Summary of the invention
To solve the above problems, the purpose of the present invention is to provide a kind of behavioral datas to collect and survey system, consider comprehensively
The each influence factor of evaluation is participated in, so that appraisal result is more objective reliable.
The present invention provides a kind of behavioral datas to collect and survey system, comprising:
Data collection module is used to collect the behavioral data of student by unified data access interface;
Wherein, behavioral data includes: the academic record of teacher's typing and record that student and teacher submit and corresponding
Picture;
Data processing module, the record content for being used to submit the academic record of teacher's typing and student and teacher into
Row data processing obtains and comments data for carrying out the comprehensive of overall merit to student, and carries out to the picture in the record of submission former
Figure storage and the storage of compressed picture;
Wherein, it is comprehensive comment data include: characterize Scores data, and submit record in characterization student other element
The data of matter;
Data display module is used to for the record and corresponding picture of student and teacher's submission being shown;
Data analysis module is used to carry out the comprehensive statistical for commenting data to all behavioral datas submitted in a period of time
Analysis, obtains the comprehensive of each student and comments scoring event, and generate the comprehensive of each student and comment report;
Wherein, data analysis module specifically includes:
First order impact factor determining module is used to determine that comprehensive comment participates in the comprehensive first order impact factor commented in data, building
The comprehensive first order impact set of factors X={ X commented1,X2,…,Xi..., Xm, wherein XiIndicate first order impact factor, i=1,2 ..., m;
Secondary influence factors determining module is used to determine that comprehensive comment participates in the comprehensive each first order impact factor institute commented in data
Corresponding Secondary influence factors construct its corresponding Secondary influence factors collection X for each first order impact factor respectivelyi={ Xi1,
Xi2,…,Xij,…,Xini, wherein XijIndicate the corresponding Secondary influence factors of first order impact factor, j=1,2 ..., ni;
Opinion rating determining module is used to construct the comprehensive evaluation indice Y={ Y commented1,Y2,…,Yk,…,Yn,
In, YkIndicate opinion rating, k=1,2 ..., n;
First order impact factor evaluation module is used to evaluate each first order impact factor, obtains single level-one shadow
The evaluations matrix of the factor of soundIn formula, ni indicates Xi
In element number, rijkIndicate Secondary influence factors XijTo opinion rating YkDegree of membership, i=1,2 ..., m, j=1,2 ...,
Ni, k=1,2 ..., n;
Weight determination module is used to determine the corresponding weight vectors W={ W of each first order impact factor1,W2,…,
Wi,…,Wm, andWherein, component WiIndicate the corresponding weight of each first order impact factor;
Meanwhile determining the corresponding weight vectors W of each Secondary influence factorsi={ Wi1,Wi2,…,Wij,…,Wini, andWherein, component WijIndicate the corresponding weight of each Secondary influence factors;
First order impact factor evaluation matrix module is used to obtain that the final evaluations matrix of single first order impact factor to be Bi
=WiRi={ bi1,bi2,…bij,…,bini, wherein i=1,2 ..., m, j=1,2 ..., ni;
Wherein,
Overall merit matrix module is used for each first order impact factor XiAs separate element, BiAs XiEvaluation to
Amount constructs the comprehensive overall merit matrix R={ B commented1,B2,…,Bi,…,Bm}T, and then obtain overall merit vector B=WR={ b1,
b2,…,bi,…,bm};
Wherein, bi=WiBi, i=1,2 ..., m;
It is comprehensive to comment total score computing module, it is used for b in overall merit vectoriCorresponding score value yiIt substitutes into and calculates, obtain
It gives birth to finally comprehensive and comments total scoreWherein, i=1,2 ..., m.
It further improves as of the invention, in data display module, in display record content, shows according to demand each
Record statistical information under kind various combination, exhibition method have pie chart, histogram, spider diagram or table.
Improved as of the invention further, in data display module, in exhibiting pictures, show according to demand original image or
Compressed picture.
It is improved as of the invention further, in data display module, the record content and figure of student and teacher to displaying
Piece is commented on and is interacted, and when having false record and picture, deletes this record and corresponding picture.
It is improved as of the invention further, in data analysis module, to comment report include: the basic of student for student comprehensive
Information, the record of submission, student submit the frequency trend graph of record on a timeline and comprehensive comment total score.
