CN114491050A - Learning ability assessment method and system based on cognitive diagnosis - Google Patents
Learning ability assessment method and system based on cognitive diagnosis Download PDFInfo
- Publication number
- CN114491050A CN114491050A CN202210147876.3A CN202210147876A CN114491050A CN 114491050 A CN114491050 A CN 114491050A CN 202210147876 A CN202210147876 A CN 202210147876A CN 114491050 A CN114491050 A CN 114491050A
- Authority
- CN
- China
- Prior art keywords
- answer
- records
- user
- labels
- cognitive diagnosis
- 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
- 230000001149 cognitive effect Effects 0.000 title claims abstract description 85
- 238000003745 diagnosis Methods 0.000 title claims abstract description 79
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000002372 labelling Methods 0.000 claims abstract description 25
- 238000011156 evaluation Methods 0.000 claims abstract description 19
- 239000011159 matrix material Substances 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 238000012360 testing method Methods 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 5
- 239000013598 vector Substances 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006998 cognitive state Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Classifications
-
- 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/103—Formatting, i.e. changing of presentation of documents
- G06F40/117—Tagging; Marking up; Designating a block; Setting of attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Databases & Information Systems (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Electrically Operated Instructional Devices (AREA)
Abstract
The invention provides a learning ability assessment method based on cognitive diagnosis, which comprises the following steps: acquiring the answer records of a user, and performing labeling pretreatment on the answer records of the user to obtain labeled answer records and unlabeled answer records; according to the answer records with the labels, clustering the answer records without the labels to obtain the labels of all the answer records; inputting all answer records and labels thereof into a cognitive diagnosis model, outputting the answer correct probability of the user by the cognitive diagnosis model, and evaluating the learning ability of the user according to the answer correct probability of the user. The invention takes different types of answer information into consideration, labels the answer records, clusters the unlabelled answer records and the labeled answer records, inputs the clustering result into the traditional serious diagnosis model, completes the evaluation of the learning ability of the user, fully utilizes a large amount of different types of data information contained in the answer records of the user, and improves the accuracy of the cognitive diagnosis model.
Description
Technical Field
The invention relates to the field of cognitive diagnosis, in particular to a learning ability assessment method and system based on cognitive diagnosis.
Background
The conventional examination test uses a single score to evaluate the learning ability of the user, and neither knows which specific knowledge the user mastered or did not master nor obtains the reason why the user made a wrong examination question to remedy. For users with the same score, the differences in cognitive state and knowledge structure that may exist between them are less likely to be obtained. The information provided by conventional examinations is less suitable for the needs of individual development. The cognitive diagnostic test is an improvement and perfection on the traditional test and evaluation, and aims to find out the state of a user in the learning process, such as the proficiency degree of the user on specific knowledge concepts so as to accurately evaluate the learning ability of the user. Cognitive diagnosis has become a fundamental problem in artificial intelligence education today.
The existing student cognitive diagnosis method comprises the steps of obtaining historical answer information of students, extracting test question texts and predefined knowledge points contained in the test question texts, and calculating knowledge point relevancy vectors of all test questions; inputting the set student parameters and the test question parameters including the knowledge point relevancy vectors of the test questions into a cognitive diagnosis model, fitting answer results, obtaining knowledge point mastery vectors of students through training, and completing the cognitive diagnosis of the students. However, the above method uses a conventional cognitive diagnosis model, only correct and incorrect answer results or the final score of a certain question are considered for the answer records of the question, and for some questions, such as a programming test question, the answer records not only include the answer results, but also include information such as submission times, interval time, programming state, number of pass cases and the like, and if the answer records are not considered, the defects of serious diagnosis and low accuracy of learning ability evaluation are caused.
Disclosure of Invention
The invention provides a learning ability evaluation method and system based on cognitive diagnosis, aiming at overcoming the defect of low evaluation accuracy when the existing cognitive diagnosis model is applied to some questions with answer records containing various types of information.
In order to solve the technical problems, the technical scheme of the invention is as follows:
in a first aspect, the present invention provides a learning ability assessment method based on cognitive diagnosis, comprising the following steps:
s1: and acquiring the answer records of the user, and performing labeling pretreatment on the answer records of the user to obtain labeled answer records and unlabeled answer records.
