CN105336235A - Score setting method used for intelligent learning system - Google Patents
Score setting method used for intelligent learning system Download PDFInfo
- Publication number
- CN105336235A CN105336235A CN201510796050.XA CN201510796050A CN105336235A CN 105336235 A CN105336235 A CN 105336235A CN 201510796050 A CN201510796050 A CN 201510796050A CN 105336235 A CN105336235 A CN 105336235A
- Authority
- CN
- China
- Prior art keywords
- setting method
- learning system
- student
- intelligent learning
- function
- 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
Classifications
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09B—EDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
- G09B7/00—Electrically-operated teaching apparatus or devices working with questions and answers
- G09B7/02—Electrically-operated teaching apparatus or devices working with questions and answers of the type wherein the student is expected to construct an answer to the question which is presented or wherein the machine gives an answer to the question presented by a student
Abstract
The invention discloses a score setting method used for an intelligent learning system. The score setting method comprises the following steps: establishing evaluation on a total grasping capability according to scores of students, and dividing evaluation results into excellent, good, moderate, passing and failed grades, wherein corresponding hundred-mark system ranges are excellent [100, 90], good [89, 80], moderate [79, 70], passing [69, 60] and failed [59, 0]; carrying out classified representation on learning results of the students; establishing two functions G(x,p) and S(y,q) which reflect difficulty levels of examination questions; estimating variables x and y of each examination question by using data digging; carrying out difficulty level estimation on the examination questions by using established function models; calculating comprehensive indexes of the difficulty levels of the examination questions; and sharing scores P according to the comprehensive indexes. According to the score setting method used for the intelligent learning system, different evaluation criteria can be automatically set according to the difficulty levels.
Description
Technical field
The invention belongs to learning system technical field, particularly relate to a kind of point value setting method for intelligent learning system.
Background technology
In an intelligent learning system, student and computing machine carry out alternately, select some knowledge points of required study, and test at different time points, finally by the study testing into regulation knowledge point.The method of usual evaluate student learning outcome (the institute's learning knowledge point namely whether grasped) is the score value adopting final test.
The existing exercise question different for complexity, arranges the necessary condition that different score values is accurate evaluation.Existing method is all manual setting, more arbitrarily; Do not consider the mean scores establishing method of exercise question complexity, in evaluate student learning outcome, there is very large instability, low precision.
Summary of the invention
The object of the present invention is to provide a kind of point value setting method for intelligent learning system, be intended to solve the existing mean scores establishing method not considering exercise question complexity, in evaluate student learning outcome, there is very large instability, the problem of low precision.
The present invention is achieved in that a kind of point value setting method for intelligent learning system, and a described point value setting method for intelligent learning system comprises:
According to the achievement of student, set up the evaluation of an overall grasp ability, evaluation result is divided into excellent, good, in, pass and fail, corresponding centesimal system scope is [100,90] excellent [89,80] good [79,70] in [69,60] qualifying [59,0] is failed, and the classification carrying out students ' learning performance represents;
Set up two function G (x, p) and the S (y, q) of reflection exercise question complexity;
Utilize data mining, each problem object variable x and y is estimated;
Utilize the function model set up, carry out the estimation of exercise question complexity respectively;
Calculate the overall target of exercise question complexity;
Score value P is shared according to overall target D.
Further, normal distyribution function is adopted to carry out modeling to G (x, p) and S (y, q):
The normal distyribution function of function G (x, p) is:
The normal distyribution function of function S (y, q) is:
Further, describedly utilize data mining, each problem object variable x and y estimated, from the angle of data analysis, carries out the estimation of each problem object variable x:
X=Sum (student i answer situation)/total number of students;
From the angle of data analysis, carry out the estimation of each problem object variable y:
Y=(total number of students-Sum (student i answer situation))/total number of students;
Further, the described function model utilizing foundation, carries out exercise question complexity respectively and estimates to comprise:
Normal distyribution function according to function G (x, p) is:
The normal distyribution function of function S (y, q) is:
Further, the overall target of described calculating exercise question complexity, the overall target D of the calculating exercise question complexity of employing is:
D=q+p。
Further, describedly score value P is shared according to overall target D.
