CN114461786B - Learning path generation method and system - Google Patents
Learning path generation method and system Download PDFInfo
- Publication number
- CN114461786B CN114461786B CN202210381561.5A CN202210381561A CN114461786B CN 114461786 B CN114461786 B CN 114461786B CN 202210381561 A CN202210381561 A CN 202210381561A CN 114461786 B CN114461786 B CN 114461786B
- Authority
- CN
- China
- Prior art keywords
- test
- learning
- test result
- knowledge
- training data
- 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.)
- Active
Links
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
- G06F16/33—Querying
- G06F16/335—Filtering based on additional data, e.g. user or group profiles
-
- 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
- G06F16/33—Querying
- G06F16/338—Presentation of query results
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
- G06Q50/205—Education administration or guidance
Abstract
The application provides a learning path generation method, which comprises the following steps: behavior characteristic data, knowledge point characteristic data and test results of a plurality of samples form a training data set; performing regression fitting calculation by combining the training data set according to the behavior characteristic data of the current student and the characteristic data of the knowledge points to be sorted to obtain a test result predicted value corresponding to the current student and the knowledge points to be sorted; and comparing the test result predicted values of a plurality of knowledge points to be sequenced of the current student to generate a learning path. The application also provides a device for realizing the method. The method and the device solve the problems that learning software in the prior art outputs learning resources when learning repeatedly, is low in efficiency and cannot adapt to the specific conditions of students.
Description
Technical Field
The application relates to the technical field of computers, in particular to a learning path generation method and a learning path generation system in an intelligent learning application scene.
Background
In the current online learning process of recording and broadcasting, all students can learn under a uniformly arranged content and progress frame, and learning feedback is mainly in a mode of spontaneous inquiry of the students. Because individual situations are different and the same time is spent, the learning effect of each student is different.
In the current learning form, efficient feedback and effective tracking are lacked, and the individual learning requirements of students are difficult to meet. Meanwhile, the relation between the content importance degree and the weak points of the student is difficult to accurately evaluate by the student, and the rhythm of learning and reviewing is reasonably arranged. Limitations in the learner's own cognition can lead to an inability to correctly select an appropriate learning path, resulting in learning inefficiencies. Therefore, there is a need to develop an intelligent learning system that can push suitable learning content for students of different levels in real time during learning, collect learning feedback in time, and calculate the optimal learning path of the students, thereby helping the students to effectively improve learning efficiency.
Disclosure of Invention
The application provides a learning path generation method and a learning path generation system, which solve the problems that learning software in the prior art is low in learning resource output efficiency and cannot adapt to specific conditions of students when learning is repeated.
In a first aspect, an embodiment of the present application provides a learning path generating method, including the following steps:
behavior characteristic data, knowledge point characteristic data and test results of a plurality of samples form a first training data set; the behavior characteristic data comprises statistical data of access behaviors to resources; the knowledge point feature data is data for quantitatively representing the knowledge points;
performing regression fitting calculation by combining the first training data set according to the behavior characteristic data of the current student and the characteristic data of the knowledge points to be sorted to obtain a first test result predicted value corresponding to the current student and the knowledge points to be sorted;
and comparing first test result predicted values of a plurality of knowledge points to be sequenced of the current student, and sequencing the knowledge points to be sequenced with the first test result predicted values larger than a first set threshold value in a descending order according to the first test result predicted values to generate a learning path.
In another embodiment of the present application, the method further comprises the steps of:
behavior characteristic data of a plurality of samples, characteristic data of knowledge points, test question characteristic data and test scores form a second training data set; the test question feature data is data for quantitatively representing the test question;
performing regression fitting calculation by combining the second training data set according to the behavior characteristic data of the current student, the characteristic data of any knowledge point of the learning path and the characteristic data of the test questions to be distributed to obtain a second test result predicted value corresponding to the current student, any knowledge point and the test questions to be distributed;
and comparing second test result predicted values of a plurality of test questions to be distributed of the current student, and determining the test questions to be distributed, wherein the difference between the second test result predicted values and the first test result predicted values of any knowledge points is greater than a second set threshold value.