As further improvement of the invention, weight determination module is specifically included:
Judgment matrix constructing module is used to using 1-9 scaling law be compared several first order impact factors two-by-two, be constructed
One m rank matrix A,Wherein, p=1,2 ..., m, q=1,2 ..., m;
Feature vector computing module is used to calculate the feature vector and normalized of m rank matrix A, is normalized
Feature vector α=[α afterwards1 α2 … αi … αm], i=1,2 ..., m;
Consistency check module is used to carry out consistency check to m rank matrix A, if upchecking, will obtain in step b
To normalization after feature vector in component be determined as the weight component α of first order impact factori, i=1,2 ..., m;
Weight component computing module determines the weight of each first order impact factor for several experts using Delphi method,
The mean value and standard deviation for the weight being calculated, then calculated result is returned into expert, expert determine again each first order impact because
The weight of element, repeatedly, until the deviation of the weight of each first order impact factor and its mean value is no more than preset threshold
Value, at this point, using the mean value of the weight of each first order impact factor as the weight component β of first order impact factori, i=1,2 ...,
m;
Processing module is weighted, is used to determine the weight W of each first order impact factor using linear weighted functioni=λ αi+(1-
λ)βi, i=1,2 ..., m;
Normalized module is used for normalized weight vector, i.e.,It is corresponding to obtain first order impact factor
Weight vectors W={ W1,W2,…,Wi,…,Wm}。
It is improved as of the invention further, for each first order impact factor, to being subordinated to the first order impact factor
All Secondary influence factors, using above-mentioned judgment matrix constructing module, feature vector computing module, consistency check module, power
Weight component computing module, weighting processing module and normalized module calculation processing, obtain the corresponding power of Secondary influence factors
Weight vector Wi={ Wi1,Wi2,…,Wij,…,Wini}。
It is further improved as of the invention, in judgment matrix constructing module, when comparing two-by-two, in m rank matrix A:
When element p is identical as the importance of element q, apq=1;
When element p and element q are slightly important, apq=3, conversely, apq=1/3;
When element p and element q are important, apq=5, conversely, apq=1/5;
As element p and element q much more significant, apq=7, conversely, apq=1/7;
When element p and element q are of crucial importance, apq=9, conversely, apq=1/9.
It is further improved as of the invention, in feature vector computing module:
M rank matrix A is respectively arranged into summation;
Sum i.e. with the element of each column divided by column is normalized to each column:Obtain one it is new
M rank matrix B, wherein ∑ apqFor the sum of each column;
To a line every in m rank matrix B summation you can get it feature vector;
To the feature vector normalized in m rank matrix B, feature vector α=[α can be obtained1 α2 … αi …
αm], i=1,2 ..., m.
It is further improved as of the invention, in consistency check module:
Calculate the Maximum characteristic root of m rank matrix A
Calculate the coincident indicator of m rank matrix A
Calculate the consistent ratio of randomness of m rank matrix AWherein, RI is constant;
As CR < 0.1, consistency check passes through.
As further improvement of the invention, λ=0.5.
The invention has the benefit that
1, participating in the comprehensive data commented includes the record data that student submits and the academic record data that teacher submits, and is avoided
The unification for evaluating data, is also abandoned in the past using school grade as judgment criteria, so that comprehensive comment result more objective;Meanwhile it is right
The record of submission joined judgment mechanism, false record be deleted, this makes data more transparent credible, improves
The comprehensive confidence level commented;
2, the comprehensive data commented are participated in and is related to multiple Secondary influence factors of multiple first order impact factors and subordinate, it can be according to need
It asks dynamically to adjust these influence factors, data are more perfect, comment more diversification so that comprehensive, more objective also more reliable;
3, for the weight of first order impact factor and Secondary influence factors, subjective expert opinion has been concentrated, power is increased
Prestige has carried out objective processing also by mathematical algorithm, reduces artificial unilateral influence, and by subjective results and objective results
Be weighted processing, effectively combine subjectivity result and objectivity as a result, it is possible to reasonably reflect evaluation expert preference,
Also can qualitative, quantitative calculate reasonable weight coefficient so that the comprehensive accuracy commented and reasonability are greatly improved;
4, the comprehensive data commented are participated in and covers teaching and extracurricular a variety of record data, complete students'learning and class
The Process-oriented evaluation of outer development promotes the learning initiative and initiative of student, also specifies for the promotion of student and improvement
Direction, excitation student's active various aspects develop oneself.
Specific embodiment
The present invention is described in further detail below by specific embodiment and in conjunction with attached drawing.
As shown in Figure 1, a kind of behavioral data of the embodiment of the present invention collects and surveys system, comprising:
Data collection module collects the behavioral data of student by unified data access interface.
Wherein, behavioral data includes: the academic record of teacher's typing and record that student and teacher submit and corresponding
Picture.
This data access interface can by directly by execl table in a manner of batch import behavioral data, can also be right
Other systems are connected to, directly directly import the data of other systems in batches.The academic record of typing can be hundred-mark system, Pyatyi
System, two-stage system etc., meet the needs of without Examination Form with this, can diversified record Students ' in terms of academic record
The achievement of acquisition.Meanwhile the academic record of typing further includes lacking achievement of examining and take make-up exam, and really meets the requirement of Process Character record,
So that it is more perfect to participate in comprehensive data commented, so make it is subsequent it is comprehensive comment it is more objective credible.
Data processing module, the record content that academic record and student and teacher to teacher's typing are submitted carry out data
Processing obtains and comments data for carrying out the comprehensive of overall merit to student, and carries out original image storage to the picture in the record of submission
It is stored with compressed picture.
Wherein, it is comprehensive comment data include: characterize Scores data, and submit record in characterization student other element
The data of matter, these data are all the data for participating in subsequent evaluation.Original image and pressure are carried out to the picture in the record of submission
Picture stores respectively after contracting, can show compressed picture according to demand, can be improved the load and service efficiency of system,
It just goes to transfer original image when needing to check original image.
The record and corresponding picture of student and teacher's submission are shown by data display module.
Student and teacher comment on and interact to the record content and picture of displaying, when having false record and picture,
This record and corresponding picture are deleted, this makes data more transparent credible, improves the comprehensive confidence level commented.In display record
Rong Shi, shows the record statistical information under various various combinations according to demand, exhibition method have pie chart, histogram, spider diagram or
Table can intuitively show the statistical information that batch processing obtains.In exhibiting pictures, after showing original image or compression according to demand
Picture.In addition, can add the record of the achievement for student for the academic record of success typing, student and teacher can also
Secondary record is commented on and interacted.
Data analysis module carries out the comprehensive statistical analysis for commenting data to all behavioral datas submitted in a period of time, obtains
It takes the comprehensive of each student to comment scoring event, and generates the comprehensive of each student and comment report.
The comprehensive of student comments report to be shown using term and academic year as the period.Student it is comprehensive comment report include: learn
Raw essential information (name and academic year, term information etc.), the record of submission, student submits the frequency of record to walk on a timeline
Gesture figure and comprehensive comment total score.When showing record, all records of corresponding first order impact factor, Yi Jicong can be shown respectively
Belong to the corresponding record of Secondary influence factors of each first order impact factor.For example, the comprehensive base for commenting report top to show student
This information, middle section are that comprehensive within selected term or academic year of student comments total score and each grading module (i.e. first order impact
Factor) under each dimension (i.e. Secondary influence factors) record strip number and it is comprehensive comment score, be then that student mentions on a timeline below
Hand over the frequency trend graph of record.The comprehensive record addition situation for commenting report not only to show student and it is comprehensive comment scoring event, also with
The original record of student's addition is associated, and student and teacher can be checked all under the dimension by the title of click dimension
Original record.
Wherein, data analysis module specifically includes:
First order impact factor determining module determines that comprehensive comment participates in the comprehensive first order impact factor commented in data, constructs comprehensive comment
First order impact set of factors X={ X1,X2,…,Xi..., Xm, wherein XiIndicate first order impact factor, i=1,2 ..., m.
Secondary influence factors determining module determines that comprehensive comment is participated in data corresponding to the comprehensive each first order impact factor commented
Secondary influence factors construct its corresponding Secondary influence factors collection X for each first order impact factor respectivelyi={ Xi1,Xi2,…,
Xij,…,Xini, wherein XijIndicate the corresponding Secondary influence factors of first order impact factor, j=1,2 ..., ni.
Opinion rating determining module constructs the comprehensive evaluation indice Y={ Y commented1,Y2,…,Yk,…,Yn, wherein YkIt indicates
Opinion rating, k=1,2 ..., n.
First order impact factor evaluation module evaluates each first order impact factor, obtains single first order impact factor
Evaluations matrixIn formula, ni indicates XiIn member
Plain number, rijkIndicate Secondary influence factors XijTo opinion rating YkDegree of membership, i=1,2 ..., m, j=1,2 ..., ni, k=
1,2,…,n。
Wherein, degree of membership can be determined using Delphi method.
Weight determination module determines the corresponding weight vectors W={ W of each first order impact factor1,W2,…,Wi,…,Wm,
AndWherein, component WiIndicate the corresponding weight of each first order impact factor;
Meanwhile determining the corresponding weight vectors W of each Secondary influence factorsi={ Wi1,Wi2,…,Wij,…,Wini, andWherein, component WijIndicate the corresponding weight of each Secondary influence factors.
Specifically, weight determination module includes:
Several first order impact factors are compared two-by-two using 1-9 scaling law, construct a m by judgment matrix constructing module
Rank matrix A,Wherein, p=1,2 ..., m, q=1,2 ..., m;
Wherein, when comparing two-by-two, in m rank matrix A:
When element p is identical as the importance of element q, apq=1;
When element p and element q are slightly important, apq=3, conversely, apq=1/3;
When element p and element q are important, apq=5, conversely, apq=1/5;
As element p and element q much more significant, apq=7, conversely, apq=1/7;
When element p and element q are of crucial importance, apq=9, conversely, apq=1/9.
Feature vector computing module calculates the feature vector and normalized of m rank matrix A, the spy after being normalized
Levy vector α=[α1 α2 … αi … αm], i=1,2 ..., m.
Specifically, the method for calculating feature vector are as follows:
M rank matrix A is respectively arranged into summation;
Sum i.e. with the element of each column divided by column is normalized to each column:Obtain one it is new
M rank matrix B, wherein ∑ apqFor the sum of each column;
To a line every in m rank matrix B summation you can get it feature vector;
To the feature vector normalized in m rank matrix B, feature vector α=[α can be obtained1 α2 … αi …
αm], i=1,2 ..., m.
Consistency check module, carrying out consistency check to m rank matrix A will return if upchecking obtained in step b
The component in feature vector after one change is determined as the weight component α of first order impact factori, i=1,2 ..., m.
Specifically, the method for consistency check are as follows:
Calculate the Maximum characteristic root of m rank matrix A
Calculate the coincident indicator of m rank matrix A
Calculate the consistent ratio of randomness of m rank matrix AWherein, RI is constant;
As CR < 0.1, consistency check passes through.
Weight component computing module, several experts determine the weight of each first order impact factor using Delphi method, calculate
The mean value and standard deviation of obtained weight, then calculated result is returned into expert, expert determines each first order impact factor again
Weight, repeatedly, until the deviation of the weight of each first order impact factor and its mean value is no more than preset threshold value, this
When, using the mean value of the weight of each first order impact factor as the weight component β of first order impact factori, i=1,2 ..., m.It should
Module is determining weight component βiIt can replace with the following method: the level-one factor of evaluation in m rank matrix A is compared two-by-two,
Row element and column element are compared, each expert obtains a m rank judgment matrix, and the m rank that L expert obtains is judged square
Corresponding element adduction in battle array, obtains the D of a new m rank judgment matrix,
Wherein, p=1,2 ..., m, q=1,2 ..., m;Calculate new m
The summation of every a line level-one factor of evaluation in rank judgment matrix DCalculate being averaged for each level-one factor of evaluation
Score valueObtain weightAnd using the weight as the weight component β of level-one factor of evaluationi, i.e.,
It weights processing module and determines the weight W of each first order impact factor using linear weighted functioni=λ αi+(1-λ)βi, i
=1,2 ..., m.
λ indicates objective preference coefficient, and value is the apparent preference coefficient of 0~1,1- λ, according to the inclined of evaluation side when specifically choosing
It is good to set, it is preferred that λ=0.5.