Further, described point C and the CA data of value setting method associated by a knowledge point for intelligent learning system, provide the computing method of parameter PG and PS, specifically carry out according to following steps:
Step one, the probability P G hit it: even if namely student does not know the probability that the knowledge point that exercise question relates to still correctly is answered;
Step 2, the probability of making a mistake (PS): namely student knows this knowledge point, but the probability of still careless erroneous answers;
Step 3, finally can arrange rational score value according to PG and PS, and score value is set to basic with reference to being: PS/PG, under the framework of total score, carries out sharing the score value obtaining each exercise question according to the PS/PG value of all exercise questions.
Further, the calculating of the probability of hitting it described in comprises:
The average of compute associations student answer matrix, namely each row is cumulative gets average, obtains a mean vector K; Mean vector K={ passes through the average answer of student at examination question i of final test | i=1 ..., m};
K is utilized to calculate the probability P G hit it:
Further, the calculating of the probability P S made a mistake described in comprises:
The average of compute associations student answer matrix, namely each row is cumulative gets average, obtains a mean vector K;
K is utilized to calculate the probability P S made a mistake:
Point value setting method for intelligent learning system provided by the invention, can set different evaluation criterions automatically according to complexity.
Accompanying drawing explanation
Fig. 1 is point value setting method process flow diagram for intelligent learning system that the embodiment of the present invention provides.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with embodiment, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.
Below in conjunction with accompanying drawing, application principle of the present invention is explained in detail.
As shown in Figure 1, point value setting method for intelligent learning system of the embodiment of the present invention comprises the following steps:
S101: according to the achievement of student, set up the evaluation of an overall grasp ability, evaluation result can be divided into excellent, good, in, pass and fail, corresponding centesimal system scope is [100,90] excellent [89,80] good [79,70] in [69,60] qualifying [59,0] is failed, and the classification carrying out students ' learning performance represents;
S102: two function G (x, p) and the S (y, q) that set up reflection exercise question complexity;
S103: utilize data mining, estimates each problem object variable x and y;
S104: utilize the function model set up, carry out the estimation of exercise question complexity respectively;
S105: the overall target calculating exercise question complexity;
S106: share score value P according to overall target D.
Below in conjunction with specific embodiment, application principle of the present invention is further described.
Embodiment 1: to judge to be entitled as example:
1. the model answer of the model answer vector C={ examination question i of test item bank | i=1 ..., m}, judges to be entitled as routine C={10111} with 5 roads.
2. by the answer of the examination question i of student's answer Matrix C A={ student j of final test | j=1 ..., h; I=1 ..., m}.
Such as CA={
11111
10011
00111
11111
}
Although be the score value determining a problem, different exercise question can relate to identical knowledge point, so these exercise questions score value roughly should be identical.So need according to knowledge point, examination question is divided into groups.We arrange the score value of one exercise question and can utilize other topic destination datas of same knowledge point like this.
A kind of intelligence for intelligent learning system of the embodiment of the present invention divides value setting method, based on data mining, according to C and the CA data associated by a knowledge point, provides the computing method of parameter PG and PS, specifically carries out according to following steps:
Step one, the probability of hitting it (PG): even if namely student does not know the probability that the knowledge point that exercise question relates to still correctly is answered;
The calculating of the probability of hitting it is divided into two steps:
1. the average of compute associations student answer matrix, namely each row is cumulative gets average, obtains a mean vector K;
Mean vector K={ passes through the average answer of student at examination question i of final test | i=1 ..., m};
For example above, K=3/4,1/2,3/4,1,1.
2. utilize K to calculate the probability P G hit it:
Step 2, the probability of making a mistake (PS): namely student knows this knowledge point, but the probability of still careless erroneous answers.
The calculating of the probability P S made a mistake is divided into two steps:
1. the average of compute associations student answer matrix, namely each row is cumulative gets average, obtains a mean vector K;
2. utilize K to calculate the probability P S made a mistake:
Finally rational score value can be set according to PG and PS:
Score value is set to basic with reference to being: PS/PG
Step 3, finally under the framework of total score, carries out sharing the score value that can obtain each exercise question according to the PS/PG value of all exercise questions.
Embodiment 2:
Be entitled as example so that 10 roads are objective, following table provides the answer situation of 8 students, and 1 expression is answered questions, and mistake is answered in 0 expression.
Table 1
The first step, the classification carrying out students ' learning performance represents.