Preferably, the step of performing a regression fit calculation uses the GBDT + LR model.
Preferably, the behavioral characteristics include at least one of: the duration, frequency, amount of resources of visiting the resources; the test question features include at least one of: test question identification, real question proportion, score statistics, difficulty and importance; the knowledge point features include at least one of: knowledge point identification, examination proportion, difficulty and importance.
Preferably, the test result includes a result value of the test performed on the sample by using a set of test questions.
Further preferably, each knowledge point and each test question to be distributed of the learning path are combined into a learning set, and the learning sets are sorted in descending order according to the difference between the second test result predicted value and the first test result predicted value.
Further preferably, in the test questions to be distributed/knowledge points to be sorted, the test questions to be distributed/knowledge points to be sorted are excluded, wherein the time length between the current time and the last access time of the current student is less than a third set threshold, and/or the time length between the current time and the last access time of the current student is greater than a fourth set threshold.
In a second aspect, the present application further provides a learning path generating apparatus, configured to implement the method according to any embodiment of the present application, where the learning path generating apparatus includes:
the system comprises a first module, a second module and a third module, wherein the first module is used for identifying resource access behaviors and generating behavior characteristic data;
the second module is used for calculating the first test result predicted value and/or the second test result predicted value;
the third module is used for accessing the resources in a descending order according to the first test result predicted value to obtain the corresponding knowledge points, and/or is used for accessing the resources in a descending order according to the difference between the second test result predicted value and the first test result predicted value to obtain the corresponding learning set;
a training database for storing the first training data set and/or the second training data set;
the resource comprises at least one of: a knowledge point database, a test question database, and an application program for reading the knowledge point database and/or the test question database.
The embodiment of the application adopts at least one technical scheme which can achieve the following beneficial effects:
the method and the system optimize the learning path generator by taking the improvement of the test score as an optimization target and combining the learning feedback characteristic and the Ebbinghaus memory curve of the student in time, further perform joint optimization on the learning path generator and the test question personalized selector, provide a reasonable and efficient personalized learning path for the student, and improve the learning efficiency and the test score.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
FIG. 1 is a flow chart of an embodiment of the method of the present application;
FIG. 2 is a flow chart of another embodiment of the method of the present application;
FIG. 3 is an embodiment of a learning path method for dynamically generating training clusters based on current trainees;
FIG. 4 is an embodiment of the apparatus of the present application;
FIG. 5 is a block diagram of an embodiment of a resource database of the present application;
FIG. 6 shows a second embodiment of the structure of the apparatus of the present application.
Detailed Description
To make the objects, technical solutions and advantages of the present application more clear, the technical solutions of the present application will be clearly and completely described below with reference to specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only a few embodiments of the present application, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.
Fig. 1 is a flow chart of an embodiment of the method of the present application.
The embodiment of the application provides a learning path generation method, which comprises the following steps:
the samples in this application may include current trainees and/or historical trainees. The related data of the sample can also be set sample data special for training.
The test result comprises a result value for testing the sample by using a set test question set.
Preferably, the behavior feature comprises at least one of the following feature types: the duration, frequency, amount of resources accessed; each feature type represents a feature value in terms of a numerical value or a range of numerical values.
Preferably, the knowledge point features comprise at least one of the following feature types: knowledge point identification, and statistics and/or evaluation values of test proportion, difficulty, importance and the like; each feature type represents a feature value in terms of a numerical value or a range of numerical values.
102, performing regression fitting calculation by combining the first training data set according to the behavior characteristic data of the current student and the characteristic data of the knowledge points to be sorted to obtain a first test result predicted value corresponding to the current student and the knowledge points to be sorted;
preferably, the step of performing a regression fit calculation uses the GBDT + LR model.
Regarding the knowledge points to be sorted, at the beginning of learning, the current learner can be tested before learning to determine the knowledge level of the learner, and then the set of the knowledge points to be sorted is determined according to the level value. Preferably, there is a preset progressive order relationship between the plurality of knowledge points.
