Remote education resource recommendation method and system

Abstract

The invention discloses a remote education resource recommendation method and system, and relates to the technical field of resource recommendation. Comprising the following steps: dynamically grouping resources according to content types and contents of learning resources, and establishing a resource grouping table for offline analysis; extracting recent learning data of students to form a learning behavior feature set; according to the learning resource preference type and the capacity adaptation degree, education resources meeting conditions are screened from a local resource library, a recommendation resource set based on time sequence arrangement is generated, and a recommendation priority is determined through a dynamic adjustment algorithm; personalized resource list display is provided on a student terminal interface, and autonomous selection of students is supported; and when the network connection is recovered, comparing the learning behavior feature set with the learning behavior feature set generated offline, updating the recommendation strategy and synchronizing the recommendation strategy to the resource index table. According to the method, the recommendation quality of the remote education resources can be remarkably improved, and more personalized and accurate learning support is provided for students.

Description

Technical Field

[0001] The present invention relates to the technical field of resource recommendation, and in particular to a method and system for recommending remote education resources. Background Art

[0002] As a flexible and efficient learning mode, distance education is increasingly favored by students, educational institutions and enterprises. Especially driven by massive open online courses (MOOCs) and mobile learning applications, online education platforms and applications have gradually become mainstream. In order to better meet the personalized learning needs of students, educational resource recommendation technology has been widely used and studied. Most of the existing distance education resource recommendation methods are based on user historical behavior data, such as learning records, resource clicks, etc., and generate recommendation lists through data mining and machine learning algorithms. Common recommendation methods include collaborative filtering, content-based recommendation, hybrid recommendation, etc. However, these traditional methods often rely on relatively simple statistical features, are easily affected by data sparsity and cold start problems, and are difficult to achieve accurate matching of students' real learning needs.

[0003] Traditional educational resource recommendation systems face many challenges, especially in terms of personalized recommendation and adaptability of resources. First, many recommendation systems rely too much on user behavior data, while ignoring the type, depth and adaptability of learning content, resulting in single and lack of pertinence in recommendation results. Secondly, most recommendation methods only consider historical learning behavior, but lack real-time tracking and updating of students' current learning status and progress, and cannot adapt to changes in students' learning needs in a timely manner. In addition, most existing technologies fail to effectively combine students' learning progress and learning preferences, and lack the ability to dynamically adjust recommendation strategies, resulting in a decrease in recommendation accuracy and student satisfaction. More seriously, due to unstable network connections, many recommendation systems are unable to obtain students' learning data and feedback in real time when offline, limiting the real-time and accuracy of recommendations. Summary of the invention

[0004] In view of the problems existing in the above background technology, the present invention proposes a distance education resource recommendation method and system.

[0005] Therefore, the problem to be solved by the present invention is how to effectively solve the above-mentioned background technology problems by introducing dynamic adjustment and real-time synchronization mechanisms.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for recommending remote education resources, which includes: S1: obtaining the downloaded education resource data and the associated learning record information in the student learning terminal, dynamically grouping the resources according to the content type and content relevance of the learning resources, establishing a resource grouping table for offline analysis, and recording the corresponding usage probability in each group of resources; the dynamic grouping of resources includes: calculating the content relevance between resources, presetting multiple groups of relevance thresholds, and dynamically grouping the resources into several groups according to the content relevance; S2: extracting the student's recent learning time period, learning frequency, learning resource preference type and usage progress characteristics to form a learning behavior feature set, wherein the feature set includes time tags, resource types, and usage progress characteristics. type weight and priority score; S3: According to the learning behavior feature set, combined with the resource index table, according to the learning resource preference type and capacity adaptability, screen qualified educational resources from the local resource library, generate a recommended resource set based on time sequence, and determine the recommendation priority through a dynamic adjustment algorithm; S4: Provide a personalized resource list display on the student terminal interface, annotate the resource priority score and expected usage time to support students' independent selection; S5: When the network connection is restored, extract the student's latest learning record information, and compare it with the learning behavior feature set generated offline, update the recommendation strategy and synchronize it to the resource index table to ensure that the recommendation strategy is consistent with the student's real-time learning status.

[0007] As a preferred solution of the remote education resource recommendation method of the present invention, S1 specifically includes the following steps: extracting the downloaded education resource data from the student learning terminal, extracting the content feature vector of each resource ,in, For resources No. content feature values, generate a content feature set by parsing the metadata of the resource; calculate the content relevance between resources based on the extracted content feature set ,in, Representation Resources and resources The correlation between the historical usage frequency and the cumulative usage time is extracted from the students' learning record information, and the historical usage frequency mean of each group of resources is calculated by combining the dynamic grouping of resources; the comprehensive usage probability of each group of resources is determined by combining the mean usage frequency, the cumulative usage time and the dynamic grouping of resources .

[0008] As a preferred solution of the remote education resource recommendation method of the present invention, the content relevance between the resources is The calculation formula is: ; in, For resources and resources The relevance of the content, is the weight of the content feature, used to represent the The contribution of each feature to the content relevance. For resources In the The value of content features, n is the total number of content features, For resources In the The value of a content feature.

