CN114943628A - Test paper difficulty coefficient evaluation method, device and equipment based on artificial intelligence - Google Patents
Test paper difficulty coefficient evaluation method, device and equipment based on artificial intelligence Download PDFInfo
- Publication number
- CN114943628A CN114943628A CN202210636889.7A CN202210636889A CN114943628A CN 114943628 A CN114943628 A CN 114943628A CN 202210636889 A CN202210636889 A CN 202210636889A CN 114943628 A CN114943628 A CN 114943628A
- Authority
- CN
- China
- Prior art keywords
- acquiring
- score
- knowledge point
- test paper
- student
- 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
- 238000012360 testing method Methods 0.000 title claims abstract description 149
- 238000011156 evaluation Methods 0.000 title claims abstract description 31
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 22
- 239000013598 vector Substances 0.000 claims abstract description 40
- 238000000034 method Methods 0.000 claims abstract description 33
- 239000011159 matrix material Substances 0.000 claims abstract description 31
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 12
- 238000011835 investigation Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 10
- 230000035945 sensitivity Effects 0.000 abstract description 8
- 230000006870 function Effects 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/242—Dictionaries
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06398—Performance of employee with respect to a job function
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Educational Administration (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Engineering & Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- General Business, Economics & Management (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Mathematical Physics (AREA)
- Evolutionary Biology (AREA)
- Pure & Applied Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Mathematics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Educational Technology (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Primary Health Care (AREA)
- Computational Linguistics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Algebra (AREA)
- Evolutionary Computation (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, in particular to a test paper difficulty coefficient evaluation method, device and equipment based on artificial intelligence. The method comprises the following steps: acquiring the class characteristics of the knowledge points; acquiring an attention index of each knowledge point according to the category characteristics of the knowledge points and the score losing rate of each test question; the score losing weight of each knowledge point is obtained through the score losing rate vector and the attention index, and the learning ability index of each student is obtained by taking the score losing weight as the weight of the score losing rate vector; clustering the learning ability indexes by using a fuzzy mean clustering algorithm to obtain the membership degree of each student belonging to the middle-grade students, and obtaining the difficulty coefficient of each historical test paper according to the historical scores and the membership degrees of all students; and constructing a feature matrix of the historical test paper, and establishing a difficulty coefficient evaluation network by using the feature matrix and the difficulty coefficient so as to evaluate the difficulty coefficient of the test paper. According to the embodiment of the invention, different students can obtain accurate difficulty coefficients by utilizing different sensitivity degrees of the students to the difficulty coefficients.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a test paper difficulty coefficient evaluation method, device and equipment based on artificial intelligence.
Background
The test paper difficulty coefficient is data reflecting the difficulty degree of the test paper, and the larger the difficulty coefficient is, the higher the failure rate of the test paper is, and the higher the test question discrimination is. When an expert gives questions to a test paper, the expert needs to accurately master the difficulty coefficient of the test paper, ensure that the test paper has a certain degree of distinction, comprehensively know and master the learning condition of students, and avoid that the excessive difficulty of the test paper damages the learning enthusiasm and confidence of the students, so an evaluation method of the difficulty coefficient of the test paper is needed to ensure the difficulty level of the test paper.
In practice, the inventors found that the above prior art has the following disadvantages:
at present, an automatic test paper generation system usually judges difficulty coefficients through average scores of all students, does not consider that the students have different learning abilities and different sensitivity degrees to test paper difficulty, possibly evaluates the difficulty of the test paper by mistake and cannot accurately and properly evaluate the learning conditions of the students.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a test paper difficulty coefficient evaluation method, a test paper difficulty coefficient evaluation device and test paper difficulty coefficient evaluation equipment based on artificial intelligence, and the adopted technical scheme is as follows:
in a first aspect, an embodiment of the present invention provides a test paper difficulty coefficient evaluation method based on artificial intelligence, including the following steps:
acquiring text information of a historical test paper, and acquiring knowledge point class characteristics by classifying the text information; acquiring an attention index of each knowledge point according to the knowledge point category characteristics and the fraction losing rate of each test question;
acquiring the historical score of each student, and acquiring the score losing condition of each student to each knowledge point according to the historical score to form a score losing rate vector; acquiring the score losing weight of each knowledge point by combining the score losing vector with the attention index, and acquiring the learning ability index of each student by taking the score losing weight as the weight of the score losing vector;
clustering the learning ability indexes by using a fuzzy mean clustering algorithm to obtain the membership degree of each student belonging to a middle-grade student, and obtaining the difficulty coefficient of each historical test paper according to the historical scores and the membership degrees of all students;
and constructing a feature matrix of the historical test paper according to the knowledge point class features, the attention index and the text information, and establishing a difficulty coefficient evaluation network by using the feature matrix and the difficulty coefficient so as to evaluate the difficulty coefficient of the test paper.