In consideration of the human brain memory characteristic curve, it is further preferable that, of the knowledge points to be sorted, whose duration between the current time and the current student's last visit time is smaller than a third set threshold, and/or the knowledge points to be sorted, whose duration between the current time and the current student's last visit time is larger than a fourth set threshold, are excluded. When the third set threshold and the fourth set threshold are used simultaneously, it is preferable that the fourth set threshold is larger than the third set threshold.
And 103, comparing first test result predicted values of a plurality of knowledge points to be sequenced of the current student, and sequencing the knowledge points to be sequenced with the first test result predicted values larger than a first set threshold value in a descending order according to the first test result predicted values to generate a learning path.
If the predicted value of the first test result is smaller than a first set threshold value, skipping the knowledge point; and according to the selection result, the optimal learning content is pushed to the trainees, and the knowledge point pushed first is the knowledge point with the maximum predicted value of the first test result.
Fig. 2 is a flowchart of another embodiment of the method of the present application, which implements further optimization of the learning path.
Step 201 (synchronization step 101), recording behavior characteristic data, knowledge point characteristic data and test results of a plurality of samples to form a first training data set; the behavior characteristic data comprises statistical data of access behaviors to resources; the knowledge point feature data is data for quantitatively representing the knowledge point; the test result comprises a result value for testing the sample by using a set test question set.
Step 202 (synchronization step 102), performing regression fitting calculation according to the behavior characteristic data of the current student and the characteristic data of the knowledge points to be sorted and by combining the first training data set to obtain a first test result predicted value corresponding to the current student and the knowledge points to be sorted.
preferably, the test question features comprise at least one of the following feature types: test question identification, and statistics and/or evaluation values such as true question ratio, score statistics, difficulty, importance and the like; each feature type represents a feature value in terms of a numerical value or a range of numerical values.
And step 204 (synchronization step 103), comparing first test result predicted values of the multiple knowledge points to be sequenced of the current student, and sequencing the knowledge points to be sequenced, of which the first test result predicted values are greater than a first set threshold value, in a descending order according to the first test result predicted values to generate a learning path.
in consideration of the human brain memory characteristic curve, it is further preferable that, of the to-be-distributed test questions, the to-be-distributed test questions whose duration between the current time and the current student's last visit time is smaller than a third set threshold value are excluded, or the to-be-distributed test questions whose duration between the current time and the current student's last visit time is larger than a fourth set threshold value are excluded.
And step 206, comparing second test result predicted values of a plurality of test questions to be distributed of the current student, and determining the test questions to be distributed, wherein the difference between the second test result predicted values and the first test result predicted values is larger than a second set threshold value.
When the second test result predicted value and the first achievement predicted value are subtracted, the samples and knowledge points corresponding to the first achievement predicted value and the second test result predicted value are the same.
Furthermore, the plurality of test questions to be distributed are arranged in a descending order according to the difference between the corresponding second test result predicted value and the corresponding first test result predicted value, so that any knowledge point on the learning path determines the corresponding best test question to be distributed (or the best test question combination to be distributed).
Further preferably, each knowledge point of the learning path and a group of test questions to be distributed (the group of test questions to be distributed may be, for example, an optimal test question to be distributed corresponding to the knowledge point or a combination of the optimal test question to be distributed) are combined into one learning set, and the learning sets are sorted in descending order according to a difference between the second test result predicted value and the first test result predicted value for each learning set, so as to generate another selectable learning path as the optimized learning path.
And according to the selection result, the optimal learning content is pushed to the trainees, and the knowledge point pushed first is the learning set with the largest difference between the predicted value of the second test result and the predicted value of the first test result.
FIG. 3 is an embodiment of a learning path method for dynamically generating training cluster clusters based on current trainees.
Preferably, in the training data set, the samples are grouped for different test question feature data. Ensuring that the sample features in each cluster are distributed the same. And (3) constructing training data by showing different knowledge points or different subjects for different students in proportion in the experimental group, counting test results of the students as prediction data, and performing regression fitting by using the GBDT + LR model modified by LOSS.
and clustering the knowledge points and the questions, and extracting the related characteristics of the knowledge points and the question clusters, including the types and the characteristic values of the knowledge points or the test question characteristics. The type of knowledge point feature, as described in step 101; the type of the test question feature, as described in step 203.
the behavior feature data, as described in step 101.
in step 303, the knowledge points to be sorted are determined according to the current progress, and knowledge points which do not meet requirements are filtered by combining an Ebingois memory curve and time.
