancy Services Limited | Method and system for detection and analysis of cognitive flow |
| CN106991627A (en) * | 2017-03-28 | 2017-07-28 | 广西师范大学 | The distributed intelligence tutoring system acted on behalf of based on domain body and more |
| CN107423851A (en) * | 2017-07-20 | 2017-12-01 | 上海理工大学 | Adaptive learning method based on learning style context aware |
Non-Patent Citations (4)
| Title |
|---|
| 全丽莉等: "情感计算在E-Learning***中的应用探索", 《华中师范大学研究生学报》 * |
| 杨一柳等: "基于情绪获取的暗示性"智能型课件"的设计方法研究", 《渤海大学学报(自然科学版)》 * |
| 苏钦等: "智慧学习环境中基于贝叶斯网络构建学生模型的研究", 《山西师范大学学报(自然科学版)》 * |
| 钟国祥等: "采用贝叶斯网络建立一种智能学习环境学生代理模型的研究", 《计算机科学》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107463609B (en) | Method for solving video question-answering by using layered space-time attention codec network mechanism | |
| Wu et al. | Exercise recommendation based on knowledge concept prediction | |
| CN107766320A (en) | A kind of Chinese pronoun resolution method for establishing model and device | |
| CN106409290A (en) | Infant intelligent voice education method based on image analysis | |
| CN109829049A (en) | The method for solving video question-answering task using the progressive space-time attention network of knowledge base | |
| CN115510814B (en) | Chapter-level complex problem generation method based on dual planning | |
| KR20210050377A (en) | Personalized learning system | |
| Figoli et al. | Artificial intelligence in the design process: The Impact on Creativity and Team Collaboration | |
| CN108388944B (en) | A kind of automatic chatting method and robot based on deep neural network | |
| Lin et al. | Intelligent decision support for new product development: a consumer-oriented approach | |
| Cabitza et al. | The knowledge-stream model-a comprehensive model for knowledge circulation in communities of knowledgeable practitioners | |
| Maroto-Gómez et al. | Active learning based on computer vision and human–robot interaction for the user profiling and behavior personalization of an autonomous social robot | |
| CN114154839A (en) | Course recommendation method based on online education platform data | |
| CN111027433A (en) | Multiple style face characteristic point detection method based on convolutional neural network | |
| Chaganti et al. | UI/UX Design for Online Learning Approach by Predictive Student Experience | |
| CN111191045B (en) | Entity alignment method and system applied to knowledge graph | |
| CN110399982A (en) | Incorporate the initialization of Bayesian Network Learning style and the correction algorithm of emotional factor | |
| Zakrzewska | Using clustering technique for students’ grouping in intelligent e-learning systems | |
| US20200380264A1 (en) | Systems and methods for video analysis tool for communications coursework scheduling and management | |
| Wang | The Impact of Animation and Film English Education Environment on Students' Psychological Health | |
| Hao et al. | Intelligent analysis on the rationalization of children’s physical education curriculum based on recurrent neural networks | |
| CN111400602A (en) | Online learning system and method based on personalized recommendation | |
| JP6910061B2 (en) | Text generator, text generator and text generator | |
| Makkreel | Metaphysics and the Hermeneutical Relevance of Worldviews | |
| Pei et al. | Visual relational reasoning for image caption |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191101 |