CN107609651A - A kind of design item appraisal procedure based on learner model - Google Patents
A kind of design item appraisal procedure based on learner model Download PDFInfo
- Publication number
- CN107609651A CN107609651A CN201710698682.1A CN201710698682A CN107609651A CN 107609651 A CN107609651 A CN 107609651A CN 201710698682 A CN201710698682 A CN 201710698682A CN 107609651 A CN107609651 A CN 107609651A
- Authority
- CN
- China
- Prior art keywords
- index
- learner
- design item
- item
- level
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to study analysis, behavioural information to perceive field, a kind of design item appraisal procedure based on learner model, including (1) learner model structure are provided, based on learner's example, its attribute and related service are analyzed and concluded, establishes property set and the related service of learner.(2) design item is formulated, determine the design item related to learner model according to demand first, then attribute extraction and classification are carried out according to the interactive object of learner and its service, form the index for assessing design item, finally according to the data attribute extracted, recursive hierarchy structure is established, forms the overall framework of the design item.(3) appraisal procedure is formulated, the design item appraisal procedure based on learner model is designed using analytic hierarchy process (AHP).The invention provides the method that big data basic model is established for learner, and a set of feasible evaluation scheme is formulated, good application foundation is provided for the analysis and excavation of learner's related service under big data environment.
Description
Technical field
The invention belongs to study analysis, behavioural information to perceive field, and in particular to a kind of design based on learner model
Item appraisal procedure.
Background technology
The expression of learner model typically uses the building method based on vector space model, and it can reflect different concepts
Significance level in learner model, and the vector operation formula of convenient use standard carries out the project of follow-up phase
With task.But the attribute of learner is more complicated, only can not fully be caught with one group of keyword；Meanwhile word lists reach itself
Intrinsic synonymy and semantic difference, and do not account for word order or linguistic context problem during expression so that based on this model
Result has ambiguousness caused by expression.
For design item appraisal procedure typically using multi objective carry out overall merit, can according to different evaluation purposes,
The corresponding evaluation form of selection, such as Information Entropy, VC Method etc., usually determine to weigh according to the correlation between index
Overall merit is carried out again.But the scene in reality calculates often by factors composition that is numerous interrelated or mutually restricting
When may lack more quantitative data, make assessment result be difficult to quantify.
At present, study analysis field is generally divided into two classes on the research of learner model, and one kind is by describing to learn
The specific object and feature of person, and sort out modeling on this basis, as Wei Shunping exists《Study analysis data model and data processing
Technique study》It is middle that the key element of learning behavior is divided into the class of learner, content, place, time, result etc. five；It is another kind of, it is pair
Learner's related data in a certain academic environment is analyzed and excavated, and is built so as to which the feature for realizing to learner carries out cluster
Mould, if Greece's open university is using content of the discussions in online forum as research object, utilize text mining and social network analysis skill
The participation model of art enquiry learning person, and learner is classified according to specific features.However, these methods are to learning model
Use be confined to describe learning characteristic inside learner and outside with model, according to model by with similar learning characteristic
Learner is classified, and learning Content, strategy and the education resource of personalization are provided for student, learner can be allowed to understand in addition
The learning state and deficiency of oneself, and then the learning behavior of oneself is corrected in advance.But how the studies above is without reference in structure
The method of the behavior and service of going research to assess learner on the basis of learner model is built, and is related to grinding for learner's assessment
Study carefully often use qualitatively to analyze, quantitative analysis is rare to be related to more.
The important learner model of four relatively conventional classes is respectively under the study analysis ken：Knowledge model, cognitive model,
Emotion model and Learner behavior model, their numbers from different directions such as knowledge, cognition, emotion, behaviors to learner respectively
According to being modeled.In terms of appraisal procedure, Fan Jie《On-line study behavior evaluation system design and reality based on data mining
It is existing》In, it is proposed that online " learning behavior-effect " model is built based on data mining technology student's on-line study is assessed
Scheme；Jiang Hua, Zhao Jie《Learning behavior evaluation model and realization based on BP neural network》In, use BP neural network
Learning behavior evaluation model is evaluated learning behavior.
On the one hand, assessed for the learner model under study analysis visual angle and its results of learning,《Based on data mining
On-line study behavior evaluation systematical design idea》With《Learning behavior evaluation model and realization based on BP neural network》In
Although relevant programme has used data mining technology or BP neural network to construct dependent evaluation model, but application field is limited
(carrying out results of learning assessments only for specific learning process data), many design items in learning process (such as are learnt
Participate in, study is satisfied, learning interest etc.) assessment it is rare be related to, do not form a kind of more comprehensive evaluation scheme；Separately
On the one hand, the studies above also fails to learner model being combined with the design such as learning process item key element.
Based on learner model (such as：Learner, visitor, consumer) assessed, predicted and recommended, it has also become
During the every profession and trade service implementation of big data epoch, the preparation work of mining analysis is carried out, therefore do not determining specific data attribute
Before scale, learner is modeled and seems and is even more important, also to study relevant design on the basis of learner model
Item appraisal procedure provides good basis.