For example, the knowledge points are divided into different chapters, the knowledge points in different chapters have similar or advanced relations, and different chapters extract different numbers of questions with different types to form chapter test rolls before learning according to chapter emphasis, difficulty and distribution in the examination questions in the past year.
For another example, the knowledge points to be sorted with the duration between the current time and the last visit time of the current trainee being less than the set threshold value are excluded from the knowledge points to be sorted; or, conversely, the knowledge points to be sorted with the time length between the current time and the current student previous azimuth time being greater than the set threshold value are excluded from the knowledge points to be sorted.
Further, or alternatively, the knowledge points with any characteristic value within a set range are excluded from the knowledge points to be ranked.
For example, a learning order label, a progression relation label (knowledge points associated with the knowledge points), a knowledge point importance label, and a content attribute label (knowledge point summary data, video explanation) are marked on all the knowledge points according to the provision of a lesson, the distribution of chapters in a calendar year examination, and statistics of questions made by a historical student in each chapter. Adding the knowledge points (belonging to a chapter) which are more than the current sequence, the knowledge points which have progressive relation with the current knowledge points and the historically learned knowledge points into candidate knowledge points according to the current learning progress of the student, filtering part of learning contents by combining business rules, the mastery degree of the knowledge points of the student and the importance of the knowledge points, and adding the rest of the learning contents belonging to the candidate knowledge points into the candidate contents.
Preferably, the plurality of training data sets used for performance prediction of the current trainee have the same range of corresponding set sample feature values. Thus, the sample feature ranges of the first training data set and the second training data set are the same.
The training database includes sample feature values including data representing sample features, such as region, age, gender, academic calendar or academic aptitude. In a training data set, the sample characteristic values are counted, samples are classified according to the sample characteristic values, and the proportion of the set sample characteristic value range in all the samples is obtained. For example, the proportion of samples from a set area to the total samples is set, and the proportion of samples in an age range to the total samples is set.
The current trainee, as one of the samples, is also described by a sample characteristic value. The behavior feature data of the current trainee can also be used for updating the training database.
The training data sets are clustered to form a plurality of training data sets. In each training data set, the proportion of the set sample characteristic value range in all samples of each training data set is the same. For example, in the first training data set, the percentage of samples within the age range to all samples in the first training data set is set to 20%, and in the second training data set, the percentage of samples within the age range to all samples in the second training data set is also set to 20%.
The sample feature value range comprises 1 or more sample feature types, and for each 1 sample feature type, the sample feature value range comprises 1 or more sample feature value ranges. For example, in addition to the age characteristics, the first training data set also includes samples from a set area, accounting for 30% of all samples in the first training data set; the second training data set also contains samples from a predetermined area, and the percentage of the samples in the second training data set is 30%.
Each first training data set of the cluster is used for predicting at least one knowledge point setting a knowledge point characteristic value range. That is, each first training data set includes knowledge point feature data including 1 or more knowledge point feature types corresponding to the plurality of samples; for each knowledge point feature type, 1 or more knowledge point feature value ranges are included. Each knowledge point feature type and the range of the knowledge point feature value are preset.
And each clustered second training data set is used for predicting the test questions to be distributed corresponding to at least one set test question characteristic value range. That is, the test question feature data included in each of the second training data sets includes 1 or more test question feature types corresponding to the plurality of samples, and includes 1 or more test question feature value ranges for each test question feature type. Each test question feature type and the range of the test question feature values are preset.
And 305, calculating a first test result predicted value of the knowledge point to be ranked by using at least one first training data set and the behavior data of the current student, generating a learning path, and synchronizing the steps 101 to 103.
In step 305, the degree of improvement of each knowledge point to the test score is predicted by combining the learning progress of the student, the learning behavior characteristics and the characteristics of the knowledge points.
And 307, optimizing the learning path, namely sequencing the learning set in a descending order according to the difference between the second test result predicted value and the first test result predicted value to generate another optional learning path.