According to the achievement of student, set up the evaluation of an overall grasp ability, evaluation result can be divided into excellent, good, in, pass and fail, corresponding centesimal system scope is [100,90] excellent [89,80] good [79,70] in, [69,60] qualifying [59,0] fails.
In table 1, student 1234 is the students that fail, and can think that these students do not grasp learned knowledge point very well.
In table 1, student 5678 is Ontario Scholars, can think that these students have grasped the knowledge point learned preferably.
Second step, sets up two function G (x, p) and the S (y, q) of reflection exercise question complexity.
X portrays " but student does not grasp learned knowledge point very well can the probability of correct answer yet " in G (x, p).
Obviously when this topic of the larger expression of x value is simpler.Usually have two kinds of situations: 1) this topic and required Knowledge Relation degree little; 2) although this topic examination be learned knowledge point, have certain guiding answering information in exercise question.So x value is larger, reflection exercise question complexity p should be less, and in S (y, q), y portrays " but student has better grasped the knowledge point learned also made a mistake and answered the probability of wrong topic ".Obviously when this topic of the larger expression of y value is more difficult.Larger this topic of y value ordinary representation carries out integrated application to it on the basis requiring student to understand knowledge point, so y value is larger, reflection exercise question complexity q should be larger.
The present invention adopts normal distyribution function to carry out modeling to G and S.
If stochastic variable X obeys, a location parameter is μ, scale parameter is the probability distribution of σ, and its probability density function is:
Then this stochastic variable is just called normal random variable, and the distribution that normal random variable is obeyed just is called normal distribution, is denoted as X ~ N (μ, σ
2).
Under normal curve, transverse axis interval (μ-σ, μ+σ) in area be 68.268949%, transverse axis interval (μ-1.96 σ, μ+1.96 σ) in area be 95.449974%, area in transverse axis interval (μ-2.58 σ, μ+2.58 σ) is 99.730020%.
Because the span of G and S is [0,1], so when adopting normal distyribution function to carry out modeling to G and S, the area in transverse axis interval (0,1) will be far longer than the area in (1 ,+∞).So σ=1/2.58 is got in unification.Be 0 and 1. can to obtain thus according to the value of G and S and score value incidence relation parameters μ further,
The normal distyribution function of function G (x, p) is:
The normal distyribution function of function S (y, q) is:
3rd step, utilizes data mining, estimates each problem object variable x and y.
From the angle of data analysis, from the data of student 1234, the estimation of each problem object variable x can be carried out.
X=Sum (student i answer situation)/total number of students
From the angle of data analysis, from the data of student 5678, the estimation of each problem object variable y can be carried out.
Y=(total number of students-Sum (student i answer situation))/total number of students
4th step, utilizes the function model that second step is set up, and carries out the estimation of exercise question complexity respectively.
Normal distyribution function according to function G (x, p) is:
The normal distyribution function of function S (y, q) is:
5th step calculates the overall target of exercise question complexity
According to the practical significance of two function G (x, p) and S (y, q), exercise question is simpler, and x value is larger, and p is less.Exercise question is more difficult, and y value is larger, and q is larger.Accordingly, the overall target D of the calculating exercise question complexity of the present invention's employing is:
D=q+p;
6th step: share score value P according to overall target D.
According to the score value again upgraded, the achievement of each student recalculates as follows:
Effect of the present invention can be learnt, can evaluate more accurately the study condition of student,
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.
Claims (9)
1. for a point value setting method for intelligent learning system, it is characterized in that, a described point value setting method for intelligent learning system comprises:
According to the achievement of student, set up the evaluation of an overall grasp ability, evaluation result is divided into excellent, good, in, pass and fail, corresponding centesimal system scope is [100,90] excellent [89,80] good [79,70] in [69,60] qualifying [59,0] is failed, and the classification carrying out students ' learning performance represents;
Set up two function G (x, p) and the S (y, q) of reflection exercise question complexity;
Utilize data mining, each problem object variable x and y is estimated;
Utilize the function model set up, carry out the estimation of exercise question complexity respectively;
Calculate the overall target of exercise question complexity;
Score value P is shared according to overall target D.