The content of the invention
It is an object of the invention to the deficiency for more than, there is provided a kind of design item appraisal procedure based on learner model,
The method that big data basic model is established for learner is this method provide, and has been formulated for its relevant design item a set of feasible
Evaluation scheme, provide good application foundation for the analysis and excavation of learner's related service under big data environment.
The present invention relates to term to be defined as follows.
Learner model：Learner model including narrow sense, and visitor's model of broad sense.
Design item：Effect caused by because of a certain activity of participation is experienced and (such as results of learning, visits satisfaction), for answering
Portrayed with the attribute of the related different dimensions in field.
Design item overall framework：For a certain design item, the hierarchical classification structure of its property set included.
Appraisal procedure：Appraisal procedure for designing item.
Interpretational criteria：That is the rule of judgment matrix importance scale, importance degree difference index is made to be obtained by calculating
Corresponding weight.
Judgment matrix：The relative importance of the design item similarly hereinafter other index of one-level is compared respectively according to interpretational criteria
Compared with the matrix after quantization represents.
The weight vector of matrix：The vector that weight corresponding to all indexs of i.e. same rank is formed.
Analytic hierarchy process (AHP)：Refer to using a complicated decision-making problem of multi-objective as a system, be more by goal decomposition
Individual target or criterion, and then some levels of multi objective (or criterion, constraint) are decomposed into, pass through qualitative index Fuzzy Quantifying
Mode of Level Simple Sequence (weight) and total sequence are calculated, to be used as target (multi objective), the systems approach of multi-scheme Optimal Decision-making.
The present invention provides a kind of design item appraisal procedure based on learner model, comprises the following steps：
(1) learner model is built, and based on learner's example, is analyzed and is concluded its attribute and related service, establish study
The property set of person and related service.
(2) design item is formulated, determines the design item related to learner model (such as learner according to demand first
Practise participation, the visit satisfaction of visitor).Then according to the interactive object (such as course, showpiece) of learner and its service (such as
Learning behavior prediction, visit behavioural analysis etc.) attribute extraction and classification are carried out, form the index for assessing design item.Formulate
Design item need to extract association attributes (such as according to the interactive object of learner：When being related to on-line study in learner's participation
Between, classroom effectively pay attention to the class the time, be related in visitor's satisfaction recommend showpiece frequency), be tectonic sieving item overall framework
Data basis is provided.Finally according to the data attribute extracted, recursive hierarchy structure is established, forms the overall frame of the design item
Frame.
(3) appraisal procedure is formulated, the design item appraisal procedure based on learner model is designed using analytic hierarchy process (AHP).
In the above-mentioned technical solutions, in step (1) on the basis of demand acquisition, UML, concept database mould can be used
The formalization such as type mode establishes learner model.Based on learner, its information is mainly with two kinds of static structure and dynamic structure
Form is present.Static structure is mainly to be established according to demographic information, this partial content essential record learner's
Personal essential information (such as user name, student number, name, sex, specialty, age, nationality, contact method, hobby), these
Information will not typically change in the whole service activity of learner.Service rows of the dynamic structure primarily directed to learner
For for, this category information be according to learner service process and dynamic establish, and can with service process deeply and
Constantly occur to change, the information of these changes is related to every aspect of the learner during service activity, this category information master
To include style and features information (the personal preference of such as learner), (different of such as learner of service process information of learner
Practise path and produce different results of learning, or the different track of visiting of visitor produces different visit effects), performance information
(such as learner classroom answer a question situation, operation performance, or visitor's participation situation etc.) and status information are (as learned
Habit person is in enquirement state, or visitor is in and showpiece interaction mode) etc..
In the above-mentioned technical solutions, the specific construction step of design item overall framework is as follows in step (2)：
(2-1) is classified to the attribute of design item extraction by it to the correlation degree for designing item.Will be direct with design item
The attribute of association is arranged to first class index Xi, XiI-th of first class index of design item is represented, such as：The study participation of learner
First class index includes：X1=classroom participates in, X2=participate in online；Or the first class index of the level of interest of visitor includes：X1=
Objective data, X2=interactive degree, X3=experience sense, X4=intellectual, X5=innovative, X6=excitant.
(2-2) is directed to first class index XiActual content, analyze it and determined by which sub- factor, select this little factor to make
For two-level index Xij, XijJ-th of two-level index under i-th of first class index of design item is represented, such as：Learner learns to participate in
Spend first first class index (X of this design item1=classroom participate in) under two-level index include：X11=work attendance, X12=pay attention to the class
Time, X13=handling situations, X14=answer a question, X15=exchange is discussed；Or the of visitor's level of interest this design item
Three first class index (X3=experience sense) under two-level index include：X31The accuracy and flexibility that=interaction is supported, X32=ring
Between seasonable, X33=left-hand seat time, X34=environmental factor.
(2-3) is examined in whether all two-level index can segment, if can continue to segment, three are filled by upper step
Level index.This process is repeated, untill all indexs can not subdivide, that is, forms last design item overall framework.