S2: and clustering the non-label answer records according to the answer records with labels to obtain the labels of all the answer records.
S3: establishing a cognitive diagnosis model, inputting all answer records and labels thereof into the cognitive diagnosis model, outputting the answer correct probability of the user by the cognitive diagnosis model, and evaluating the learning ability of the user according to the answer correct probability of the user.
Preferably, in S1, the answer record includes answer state, use case passing rate, answer interval time and answer times.
Preferably, S1 specifically includes the following steps:
s1.1: counting the answer records of the user, acquiring the case passing rate of each answer state of the user in each question, and calculating the scoring rate g of each answer state of the user in each question according to the case passing rate of each answer stateihThe calculation formula is as follows:
wherein S isihAll answer record sets with the answer state H in the ith question of the user are represented, H represents the answer state, m represents the total answer state number, H represents the total number of the questions, and x represents the total number of the questionsiThe number of scoring use cases of the ith topic is represented;
s1.2: the method comprises the following steps of arranging each answer record of the same question according to a time sequence for a user, and calculating the score of the question in each answer state according to an arrangement result and a score, wherein a calculation formula is as follows:
wherein beta is a preset attenuation rate,
the score of the user on the nth answer in the ith answer state h is represented;
s1.3: setting a first score threshold and a second score threshold, labeling the answer records according to the scores in each answer state, labeling the answer records with the scores lower than the first score threshold with a careless label, labeling the answer records with the scores higher than the second score threshold with a proficiency label, and obtaining the answer records with the proficiency label, the answer records with the careless label and the non-label answer records with the scores higher than the first score threshold and lower than the second score threshold.
As a preferred scheme, clustering the non-label answer states according to the answer states with skilled labels and the answer states with the livingthe labels by using a K-means algorithm; the method specifically comprises the following steps: respectively taking the answering state with skilled labels and the answering state with the life and mind labels as initial clusters, and calculating the center point set of each cluster
Wherein the content of the first and second substances,
b represents the number of clusters, UbSet of answer states representing all unlabeled in cluster b, yiRepresenting the ith non-label answering state;
assigning an unlabeled answer state to the cluster b 'until b' satisfies the following condition:
wherein v represents the number of iterations;
then, the central point of each cluster is recalculated, the iteration times are updated, all the non-label answer states are distributed into the class b' until the sum of the distances from all the answer states to the central point of each class of clusters converges, and the expression of the objective function of the K-means cluster is as follows:
yi＝(fi,ti)
wherein K represents the number of the answer state categories without labels, fiShowing the number of times of answer of the ith question, tiShowing answer interval time, y, of the ith questioniAnd
the euclidean distance of (c).
Preferably, the programmed answer state includes a programmed answer state including a programmed pass, a programmed error, a runtime error, and an answer error.
As a preferred scheme, extracting questions and knowledge points contained in the questions according to answer records of users, and constructing a question-knowledge point matrix by using the questions and the knowledge points contained in the questions; and inputting all answer records, labels thereof and the question-knowledge point matrix into a cognitive diagnosis model, and outputting the answer correct probability of the user by the cognitive diagnosis model.
Preferably, the cognitive diagnostic model comprises an IRT model and an NCD model.
In a second aspect, the present invention provides a learning ability evaluation system based on cognitive diagnosis, which is applied to the learning ability evaluation method based on cognitive diagnosis according to any of the above aspects, and includes:
the system comprises a preprocessing module, a labeling module and a processing module, wherein the preprocessing module is used for acquiring the answer records of a user and performing labeling preprocessing on the answer records of the user to obtain labeled answer records and unlabeled answer records;
the clustering module is used for clustering the non-label answer records according to the answer records with labels to obtain the labels of all the answer records;
and the cognitive diagnosis model is used for outputting the correct answer probability of the user according to all input answer records and labels thereof and evaluating the learning ability of the user according to the correct answer probability of the user.
Preferably, the cognitive diagnostic model comprises an IRT model and an NCD model.
In a third aspect, the present invention further provides a computer system, including a memory and a processor, where the memory stores a computer program, and the processor implements the steps of any one of the learning ability assessment methods based on cognitive diagnosis when executing the computer program in the memory.