Preferably, the method for acquiring the loss rate is as follows:
and acquiring the average score of each test question of all students, and acquiring the losing rate according to the average score.
Preferably, the step of acquiring the attention index includes:
acquiring the investigation times of a target knowledge point in a certain historical test paper according to the knowledge point category characteristics, and acquiring an initial attention index of the target knowledge point in the test paper according to the failure rate and the investigation times of all test questions of the investigation target knowledge point;
and acquiring the attention index of the target knowledge point according to the initial attention index of the target knowledge point in all historical test papers.
Preferably, the method for obtaining the vector of the fraction loss is as follows:
acquiring a first losing rate of each test question according to the historical score of each student, and taking the average value of the first losing rates of the students on the historical test questions for investigating the target knowledge point as a second losing rate; and the second fraction loss rates of all the target knowledge points form the fraction loss rate vector.
Preferably, the step of obtaining the score losing weight comprises:
obtaining a score losing weight factor of the target knowledge point according to the difference of the second score losing rates of all students;
and acquiring the score losing weight according to the attention index of the target knowledge point and the score losing weight factor.
Preferably, the learning ability index acquiring step includes:
acquiring the learning ability index of each student to the target knowledge point according to the score losing weight and the second score losing rate;
and acquiring the learning ability index of each student according to the learning ability indexes of all the target knowledge points.
Preferably, the step of obtaining the membership degree includes:
respectively acquiring initial clustering centers of high-grade students, medium-grade students and low-grade students according to the learning capacity indexes;
randomly distributing an initial membership degree belonging to the initial clustering center for all students, and updating the initial membership degree by using the fuzzy mean clustering algorithm to obtain a membership degree matrix;
and acquiring the membership degree of each student belonging to the middle-grade student according to the membership degree matrix.
Preferably, the difficulty coefficient obtaining method includes:
and normalizing the membership degrees of all students to obtain difficulty weights, and obtaining the difficulty coefficients according to the test paper score of all students and the difficulty weights.
In a second aspect, another embodiment of the present invention provides an apparatus for evaluating a difficulty coefficient of a test paper based on artificial intelligence, including the following modules:
the attention index acquisition module is used for acquiring text information of the historical test paper and acquiring the class characteristics of the knowledge points by classifying the text information; acquiring an attention index of each knowledge point according to the knowledge point category characteristics and the fraction losing rate of each test question;
the learning ability index acquisition module is used for acquiring the historical score of each student, acquiring the score losing condition of each student to each knowledge point according to the historical score and forming a score losing rate vector; acquiring the score losing weight of each knowledge point by combining the score losing vector with the attention index, and acquiring the learning ability index of each student by taking the score losing weight as the weight of the score losing vector;
the difficulty coefficient acquisition module is used for clustering the learning ability indexes by using a fuzzy mean clustering algorithm, acquiring the membership degree of each student belonging to a middle-grade student, and acquiring the difficulty coefficient of each historical test paper according to the historical scores and the membership degrees of all students;
and the difficulty coefficient evaluation network establishing module is used for establishing a characteristic matrix of the historical test paper according to the knowledge point class characteristics, the attention degree indexes and the text information, and establishing a difficulty coefficient evaluation network by using the characteristic matrix and the difficulty coefficient so as to evaluate the difficulty coefficient of the test paper.