In the above-mentioned technical solutions, comprising the following steps that for appraisal procedure is formulated in step (3)：
(3-1) Judgement Matricies
After to being contrasted two-by-two between each index at the same level, dividing position by n, (statistical indicator number n general recommendations is not
More than 9) ratio is ranked the relative superior or inferior order of each evaluation index, the judgment matrix D of evaluation index is constructed successively.Construction is sentenced
Interpretational criteria used in disconnected matrix, is specifically shown in Table 1.Such as：Index 1 is just 1 compared to no less important to index 2, and somewhat important is 3
Deng.
The importance scale implication table of table 1
The weight vector of (3-2) matrix calculates
Using feature vector method, the eigenvalue of maximum of the judgment matrix is calculated using MATLAB softwares, draws corresponding spy
Sign vector.
(3-3) Consistency Check in Judgement Matrix
In order that obtained judgment matrix will not produce logic error, and meet the uniformity and transitivity between element,
Reliable input is provided by the calculating of follow-up weight, it is necessary to carry out consistency check to the judgment matrix constructed.
Consistency check rule is as follows：Consistency rationWork as CR<When 0.1, it is believed that the uniformity of judgment matrix is
Acceptable, otherwise need suitably to correct judgment matrix.Wherein, CI represents coincident indicator, and RI represents to search coincident indicator.
Coincident indicator CI calculation formula isWherein n is the index number compared, λmaxIt is special for the maximum of judgment matrix
Levy root；Coincident indicator RI is searched, need to be provided with reference to practical experience, the Aver-age Random Consistency Index value RI used here, is had
Body is as shown in table 2.The present invention using the method for MATLAB programming realizations, examines whether the judgment matrix meets uniformity simultaneously.
The Aver-age Random Consistency Index RI of table 2. value
| n | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
| RI | 0 | 0 | 0.58 | 0.90 | 1.12 | 1.24 | 1.32 | 1.41 | 1.45 |
(3-4) calculates weight
After judgment matrix is by consistency check, it will determine that the corresponding characteristic vector of matrix eigenvalue of maximum institute carries out normalizing
Change is handled, and obtains the weight of each index.Weight corresponding to first class index is Wi, WiRepresent the power of i-th of first class index of design item
Weight, such as：W in this design phase of the study participation of learner1, represent index X1Weight, i.e., classroom participate in weight；Or
W in this design item of visitor's level of interest1, the weight of objective data is represented, weight corresponding to two-level index is Wij, WijRepresent
Design weight corresponding to j-th of two-level index under i-th of first class index of item, the method for expressing of three-level index respective weights with
This analogizes.
The scoring of (3-5) each index.The threshold value of indexs at different levels is set according to the available accuracy demand of design item and obtains minute mark
Standard, and the property value to designing item indexs at different levels unifies dimension.Scoring is Y corresponding to first class indexi, YiRepresent the i-th of design item
Scoring corresponding to individual first class index, scoring corresponding to two-level index is Yij,YijRepresent i-th of first class index of design item
Under j-th of two-level index corresponding to score, the method for expressing of the corresponding scoring of three-level index is by that analogy.
(3-6) calculates design item assessment result.The category after weight and unification dimension by designing item evaluation indexes at different levels
Property value, is calculated the final assessment formula of the design itemThis formula eliminates the design three-level of item and following
Index.
This assesses summary and refinement of the formula as appraisal procedure, makes whole appraisal procedure more succinct understandable, and be easy to
Calculate.Meanwhile also clearly describe the hierarchical structure of design item with the mode of symbol so that from the formulation of design item to assessment
This process of the formulation of method becomes more directly perceived and smooth.
The inventive method has advantages below compared with prior art：
1. the assessment result of the present invention can quantify.It is of the invention from the angle of design item relative to existing model evaluation method
Degree is set out, and learner model is carried out comprehensively, hierarchically to analyze, and derives that the assessment of design item is public using analytic hierarchy process (AHP)
Formula, to quantify the dependent evaluation based on learner.
2. model describes unambiguity.The present invention is based on learner's example, from learner's attribute and the angle of related service
Set out, establish learner model.Modelling fine size, the rank of primary attribute is reached so that constructed learner's mould
Accurately and unambiguously, concept database model easy to use, uml model etc. are realized for type definition.
3. design item overall framework has higher flexibility.Modularization is used during formulating design item overall framework
The thought of design, other design items can be used as certain grade of index of the design item, without the overall frame to designing item
Frame makes the modification of essence.Correlation is smaller between designing the index of item overall framework different levels, designs item overall framework
The logicality and estimation flow for not interfering with design item overall framework are replaced in the additions and deletions of lower certain grade of index.
4. the present invention has higher applicability.The present invention is based on learner's example, from learner's attribute and related service
Angle set out, to establish learner model.It designs the quantization and analysis that property set is paid attention in the assessment of item, reduces to specialty
The degree of dependence of knowledge, the learner of broad sense is applicable not only to, is also applied for the learner (such as visitor, consumer) of narrow sense.
Brief description of the drawings
Fig. 1 is recruitment evaluation achievement and real result comparison diagram in embodiment one.
Fig. 2 is that the visit recruitment evaluation in embodiment two visits effect contrast figure with actual.
Embodiment
The inventive method is applied in study analysis field and behavioural information perception field based on visitor, it is caused
Embodiment is as follows.