Compared with the prior art, the technical scheme of the invention has the beneficial effects that: the invention considers a plurality of different types of answer information contained in some answer records, labels different answer records of a user part according to the answer information, and inputs a clustering result into a traditional carefully diagnostic model by clustering unlabelled answer records and labeled answer records based on the thought of semi-supervised learning, thereby completing the evaluation of the learning ability of the user, fully utilizing a large amount of different types of data information contained in the answer records of the user, and finally effectively improving the accuracy of the cognitive diagnostic model.
Drawings
Fig. 1 is a flowchart of a learning ability evaluation method based on cognitive diagnosis.
FIG. 2 is a schematic diagram of the NCD model.
Fig. 3 is an architecture diagram of a learning ability evaluation system based on cognitive diagnosis.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent;
the technical solution of the present invention is further described below with reference to the accompanying drawings and examples.
Example 1
Referring to fig. 1, the present embodiment provides a learning ability assessment method based on cognitive diagnosis, including the following steps:
s1: and acquiring the answer records of the user, and performing labeling pretreatment on the answer records of the user to obtain labeled answer records and unlabeled answer records.
In the embodiment, the answer records of the user are counted to obtain the case passing rate in each answer state, and the scoring rate of each answer state is calculated according to the case passing rate in each answer state; arranging each answer record of the same question according to a time sequence aiming at a user, and calculating the score of the question in each answer state according to the arrangement result and the score; setting a first score threshold value and a second score threshold value, labeling the answer records according to the score under each answer state, labeling the answer records with the score lower than the first score threshold value with a lifelong label, labeling the answer records with the score higher than the second score threshold value with a proficiency label, and obtaining the answer records with the proficiency label, the answer records with the lifelong label and the non-label answer records with the score higher than the first score threshold value and lower than the second score threshold value.
S2: and clustering the non-label answer records according to the answer records with labels to obtain the labels of all the answer records.
This embodiment assigns all the non-labeled answer states to two clusters of the skillful-labeled answer records and the lifelong-labeled answer records until the sum of the distances from all the answer states to the center point of each cluster converges.
S3: establishing a cognitive diagnosis model, inputting all answer records and labels thereof into the cognitive diagnosis model, outputting the answer correct probability of the user by the cognitive diagnosis model, and evaluating the learning ability of the user according to the answer correct probability of the user.
In the embodiment, according to the answer records of the user, the questions and the knowledge points contained in the questions are extracted, and a question-knowledge point matrix is constructed by using the questions and the knowledge points; inputting all answer records, labels thereof and the question-knowledge point matrix into a cognitive diagnosis model, respectively outputting answer correct probabilities of different users by the cognitive diagnosis model according to the answer records with different labels, and analyzing and evaluating the learning ability of the users according to the answer correct probabilities of the users; when the answer correct probability of the user is higher, the learning ability of the user is considered to be higher; when the answer correctness probability of the user is low, the learning ability of the user is considered to be relatively low.
The invention considers a plurality of different types of answer information contained in some answer records, labels different answer records of a user part according to the answer information, and inputs a clustering result into a traditional carefully diagnostic model by clustering unlabelled answer records and labeled answer records based on the thought of semi-supervised learning, thereby completing the evaluation of the learning ability of the user, fully utilizing a large amount of different types of data information contained in the answer records of the user, and finally effectively improving the accuracy of the cognitive diagnostic model.
Example 2
In this embodiment, taking the evaluation of the programmed learning ability of the user as an example, on the basis of embodiment 1, a learning ability evaluation method based on cognitive diagnosis is provided, which includes the following steps:
s1: and acquiring the answer records of the user, and performing labeling pretreatment on the answer records of the user to obtain labeled answer records and unlabeled answer records.
The programmed answer record comprises the final answer state, and also comprises information such as answer times, answer interval time, case passing rate and the like. In addition, unlike the conventional cognitive diagnosis question which has only correct and wrong answer states, the return result of the programming question is more diversified, and even in the minimum case, the programmed answer states include AC (pass), CE (programming error), RTE (run-time error) and WA (answer error). For programming titles, in most cases, a user will do the title until the title passes, and if whether the title finally passes or not is used as prediction data, the knowledge mastering degree of the user is predicted to be high. Therefore, if only correct or incorrect questions are input into the cognitive diagnosis model as final answer states, a prediction result higher than the actual programming ability level of the user is often obtained.