In a third aspect, another embodiment of the present invention provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the above-mentioned artificial intelligence-based test paper difficulty coefficient assessment method when executing the computer program.
The embodiment of the invention at least has the following beneficial effects:
according to the embodiment of the invention, students are classified based on learning ability, the sensitivity degree of the students to the difficulty coefficient is reflected by utilizing the membership degree of the students to different learning ability classes, the sensitivity degree of the students to the difficulty coefficient is different in consideration of different learning ability and historical score of the students in actual conditions, namely the influence degree of the change of the difficulty coefficient of the test paper on the score is different, and the accurate difficulty coefficient is obtained.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flowchart of a test paper difficulty coefficient evaluation method based on artificial intelligence according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating steps of a method for evaluating difficulty factors of test paper based on artificial intelligence according to an embodiment of the present invention;
FIG. 3 is a block diagram of an apparatus for evaluating difficulty coefficients of test paper based on artificial intelligence according to an embodiment of the present invention;
fig. 4 is an internal structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description will be given to a method, an apparatus and a device for evaluating difficulty coefficient of test paper based on artificial intelligence according to the present invention, with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of a test paper difficulty coefficient evaluation method, device and equipment based on artificial intelligence in detail with reference to the accompanying drawings.
Referring to fig. 1 and fig. 2, fig. 1 is a flowchart illustrating a test paper difficulty coefficient evaluation method based on artificial intelligence according to an embodiment of the present invention; fig. 2 is a flowchart illustrating steps of a test paper difficulty coefficient evaluation method based on artificial intelligence according to an embodiment of the present invention, where the method includes the following steps:
step S001, acquiring text information of a historical test paper, and acquiring knowledge point class characteristics by classifying the text information; and acquiring the attention index of each knowledge point according to the category characteristics of the knowledge points and the score losing rate of each test question.
In the examination, because the examination papers are limited in space and the examination of the knowledge points is biased, the attention degree of the knowledge points is obtained according to the occurrence frequency and the score losing rate of each knowledge point.
The method comprises the following specific steps:
1) and constructing a knowledge point word library.
The method comprises the steps of firstly creating n dictionaries, wherein one dictionary corresponds to one knowledge point, and manually storing keywords corresponding to the knowledge point in each dictionary, wherein the keywords are words with definite classification effect.
As an example, the keyword of the knowledge point Newton's second law may be set as: acceleration, force, etc.
2) And acquiring the class characteristics of the knowledge points.
And acquiring the text information of the historical test paper, including the text information of each test question. Performing word segmentation on the text information of the test question with the number i, matching word vectors after word segmentation with a knowledge point word library to obtain knowledge points for examination of the test question with the number i, and obtaining the class characteristics of the knowledge points
In the embodiment of the invention, the historical test paper is the test paper data of the same course in the last five years.
It should be noted that, more than one knowledge point is examined for one test question, the number n of knowledge points is determined according to the scope of the test,
is a vector with 1 line and n columns, each dimension corresponds to a knowledge point, and the test question with the number i contains a certain knowledge point
And setting the value of the corresponding dimension to be 1, otherwise, setting the value to be 0.
3) And acquiring the loss rate of each test question.
And acquiring the average score of each test question of all students, and acquiring the losing rate according to the average score.
The method for calculating the loss rate α is as follows:
wherein S is i Indicates the full score value of the test question numbered i,
the average score of the student for the test question numbered i is shown.
And taking the score losing rate of the test questions as the score losing rate of the knowledge points, wherein the larger the numerical value is, the larger the difficulty of the investigation of the knowledge points is.
Obtaining a knowledge point score vector according to the score losing rate of each test question
When the test question with the number i contains a certain knowledge point, the test question is
And setting the value of the corresponding dimension as alpha, otherwise, setting the value as 0.
4) And acquiring the attention index of each knowledge point.
Specifically, the method comprises the following steps:
a. and acquiring the investigation times of the target knowledge point in a certain historical test paper according to the class characteristics of the knowledge point, and acquiring the initial attention index of the target knowledge point in the test paper according to the loss rate and the investigation times of all the test questions of the investigation target knowledge point.