The study analysis learning recruitment evaluation of embodiment one
The first step, learner model structure.Based on learner's example, its attribute and related service are analyzed and concluded.Pass through
Analysis, the information of learner's example mainly exist in the form of two kinds of static structure and dynamic structure.The information of static structure is main
For essential information (such as user name, student number, name, sex, specialty, age, nationality, contact method, hobby)；Dynamic is tied
The information of structure is mainly achievement corresponding to behavioural information in learning process (such as preference information, on-line study performance etc.) and learning outcome
Imitate information (such as performance is summarized, performance of the test).Learner's attribute is further filled according to the actual requirements.It is specific as shown in table 3.
The learner model of table 3.
Second step, formulate design item.Determine the design item related to learner model, the study analysis according to demand first
, it is necessary to assess results of learning in scene, therefore results of learning are selected as design item.Then according to the interaction pair of learner
As (course) and its learning process progress attribute extraction and classification, results of learning can be subdivided into learning ability, study here
Four satisfaction, study participation, learning interest first class index.Extract four one-levels and refer to the caused attribute in service process,
To meet evaluation requirement.Finally according to the data attribute extracted, recursive hierarchy structure is established, forms design item (the study effect
Fruit) overall framework.
The specific construction step for designing item (results of learning) overall framework is as follows：(2-1) carries to design item (results of learning)
The attribute taken is classified by it to the correlation degree for designing item.It will be set with designing the attribute of item (results of learning) direct correlation
For first class index Xi, i.e. learning ability X1, Learning satisfaction X2, study participation X3, learning interest X4.(2-2) refers to for one-level
Mark learning ability X1Actual content, analyze it and determined by three know-how, mental factor, nonintellectual factor sub- factors, select
This little factor is selected as two-level index, uses X respectively11、X12、X13Represent.(2-3) analysis knowledge horizontal X11, can continue to segment
For six three-level indexs, respectively very advanced X111, advanced X112, more advanced X113, intermediate X114, primary X115, beginner X116。
Other two-level index refine successively according to the method described above.Above-mentioned three-level index X111、X112、X113、X114、X115、X116Can not
Subdivide, so far two-level index know-how X11Structure finishes.Other two-level index are built in the same way.Final design
The overall frame of item (results of learning) finishes.The total framework of specific design item (eliminates three-level and following as space is limited,
Index) as shown in table 4.
The results of learning of table 4. design item overall framework
3rd step, the formulation of appraisal procedure.Design item (results of learning) is assessed using analytic hierarchy process (AHP), its flow
It is specific as follows：
(3-1) Judgement Matricies.
Comprise the following steps that：
(3-1-1) is directed to first class index Xi(learning ability X1, Learning satisfaction X2, study participation X3, learning interest X4),
Judge the relative importance between four indexs.Corresponding importance scale a is provided according to table 1ij,a11Refer to learning ability X1
With learning ability X1The importance scale compared, it is of equal importance to be entered as 1, it is designated as a11=1；a12Refer to learning ability X1With study
Satisfaction X2The importance scale compared, is designated as a12=2；a21Refer to Learning satisfaction X2With learning ability X1The importance compared
Scale, it is X1With X2The inverse compared, is designated as a21=1/2；It can similarly obtain, a13=4, a14=5, a21=1/2, a22=1, a23=
3,a24=4, a31=1/4, a32=1/3, a33=1, a34=3, a41=1/5, a42=1/4, a43=1/3, a44=1；Designed
The judgment matrix of all first class index of item is as follows：
(3-1-2) calculates judgment matrix D corresponding to all two-level index under each first class index successivelyi(Di, represent i-th
The judgment matrix of all two-level index under individual first class index), for first first class index learning ability X1, by its corresponding two
Level index know-how X11, mental factor X12, nonintellectual factor X13Judgement Matricies D1It is as follows：
Learning satisfaction X is calculated successively2, study participation X3, learning interest X4Corresponding Judgement Matricies D2, D3, D4。
(3-1-3) calculates judgment matrix D corresponding to three-level index successivelyij, DijRepresent under i-th of first class index j-th two
The judgment matrix of three-level index corresponding to level index.
The weight vector of (3-2) matrix calculates.
Using feature vector method, the eigenvalue of maximum of the judgment matrix is calculated using MATLAB softwares, draws corresponding spy
Sign vector, the characteristic vector as corresponding to judgment matrix D eigenvalue of maximum is C=(0.82,0.51,0.23,0.12)T。
(3-3) Consistency Check in Judgement Matrix.
Using the consistency check algorithm of MATLAB programming realizations, to calculate the consistency ration CR of corresponding judgment matrix.
D consistency ration CR=0.0424 is such as calculated<0.1, therefore the judgment matrix meets uniformity.If it is unsatisfactory for uniformity
Inspection then needs to adjust corresponding judgment matrix untill consistency check is met.
(3-4) calculates weight.
After judgment matrix is by consistency check, it will determine that the characteristic vector corresponding to matrix eigenvalue of maximum carries out normalizing
Change is handled, and obtains the weight of all elements under this grade of index.Such as it is to the normalized results of matrix-vector C (weight)：
(0.4869,0.3031,0.1395,0.0705)T。
The scoring of (3-5) each index.