In this embodiment, the method for pre-processing the programmed answer records of the user in a labeling manner according to the answer states of the programmed questions, the case passing rate, the answer interval time and the answer times to obtain labeled answer records and unlabeled answer records specifically includes the following steps:
s1.1: counting the answer records of the user, acquiring the case passing rate of each answer state of the user in each question, and calculating the scoring rate g of each answer state of the user in each question according to the case passing rate of each answer stateihThe calculation formula is as follows:
wherein S isihAll answer record sets with the answer state H in the ith question of the user are represented, H represents the answer state, m represents the total answer state number, H represents the total number of the questions, and x represents the total number of the questionsiThe number of scoring use cases of the ith topic is represented;
s1.2: the method comprises the following steps of arranging each answer record of the same question according to a time sequence for a user, and calculating the score of the question in each answer state according to an arrangement result and a score, wherein a calculation formula is as follows:
wherein beta is a preset attenuation rate,
the score of the user on the nth answer in the ith answer state h is represented;
s1.3: setting a first score threshold and a second score threshold, labeling the answer records according to the scores in each answer state, labeling the answer records with the scores lower than the first score threshold with a careless label, labeling the answer records with the scores higher than the second score threshold with a proficiency label, and obtaining the answer records with the proficiency label, the answer records with the careless label and the non-label answer records with the scores higher than the first score threshold and lower than the second score threshold.
S2: and clustering the non-label answer records according to the answer records with labels to obtain the labels of all the answer records.
Clustering the non-label answer states by using a K-means algorithm according to the answer states with the skilled labels and the answer states with the life and thin labels; the method specifically comprises the following steps: respectively taking the answering state with skilled labels and the answering state with the life and mind labels as initial clusters, and calculating the center point set of each cluster
Wherein the content of the first and second substances,
b represents the number of clusters, UbSet of answer states representing all unlabeled in cluster b, yiRepresenting the ith non-label answering state;
assigning an unlabeled answer state to the cluster b 'until b' satisfies the following condition:
wherein v represents the number of iterations;
then, the central point of each cluster is recalculated, the iteration times are updated, all the non-label answer states are distributed into the class b' until the sum of the distances from all the answer states to the central point of each class of clusters converges, and the expression of the objective function of the K-means cluster is as follows:
yi＝(fi，t；)
wherein K represents the number of the answer state categories without labels, fiShowing the number of times of answer of the ith question, tiShowing answer interval time, y, of the ith questioniAnd
the euclidean distance of (c).
S3: establishing a cognitive diagnosis model, inputting all answer records and labels thereof into the cognitive diagnosis model, outputting the answer correct probability of the user by the cognitive diagnosis model, and evaluating the learning ability of the user according to the answer correct probability of the user.
The present embodiment applies a conventional cognitive diagnostic model to evaluate the learning ability of the user. The learning ability assessment of the user is influenced by the user factor, the subject factor and the knowledge point factor. The cognitive diagnosis model needs to learn according to input user answer records and relations between knowledge points and questions, update parameters of the cognitive diagnosis model, and finally output an evaluation result.
Assume that the cognitive diagnostic model includes a set of users O ═ { O ═ O1,o2,...,onItem set P ═ P1,p2,...,pnD ═ D of knowledge points1,d2,...,dn}; and programming answer records R (o, p, R, f, t), wherein the programming answer records comprise a user number o, a question number p, an answer score R, answer times f and answer interval time t.
Extracting the question and the knowledge points contained in the question according to the answer record of the user, and constructing a question-knowledge point matrix Q by using the question and the knowledge points, wherein Q is { Q ═ Q { (Q) }ij}P×DWhen Q isijWhen 1, indicates the topic piAnd knowPoint of identification djCorrelation when Q isijWhen 0, indicates the topic piAnd knowledge point djIrrelevant;
inputting all answer records and labels thereof and the question-knowledge point matrix Q into a cognitive diagnosis model, outputting the answer correct probability of the user by the cognitive diagnosis model, and finally analyzing and evaluating the learning ability of the user based on the answer correct probability of the user.