FIG. 4~6 is an embodiment of the apparatus of the present application.
In a second aspect, the present application further provides a learning path generating apparatus, configured to implement the method according to any embodiment of the present application, including a first module 41, a second module 42, a third module 43, a training database 44, and a resource database 45, as shown in fig. 4.
The first module is used for identifying resource access behaviors and generating behavior characteristic data. For example, in a smart learning software or internet online learning application system, a first module is associated with the user application module 46, by which behavior profile data is generated by identifying resource access behavior in response to user actions acting on the user application.
It should be noted that, during the learning process, any user, including the current learner or the historical learner, accesses the knowledge point database or the test question database through the audio/video playing module and the text or graphic display module with the user operation interface. The resource comprises at least one of: knowledge point data, test question data, and an application program for reading the knowledge point database and/or the test question database. As shown in fig. 5, the resource database 45 is a general term for resources and media for storing or running the resources, and includes a knowledge point database 51, a test question database 52, and an application file database 53 for reading the knowledge point database and/or the test question database. Access to the resource by any user triggers the first module.
In order to support the current user to form a training data set, the resource database further comprises a resource classifier 54 for performing cluster analysis on knowledge points and topics, for example, the system subdivides the nodes of the learning path into five types, namely a stage test, a chapter test, a knowledge point video, a knowledge point image-text and a knowledge point question test, and the overall optimization goal is to improve the scores of the chapter test. Marking the knowledge point video and knowledge point data through a label system, and determining characteristic values representing learning sequence, importance degree, belonging knowledge clusters and the like; marking the test questions, and determining the characteristic values of representing importance degree, examination frequency, difficulty, knowledge point relevance, knowledge point coverage rate and the like.
And in the learning process of the student, the first module extracts a specific characteristic of the learning behavior of the student and uses the specific characteristic as an input for performance prediction of the second module. For example, the content to be learned is predicted to be a knowledge point test question test, and furthermore, the behavior characteristics of the test questions made by the trainee can also be used as historical data to enter a training data set to be used as the input of the next prediction.
And the second module is used for calculating the first test result predicted value and/or the second test result predicted value to form a learning path generator. When the second module is only used for calculating the second test result predicted value, the second module also forms a test question personalized selector. Preferably, as shown in fig. 6, the second module further comprises a test question selector 61 and a learning path generator 62, and the learning path generator is connected with the test question selector. The test question selector is used for calculating a second test result predicted value and sorting according to the size of the second test result predicted value; and the learning path generator is used for calculating the first test result predicted value and sequencing according to the size of the first test result predicted value or sequencing the difference value of the second test result predicted value and the first test result predicted value.
The training database is used for storing the first training data set and/or the second training data set. The test result comprises a result value for testing the sample by using a set test question set. For example, in the intelligent learning software or the internet online learning application system, all test questions are grouped into stage tests, chapter tests, and the like, and each test is applied to one test question set.
Preferably, the second module is further configured to collect basic information related to the trainee after the trainee starts learning, including information such as learning goal, learning time, current level, and personal characteristics. The second module combines the student basic information, the learning behavior characteristics and the Einbinghaus memory curve, in the running process of the user application program for realizing learning, aiming at the current student, on one hand, the resources are called, the mastering degree is timely evaluated in a mode of knowledge point test question testing, chapter testing and the like, sample data is generated, and a training database is updated; on the other hand, at the beginning of learning, the current trainees are tested before learning to determine the knowledge levels of the trainees, and then the set of knowledge points to be sorted is determined according to the level values. The pre-learning test can be, for example, dividing the knowledge points into different chapters, wherein the knowledge points in different chapters have similar or advanced relations, and extracting different chapters to form chapter test volumes before learning according to the chapter emphasis, difficulty and distribution in examination questions of all years, wherein the different chapters extract questions with different numbers and types. Preferably, the second module further comprises a preselector 63 for collecting the current student basic information, performing pre-learning tests and determining a knowledge point set to be ranked. The pre-selector is further connected to the learning path generator.
And the third module is used for accessing the resources in a descending order according to the first test result predicted value to obtain the corresponding knowledge points, and/or is used for accessing the resources in a descending order according to the difference between the second test result predicted value and the first test result predicted value to obtain the corresponding learning set.