(3-5-1) sets the transformation rule of the attribute corresponding to all indexs according to the available accuracy demand of design item.Such as
Learn the work attendance under participation, its transformation rule is：(arrive number/number need to be arrived) * 100.
(3-5-2) obtains the property value of corresponding index, enters line discipline conversion, is scored.As certain learner's halves needs
It is 20 to number, its actual arrival number is 15, then its work attendance is scored at 75.
(3-6) calculates design item assessment result.
After obtaining scoring corresponding to each index and weight, formula is assessed according to design itemIt is calculated
The final assessment result of the design item, formula are as follows：
Results of learning=48.69%* study participation+30.31%* learning ability+13.95%* learning interests+
7.05%* Learning satisfactions
Design the using effect that item is assessed
The contextual data comes from " C language " class selected by the first term of 41 learners of certain domestic freshman year class
The learning process of journey.The course opened up time span as 11 weeks, and learner can also log in this in addition to normal classroom learning
The cloud classroom platform in school carries out on-line study.Under the background for being completely in natural experiment, what extraction obtained 10 learners should
Course all on-line learning behavior daily record, and pass through the hands such as classroom hand-kept, survey, video recording recording and Expression analysis
Section, the classroom learning data in terms of being got by quantification manner comprising students psychology and cognition.Extract learner's results of learning
The property value of this design item index at different levels, and distinguish corresponding multiplied by weight after being normalized, it can obtain
The scoring assessed on these learner's results of learning, it is specific as shown in table 5.In table 5, also this 10 study are given simultaneously
The real result of person's final examination is to compare.
The learner's results of learning of table 5. are assessed and final examination achievement
| Learner | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
| Results of learning are assessed | 73 | 63 | 50 | 60 | 71 | 50 | 62 | 64 | 68 | 78 |
| Final examination achievement | 88 | 84 | 70 | 73 | 81 | 68 | 73 | 71 | 76 | 92 |
High-level programming language (such as Python) programming realization correlation coefficient function can be used, is calculated after substituting into data
The coefficient correlation of both results of learning evaluation score and real result is 82.85%, and this shows to comment by learner's results of learning
It is high-positive correlation (increasing and decreasing as synchronous) to estimate achievement that the index system of model is calculated with learner's final examination achievement,
With higher uniformity, it is capable of the actual learning situation of preferable representative learning person.Meanwhile according to the design of karr Pearson cames
Coefficient correlation, inquire about related-coefficient test tables of critical values, it is known that both are significantly correlated in 0.01 level, and both are related
Relation is as shown in Figure 1.
Behavioural information perceptual interest scale evaluation of the embodiment two based on visitor
The first step, learner model structure.Based on visitor's example, its attribute and related service are analyzed and concluded.Pass through
Analysis, the information of visitor's example mainly exist in the form of two kinds of static structure and dynamic structure.The information of static structure is main
For identity information (such as name, sex, specialty, age, identification card number, nationality, contact method, hobby)；Dynamic structure
Information predominantly visit during behavioural information and body feeling interaction information, subjective survey data, score data.According to actual need
Seek further filling visitor's attribute.It is specific as shown in table 6.
The visitor's model of table 6.
Second step, formulate design item.
First, the design item related to visitor's model is determined according to demand.In visitor's scene, it is necessary to interest journey
Degree is assessed, therefore selects level of interest as design item.
Then, attribute extraction and classification are carried out according to the interactive object (showpiece) of visitor and its service, can incited somebody to action here
Level of interest is subdivided into objective data, interactive degree, experience property, intellectual, novelty, excitant, scientific seven one-levels and referred to
Mark.Seven first class index caused attribute in service process is extracted, to meet evaluation requirement.
Finally, according to the data attribute extracted, recursive hierarchy structure is established, forms the total of the design item (level of interest)
Body framework.
The specific construction step for designing item (level of interest) overall framework is as follows：
(2-1) is classified to the attribute of design item (level of interest) extraction by it to the correlation degree for designing item.Will be with
The attribute of design item (level of interest) direct correlation is arranged to first class index Xi, objective data X1, interactive degree X2, experience property X3、
Excitant X4, intellectual X5, innovative X6。
(2-2) is directed to first class index X1Actual content, analyze its by taking pictures, video recording behavior, experience number, be resident when
Between, splitting glass opaque, queuing time, ask for help or inquire that the sub- factor of number six determines, select this little factor as two level
Index, X is used respectively11、X12、X13、X14、X15、X16Represent.
(2-3) analysis is taken pictures, record a video behavior X11, it is not possible to continue to segment three-level index.So far take pictures, record a video behavior X11
Structure finishes.Other two-level index X is constructed in the same way12、X13、X14、X15、X16.Final design item (interest journey
Degree) overall frame finish.Specific design item overall framework is as shown in table 7.