In this embodiment, the cognitive diagnosis model sets corresponding weight parameters according to answer records with different labels. Inputting the answer records with the skilled labels into a cognitive diagnosis model, and outputting a relatively high answer correct probability of the user by the cognitive diagnosis model; and inputting the answer records with the liberal and sparse labels into the cognitive diagnosis model, and outputting the answer correct probability of the user with relatively low output by the cognitive diagnosis model. When the answer correct probability of the user is higher, the learning ability of the user is considered to be higher; when the answer correctness probability of the user is low, the learning ability of the user is considered to be relatively low.
Example 3
This example is an improvement of the learning ability evaluation method based on cognitive diagnosis proposed in example 2.
In practical applications, if a simple and conveniently arranged cognitive diagnostic model is required, an IRT model may be used. The expression of the IRT model is as follows:
wherein e represents the base of the natural logarithm, D represents the constant 1.7, apA discrimination parameter representing the p-th question, bpRepresenting the difficulty parameter of the p question, wherein o is a user number; thetaoA capability value representing user o; p represents the title number, Ppo(θo) Representing the probability of the user o answering the question i.
In this embodiment, θoAnd (4) iteratively obtaining by using an EM (Expectation maximization) algorithm according to the programmed answer recording information. a is apAnd bpTo be provided withAnd (4) determining parameters.
And after all answer records and labels thereof are input into the IRT model, the IRT model outputs the probability of answering the question i by the user o.
Example 4
The present embodiment improves on the learning ability evaluation method based on cognitive diagnosis proposed in embodiment 3.
In this embodiment, the cognitive diagnosis model adopts an NCD model combined with a neural network, as shown in fig. 2, fig. 2 is a schematic diagram of the NCD model; the NCD model is based on the IRT model and combines the neural network technology, in the concrete implementation process, the programming answer records are input into the NCD model, firstly matrix decomposition is carried out on the programming answer records, the programming answer records are decomposed into three dimensions of knowledge proficiency, knowledge difficulty and question discrimination, then the knowledge proficiency, the knowledge difficulty and the question discrimination are subjected to exponential vector multiplication and subtraction preliminary operation, then the operation results are input into the neural network and respectively pass through an input layer, a hidden layer and an output layer of the neural network, and finally the answer correct probability of a user is output. The NCD model incorporates a neural network with a strong fitting ability, so that an effect more accurate than that of the IRT model can be obtained.
Example 5
Referring to fig. 3, the present embodiment provides a learning ability evaluation system based on cognitive diagnosis, which includes a preprocessing module, a clustering module, and a cognitive diagnosis model.
In a specific implementation process, the preprocessing module counts the answer records of the user, obtains the case passing rate in each answer state, and calculates the scoring rate of each answer state according to the case passing rate in each answer state; arranging each answer record of the same question according to a time sequence aiming at a user, and calculating the score of the question in each answer state according to the arrangement result and the score; setting a first score threshold and a second score threshold, labeling the answer records according to the scores in each answer state, labeling the answer records with the scores lower than the first score threshold with a careless label, labeling the answer records with the scores higher than the second score threshold with a proficiency label, and obtaining the answer records with the proficiency label, the answer records with the careless label and the non-label answer records with the scores higher than the first score threshold and lower than the second score threshold.
And the clustering module performs semi-supervised K-means clustering on the non-labeled answer records according to the answer records with the skilled labels and the answer records with the sparse labels, and inputs the clustered results into the cognitive diagnosis model.
Extracting questions and knowledge points contained in the questions according to answer records of users to construct a question-knowledge point matrix; and inputting all answer records, labels thereof and the question-knowledge point matrix into a cognitive diagnosis model, outputting correct answer probabilities of different users by the cognitive diagnosis model according to different types of labels carried by the answer records, and evaluating the learning ability of the users according to the correct answer probabilities of the different users.
In the embodiment, the answer records with the skilled labels are input into the cognitive diagnosis model, and the cognitive diagnosis model outputs relatively high answer correct probability of the user; and inputting the answer records with the liberal and sparse labels into the cognitive diagnosis model, and outputting the answer correct probability of the user with relatively low output by the cognitive diagnosis model. When the answer correct probability of the user is higher, the learning ability of the user is considered to be higher; when the answer correctness probability of the user is low, the learning ability of the user is considered to be relatively low.