2. as claimed in claim 1 for point value setting method of intelligent learning system, it is characterized in that, adopt normal distyribution function to carry out modeling to G (x, p) and S (y, q):
The normal distyribution function of function G (x, p) is:
The normal distyribution function of function S (y, q) is:
3. as claimed in claim 1 for point value setting method of intelligent learning system, it is characterized in that, describedly utilize data mining, each problem object variable x and y is estimated, from the angle of data analysis, carries out the estimation of each problem object variable x:
X=Sum (student i answer situation)/total number of students;
From the angle of data analysis, carry out the estimation of each problem object variable y:
Y=(total number of students-Sum (student i answer situation))/total number of students.
4. as claimed in claim 1 for point value setting method of intelligent learning system, it is characterized in that, the described function model utilizing foundation, carry out exercise question complexity respectively and estimate to comprise:
Normal distyribution function according to function G (x, p) is:
The normal distyribution function of function S (y, q) is:
5., as claimed in claim 1 for point value setting method of intelligent learning system, it is characterized in that, the overall target of described calculating exercise question complexity, the overall target D of the calculating exercise question complexity of employing is:
D=q+p。
6. as claimed in claim 1 for point value setting method of intelligent learning system, it is characterized in that, describedly share score value P according to overall target D.
7. as claimed in claim 1 for point value setting method of intelligent learning system, it is characterized in that, described point C and the CA data of value setting method associated by a knowledge point for intelligent learning system, provide the computing method of parameter PG and PS, specifically carry out according to following steps:
Step one, the probability P G hit it: even if namely student does not know the probability that the knowledge point that exercise question relates to still correctly is answered;
Step 2, the probability of making a mistake (PS): namely student knows this knowledge point, but the probability of still careless erroneous answers;
Step 3, finally can arrange rational score value according to PG and PS, and score value is set to basic with reference to being: PS/PG, under the framework of total score, carries out sharing the score value obtaining each exercise question according to the PS/PG value of all exercise questions.
8., as claimed in claim 7 for point value setting method of intelligent learning system, it is characterized in that, described in the calculating of probability of hitting it comprise:
The average of compute associations student answer matrix, namely each row is cumulative gets average, obtains a mean vector K; Mean vector K={ passes through the average answer of student at examination question i of final test | i=1 ..., m};
K is utilized to calculate the probability P G hit it:
9., as claimed in claim 7 for point value setting method of intelligent learning system, it is characterized in that, described in the calculating of probability P S of making a mistake comprise:
The average of compute associations student answer matrix, namely each row is cumulative gets average, obtains a mean vector K;
K is utilized to calculate the probability P S made a mistake:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510796050.XA CN105336235A (en) | 2015-11-18 | 2015-11-18 | Score setting method used for intelligent learning system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510796050.XA CN105336235A (en) | 2015-11-18 | 2015-11-18 | Score setting method used for intelligent learning system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105336235A true CN105336235A (en) | 2016-02-17 |
Family
ID=55286729
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510796050.XA Pending CN105336235A (en) | 2015-11-18 | 2015-11-18 | Score setting method used for intelligent learning system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105336235A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106780222A (en) * | 2017-01-12 | 2017-05-31 | 陈星� | Analog synthesis achievement preparation method and device |
CN107169903A (en) * | 2017-07-25 | 2017-09-15 | 山东工商学院 | Learning behavior evaluation method and system based on college teaching big data |
CN109492896A (en) * | 2018-11-02 | 2019-03-19 | 福建书香伟业教育科技有限公司 | The method and computer equipment of selection competition student when a kind of interschool match |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5586022A (en) * | 1990-02-14 | 1996-12-17 | Hitachi, Ltd. | Method of evaluating easiness of works and processings performed on articles and evaluation apparatus |
CN1474339A (en) * | 2002-08-09 | 2004-02-11 | 明日工作室股份有限公司 | Dynamic grading type test system and method |