The level of interest of table 7. designs item overall framework
3rd step, the formulation of appraisal procedure.Use analytic hierarchy process (AHP) commenting as design item (level of interest) analyzed
Estimate method, its steps flow chart is specific as follows shown：
(3-1) Judgement Matricies
Comprise the following steps that：
(3-1-1) is directed to first class index Xi, objective data X1, interactive degree X2, experience property X3, intellectual X4, innovative X5、
Excitant X6, judge the relative importance between six indexs, corresponding importance scale a provided according to table 1ij,a11Refer to
Objective data X1With objective data X1The importance scale compared, it is of equal importance to be entered as 1, it is designated as a11=1；a12Refer to objective number
According to X1With interactive degree X2The importance scale compared, is designated as a12=2；A21 refers to interactive degree X2 compared with objective data X1
Importance scale, be inverses of the X1 compared with X2, be designated as a21=1/2；Similarly, it is as follows that judgment matrix D can be obtained：
(3-1-2) calculates judgment matrix D corresponding to all two-level index under each first class index successivelyi, for first
Individual first class index objective data X1, by its corresponding two-level index expression sound X11, take pictures video recording behavior X12, experience number X13、
Residence time X14, history visitor number ranking X15, splitting glass opaque X16, queuing time X17Judgement Matricies D1It is as follows：
Interactive degree X is calculated successively2, experience property X3, intellectual X4, innovative X5, excitant X6, scientific X7Corresponding structure
Make judgment matrix D2, D3, D4, D5, D6, D7。
(3-1-3) calculates judgment matrix D corresponding to three-level index successivelyij, DijRepresent under i-th of first class index j-th two
The judgment matrix of three-level index corresponding to level index.
The weight vector of (3-2) matrix calculates.Using feature vector method, the judgment matrix is calculated most using MATLAB softwares
Big characteristic value, corresponding characteristic vector is drawn, the characteristic vector as corresponding to judgment matrix D eigenvalue of maximum is C=
(0.7745,0.5103,0.3122,0.1593,0.1086,0.0715)T。
(3-3) Consistency Check in Judgement Matrix.It is corresponding to calculate using the consistency check algorithm of MATLAB programming realizations
The consistency ration CR of judgment matrix.D consistency ration CR=0.0219 is such as calculated<0.1, therefore the judgment matrix is expired
Sufficient uniformity.Judgment matrix is untill consistency check is met corresponding to need to adjust if being unsatisfactory for consistency check.
(3-4) calculates weight.After judgment matrix is by consistency check, the spy corresponding to matrix eigenvalue of maximum will determine that
Sign vector is normalized, and obtains the weight of all elements under this grade of index.Such as to the normalized results of matrix-vector C
(weight) is：(0.3999,0.2635,0.1612,0.0823,0.0561,0.0369)T。
The scoring of (3-5) each index.
(3-5-1) sets the transformation rule of the attribute corresponding to all indexs according to the available accuracy demand of design item.Such as
The work attendance of left-hand seat time under experience sense, its transformation rule are：The left-hand seat time<=10 seconds, it is scored at 90；10<The left-hand seat time<=30
Second, it is scored at 60；Otherwise score=0.
(3-5-2) obtains the property value of corresponding index, enters line discipline conversion, is scored.As certain visitor one visited
Cheng Zhong, the left-hand seat time interacted with certain showpiece are 12 seconds, then are scored at 60.
(3-6) calculates design item assessment result.After obtaining scoring corresponding to each index and weight, assessed according to design item public
FormulaThe final assessment result of the design item is calculated, formula is as follows
Level of interest=39.99%* objective data+26.35%* interaction degree+16.12%* experience senses+8.23%* thorns
It is innovative to swash property+5.61%* intellectual+3.69%*
Design the using effect that item is assessed
The contextual data comes from visit of all visitors to science and technology center's showpiece exhibition item in certain domestic science and technology center 1 year
Learning process.Data caused by the whole visit process of visitor are mainly adopted by infrastructure (camera, sensor etc.)
Collection is obtained, while another part data are obtained in a manner of subjective survey.
Relative to traditional questionnaire method is used, proposed level of interest appraisal procedure is effective, is tested
Card mode is as follows：(1) first by survey, the mode of comprehensive grading, obtain each visitor during visit for
The actual interest degree of showpiece exhibition item；(2) the level of interest assessment result of visitor is calculated by context of methods；(3) to this
Both values carry out relevance verification, and concrete outcome is shown in Table 8 (coefficient correlation obtained by calculating is 78.68%), this table
Bright visitor's actual interest degree and visitor's level of interest assessment result are high-positive correlation (such as synchronous increases and decreases), with compared with
High uniformity, it can preferably characterize the actual visit situation of visitor.Meanwhile according to the design phase relation of karr Pearson cames
Number, inquire about related-coefficient test tables of critical values, it is known that both are significantly correlated in 0.01 level, and both dependency relations are as schemed
Shown in 2.
The visitor's level of interest of table 8. is assessed and actual interest degree
| Visitor | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
| Actual interest degree | 65 | 60 | 65 | 50 | 60 | 70 | 75 | 70 | 60 | 70 |
| Level of interest is assessed | 85 | 66 | 81 | 52 | 90 | 92 | 87 | 79 | 70 | 86 |
The content not being described in detail in this specification, belong to prior art known to those skilled in the art.