Initial attention index of target knowledge point b in jth test paper
The calculation formula of (a) is as follows:
wherein N is 0 Indicates the number of subjects in the jth test paper,
representing a vector
Value of the middle b-th dimension, N b The question number of the investigation target knowledge points b in the jth test paper is shown,
representing a vector
The dimension b represents the corresponding dimension of the target knowledge point b.
b. And acquiring the attention index of the target knowledge point according to the initial attention index of the target knowledge point in all the historical test papers.
Attention index A of target knowledge point b b Comprises the following steps:
wherein N is * Indicating the number of test papers of the history test paper.
According to the same method, attention indexes of n knowledge points can be obtained, and the value range of the attention indexes is [0, 1 ].
Step S002, obtaining historical scores of each student, obtaining the score losing situation of each student on each knowledge point according to the historical scores, and forming a score losing rate vector; and acquiring the score losing weight of each knowledge point by combining the score losing rate vector with the attention index, and acquiring the learning ability index of each student by taking the score losing weight as the weight of the score losing rate vector.
The method comprises the following specific steps:
1) and obtaining the score loss rate vector of each student.
Acquiring a first losing rate of each test question according to the historical score of each student, and taking the average value of the first losing rates of each student on the historical test questions of the investigation target knowledge point as a second losing rate; the second fraction loss rates of all target knowledge points constitute a fraction loss rate vector.
Taking the d-th student as an example, obtaining a first losing rate of each test question of the student according to the score of each test question by using the same method as the step S001, wherein correspondingly, the losing rates of knowledge points investigated by each test question are the first losing rates; the average value of the first fraction loss rates of the test questions of all the investigation target knowledge points is the second fraction loss rate of the target knowledge points; the n-dimensional score loss vector Q of the d-th student is composed of the second score loss of all the target knowledge points.
2) And acquiring the losing weight.
Specifically, the method comprises the following steps:
a. and acquiring a score losing weight factor of the target knowledge point according to the difference of the second score losing rates of all students.
Obtaining second fraction loss of all students to the target knowledge point b, recording the total number of the students as Num, obtaining a data set consisting of Num data, and calculating the variance sigma of the data set b The larger the variance is, the larger the difference of the second score of the student with respect to the target knowledge point b is, i.e., the more the target knowledge point can reflect the difference of the learning ability of the student.
Obtaining a score losing weight factor w of the target knowledge point b b1 :
b. And acquiring the score losing weight according to the attention index and the score losing weight factor of the target knowledge point.
The attention index of the knowledge points reflects the attention degree of the examination paper to each knowledge point and the score losing weight W of the target knowledge point b b Comprises the following steps:
W b ＝A b ×w b1
3) and calculating the learning ability index of the student.
Obtaining the learning ability index of each student to the target knowledge point according to the score losing weight and the second score losing rate; and acquiring the learning ability index of each student according to the learning ability indexes of all the target knowledge points.
The calculation method of the learning ability index comprises the following steps:
wherein β represents a learning ability index, α b A second failure rate representing the target knowledge point b.
The larger β indicates the higher learning ability, and the better the knowledge points are grasped.
It should be noted that the learning ability index is a relative concept, and if the score losing rate of all students in a knowledge point is very low, the knowledge point cannot reflect the learning ability, so the embodiment of the invention acquires the learning ability index through the score losing weight and the score losing rate of the students.
And S003, clustering the learning ability indexes by using a fuzzy mean clustering algorithm to obtain the membership degree of each student belonging to the middle-grade students, and obtaining the difficulty coefficient of each historical test paper according to the historical scores and the membership degrees of all the students.
The method comprises the following specific steps:
1) and acquiring the membership degree of each student belonging to the middle-grade students.
Specifically, initial clustering centers of high-grade students, medium-grade students and low-grade students are respectively obtained according to the learning capacity index; randomly distributing an initial membership degree belonging to an initial clustering center for all students, and updating the initial membership degree by using a fuzzy mean clustering algorithm to obtain a membership degree matrix; and acquiring the membership degree of each student belonging to the middle-grade students according to the membership degree matrix.