The third module is associated with a user application, and is responsive to a knowledge point access request from a user to form a learning resource distributor.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application therefore also proposes a computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method according to any one of the embodiments of the present application.
Further, the present application also proposes an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the method according to any of the embodiments of the present application.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, or loaded onto a computer or other programmable apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the functions specified in the flowchart flow or flows and/or block diagram block or blocks are implemented.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory. The memory may include forms of volatile memory in a computer readable medium, random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (10)
1. A learning path generation method characterized by comprising the steps of:
behavior characteristic data, knowledge point characteristic data and test results of a plurality of samples form a first training data set; the behavior characteristic data comprises statistical data of access behaviors to resources; the knowledge point feature data is used for clustering knowledge points to extract cluster-like related features, and comprises types and feature values of the knowledge point features; the test result comprises a result value for testing the sample by using a set test question set;
behavior characteristic data of a plurality of samples, characteristic data of knowledge points, test question characteristic data and test scores form a second training data set; the test question feature data is used for quantitatively representing the test questions, clustering the questions and extracting cluster-like related features, and comprises types and feature values of the test question features;
counting test results as prediction data, and performing regression fitting calculation by combining the first training data set according to the behavior characteristic data of the current student and the characteristic data of the knowledge points to be sorted to obtain a first test result prediction value corresponding to the current student and the knowledge points to be sorted;
comparing first test result predicted values of a plurality of knowledge points to be sequenced of the current student, and sequencing the knowledge points to be sequenced, of which the first test result predicted values are greater than a first set threshold value, in a descending order according to the first test result predicted values to generate a learning path;
performing regression fitting calculation by combining the second training data set according to the behavior characteristic data of the current student, the characteristic data of any knowledge point of the learning path and the characteristic data of the test questions to be distributed to obtain a second test result predicted value corresponding to the current student, any knowledge point and the test questions to be distributed;
and comparing second test result predicted values of a plurality of test questions to be distributed of the current student, and determining the test questions to be distributed, wherein the difference between the second test result predicted values and the first test result predicted value of any knowledge point is larger than a second set threshold value.
2. The learning path generation method according to claim 1,
the step of performing a regression fit calculation uses the GBDT + LR model.
3. The learning path generation method according to claim 1,
the behavioral characteristics include at least one of: the duration, frequency, amount of resources accessed;
the test question features include at least one of: test question identification, real question proportion, score statistics, difficulty and importance;
the knowledge point features include at least one of: knowledge point identification, examination proportion, difficulty and importance.
4. The learning path generation method according to claim 1,
clustering training data, and determining a first training data set for the current trainee and the knowledge points to be ranked;
the training data sets are used for performing score prediction on the current trainees, and the corresponding set sample characteristic value ranges are the same; the sample characteristic value range comprises 1 or more sample characteristic types, and for each 1 sample characteristic type, the sample characteristic value range comprises 1 or more sample characteristic value ranges;
each first training data set of the cluster is used for predicting at least one knowledge point setting a knowledge point characteristic value range.
5. The learning path generation method according to claim 4,
clustering training data, and determining a second training data set for the current trainee, the knowledge points to be sorted and the test questions to be distributed;
and each clustered second training data set is used for predicting the test questions to be distributed corresponding to at least one set test question characteristic value range.
6. The learning path generation method according to claim 1, further comprising the steps of:
and combining each knowledge point of the learning path and each test question to be distributed into a learning set, and sequencing the learning sets according to the difference between the second test result predicted value and the first test result predicted value in a descending order.
7. The method of claim 6, further comprising the step of:
and in the test questions to be distributed/knowledge points to be sorted, eliminating the test questions to be distributed/knowledge points to be sorted of which the time length between the current time and the last access time of the current student is less than a third set threshold value, and/or eliminating the test questions to be distributed/knowledge points to be sorted of which the time length between the current time and the last access time of the current student is greater than a fourth set threshold value.