The foregoing is only presently preferred embodiments of the present invention, be not intended to limit the invention, it is all the present invention spirit and
Within principle, any modification, equivalent substitution and improvements made etc., it should be included in the scope of the protection.
Claims (4)
1. a kind of design item appraisal procedure based on learner model, it is characterised in that this method comprises the following steps：
(1) learner model is built, and based on learner's example, is analyzed and is concluded its attribute and related service, establish learner's
Property set and related service；
(2) design item is formulated, determines the design item related to learner model according to demand, the visitor then acted on according to learner
See object and its service carries out attribute extraction and classification, the index for assessing design item is formed, finally, according to the number extracted
According to attribute, recursive hierarchy structure is established, forms the overall framework of the design item；
(3) appraisal procedure is formulated, the design item appraisal procedure based on learner model is designed using analytic hierarchy process (AHP).
2. the design item appraisal procedure according to claim 1 based on learner model, it is characterised in that：In step (1)
Learner model is established using UML or concept database model；Based on learner, its information is with static structure and dynamic structure
Two kinds of forms are present, and static structure is established according to demographic information, have recorded the personal essential information of learner, dynamic
Structure be for the service behavior of learner, including the style and features information of learner, service process information, performance letter
Breath and status information.
3. the design item appraisal procedure according to claim 1 based on learner model, it is characterised in that step is set in (2)
The specific construction step for counting item overall framework is as follows：
(2-1) is classified to the attribute of design item extraction by it to the correlation degree for designing item, will be with designing item direct correlation
Attribute be arranged to first class index Xi, Xi, represent i-th of first class index for designing item；
(2-2) is directed to first class index XiActual content, analyze it and determined by which sub- factor, select this little factor as two
Level index Xij, XijRepresent j-th of two-level index under i-th of first class index of design item；
(2-3) is examined in whether all two-level index can segment, if can continue to segment, refers to by upper step filling three-level
Mark, this process is repeated, untill all indexs can not subdivide, that is, forms last design item overall framework.
4. the design item appraisal procedure according to claim 1 based on learner model, it is characterised in that made in step (3)
Determine comprising the following steps that for appraisal procedure：
(3-1) Judgement Matricies
After to being contrasted two-by-two between each index at the same level, the relative superior or inferior for each evaluation index that is ranked by 9 points of position ratios
Sequentially, the judgment matrix D of evaluation index is constructed successively；
The weight vector of (3-2) matrix calculates
Using feature vector method, judgment matrix D eigenvalue of maximum is calculated using MATLAB softwares, draw corresponding to feature to
Amount；
(3-3) Consistency Check in Judgement Matrix
Consistency check rule is as follows：Consistency rationWork as CR<When 0.1, it is believed that the uniformity of judgment matrix is can be with
Receive, otherwise need suitably to correct judgment matrix；Wherein, CI represents coincident indicator, and RI represents to search coincident indicator, unanimously
Property index CI calculation formula isWherein n is the index number compared, λmaxFor the Maximum characteristic root of judgment matrix；
Coincident indicator RI is searched, need to be provided with reference to practical experience；
(3-4) calculates weight
It will determine that the characteristic vector corresponding to matrix eigenvalue of maximum is normalized, obtain the weight of each index, one-level
Weight corresponding to index is Wi, WiThe weight of i-th of first class index of design item is represented, weight corresponding to two-level index is Wij,
WijRepresent weight corresponding to j-th of two-level index under i-th of first class index of design item, the expression of three-level index respective weights
Method is by that analogy；
The scoring of (3-5) each index
The transformation rule and scoring criteria of indexs at different levels are set according to the available accuracy demand of design item, and to designing item fingers at different levels
Target property value unifies dimension, its data is all concentrated in an identical span, draws the scoring of each index；One-level
Scoring is Y corresponding to indexi, YiThe scoring corresponding to i-th of first class index of design item is represented, is commented corresponding to two-level index
It is divided into Yij, YijRepresent and scored corresponding to j-th of two-level index under i-th of first class index of design item, three-level index is correspondingly commented
The method for expressing divided is by that analogy；
(3-6) calculates design item assessment result
By design item evaluation indexes at different levels weight and unified dimension after property value, be calculated that the design item is final to be commented
Estimate formulaThis formula eliminates the three-level of design item and following index.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710698682.1A CN107609651A (en) | 2017-08-15 | 2017-08-15 | A kind of design item appraisal procedure based on learner model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710698682.1A CN107609651A (en) | 2017-08-15 | 2017-08-15 | A kind of design item appraisal procedure based on learner model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107609651A true CN107609651A (en) | 2018-01-19 |