The fuzzy C-means clustering algorithm comprises the following steps: selecting the number of columns as 3, initializing a membership matrix U, and realizing clustering by minimizing an objective function J:
wherein beta is d Denotes a learning ability index of the d-th student, c μ Represents the μ cluster center; num represents the number of samples, namely the number of students; u. of μd Representing the degree of membership of the d-th student to the μ -th cluster center.
To be explainedThe constraint condition of the objective function is
I.e. the sum of the membership of three classes per student is 1.
2) And acquiring the difficulty coefficient of each historical test paper.
And normalizing the membership degrees of all students to obtain difficulty weights, and obtaining difficulty coefficients according to the test paper score and the difficulty weights of all students.
Normalizing the membership degree of each student belonging to the middle-grade students to obtain a normalization coefficient, namely the difficulty weight, of the d-th student
The calculation formula of (c) is:
wherein, gamma is d Indicating that the d-th student belongs to the membership of the middle-ranked student,
representing the sum of membership of all students belonging to the middle-ranked students.
Obtaining the difficulty coefficient delta of each historical test paper according to the normalization coefficient and the examination score of each student:
wherein, Score d Score, representing the examination Score of a given test paper of the d-th student 0 The total score of the test paper is shown, and delta is the difficulty coefficient of the test paper.
And obtaining the difficulty coefficient delta of each test paper in the historical test paper information base.
In the prior art, the difficulty coefficient is evaluated by solving the mean value of all students, the difference between the historical scores and the learning capacity of the students is ignored, the student scores with strong learning capacity and poor learning capacity are not greatly influenced by the change of the difficulty coefficient, the student scores with the learning capacity at the middle level can reflect the change of the difficulty of the test paper, namely, the learning capacities of different students have different sensitivity degrees to the difficulty coefficient, the sensitivity degree of the student scores to the difficulty of the test paper can be obtained based on fuzzy C mean value clustering, and the accurate test paper difficulty coefficient can be obtained.
And step S004, constructing a feature matrix of the historical test paper according to the class features of the knowledge points, the attention indexes and the text information, and establishing a difficulty coefficient evaluation network by using the feature matrix and the difficulty coefficient so as to evaluate the difficulty coefficient of the test paper. According to the embodiment of the invention, different students can obtain accurate difficulty coefficients by utilizing different sensitivity degrees of the students to the difficulty coefficients.
The method comprises the following specific steps:
in step S003, the accurate difficulty coefficient δ of each historical test paper is obtained, and a neural network is trained by using the coefficient as label data and using the feature matrix of the corresponding historical test paper as training data. The input of the neural network is a characteristic matrix of the test paper, and the output is a difficulty coefficient of the test paper.
The construction method of the test paper characteristic matrix comprises the following steps:
1) segmenting the test paper information according to the test question numbers, obtaining knowledge points investigated by each question according to the method of the step S001, and obtaining the category characteristic vectors of the knowledge points
Computing the L0 norm of a vector
Reflecting the number of knowledge points for topic investigation.
2) To obtain
Corresponding to the attention index of the knowledge point, taking the maximum value of the attention index of each knowledge point and recording as A max 。
3) Aim to solve the problemExtracting semantic information from text information by using word vectors, obtaining semantic vectors with fixed dimension T from the extracted semantic information by using dimension reduction technology, and obtaining feature vectors of a question
4) And stacking according to the sequence of the test question numbers to obtain a feature matrix of the test paper.
5) Obtaining the test paper characteristic matrix according to the method 1) to 4) for each historical test paper.
The neural network is of an Encoder + FC structure, a test paper characteristic matrix of each test paper in a historical test paper information base is used as a training set, a difficulty coefficient delta is used as a label, a loss function is a mean square error function, and a gradient descent method is used for updating network parameters to complete training.
In the difficulty coefficient evaluation stage, the feature matrix of the test paper to be evaluated is obtained by the same method, and the feature matrix is input into the neural network to obtain the difficulty coefficient corresponding to the test paper.