The terms describing positional relationships in the drawings are for illustrative purposes only and are not to be construed as limiting the patent;
it should be understood that the above-described embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the claims of the present invention.
Claims (10)
1. A learning ability assessment method based on cognitive diagnosis is characterized by comprising the following steps:
s1: acquiring the answer records of a user, and performing labeling pretreatment on the answer records of the user to obtain labeled answer records and unlabeled answer records;
s2: according to the answer records with the labels, clustering the answer records without the labels to obtain the labels of all the answer records;
s3: establishing a cognitive diagnosis model, inputting all answer records and labels thereof into the cognitive diagnosis model, outputting the answer correct probability of the user by the cognitive diagnosis model, and evaluating the learning ability of the user according to the answer correct probability of the user.
2. The learning ability assessment method according to claim 1, wherein the answer records include answer states, use case passing rates, answer interval times and answer times in S1.
3. The learning ability assessment method based on cognitive diagnosis according to claim 2, wherein S1 specifically comprises the following steps:
s1.1: counting the answer records of the user, acquiring the case passing rate of each answer state of the user in each question, and calculating the scoring rate g of each answer state of the user in each question according to the case passing rate of each answer stateihThe calculation formula is as follows:
wherein S isihAll answer record sets with the answer state H in the ith question of the user are represented, H represents the answer state, m represents the total answer state number, H represents the total number of the questions, and x represents the total number of the questionsiThe number of scoring use cases of the ith topic is represented;
s1.2: the method comprises the following steps of arranging each answer record of the same question according to a time sequence for a user, and calculating the score of the question in each answer state according to an arrangement result and a score, wherein a calculation formula is as follows:
wherein beta is a preset attenuation rate,
the score of the user on the nth answer in the ith answer state h is represented;
s1.3: setting a first score threshold value and a second score threshold value, labeling the answer records according to the score under each answer state, labeling the answer records with the score lower than the first score threshold value with a lifelong label, labeling the answer records with the score higher than the second score threshold value with a proficiency label, and obtaining the answer records with the proficiency label, the answer records with the lifelong label and the non-label answer records with the score higher than the first score threshold value and lower than the second score threshold value.
4. The learning ability evaluation method based on cognitive diagnosis according to claim 3, wherein the unlabeled answer states are clustered using a K-means algorithm according to the skillful labeled answer state and the lifelong labeled answer state; the method specifically comprises the following steps: respectively taking the answering state with skilled labels and the answering state with the life and mind labels as initial clusters, and calculating the center point set of each cluster
Wherein, the first and the second end of the pipe are connected with each other,
b represents the number of clusters, UbSet of answer states representing all unlabeled in cluster b, yiRepresenting the ith non-label answering state;
assigning an unlabeled answer state to the cluster b 'until b' satisfies the following condition:
wherein v represents the number of iterations;
then, the central point of each cluster is recalculated, the iteration times are updated, all the non-label answer states are distributed into the class b' until the sum of the distances from all the answer states to the central point of each class of clusters converges, and the expression of the objective function of the K-means cluster is as follows:
yi＝(fi,ti)
wherein K represents the number of the answer state categories without labels, fiShowing the number of times of answer of the ith question, tiShowing the answer interval time of the ith question, yiAnd
the euclidean distance of (c).
5. The cognitive diagnosis-based learning ability assessment method according to claim 2, wherein the answer state comprises a programmed answer state, and the programmed answer state comprises a programmed pass, a programmed error, a runtime error, and an answer error.
6. The learning ability assessment method based on cognitive diagnosis according to claim 1, wherein a question and knowledge points contained therein are extracted according to a user's answer record, and a question-knowledge point matrix is constructed by using the question and the knowledge points contained therein; and inputting all answer records, labels thereof and the question-knowledge point matrix into a cognitive diagnosis model, and outputting the answer correct probability of the user by the cognitive diagnosis model.
7. The cognitive diagnosis-based learning ability assessment method according to any one of claims 1 to 6, wherein the cognitive diagnosis model comprises an IRT model and an NCD model.