Family
ID=61064337
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710698682.1A Pending CN107609651A (en) | 2017-08-15 | 2017-08-15 | A kind of design item appraisal procedure based on learner model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107609651A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108766464A (en) * | 2018-06-06 | 2018-11-06 | 华中师范大学 | Digital audio based on mains frequency fluctuation super vector distorts automatic testing method |
| CN109784671A (en) * | 2018-12-24 | 2019-05-21 | 远光软件股份有限公司 | A kind of user experience quality appraisal procedure and system based on user behavior analysis |
| CN110097099A (en) * | 2019-04-19 | 2019-08-06 | 北京中庆现代技术股份有限公司 | A kind of association analysis method based on student classroom performance sexual behaviour and achievement |
| CN110782375A (en) * | 2019-09-05 | 2020-02-11 | 华南师范大学 | Online learning overall process dynamic analysis method and system based on data |
| CN111221915A (en) * | 2019-04-18 | 2020-06-02 | 江苏大学 | Online learning resource quality analysis method based on CWK-means |
| CN112419810A (en) * | 2020-11-13 | 2021-02-26 | 雷恩友力数据科技南京有限公司 | Intelligent education method for accurate control based on adaptive cognitive interaction |
| CN113505952A (en) * | 2020-12-04 | 2021-10-15 | 广东省建筑设计研究院有限公司 | Masonry structure building group safety assessment method based on geometric and performance similarity |
-
2017
- 2017-08-15 CN CN201710698682.1A patent/CN107609651A/en active Pending
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108766464A (en) * | 2018-06-06 | 2018-11-06 | 华中师范大学 | Digital audio based on mains frequency fluctuation super vector distorts automatic testing method |
| CN108766464B (en) * | 2018-06-06 | 2021-01-26 | 华中师范大学 | Digital audio tampering automatic detection method based on power grid frequency fluctuation super vector |
| CN109784671A (en) * | 2018-12-24 | 2019-05-21 | 远光软件股份有限公司 | A kind of user experience quality appraisal procedure and system based on user behavior analysis |
| CN111221915A (en) * | 2019-04-18 | 2020-06-02 | 江苏大学 | Online learning resource quality analysis method based on CWK-means |
| CN111221915B (en) * | 2019-04-18 | 2024-01-09 | 西安睿德培欣教育科技有限公司 | Online learning resource quality analysis method based on CWK-means |
| CN110097099A (en) * | 2019-04-19 | 2019-08-06 | 北京中庆现代技术股份有限公司 | A kind of association analysis method based on student classroom performance sexual behaviour and achievement |
| CN110782375A (en) * | 2019-09-05 | 2020-02-11 | 华南师范大学 | Online learning overall process dynamic analysis method and system based on data |
| CN112419810A (en) * | 2020-11-13 | 2021-02-26 | 雷恩友力数据科技南京有限公司 | Intelligent education method for accurate control based on adaptive cognitive interaction |
| CN113505952A (en) * | 2020-12-04 | 2021-10-15 | 广东省建筑设计研究院有限公司 | Masonry structure building group safety assessment method based on geometric and performance similarity |
| CN113505952B (en) * | 2020-12-04 | 2024-02-02 | 广东省建筑设计研究院有限公司 | Masonry structure building group safety assessment method based on geometric and performance similarity |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107609651A (en) | A kind of design item appraisal procedure based on learner model | |
| CN107230174B (en) | Online interactive learning system and method based on network | |
| Freedman | Statistical models and causal inference: a dialogue with the social sciences | |
| Chen et al. | Combining fuzzy AHP with MDS in identifying the preference similarity of alternatives | |
| CN109582875B (en) | Personalized recommendation method and system for online medical education resources | |
| US20160042282A1 (en) | Relationship evaluator | |
| Danhier et al. | Comparing compositional effects in two education systems: The case of the Belgian communities | |
| Ferrari et al. | Identification of cross-domain ambiguity with language models | |
| CN111724881A (en) | Psychological sand table analysis method and system | |
| Egilmez et al. | Application of artificial neural networks to assess student happiness | |
| Herlambang et al. | Intelligent computing system to predict vocational high school student learning achievement using Naï ve Bayes algorithm | |
| Vieira et al. | Definition of a model based on bibliometric indicators for assessing applicants to academic positions | |
| Marshall | Contribution of open-ended questions in student evaluation of teaching | |
| Palacios et al. | Extending a simple genetic cooperative-competitive learning fuzzy classifier to low quality datasets | |
| Simon et al. | Utilizing data driven methods to identify gender bias in LinkedIn profiles | |
| Alshraideh et al. | Beyond the scoreboard: a machine learning investigation of online games’ influence on Jordanian university students’ grades | |
| Yilmaz İnce et al. | Automatic e-content sequencing system for personalised learning environments by using fuzzy AHP based on multiple intelligences | |
| Tahmasebi et al. | Implementation and evaluation of a resource-based learning recommender based on learning style and web page features | |
| Sabnani et al. | Prediction of student enrolment using data mining techniques | |
| Din et al. | Creation of a knowledge society via the use of mobile blog: a model of integrated meaningful hybrid e-training | |
| Zhou et al. | A new information theory-based serendipitous algorithm design | |
| Lee et al. | Prescribing deep attentive score prediction attracts improved student engagement | |
| Ochirbat et al. | Occupation recommendation with major programs for adolescents | |
| Zhou | Research on teaching resource recommendation algorithm based on deep learning and cognitive diagnosis | |
| Lan | Research on the construction and reform path of online and offline mixed English teaching model in the internet era |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180119 |
|
| RJ01 | Rejection of invention patent application after publication |