As an example, the Encoder in the embodiment of the present invention adopts a ResNet structure, and in other embodiments, a network structure that can achieve the same effect, such as send, may also be adopted.
In summary, the embodiment of the present invention obtains the text information of the historical test paper, and obtains the class characteristics of the knowledge points by classifying the text information; acquiring an attention index of each knowledge point according to the category characteristics of the knowledge points and the score losing rate of each test question; acquiring the historical score of each student, and acquiring the score losing condition of each student to each knowledge point according to the historical score to form a score losing rate vector; the score losing weight of each knowledge point is obtained through the score losing rate vector and the attention index, and the learning ability index of each student is obtained by taking the score losing weight as the weight of the score losing rate vector; clustering the learning ability indexes by using a fuzzy mean clustering algorithm to obtain the membership degree of each student belonging to the middle-grade students, and obtaining the difficulty coefficient of each historical test paper according to the historical scores and the membership degrees of all students; and constructing a feature matrix of the historical test paper according to the knowledge point class features, the attention index and the text information, and establishing a difficulty coefficient evaluation network by using the feature matrix and the difficulty coefficient so as to evaluate the difficulty coefficient of the test paper. According to the embodiment of the invention, different students can obtain accurate difficulty coefficients by utilizing different sensitivity degrees of the students to the difficulty coefficients.
Based on the same inventive concept as the method, another embodiment of the present invention provides an artificial intelligence-based test paper difficulty coefficient evaluation device, please refer to fig. 3, which includes the following modules:
an attention index obtaining module 1001, a learning ability index obtaining module 1002, a difficulty coefficient obtaining module 1003 and a difficulty coefficient evaluation network establishing module 1004.
The attention index acquisition module is used for acquiring text information of the historical test paper and acquiring the class characteristics of the knowledge points by classifying the text information; acquiring an attention index of each knowledge point according to the category characteristics of the knowledge points and the score losing rate of each test question; the learning ability index acquisition module is used for acquiring the historical score of each student, acquiring the losing score condition of each student on each knowledge point according to the historical score, and forming a losing score rate vector; the score losing weight of each knowledge point is obtained through the score losing rate vector and the attention index, and the learning ability index of each student is obtained by taking the score losing weight as the weight of the score losing rate vector; the difficulty coefficient acquisition module is used for clustering the learning ability indexes by using a fuzzy mean clustering algorithm, acquiring the membership degree of each student belonging to a middle-grade student, and acquiring the difficulty coefficient of each historical test paper according to the historical scores and the membership degrees of all students; the difficulty coefficient evaluation network establishing module is used for establishing a feature matrix of the historical test paper according to the knowledge point class features, the attention index and the text information, and establishing a difficulty coefficient evaluation network by using the feature matrix and the difficulty coefficient so as to evaluate the difficulty coefficient of the test paper.
Another embodiment of the present invention provides an electronic device based on the same inventive concept as the above-described method.
Referring to fig. 4, a schematic diagram of an electronic device according to an embodiment of the invention is shown. The electronic device of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. The processor implements the steps of the above-mentioned artificial intelligence-based test paper difficulty coefficient evaluation method embodiment when executing the computer program, for example, the steps shown in fig. 2. Or the processor realizes the functions of the modules in the test paper difficulty coefficient evaluation device based on the artificial intelligence when executing the computer program.
Illustratively, the computer program may be partitioned into one or more modules, where one or more modules are stored in the memory and executed by the processor to implement the present invention. One or more of the modules may be a series of computer program instruction segments capable of performing certain functions, which are used to describe the execution of the computer program in the electronic device.
The electronic device may be a desktop computer, a notebook, a palm top computer, a cloud server, or other computing device. The electronic device may include, but is not limited to, a processor, a memory. Those skilled in the art will appreciate that the schematic diagrams are merely examples of an electronic device and do not constitute a limitation of an electronic device, and may include more or fewer components than those shown, or combine certain components, or different components, e.g., an electronic device may also include input-output devices, network access devices, buses, etc.