KR20060097101A (en) * | 2006-08-24 | 2006-09-13 | 김현채 | Relitive allotting method |
CN101000600A (en) * | 2006-12-30 | 2007-07-18 | 南京凌越教育科技服务有限公司 | Study management system and method |
CN102693654A (en) * | 2012-05-18 | 2012-09-26 | 苏州慧飞信息科技有限公司 | Examination paper generating system |
CN102831558A (en) * | 2012-07-20 | 2012-12-19 | 桂林电子科技大学 | System and method for automatically scoring college English compositions independent of manual pre-scoring |
CN103955874A (en) * | 2014-03-31 | 2014-07-30 | 西南林业大学 | Automatic subjective-question scoring system and method based on semantic similarity interval |
CN103971555A (en) * | 2013-01-29 | 2014-08-06 | 北京竞业达数码科技有限公司 | Multi-level automated assessing and training integrated service method and system |
-
2015
- 2015-11-18 CN CN201510796050.XA patent/CN105336235A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5586022A (en) * | 1990-02-14 | 1996-12-17 | Hitachi, Ltd. | Method of evaluating easiness of works and processings performed on articles and evaluation apparatus |
CN1474339A (en) * | 2002-08-09 | 2004-02-11 | 明日工作室股份有限公司 | Dynamic grading type test system and method |
KR20060097101A (en) * | 2006-08-24 | 2006-09-13 | 김현채 | Relitive allotting method |
CN101000600A (en) * | 2006-12-30 | 2007-07-18 | 南京凌越教育科技服务有限公司 | Study management system and method |
CN102693654A (en) * | 2012-05-18 | 2012-09-26 | 苏州慧飞信息科技有限公司 | Examination paper generating system |
CN102831558A (en) * | 2012-07-20 | 2012-12-19 | 桂林电子科技大学 | System and method for automatically scoring college English compositions independent of manual pre-scoring |
CN103971555A (en) * | 2013-01-29 | 2014-08-06 | 北京竞业达数码科技有限公司 | Multi-level automated assessing and training integrated service method and system |
CN103955874A (en) * | 2014-03-31 | 2014-07-30 | 西南林业大学 | Automatic subjective-question scoring system and method based on semantic similarity interval |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106780222A (en) * | 2017-01-12 | 2017-05-31 | 陈星� | Analog synthesis achievement preparation method and device |
CN106780222B (en) * | 2017-01-12 | 2020-03-27 | 陈星� | Method and device for acquiring simulated comprehensive scores |
CN107169903A (en) * | 2017-07-25 | 2017-09-15 | 山东工商学院 | Learning behavior evaluation method and system based on college teaching big data |
CN109492896A (en) * | 2018-11-02 | 2019-03-19 | 福建书香伟业教育科技有限公司 | The method and computer equipment of selection competition student when a kind of interschool match |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
van De Sande | Properties Of The Bayesian Knowledge Tracing Model. | |
CN108694501A (en) | A kind of individualized learning effect analysis system and method towards xAPI | |
CN106875769A (en) | A kind of mathematics practice question-setting system | |
CN107544973A (en) | A kind of method and apparatus that data are handled | |
CN105976282A (en) | Test question difficulty quantification method and system | |
CN106095812A (en) | Intelligent test paper generation method based on similarity measurement | |
Srisawasdi et al. | Technology-enhanced learning in science, technology, and mathematics education: results on supporting student learning | |
CN105975799A (en) | Method and system for calculating carbon emissions | |
CN109272160A (en) | Score on Prediction system and prediction technique | |
CN112905784A (en) | Personalized test question recommendation method based on student portrait | |
CN105336235A (en) | Score setting method used for intelligent learning system | |
CN111626372A (en) | Online teaching supervision management method and system | |
Pardos et al. | The Real World Significance of Performance Prediction. | |
CN105427694A (en) | Computer-aided testing method for intelligent teaching | |
Ahmad et al. | An improved course assessment measurement for analyzing learning outcomes performance using Rasch model | |
CN106296502A (en) | Evaluation of teacher's method and system | |
Steger et al. | Teaching battery basics in laboratories: Comparing learning outcomes of hands-on experiments and computer-based simulations | |
CN114708123A (en) | Method and system for measuring mastery degree of knowledge point | |
Al-Sarem et al. | Mining concepts' relationship based on numeric grades | |
Stacey et al. | A case of the inapplicability of the Rasch model: Mapping conceptual learning | |
CN111815489B (en) | Big data-based algorithm analysis design teaching method and device | |
CN108921434A (en) | A method of user capability prediction is completed by human-computer interaction | |
Monterola et al. | Characterizing the effect of seating arrangement on classroom learning using neural networks | |
Chen et al. | Research on Cognitive Diagnostic Model Based on BP Neural Network | |
Patikorn et al. | Using a Single Model Trained across Multiple Experiments to Improve the Detection of Treatment Effects. |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160217 |