8. A learning ability evaluation system based on cognitive diagnosis, comprising:
the system comprises a preprocessing module, a labeling module and a processing module, wherein the preprocessing module is used for acquiring the answer records of a user and performing labeling preprocessing on the answer records of the user to obtain labeled answer records and unlabeled answer records;
the clustering module is used for clustering the non-label answer records according to the answer records with labels to obtain the labels of all the answer records;
and the cognitive diagnosis model is used for outputting the correct answer probability of the user according to all input answer records and labels thereof and evaluating the learning ability of the user according to the correct answer probability of the user.
9. The cognitive diagnosis-based learning ability assessment system of claim 8, the cognitive diagnosis model comprising an IRT model and an NCD model.
10. A computer system comprising a memory having a computer program stored thereon and a processor that when executed in the memory performs the steps of the cognitive diagnosis based learning ability assessment method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210147876.3A CN114491050A (en) | 2022-02-17 | 2022-02-17 | Learning ability assessment method and system based on cognitive diagnosis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210147876.3A CN114491050A (en) | 2022-02-17 | 2022-02-17 | Learning ability assessment method and system based on cognitive diagnosis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114491050A true CN114491050A (en) | 2022-05-13 |
Family
ID=81481812
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210147876.3A Pending CN114491050A (en) | 2022-02-17 | 2022-02-17 | Learning ability assessment method and system based on cognitive diagnosis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114491050A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116883794A (en) * | 2023-09-07 | 2023-10-13 | 北京智精灵科技有限公司 | Spatial structure cognitive ability evaluation method and system based on graph attention network |
| CN116883794B (en) * | 2023-09-07 | 2024-05-31 | 北京智精灵科技有限公司 | Spatial structure cognitive ability evaluation method and system based on graph attention network |
-
2022
- 2022-02-17 CN CN202210147876.3A patent/CN114491050A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116883794A (en) * | 2023-09-07 | 2023-10-13 | 北京智精灵科技有限公司 | Spatial structure cognitive ability evaluation method and system based on graph attention network |
| CN116883794B (en) * | 2023-09-07 | 2024-05-31 | 北京智精灵科技有限公司 | Spatial structure cognitive ability evaluation method and system based on graph attention network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111460249B (en) | Personalized learning resource recommendation method based on learner preference modeling | |
| Huo et al. | Knowledge modeling via contextualized representations for LSTM-based personalized exercise recommendation | |
| CN107230174B (en) | Online interactive learning system and method based on network | |
| US10290221B2 (en) | Systems and methods to customize student instruction | |
| CN110941723A (en) | Method, system and storage medium for constructing knowledge graph | |
| CN112508334B (en) | Personalized paper grouping method and system integrating cognition characteristics and test question text information | |
| Lindsey et al. | Automatic discovery of cognitive skills to improve the prediction of student learning | |
| CN114913729B (en) | Question selecting method, device, computer equipment and storage medium | |
| CN114201684A (en) | Knowledge graph-based adaptive learning resource recommendation method and system | |
| CN113902129A (en) | Multi-mode unified intelligent learning diagnosis modeling method, system, medium and terminal | |
| CN114021722A (en) | Attention knowledge tracking method integrating cognitive portrayal | |
| CN115238036A (en) | Cognitive diagnosis method and device based on graph attention network and text information | |
| CN113283488B (en) | Learning behavior-based cognitive diagnosis method and system | |
| Hoq et al. | Analysis of an Explainable Student Performance Prediction Model in an Introductory Programming Course. | |
| Wu et al. | Student achievement analysis and prediction based on the whole learning process | |
| Štajduhar et al. | Using string similarity metrics for automated grading of SQL statements | |
| Rana et al. | Slow learner prediction using multi-variate naïve Bayes classification algorithm | |
| Dai et al. | Knowledge tracing: A review of available technologies | |
| CN116049349A (en) | Small sample intention recognition method based on multi-level attention and hierarchical category characteristics | |
| Abd El-Haleem et al. | A generic ai-based technique for assessing student performance in conducting online virtual and remote controlled laboratories | |
| CN114491050A (en) | Learning ability assessment method and system based on cognitive diagnosis | |
| CN112231373A (en) | Method, device and equipment for processing knowledge point data and computer readable medium | |
| Zhang et al. | Using problem similarity-and orderbased weighting to model learner performance in introductory computer science problems | |
| Stretcu | Curriculum learning | |
| Schwartz et al. | An automated sql query grading system using an attention-based convolutional neural network |
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 |