8. A learning path generation apparatus for implementing the method of any one of claims 1~7 comprising:
the system comprises a first module, a second module and a third module, wherein the first module is used for identifying resource access behaviors and generating behavior characteristic data;
the second module is used for calculating the first test result predicted value and/or the second test result predicted value;
the third module is used for accessing the resources in a descending order according to the first test result predicted value to obtain the corresponding knowledge points, and/or is used for accessing the resources in a descending order according to the difference between the second test result predicted value and the first test result predicted value to obtain the corresponding learning set;
a training database for storing the first training data set and/or the second training data set;
the resources include at least one of: the system comprises a knowledge point database, a test question database and an application program for reading the knowledge point database and/or the test question database.
9. A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the method of any one of claims 1~7.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method of any of claims 1~7 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210381561.5A CN114461786B (en) | 2022-04-13 | 2022-04-13 | Learning path generation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210381561.5A CN114461786B (en) | 2022-04-13 | 2022-04-13 | Learning path generation method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114461786A CN114461786A (en) | 2022-05-10 |
| CN114461786B true CN114461786B (en) | 2022-10-21 |
Family
ID=81418601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210381561.5A Active CN114461786B (en) | 2022-04-13 | 2022-04-13 | Learning path generation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114461786B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116777696B (en) * | 2023-06-30 | 2024-04-30 | 黑龙江大学 | Legal knowledge automatic learning system and method |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10546507B2 (en) * | 2014-09-18 | 2020-01-28 | International Business Machines Corporation | Recommending a set of learning activities based on dynamic learning goal adaptation |
| CN107274020B (en) * | 2017-06-15 | 2020-05-22 | 北京师范大学 | Learner subject total measured result prediction system and method based on collaborative filtering thought |
| CN110472060B (en) * | 2019-07-05 | 2020-12-01 | 平安国际智慧城市科技股份有限公司 | Topic pushing method and device, computer equipment and storage medium |
| CN113268611B (en) * | 2021-06-24 | 2022-11-01 | 北京邮电大学 | Learning path optimization method based on deep knowledge tracking and reinforcement learning |
-
2022
- 2022-04-13 CN CN202210381561.5A patent/CN114461786B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN114461786A (en) | 2022-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107230174B (en) | Online interactive learning system and method based on network | |
| CN108182489B (en) | Personalized learning recommendation method based on online learning behavior analysis | |
| Behr et al. | Motives for dropping out from higher education—An analysis of bachelor's degree students in Germany | |
| CN111651676B (en) | Method, device, equipment and medium for performing occupation recommendation based on capability model | |
| JP2022008867A (en) | Machine learning method, apparatus, and computer program for providing individually customized educational content based on learning efficiency | |
| Hussain et al. | Mining moodle data to detect the inactive and low-performance students during the moodle course | |
| CN114461786B (en) | Learning path generation method and system | |
| US20120329028A1 (en) | Method for intelligent personalized learning service | |
| KR20100000017A (en) | An intelligent customized learning service method | |
| JP2024511216A (en) | Systems and methods for tracking and assessing learner growth | |
| CN116028702A (en) | Learning resource recommendation method and system and electronic equipment | |
| Drăgulescu et al. | Predicting assignment submissions in a multi-class classification problem | |
| EL MEZOUARY et al. | An evaluation of learner clustering based on learning styles in MOOC course | |
| CN112419815A (en) | Planning method and device for learning path | |
| CN114461787B (en) | Test question distribution method and system | |
| CN112734142B (en) | Resource learning path planning method and device based on deep learning | |
| Çiftçi | The Studies on Educational Digital Games Regarding Children: A New Word Analysis Approach. | |
| CN113869569A (en) | Learning score prediction and personalized intervention method based on decision tree | |
| CN111815491A (en) | Adaptive learning plan making method and device, readable storage medium and terminal | |
| Valsamidis et al. | A combined algorithm for LMS usage assessment | |
| CN110807179A (en) | User identification method, device, server and storage medium | |
| CN112506855A (en) | Question bank construction method based on cognitive hierarchy | |
| KR20200084694A (en) | Method and apparatus for collecting educational materials | |
| CN111382263A (en) | Data processing method, device and system | |
| CN117726485B (en) | Intelligent adaptation education learning method and system based on big data |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |