CN106803123A - A kind of automatic volume group method for online exam - Google Patents
A kind of automatic volume group method for online exam Download PDFInfo
- Publication number
- CN106803123A CN106803123A CN201611184242.6A CN201611184242A CN106803123A CN 106803123 A CN106803123 A CN 106803123A CN 201611184242 A CN201611184242 A CN 201611184242A CN 106803123 A CN106803123 A CN 106803123A
- Authority
- CN
- China
- Prior art keywords
- paper
- constraint
- group
- individuality
- topic type
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000010353 genetic engineering Methods 0.000 claims abstract description 16
- 239000011159 matrix material Substances 0.000 claims description 21
- 230000035772 mutation Effects 0.000 claims description 20
- 238000012360 testing method Methods 0.000 claims description 16
- 230000035484 reaction time Effects 0.000 claims description 12
- 238000011156 evaluation Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 230000001149 cognitive effect Effects 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 230000006978 adaptation Effects 0.000 claims description 3
- 238000011017 operating method Methods 0.000 claims description 3
- SBNFWQZLDJGRLK-UHFFFAOYSA-N phenothrin Chemical compound CC1(C)C(C=C(C)C)C1C(=O)OCC1=CC=CC(OC=2C=CC=CC=2)=C1 SBNFWQZLDJGRLK-UHFFFAOYSA-N 0.000 claims description 3
- 238000005096 rolling process Methods 0.000 claims description 3
- 230000004083 survival effect Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 229910017435 S2 In Inorganic materials 0.000 claims 1
- 230000002068 genetic effect Effects 0.000 abstract description 19
- 108090000623 proteins and genes Proteins 0.000 description 7
- 230000002028 premature Effects 0.000 description 6
- 230000008901 benefit Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000003362 replicative effect Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000009395 breeding Methods 0.000 description 1
- 230000001488 breeding effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
- G06Q50/205—Education administration or guidance
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Educational Technology (AREA)
- Educational Administration (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Physiology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Genetics & Genomics (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of automatic volume group method for online exam, including：(1) setting group volume constraints；(2) initial population P0 is generated according to constraints；(3) each described individual fitness function value is calculated；(4) judge individuality fitness function value whether satisfaction group volume condition, generated if meeting and paper and terminate the automatic volume group method；Judge whether to reach maximum iteration if being unsatisfactory for；If be not reaching to, iterations plus 1, carry out genetic manipulation；(5) new colony P1, and return to step (3) are generated by the genetic manipulation.The present invention completes automatic volume group using genetic algorithm, can quickly obtain the paper of a science, solves the problems, such as that examination system paper randomness in group licks journey is big, unstable, improves fairness, the reasonability of examination.
Description
Technical field
The present invention relates to one kind group volume method, more particularly to a kind of automatic volume group method for online exam.
Background technology
At present, the upper various examinations of school and society mostly use traditional test way.In this manner, organize once complete
Whole examination will at least experience following four step：Manually make the test, tissue examination, group signature, examination result and examination paper analysis.
With various types of examinations be continuously increased with growing examinee's quantity, the similar work about test amount of tissue can be increasing,
Efficiency also can be more and more lower.The trend and method of traditional test way are solved, is exactly to form one using computing technique, network technology
It is online exam to plant new examination mode.A current development trend is using the computer online testing of extensive test item bank
Pattern.The group volume of tradition examination, sentence volume mode and student's score management etc. huge change just occurs.Therefore, how to make
Examination process become it is convenient, efficient, quick, just be modern education an important topic.On-line Examining system is that tradition is examined
The extension of field, it can utilize the wide immensity of network, and realization take an exam to student, enormously simplify whenever and wherever possible
The process of tradition examination, reduces examination cost.The form of online testing has the advantages that its science, timely, accurate, justice,
The advantage that cannot be substituted with traditional Examination Form and compared.Therefore On-line Examining system be under E-learning in the urgent need to
Product.
The more extensive test papers algorithm of network test system application is including randomly selecting, precedence algorithm, error compensation are calculated
The methods such as method, backtracking heuristic.Precedence algorithm can be recalculated and sorted, and cause algorithm complex to increase, and effect is not
It is good.Random algorithm is it is not intended that optimization problem, and group volume randomness is very big, uncertain very big, cannot guarantee that paper is reasonable
Property and science.Automatic volume group species is more in examination system, and each advantageous and inferior position.For the need of the application of examination
Ask, take one thing with another as far as possible, the solution of diversification can be provided.Therefore, in systematical design idea, employ above-mentioned
Outside algorithm, propose to add a kind of new genetic algorithm to solve Shortcomings in traditional algorithm, merge new research algorithm and open
Interface is put to develop and realize.
Basic genetic algorithm has three kinds of basic operations：Selection, intersect and make a variation, be the most frequently used in genetic algorithm three kinds
Algorithm；
Selection operation is exactly to replicate operation, is only conditional on replicating, and can follow the principle of " advantage is bad to eliminate " to replicate, and
The purpose of selection is exactly, in order to excellent individuality chooses from current group, to be bred down as the male parent for having excellent genes
A generation, so duplication obtain good follow-on probability will be larger.Such as according to the size of individual fitness function value
To select, examination system group volume in be exactly exercise question fitness more it is high it is easier be selected.
Crossover operation is most important genetic manipulation in genetic algorithm, and crossover operation purpose is desirable to can by crossover operation
To obtain individuality of new generation, and new individual can be with the good characteristic of hereditary parent individuality, so as to obtain more excellent new individual.
During practical operation, crossover operation is that intragroup each individuality is intersected at random according to certain probability, new so as to obtain two
It is individual.Crossover operation process is in examination system group volume using such as：Individual P1 and P2, after exchanging at random individual internal body portion examination question
Obtain new G1 and G2.
Mutation operation is for the situation that morphed when following biological gene heredity, to produce new individuality to provide chance.
Mutation operation is that one or some individualities are randomly choosed first in current group in step, random according to definitive variation probability
Change some or some genic values, so as to obtain new individuality.It is specifically certain portion volume of selection in examination system group volume
The random part exercise question for changing paper, so as to obtain a paper.
Although examination system is had many good qualities using genetic algorithm, can efficient computing parallel, in practical application
During or genetic algorithm Premature Convergence can often occur, be absorbed in the phenomenon of local value.This phenomenon is also referred to as algorithm
Premature convergence problem, which greatly limits genetic algorithm in network test system popularization and application.
Why genetic algorithm can produce the phenomenon of Premature Convergence, its reason mainly to have：
1st, because such as improper use of selection, intersection and mutation operator of the operator in genetic algorithm can cause Premature Convergence to show
As, and the improper generation for also resulting in Premature Convergence of corresponding selection of control parameter.
2nd, the finiteness of population size.Because Population Size is limited, it is also limited to cause to search for optimal solution space, easy close relative
Breeding so that population pattern quickly tends to a single state, so as to be absorbed in locally optimal solution.
3rd, the loss of pattern is also to cause genetic algorithm Premature convergence in colony.If be not controlled by, population is allowed
Constantly selected at random, intersected and mutation operation, may result in the loss of population effective gene.If effective gene is lost
When losing certain proportion degree, will result in algorithm and do not enter in locally stopping, it is impossible to effectively being evolved.
Therefore, it is necessary to a kind of new automatic volume group method is provided with improve computer system intelligent Auto-generating Test Paper practicality and
Validity.
The content of the invention
The purpose of the present invention is directed to the deficiencies in the prior art, there is provided a kind of automatic volume group method for online exam.
The present invention is achieved by the following technical solutions：
A kind of automatic volume group method for online exam, comprises the following steps：
S1. setting group rolls up constraints；
S2. generating initial population P0, the initial population P0 according to constraints includes individuality, and wherein N is just whole
Number (also known as population scale, general recommendations N is 50, genetic manipulation number of times has been lacked greatly few, it is impossible to reach the effect of genetic evolution)；
S3. each described individual fitness function F (X) value is calculated, solving F (X) value can be converted to solution target letter
Number f (x), solution procedure is as follows：
F (x) is object function, it is necessary to the attribute according to every part of paper per problem calculates evaluation；Attribute per problem includes
Question number, topic type, examination question total score, degree-of-difficulty factor, knowledge point fraction, Reaction time, cognitive classification and discrimination；According to above-mentioned category
Property, determining a matrix of n*8, wherein n is the exercise question number contained by paper, and 8 attributes per problem determine the row elements of matrix s mono-
Value, be described in detail below：
In above matrix, the declaration of will of i-th row element of A is the question number ai1 of the i-th problem, and topic type ai2, examination question is total
Score value ai3, degree-of-difficulty factor ai4, knowledge point fraction ai5, Reaction time ai6, cognitive classification ai7, discrimination ai8, this matrix
The dbjective state matrix being to solve for, the distribution of matrix element corresponds to requirement of the user to each different aspect of paper respectively, when
Matrix error is met group paper for rolling up each requirement when minimum；
In model of organizing test paper, there is restriction relation between each attribute, be the group of multiple constraintss on group volume question essence
Optimization problem is closed, a qualified paper is generated it is necessary to meet the pact of the multiple condition such as topic type, score value, difficulty and knowledge point
Beam；Certain weight is assigned by the constraint of different condition, f (x) as paper global index error is set, for 8 category of concentrated expression
Property the desired error of index and user, then the ATTRIBUTE INDEX of whole paper is exactly that 8 ATTRIBUTE INDEXs are multiplied by their own power
The sum of weight, the solution expression formula of F (X) is as follows：
Wherein Wi is the i-th road of correspondence group volume factor to group a weight for volume significance level, fi(X) be correspondence ai property distribution
Error；
S4. the process for organizing volume is the process that object function optimal solution is sought under constraints；Judge individual fitness
Functional value whether satisfaction group volume condition, i.e., whether meet and the fitness value difference of an individual had in iterations less than setting in advance
The target fitness value put and the difference of actual fitness value, paper are generated if meeting and terminate the automatic volume group method；If
It is unsatisfactory for, continues step S5；
S5. judge whether to reach maximum iteration；If be not reaching to, iterations plus 1, carry out genetic manipulation；
Iterations typically takes 100-200, sets iteration 100 or 200 end；Iterations more matter of fundamental importance evaluation time is more long, heredity
Number of operations is more, also closer to optimal solution；The genetic manipulation includes selection operation, crossover operation and mutation operation；
S6. the genetic manipulation according to S5 generates new colony P1, and return to step S3.
Further, each individual generation method is as follows in the step S2：By the examination question in exam pool according to setting
Topic shape parameter in constraints forms multiple topic type collection merging and the examination question in each topic type set sorts according to real number mode,
The examination question of setting quantity is chosen using random algorithm from each topic type set, and according to the examination question sequence number in identical topic type set
The sequence number of selected examination question is constituted sequence of real numbers by adjacent mode, and the sequence of real numbers is the individuality in colony.
Further, the step S2 includes the constraints according to setting, and generation is multiple individual, and the plurality of individuality is shape
Into initial population.
Specifically, in model of organizing test paper, the constraints between each attribute includes：The constraint of topic type, score value constraint, difficulty point
Cloth constraint, the constraint of topic type fraction, the constraint of knowledge point fraction, discrimination constraint, Reaction time constrain and repeat frequency constraint.
Further, between the attribute constraints specifically sets as follows：
1) topic type constraint
Topic type is provided with multiple-choice question, True-False, gap-filling questions, simple answer, operation questions and analysis topic, is often covered during group volume
The topic type structure of paper is usually fixed, using following expression：
TX=(TX1,TX2,…TXn)
Wherein TXi is every kind of topic type exercise question quantity；
2) score value constraint
Paper total score is expressed as：General acquiescence is 100, it is also possible to be set by the user；When a set of paper
Total score and target score S*When difference is not very big, paper can be receiving, therefore the function of score value assessment is as follows：
f1(X)=abs (S-S*)/S*
3) Distribution of difficulty constraint
ND=∑s ai3ai4/ S, wherein S are total score；
4) topic type fraction constraint
Jth topic topic type fraction, wherein
5) knowledge point fraction constraint
Jth inscribes knowledge point fraction, wherein
6) discrimination constraint
A in formulai3--- the total score of the i-th problem；ai8--- the discrimination of the i-th problem；S total scores --- the full marks of full volume
Value, typically 100 points；
7) Reaction time constraint
8) frequency constraint is repeated
In order to prevent same knowledge point frequency of occurrences in portion paper too high, the phenomenon for causing knowledge point to repeat can
Can not be more than 3 times, i.e., with the frequency that prespecified knowledge point uses in same paper
F=ais≤3。
Further, the condition of satisfaction group volume is to actually obtain meeting situation and obtaining in theory for each index in paper
Paper in each index meet the difference of situation and try one's best and reach minimum, that is, deviation is minimum, thus group volume target
Function can be expressed as the minimum of the formula for seeking F (x), i.e.,：
Min F (X),
Fitness function is generally directly converted by object function, the group volume principle more than, and object function is to get over
It is small better, and fitness function is then the bigger the better, therefore in order to avoid being absorbed in locally optimal solution, fitness function is carried out linearly
Conversion, is converted to following expression：
F (X)=α f (x)+β
F (X) is fitness function value in formula, and f (x) is object function, and α is normal number, and β is constant coefficient；According to above-mentioned
Step completes the evaluation to F (X), and obtains each individual fitness.
Specifically, the selection operation essence is exactly to carry out duplication operation, simply replicates and is conditional on replicating, it then follows
The principle of " survival of the fittest " is replicated, it is therefore an objective to select excellent individuality hereditary；Excellent individuality is based on fitness, to adopt
Object is replicated with roulette mode, is selected according to the size of individual adaptation degree function when choosing individual, the bigger quilt of fitness
The probability chosen is higher.
Further, the crossover operation purpose is desirable to obtain individuality of new generation by crossover operation, and a new generation is individual
The body heredity good characteristic of parent individuality, so as to obtaining more excellent new individual；In practical operation, to group during crossover operation
Each internal individuality intersects at random according to certain probability P c, so as to obtain two new individualities, such as to two individuality P1 and P2
Identical topic type carries out coding bunchiness：
It is 3 according to Pc random numbers, by two low three exchanges of individuality, obtains new individual G1 and G2, it is complete by that analogy
Into the crossover operation of all topic types.
Further, the mutation operation purpose is desirable to, by changing Individual genes, obtain new individuality；Operating procedure
It is：One is randomly choosed first in current group or some are individual, some is changed at random according to certain mutation probability
Or some genic values, so as to obtain new individuality；Question number to selecting some topic types in paper becomes according to coding bunchiness
ETTHER-OR operation：
Aforesaid operations are Pm=4 according to mutation probability, mutation operation are proceeded by the 4th from right to left, by original 1
It is changed into 0, so as to obtain new individuality.
The beneficial effects of the invention are as follows：
(1) present invention completes automatic volume group using genetic algorithm, can quickly obtain a science, accurate paper, solves
Determine examination system big, unstable problem of paper randomness in group licks journey, improve fairness, the reasonability of examination；
(2) selection of genetic algorithm of the present invention, intersection, three operators of variation are constantly changed during practical application
Enter, by adjusting genetic operator, so that constantly improve and optimized algorithm, make it have ability of self-teaching；
(3) present invention is coding specification according to topic type, knowledge point etc., initial population is selected according to classification, is intersected,
The operation of variation obtains new individual, is easy to quickly be met the paper of constraint, while can carry out the operation of heredity again, allows
Satisfactory examination question hands down by heredity.
Brief description of the drawings
In order to illustrate more clearly of technical scheme, below will be to being wanted needed for embodiment or description of the prior art
The accompanying drawing for using is briefly described, it should be apparent that, drawings in the following description are only embodiments of the invention, for this
For the those of ordinary skill of field, on the premise of not paying creative work, other can also be obtained according to these accompanying drawings attached
Figure.
Fig. 1 is a kind of automatic volume group method flow diagram for online exam of the invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the present invention, the technical scheme in the embodiment of the present invention is clearly and completely retouched
State, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.Based on this hair
Embodiment in bright, the every other reality that those of ordinary skill in the art are obtained on the premise of creative work is not made
Example is applied, the scope of protection of the invention is belonged to.
Embodiment 1
Refer to Fig. 1.As illustrated, the invention discloses a kind of automatic volume group method for online exam, including such as
Lower step：
S1. setting group rolls up constraints；
S2. generating initial population P0, the initial population P0 according to constraints includes individuality, and wherein N is just whole
(initial population is 50 also known as population scale, general recommendations N to number, genetic manipulation number of times has been lacked greatly few, it is impossible to reach genetic evolution
Effect), each individual generation method is as follows：The topic shape parameter in constraints by the examination question in exam pool according to setting is formed
The examination question in each topic type set is sorted in multiple topic type collection merging according to real number mode, using random from each topic type set
Algorithm picks set the examination question of quantity, and according to the adjacent mode of the examination question sequence number in identical topic type set by selected examination question
Sequence number constitutes sequence of real numbers, and the sequence of real numbers is the individuality in colony；According to the constraints of setting, generation it is many each and every one
Body, the plurality of individuality forms initial population；
S3. fitness function F (X) value of each individuality (obtaining a paper) is calculated, solving F (X) value can be with
Solution object function f (x) is converted to, solution procedure is as follows：
F (x) is object function, it is necessary to the attribute according to every part of paper per problem calculates evaluation；Attribute per problem includes
Question number, topic type, examination question total score, degree-of-difficulty factor, knowledge point fraction, Reaction time, cognitive classification and discrimination；According to above-mentioned category
Property, determining a matrix of n*8, wherein n is the exercise question number contained by paper, and 8 attributes per problem determine the row elements of matrix s mono-
Value, be described in detail below：
In above matrix, the declaration of will of i-th row element of A is the question number ai1 of the i-th problem, and topic type ai2, examination question is total
Score value ai3, degree-of-difficulty factor ai4, knowledge point fraction ai5, Reaction time ai6, cognitive classification ai7, discrimination ai8, this matrix
The dbjective state matrix being to solve for, the distribution of matrix element corresponds to requirement of the user to each different aspect of paper respectively, when
Matrix error is met group paper for rolling up each requirement when minimum；
In model of organizing test paper, the constraints between each attribute includes：The constraint of topic type, score value are constrained, Distribution of difficulty is constrained,
The constraint of topic type fraction, the constraint of knowledge point fraction, discrimination constraint, Reaction time are constrained and repeat frequency constraint, specifically such as
Under：
1) topic type constraint
Topic type is provided with multiple-choice question, True-False, gap-filling questions, simple answer, operation questions and analysis topic, is often covered during group volume
The topic type structure of paper is usually fixed, using following expression：
TX=(TX1,TX2,…TXn)
Wherein TXi is every kind of topic type exercise question quantity；
2) score value constraint
Paper total score is expressed as：General acquiescence is 100, it is also possible to be set by the user；When a set of paper
Total score and target score S*When difference is not very big, paper can be receiving, therefore the function of score value assessment is as follows：
f1(X)=abs (S-S*)/S*
3) Distribution of difficulty constraint
ND=∑s ai3ai4/ S, wherein S are total score；
4) topic type fraction constraint
Jth topic topic type fraction, wherein
5) knowledge point fraction constraint
Jth inscribes knowledge point fraction, wherein
6) discrimination constraint
A in formulai3--- the total score of the i-th problem；ai8--- the discrimination of the i-th problem；S total scores --- the full marks of full volume
Value, typically 100 points；
7) Reaction time constraint
8) frequency constraint is repeated
In order to prevent same knowledge point frequency of occurrences in portion paper too high, the phenomenon for causing knowledge point to repeat can
Can not be more than 3 times, i.e., with the frequency that prespecified knowledge point uses in same paper
F=ais≤3；
Analyzed more than, be the combinatorial optimization problem of multiple constraintss on group volume question essence, to generate portion
Qualified paper it is necessary to meet topic type, score value, difficulty and knowledge point etc. multiple conditions constraint；The constraint of different condition is assigned
Certain weight is given, f (x) as paper global index error is set, for the mistake that 8 ATTRIBUTE INDEXs of concentrated expression and user require
Difference, then the ATTRIBUTE INDEX of whole paper is exactly the sum that 8 ATTRIBUTE INDEXs are multiplied by their own weight, the solution expression of F (X)
Formula is as follows：
Wherein Wi is that the i-th road of correspondence group rolls up factor to a group weight for volume significance level, and fi (X) is the property distribution of correspondence ai
Error；The condition of satisfaction group volume is that to actually obtain meeting for each index in paper each in situation and the paper for obtaining in theory
The difference that individual index meets situation is tried one's best and reaches minimum, that is, deviation is minimum, therefore the object function of group volume can be represented
To seek the minimum of the formula of F (x), i.e.,：
Min F (X),
Fitness function is generally directly converted by object function, the group volume principle more than, and object function is to get over
It is small better, and fitness function is then the bigger the better, therefore in order to avoid being absorbed in locally optimal solution, fitness function is carried out linearly
Conversion, is converted to following expression：
F (X)=α f (x)+β
F (X) is fitness function value in formula, and f (x) is object function, and α is normal number, and β is constant coefficient；According to above-mentioned
Step completes the evaluation to F (X), and obtains each individual fitness.
S4. the process for organizing volume is the process that object function optimal solution is sought under constraints；Judge individual fitness
Functional value whether satisfaction group volume condition, i.e., whether meet and the fitness value difference of an individual had in iterations less than setting in advance
The target fitness value put and the difference of actual fitness value, paper are generated if meeting and terminate the automatic volume group method；If
It is unsatisfactory for, continues step S5；
S5. judge whether to reach maximum iteration；If be not reaching to, iterations plus 1, carry out genetic manipulation；
Iterations typically takes 100-200, sets iteration 100 or 200 end；Can be according to calculating speed and Examination Papers ' Quality requirement
Setting.Iterations more matter of fundamental importance evaluation time is more long, and genetic manipulation number of times is more, also closer to optimal solution；The genetic manipulation bag
Include selection operation, crossover operation and mutation operation；
(1) selection operation
Selection operation essence is exactly to carry out duplication operation, simply replicates and is conditional on replicating, it then follows " survival of the fittest "
Principle is replicated, it is therefore an objective to select excellent individuality hereditary；Excellent individuality is based on fitness, using roulette mode
Object is replicated, is selected according to the size of individual adaptation degree function when choosing individual, the more big selected probability of fitness is more
It is high.
(2) crossover operation
Crossover operation purpose is desirable to obtain individuality of new generation by crossover operation, and individual inheritance of new generation parent
The good characteristic of body, so as to obtain more excellent new individual；In practical operation, to intragroup each individuality during crossover operation
Intersect at random according to certain probability P c, so as to obtain two new individualities, such as two individuality P1 topic type identical with P2 is compiled
Code bunchiness：
It is 3 according to Pc random numbers, by two low three exchanges of individuality, obtains new individual G1 and G2, it is complete by that analogy
Into the crossover operation of all topic types.
(3) mutation operation
Mutation operation purpose is desirable to, by changing Individual genes, obtain new individuality；Operating procedure is：First current
One being randomly choosed in colony or some being individual, some or some genes is changed at random according to certain mutation probability
Value, so as to obtain new individuality；Question number to selecting some topic types in paper carries out mutation operation according to coding bunchiness：
Aforesaid operations are Pm=4 according to mutation probability, mutation operation are proceeded by the 4th from right to left, by original 1
It is changed into 0, so as to obtain new individuality.
S6. the genetic manipulation according to S5 generates new colony P1, and return to step S3.
The beneficial effects of the invention are as follows：
(1) present invention completes automatic volume group using genetic algorithm, can quickly obtain a science, accurate paper, solves
Determine examination system big, unstable problem of paper randomness in group licks journey, improve fairness, the reasonability of examination；
(2) selection of genetic algorithm of the present invention, intersection, three operators of variation are constantly changed during practical application
Enter, by adjusting genetic operator, so that constantly improve and optimized algorithm, make it have ability of self-teaching；
(3) present invention is coding specification according to topic type, knowledge point etc., initial population is selected according to classification, is intersected,
The operation of variation obtains new individual, is easy to quickly be met the paper of constraint, while can carry out the operation of heredity again, allows
Satisfactory examination question hands down by heredity.
The above is the preferred embodiment of the present invention, it should be noted that for those skilled in the art
For, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications are also considered as
Protection scope of the present invention.
Claims (9)
1. a kind of automatic volume group method for online exam, it is characterised in that comprise the following steps：
S1. setting group rolls up constraints；
S2. generating initial population P0, the initial population P0 according to constraints includes individuality, and wherein N is positive integer；
S3. each described individual fitness function F (X) value is calculated, solving F (X) value can be converted to solution object function f
X (), solution procedure is as follows：
F (x) is object function, it is necessary to the attribute according to every part of paper per problem calculates evaluation；Per problem attribute include question number,
Topic type, examination question total score, degree-of-difficulty factor, knowledge point fraction, Reaction time, cognitive classification and discrimination；According to above-mentioned attribute, really
A fixed matrix of n*8, wherein n is the exercise question number contained by paper, and 8 attributes per problem determine the value of the row elements of matrix s mono-,
It is described in detail below：
In above-mentioned matrix, the declaration of will of i-th row element of A is the question number a of the i-th problemi1, topic type ai2, examination question total score
ai3, degree-of-difficulty factor ai4, knowledge point fraction ai5, Reaction time ai6, cognitive classification ai7, discrimination ai8, what this matrix was also to solve for
Dbjective state matrix, the distribution of matrix element corresponds to requirement of the user to each different aspect of paper respectively, when matrix error
Group paper for rolling up each requirement is met when minimum；
In model of organizing test paper, there is restriction relation between each attribute, be that the combination of multiple constraintss is excellent on group volume question essence
Change problem, will generate a qualified paper it is necessary to meet the constraint of the multiple condition such as topic type, score value, difficulty and knowledge point；Will
The constraint of different condition assigns certain weight, f (x) as paper global index error is set, for 8 ATTRIBUTE INDEXs of concentrated expression
The error required with user, then the ATTRIBUTE INDEX of whole paper is exactly the sum that 8 ATTRIBUTE INDEXs are multiplied by their own weight,
The solution expression formula of F (X) is as follows：
Wherein WiIt is that the i-th road of correspondence group rolls up factor to group a weight for volume significance level, fi(X) it is correspondence aiProperty distribution error；
S4. the process for organizing volume is the process that object function optimal solution is sought under constraints；Judge individual fitness function
Value whether satisfaction group volume condition, i.e., whether meet and the fitness value difference of an individual is had in iterations be less than and set in advance
Target fitness value and the difference of actual fitness value, paper are generated if meeting and terminate the automatic volume group method；If discontented
Sufficient then continuation step S5；
S5. judge whether to reach maximum iteration；If be not reaching to, iterations plus 1, carry out genetic manipulation；Iteration
Number of times typically takes 100-200, sets iteration 100 or 200 end；Iterations more matter of fundamental importance evaluation time is more long, genetic manipulation
Number of times is more, also closer to optimal solution；The genetic manipulation includes selection operation, crossover operation and mutation operation；
S6. the genetic manipulation according to S5 generates new colony P1, and return to step S3.
2. a kind of automatic volume group method for online exam according to claim 1, it is characterised in that the step S2
In each individual generation method it is as follows：The topic shape parameter in constraints by the examination question in exam pool according to setting forms multiple
Topic type set, and the examination question in each topic type set is sorted according to real number mode, using random calculation from each topic type set
Method chooses the examination question of setting quantity, and according to the adjacent mode of the examination question sequence number in identical topic type set by the sequence of selected examination question
Number composition sequence of real numbers, the sequence of real numbers is the individuality in colony.
3. a kind of automatic volume group method for online exam according to claim 2, it is characterised in that the step S2
Including the constraints according to setting, generation is multiple individual, and the plurality of individuality forms initial population.
4. a kind of automatic volume group method for online exam according to claim 3, it is characterised in that model of organizing test paper
In, the constraints between each attribute includes：The constraint of topic type, score value constraint, Distribution of difficulty constraint, topic type fraction are constrained, known
Know the constraint of point fraction, discrimination constraint, Reaction time constraint and repeat frequency constraint.
5. a kind of automatic volume group method for online exam according to claim 4, it is characterised in that the attribute it
Between constraints specifically set it is as follows：
1) topic type constraint
Topic type is provided with multiple-choice question, True-False, gap-filling questions, simple answer, operation questions and analysis topic, and paper is often covered during group volume
Topic type structure be usually fixed, using following expression：
TX=(TX1,TX2,…TXn)
Wherein TXi is every kind of topic type exercise question quantity；
2) score value constraint
Paper total score is expressed as：General acquiescence is 100, it is also possible to be set by the user；The function of score value assessment is such as
Under：
f1(X)=abs (S-S*)/S*
3) Distribution of difficulty constraint
ND=∑s ai3ai4/ S, wherein S are total score；
4) topic type fraction constraint
Jth topic topic type fraction, wherein
5) knowledge point fraction constraint
Jth inscribes knowledge point fraction, wherein
6) discrimination constraint
A in formulai3--- the total score of the i-th problem；ai8--- the discrimination of the i-th problem；S total scores --- the full marks value of full volume, leads to
It is often 100 points；
7) Reaction time constraint
8) frequency constraint is repeated
In order to prevent same knowledge point frequency of occurrences in portion paper too high, the phenomenon for causing knowledge point to repeat is advised in advance
Determining frequency that knowledge point uses in same paper can not be more than 3 time, i.e.,
F=ais≤3。
6. a kind of automatic volume group method for online exam according to claim 5, it is characterised in that satisfaction group volume
Condition is that the situation that meets for actually obtaining each index in paper meets situation with each index in the paper for obtaining in theory
Difference try one's best and reach minimum, that is, deviation is minimum, therefore the object function of group volume can be expressed as asking the formula of F (x)
Minimum, i.e.,：
Min F (X),
Fitness function is generally directly converted by object function, the group volume principle more than, and object function is smaller getting over
It is good, and fitness function is then the bigger the better, therefore in order to avoid being absorbed in locally optimal solution, fitness function is linearly become
Change, be converted to following expression：
F (X)=α f (x)+β
F (X) is fitness function value in formula, and f (x) is object function, and α is normal number, and β is constant coefficient；According to above-mentioned steps
The evaluation to F (X) is completed, and obtains each individual fitness.
7. a kind of automatic volume group method for online exam according to claim 1 or 6, it is characterised in that the choosing
It is exactly to carry out duplication operation to select operation essence, it then follows the principle of " survival of the fittest " is replicated, it is therefore an objective to select excellent individuality to lose
Pass；Excellent individuality is based on fitness, object to be replicated using roulette mode, when choosing individual according to ideal adaptation
Spend the size of function to select, the more big selected probability of fitness is higher.
8. a kind of automatic volume group method for online exam according to claim 1 or 6, it is characterised in that the friendship
The purpose for pitching operation is to obtain individuality of new generation by crossover operation, and the individual inheritance of new generation excellent spy of parent individuality
Property, so as to obtain more excellent new individual；In practical operation, to intragroup each individuality according to certain general during crossover operation
Rate Pc intersects at random, so as to obtain two new individualities, such as carries out coding bunchiness to two individuality P1 topic type identical with P2：
It is 3 according to Pc random numbers, by two low three exchanges of individuality, obtains new individual G1 and G2, institute is completed by that analogy
There is the crossover operation of topic type.
9. a kind of automatic volume group method for online exam according to claim 1 or 6, it is characterised in that the change
The operating procedure of ETTHER-OR operation is as follows：One is randomly choosed first in current group or some are individual, according to certain variation
Probability changes some or some genic values at random, so as to obtain new individuality；Question number to selecting some topic types in paper
Mutation operation is carried out according to coding bunchiness：
Aforesaid operations are Pm=4 according to mutation probability, and mutation operation is proceeded by the 4th from right to left, and original 1 is changed into
0, so as to obtain new individuality.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611184242.6A CN106803123A (en) | 2016-12-20 | 2016-12-20 | A kind of automatic volume group method for online exam |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611184242.6A CN106803123A (en) | 2016-12-20 | 2016-12-20 | A kind of automatic volume group method for online exam |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106803123A true CN106803123A (en) | 2017-06-06 |
Family
ID=58983959
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611184242.6A Pending CN106803123A (en) | 2016-12-20 | 2016-12-20 | A kind of automatic volume group method for online exam |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106803123A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107194743A (en) * | 2017-06-19 | 2017-09-22 | 腾讯科技（深圳）有限公司 | A kind of network surveying questionnaire generation method and device |
| CN108898170A (en) * | 2018-06-19 | 2018-11-27 | 江苏中盈高科智能信息股份有限公司 | A kind of intelligent Auto-generating Test Paper method based on fuzzy cluster analysis |
| CN108920663A (en) * | 2018-07-04 | 2018-11-30 | 北京希子教育科技有限公司 | A kind of method of automatic recommendation item content |
| CN108932546A (en) * | 2018-05-28 | 2018-12-04 | 青岛高校信息产业股份有限公司 | A kind of group of volume method and system |
| CN109118402A (en) * | 2018-08-31 | 2019-01-01 | 中国人民解放军军事科学院评估论证研究中心 | A kind of army's cadre's on-line study examination assessment system and method |
| CN109471940A (en) * | 2018-11-02 | 2019-03-15 | 福建书香伟业教育科技有限公司 | A kind of generation method and computer readable storage medium of PK match exam pool |
| CN109684366A (en) * | 2018-12-20 | 2019-04-26 | 国家计算机网络与信息安全管理中心 | A kind of knowledge base group volume method for industrial control system risk assessment |
| CN110097241A (en) * | 2018-01-30 | 2019-08-06 | 北大方正集团有限公司 | On-line testing learning method, system, computer equipment and storage medium |
| CN110147953A (en) * | 2019-05-16 | 2019-08-20 | 电子科技大学 | A kind of questionnaire automatic generation method |
| CN111242267A (en) * | 2019-12-31 | 2020-06-05 | 国网北京市电力公司 | Examination paper composing method and examination paper composing device for on-line examination |
| CN111667070A (en) * | 2020-05-06 | 2020-09-15 | 华东师范大学 | Intelligent volume combination method based on genetic algorithm |
| CN111737448A (en) * | 2020-07-21 | 2020-10-02 | 江西理工大学南昌校区 | Question selection method and system based on basic subject short answer of answer duration |
| CN114780679A (en) * | 2022-04-08 | 2022-07-22 | 北京华航唯实机器人科技股份有限公司 | Method and device for generating test paper |
-
2016
- 2016-12-20 CN CN201611184242.6A patent/CN106803123A/en active Pending
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107194743A (en) * | 2017-06-19 | 2017-09-22 | 腾讯科技（深圳）有限公司 | A kind of network surveying questionnaire generation method and device |
| CN107194743B (en) * | 2017-06-19 | 2019-04-09 | 腾讯科技（深圳）有限公司 | A kind of network surveying questionnaire generation method and device |
| CN110097241A (en) * | 2018-01-30 | 2019-08-06 | 北大方正集团有限公司 | On-line testing learning method, system, computer equipment and storage medium |
| CN108932546A (en) * | 2018-05-28 | 2018-12-04 | 青岛高校信息产业股份有限公司 | A kind of group of volume method and system |
| CN108898170A (en) * | 2018-06-19 | 2018-11-27 | 江苏中盈高科智能信息股份有限公司 | A kind of intelligent Auto-generating Test Paper method based on fuzzy cluster analysis |
| CN108898170B (en) * | 2018-06-19 | 2022-02-01 | 江苏中盈高科智能信息股份有限公司 | Intelligent volume-forming method based on fuzzy clustering analysis |
| CN108920663A (en) * | 2018-07-04 | 2018-11-30 | 北京希子教育科技有限公司 | A kind of method of automatic recommendation item content |
| CN109118402A (en) * | 2018-08-31 | 2019-01-01 | 中国人民解放军军事科学院评估论证研究中心 | A kind of army's cadre's on-line study examination assessment system and method |
| CN109118402B (en) * | 2018-08-31 | 2021-12-07 | 中国人民解放军军事科学院评估论证研究中心 | Army cadre online learning, examination and evaluation system and method |
| CN109471940A (en) * | 2018-11-02 | 2019-03-15 | 福建书香伟业教育科技有限公司 | A kind of generation method and computer readable storage medium of PK match exam pool |
| CN109684366A (en) * | 2018-12-20 | 2019-04-26 | 国家计算机网络与信息安全管理中心 | A kind of knowledge base group volume method for industrial control system risk assessment |
| CN110147953A (en) * | 2019-05-16 | 2019-08-20 | 电子科技大学 | A kind of questionnaire automatic generation method |
| CN110147953B (en) * | 2019-05-16 | 2023-01-10 | 电子科技大学 | Automatic questionnaire generation method |
| CN111242267A (en) * | 2019-12-31 | 2020-06-05 | 国网北京市电力公司 | Examination paper composing method and examination paper composing device for on-line examination |
| CN111667070A (en) * | 2020-05-06 | 2020-09-15 | 华东师范大学 | Intelligent volume combination method based on genetic algorithm |
| CN111737448A (en) * | 2020-07-21 | 2020-10-02 | 江西理工大学南昌校区 | Question selection method and system based on basic subject short answer of answer duration |
| CN111737448B (en) * | 2020-07-21 | 2020-11-17 | 江西理工大学南昌校区 | Question selection method and system based on basic subject short answer of answer duration |
| CN114780679A (en) * | 2022-04-08 | 2022-07-22 | 北京华航唯实机器人科技股份有限公司 | Method and device for generating test paper |
| CN114780679B (en) * | 2022-04-08 | 2024-05-14 | 北京华航唯实机器人科技股份有限公司 | Method and device for generating test paper |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106803123A (en) | A kind of automatic volume group method for online exam | |
| Taconis et al. | Teaching science problem solving: An overview of experimental work | |
| Buckley et al. | Model-based teaching and learning with BioLogica™: What do they learn? How do they learn? How do we know? | |
| Lewis et al. | Finding feasible timetables using group-based operators | |
| Garg et al. | Symbolic network: generalized neural policies for relational MDPs | |
| Xia | Big data based research on the management system framework of ideological and political education in colleges and universities | |
| Wang et al. | Automatic generation of raven’s progressive matrices | |
| CN109858797A (en) | The various dimensions information analysis of the students method of knowledge based network exact on-line education system | |
| Krouska et al. | An enhanced genetic algorithm for heterogeneous group formation based on multi-characteristics in social-networking-based learning | |
| Hwang et al. | A tabu search approach to generating test sheets for multiple assessment criteria | |
| Batzli et al. | Beyond Punnett squares: student word association and explanations of phenotypic variation through an integrative quantitative genetics unit investigating anthocyanin inheritance and expression in Brassica rapa Fast Plants | |
| Kragelund | Solving a timetabling problem using hybrid genetic algorithms | |
| Wang et al. | A greedy and genetic fusion algorithm for solving course timetabling problem | |
| Ahmad et al. | A novel adaptive learning path method | |
| Waller | The synthesis of hierarchical structures: Technique and applications | |
| Verschoor | Genetic algorithms for automated test assembly | |
| Nelson | Evolutionary modelling of economic change | |
| Móczár | Arrow-Debreu Model versus kornai-critique | |
| Peng et al. | Construction and Application of Farmers’ Practical Teaching System in Vocational Education Based on Big Data Mining Technology | |
| Jiang et al. | Intelligent Optimization of curriculum system based on Washington accord graduate attribute | |
| Kazaz | Application of an expert system on the fracture mechanics of concrete | |
| Chen et al. | Research on Cognitive Diagnostic Model Based on BP Neural Network | |
| Zhu | A Self-learning Ability Assessment Method Based on Weight-Optimised Dfferential Evolutionary Algorithm | |
| Mendoza et al. | Decision theory | |
| Neville et al. | Developing a generic genetic algorithm |
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: 20170606 |