The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like that is the control center for an electronic device and that connects the various parts of the overall electronic device using various interfaces and wires.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. The test paper difficulty coefficient evaluation method based on artificial intelligence is characterized by comprising the following steps:
acquiring text information of a historical test paper, and acquiring knowledge point category characteristics by classifying the text information; acquiring an attention index of each knowledge point according to the knowledge point category characteristics and the fraction losing rate of each test question;
acquiring the historical score of each student, and acquiring the score losing condition of each student to each knowledge point according to the historical score to form a score losing rate vector; acquiring the score losing weight of each knowledge point by combining the score losing vector with the attention index, and acquiring the learning ability index of each student by taking the score losing weight as the weight of the score losing vector;
clustering the learning ability indexes by using a fuzzy mean clustering algorithm to obtain the membership degree of each student belonging to a middle-grade student, and obtaining the difficulty coefficient of each historical test paper according to the historical scores and the membership degrees of all students;
and constructing a feature matrix of the historical test paper according to the knowledge point class features, the attention index and the text information, and establishing a difficulty coefficient evaluation network by using the feature matrix and the difficulty coefficient so as to evaluate the difficulty coefficient of the test paper.
2. The method according to claim 1, wherein the method for obtaining the fraction loss is as follows:
and acquiring the average score of each test question of all students, and acquiring the losing rate according to the average score.
3. The method according to claim 1, wherein the step of obtaining the attention index comprises:
acquiring the investigation times of a target knowledge point in a certain historical test paper according to the knowledge point category characteristics, and acquiring an initial attention index of the target knowledge point in the test paper according to the failure rate and the investigation times of all test questions of the investigation target knowledge point;
and acquiring the attention index of the target knowledge point according to the initial attention index of the target knowledge point in all historical test papers.
4. The method of claim 3, wherein the vector of fraction lost is obtained by:
acquiring a first losing rate of each test question according to the historical score of each student, and taking the average value of the first losing rates of the students on the historical test questions of the target knowledge point as a second losing rate; and the second fraction loss rates of all the target knowledge points form the fraction loss rate vector.
5. The method of claim 4, wherein the step of obtaining the scoring weight comprises:
obtaining a score losing weight factor of the target knowledge point according to the difference of the second score losing rates of all students;
and acquiring the score losing weight according to the attention index of the target knowledge point and the score losing weight factor.
6. The method according to claim 4, wherein the learning ability index obtaining step includes:
obtaining the learning ability index of each student to the target knowledge point according to the score losing weight and the second score losing rate;
and acquiring the learning ability index of each student according to the learning ability indexes of all the target knowledge points.
7. The method of claim 4, wherein the step of obtaining the degree of membership comprises:
respectively acquiring initial clustering centers of high-grade students, medium-grade students and low-grade students according to the learning capacity indexes;
randomly distributing an initial membership degree belonging to the initial clustering center for all students, and updating the initial membership degree by using the fuzzy mean clustering algorithm to obtain a membership degree matrix;
and acquiring the membership degree of each student belonging to the middle-grade student according to the membership degree matrix.
8. The method of claim 1, wherein the difficulty coefficient is obtained by:
and normalizing the membership degrees of all students to obtain difficulty weights, and obtaining the difficulty coefficient according to the test paper score of all students and the difficulty weights.
9. Examination paper degree of difficulty coefficient evaluation device based on artificial intelligence, its characterized in that, the device includes following module:
the attention index acquisition module is used for acquiring text information of the historical test paper and acquiring the class characteristics of the knowledge points by classifying the text information; acquiring an attention index of each knowledge point according to the knowledge point category characteristics and the fraction losing rate of each test question;
the learning ability index acquisition module is used for acquiring the historical score of each student, acquiring the score losing condition of each student to each knowledge point according to the historical score and forming a score losing rate vector; acquiring the score losing weight of each knowledge point by combining the score losing vector with the attention index, and acquiring the learning ability index of each student by taking the score losing weight as the weight of the score losing vector;
the difficulty coefficient acquisition module is used for clustering the learning ability indexes by using a fuzzy mean clustering algorithm, acquiring the membership degree of each student belonging to a middle-grade student, and acquiring the difficulty coefficient of each historical test paper according to the historical scores and the membership degrees of all students;
and the difficulty coefficient evaluation network establishing module is used for establishing a characteristic matrix of the historical test paper according to the knowledge point class characteristics, the attention degree indexes and the text information, and establishing a difficulty coefficient evaluation network by using the characteristic matrix and the difficulty coefficient so as to evaluate the difficulty coefficient of the test paper.
10. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 8 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210636889.7A CN114943628A (en) | 2022-06-07 | 2022-06-07 | Test paper difficulty coefficient evaluation method, device and equipment based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210636889.7A CN114943628A (en) | 2022-06-07 | 2022-06-07 | Test paper difficulty coefficient evaluation method, device and equipment based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114943628A true CN114943628A (en) | 2022-08-26 |
Family
ID=82908284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210636889.7A Pending CN114943628A (en) | 2022-06-07 | 2022-06-07 | Test paper difficulty coefficient evaluation method, device and equipment based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114943628A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117079504A (en) * | 2023-10-13 | 2023-11-17 | 山东金榜苑文化传媒有限责任公司 | Wrong question data management method of big data accurate teaching and reading system |
-
2022
- 2022-06-07 CN CN202210636889.7A patent/CN114943628A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117079504A (en) * | 2023-10-13 | 2023-11-17 | 山东金榜苑文化传媒有限责任公司 | Wrong question data management method of big data accurate teaching and reading system |
| CN117079504B (en) * | 2023-10-13 | 2024-01-05 | 山东金榜苑文化传媒有限责任公司 | Wrong question data management method of big data accurate teaching and reading system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20200134263A1 (en) | Non-factoid question-answering device | |
| US20230058194A1 (en) | Text classification method and apparatus, device, and computer-readable storage medium | |
| CN111630518A (en) | ESG-based enterprise evaluation execution device and operation method thereof | |
| CN102214233B (en) | Method and device for classifying texts | |
| WO2021237934A1 (en) | Answer selection method and apparatus, computer device, and computer readable storage medium | |
| CN111767366A (en) | Question and answer resource mining method and device, computer equipment and storage medium | |
| CN111414746A (en) | Matching statement determination method, device, equipment and storage medium | |
| CN106779050A (en) | The optimization method and device of a kind of convolutional neural networks | |
| CN116386862A (en) | Multi-modal cognitive impairment evaluation method, device, equipment and storage medium | |
| CN114155397B (en) | Small sample image classification method and system | |
| Ferrari et al. | Identification of cross-domain ambiguity with language models | |
| CN111460101A (en) | Knowledge point type identification method and device and processor | |
| Shoohi et al. | DCGAN for Handling Imbalanced Malaria Dataset based on Over-Sampling Technique and using CNN. | |
| Lee et al. | Fair selective classification via sufficiency | |
| CN112836027A (en) | Method for determining text similarity, question answering method and question answering system | |
| CN114943628A (en) | Test paper difficulty coefficient evaluation method, device and equipment based on artificial intelligence | |
| CN117556034A (en) | Data processing system for standardizing output results of electronic medical record question-answering model | |
| CN113486670B (en) | Text classification method, device, equipment and storage medium based on target semantics | |
| CN113312907B (en) | Remote supervision relation extraction method and device based on hybrid neural network | |
| CN116521872A (en) | Combined recognition method and system for cognition and emotion and electronic equipment | |
| CN113392086B (en) | Medical database construction method, device and equipment based on Internet of things | |
| CN115600595A (en) | Entity relationship extraction method, system, equipment and readable storage medium | |
| Tebbe et al. | Is natural language processing the cheap charlie of analyzing cheap talk? A horse race between classifiers on experimental communication data | |
| CN116050428B (en) | Intention recognition method, device, equipment and storage medium | |
| CN113793191B (en) | Commodity matching method and device and electronic equipment |
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 |
Application publication date: 20220826 |
|
| WD01 | Invention patent application deemed withdrawn after publication |