Normalized module, normalized weight vector, i.e.,Obtain the corresponding weight of first order impact factor to
Measure W={ W1,W2,…,Wi,…,Wm}。
For each first order impact factor, to all Secondary influence factors for being subordinated to the first order impact factor, in use
State judgment matrix constructing module, feature vector computing module, consistency check module, weight component computing module, weighting processing
Module and normalized module calculation processing obtain the corresponding weight vectors W of Secondary influence factorsi={ Wi1,Wi2,…,
Wij,…,Wini, I will not elaborate.Entire weight treatment process, had both considered the expert opinion of authority, at mathematics
Reason has carried out quantum chemical method to influence factor, reduce it is unilateral think to influence, while joined weighting processing and multiple normalizing
Change processing, it is ensured that the accuracy of weight.
First order impact factor evaluation matrix module show that the final evaluations matrix of single first order impact factor is Bi=WiRi
={ bi1,bi2,…bij,…,bini, wherein i=1,2 ..., m, j=1,2 ..., ni;
Wherein,
Overall merit matrix module, by each first order impact factor XiAs separate element, BiAs XiEvaluation vector, structure
Build the comprehensive overall merit matrix R={ B commented1,B2,…,Bi,…,Bm}T, and then obtain overall merit vector B=WR={ b1,b2,…,
bi,…,bm};
Wherein, bi=WiBi, i=1,2 ..., m.
It is comprehensive to comment total score computing module, by b in overall merit vectoriCorresponding score value yiIt substitutes into and calculates, it is final to obtain the student
Comprehensive comment total scoreWherein, i=1,2 ..., m.
In the present embodiment, first order impact factor can be set and be respectively as follows: X1Sincere morals, X2Academic record, X3Physical and mental health,
X4Artistic accomplishment, X5Ability to organize and coordinate, X6Activity practice, X7Personal growth, X8Group incentive, X9Other business.Wherein, subordinate
In first order impact factor X1The Secondary influence factors of sincere morals are respectively X11Morals reward, X12Disciplinary punishment, X13Public and social interest
And voluntary service, X14Residents' civility, X15Rally performance;It is subordinated to first order impact factor X2The Secondary influence factors of academic record point
It Wei not X21Academic record hundred-mark system, X22Academic record Pyatyi system, X23Academic record two-stage system, X24Operation performance, X25Classroom table
It is existing, X26Check class attendance, X27School work reward, X28Nationwide examination for graduation qualification achievement;It is subordinated to first order impact factor X3Physically and mentally healthy second level influence because
Element is respectively X31National student physical health standard, X32Sport reward;It is subordinated to first order impact factor X4The second level of artistic accomplishment
Influence factor is respectively X41Intelligence and art reward;It is subordinated to first order impact factor X5The Secondary influence factors of ability to organize and coordinate are respectively
X51Tenure, X in class52The tenure of students' union of school Youth League committee, X53School community tenure;It is subordinated to first order impact factor X6Activity practice
Secondary influence factors are respectively X61Activity practice reward, X62Activity of the Party and the League, X63Community activity, X64Military training, X65Visiting and learning, X66
Social investigation;It is subordinated to first order impact factor X7The Secondary influence factors of personal growth are respectively X71Learn interest and preference hair
Exhibition, X72Artistic accomplishment and speciality culture, X73Physical health and physical training, X74Stage brief summary and personal self-examination;It is subordinated to level-one
Influence factor X8The Secondary influence factors of group incentive are respectively X81Class collectivity reward, X82Corporations' group incentive;It is subordinated to level-one
Influence factor X9The Secondary influence factors of other business are respectively X91Good people and good deeds.
The record of submission adhere to separately from different first order impact factors and its corresponding Secondary influence factors, in display record
When, it can be shown by pie chart, histogram, spider diagram or table, dynamic quantization will be recorded, more intuitively.Certainly, and not only limit
In above-mentioned influence factor, these influence factors can be adjusted dynamically according to demand, to adapt to the student of different regions difference school.
In addition, thus can see, participate in the comprehensive first order impact factor and subordinate commented Secondary influence factors be related to it is multiple so that evaluation is more
Diversification, it is more objective also more reliable.
In the present embodiment, opinion rating is respectively Y1Outstanding, Y2Well, Y3Generally, Y4Poor, Y5Difference, corresponding score value can
It is set as 1,0.8,0.6,0.4,0.2.Opinion rating and its corresponding score value can of course be adjusted, according to actual needs with suitable
Answer the student of different regions difference school.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field
For art personnel, the invention may be variously modified and varied.All within the spirits and principles of the present invention, made any to repair
Change, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.