[0009] As a preferred solution of the remote education resource recommendation method of the present invention, the comprehensive usage probability of each group of resources is The calculation formula is: ; in, Grouping tables for resources The comprehensive probability of use, For resources The historical frequency of use, For resources The cumulative usage time, Grouping tables for resources The historical average frequency of use, is the nonlinear adjustment parameter, is the offset parameter of the used probability distribution.

[0010] As a preferred solution of the remote education resource recommendation method of the present invention, the method of determining the recommendation priority by dynamically adjusting the algorithm includes: the updated resource set is based on the time tag of the resource Sorting, generating a preliminary recommended resource set, and combining it with the learning behavior feature set Priority scoring in , dynamically adjust the preliminary recommended resource set, and calculate the final priority score of each resource according to the priority scoring formula; the priority scoring formula is: ; in, ~ is the weight coefficient, For resources Capacity adaptability.

[0011] As a preferred solution of the remote education resource recommendation method of the present invention, the comparison with the learning behavior feature set generated offline includes: extracting the latest learning record information Compare with the offline generated learning behavior feature set, and calculate the difference between the offline behavior feature set and the real-time behavior feature set through similarity measurement. ;if If the set threshold is exceeded, the learning behavior is considered to have changed, and the learning behavior feature set must be updated to reflect the student's real-time learning status.

[0012] As a preferred solution of the remote education resource recommendation method of the present invention, wherein: The calculation formula is as follows: ; in, is the number of dimensions of the feature set; and They are the first and second feature values.

[0013] In a second aspect, the present invention provides a remote education resource recommendation system, which includes: a learning resource data acquisition module, which is used to obtain the downloaded education resource data and related learning record information in the student learning terminal, dynamically group the resources according to the content type and content relevance of the learning resources, establish a resource grouping table for offline analysis, and record the corresponding usage probability in each group of resources; a learning behavior analysis module, which is used to extract the student's recent learning time period, learning frequency, learning resource preference type and usage progress characteristics to form a learning behavior feature set; a screening and recommendation module, which is used to screen qualified education resources from the local resource library according to the learning behavior feature set, combined with the resource index table, according to the learning resource preference type and capacity adaptation, generate a recommended resource set based on time sequence arrangement, and determine the recommendation priority through a dynamic adjustment algorithm; a personalized resource display module, which is used to provide a personalized resource list display on the student terminal interface, annotate the resource priority score and the expected usage time to support students' independent selection; a network recovery and synchronization module, which is used to extract the student's latest learning record information when the network connection is restored, and compare it with the learning behavior feature set generated offline, update the recommendation strategy and synchronize it to the resource index table to ensure that the recommendation strategy is consistent with the student's real-time learning status.

[0014] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the distance education resource recommendation method as described in the first aspect of the present invention are implemented.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: the computer program instructions, when executed by a processor, implement the steps of the distance education resource recommendation method as described in the first aspect of the present invention.

[0016] The beneficial effects of the present invention are as follows: the present invention can generate a personalized recommendation list based on the student's learning time period, frequency, resource preference and other characteristics by introducing a learning behavior feature set in the recommendation process and combining it with a resource index table for dynamic resource grouping and screening. At the same time, the recommendation algorithm proposed by the present invention can be dynamically adjusted based on the student's learning progress and priority score, further improving the recommendation accuracy and user satisfaction. In addition, the present invention solves the problem that traditional technologies are difficult to update in an offline state by synchronizing the latest learning records when the network is restored. Therefore, the method of the present invention can significantly improve the recommendation quality of distance education resources and provide students with more personalized and accurate learning support. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 Flowchart of recommended methods for distance education resources; Figure 2 This is the structural diagram of the distance education resource recommendation system. DETAILED DESCRIPTION

[0019] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0020] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0021] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0022] Example 1, reference Figure 1~Figure 2 , which is the first embodiment of the present invention, and provides a remote education resource recommendation method, comprising: S1: Obtain the downloaded educational resource data and associated learning record information in the student learning terminal, dynamically group the resources according to the content type and content relevance of the learning resources, establish a resource grouping table for offline analysis, and record the corresponding usage probability in each group of resources.

[0023] S1.1: Extract the downloaded educational resource data from the student learning terminal and extract the content feature vector of each resource ,in, For resources No. A content feature set is generated by parsing the metadata of resources (such as subject tags, keywords, and document structure).

[0024] S1.2: Calculate the content relevance between resources based on the extracted content feature set ,in, Representation Resources and resources of relevance.

[0025] Multiple groups of relevance thresholds are preset, resources are dynamically grouped into several groups according to content relevance, and grouping information and resource identifiers are recorded.

[0026] Content relevance between resources The calculation formula is: ; in, For resources and resources The relevance of the content, is the weight of the content feature, used to represent the The contribution of each feature to the content relevance. For resources In the The value of content features, n is the total number of content features, For resources In the The value of a content feature.

[0027] S1.3: Extract the historical usage frequency and cumulative usage time from the students’ learning record information, combine it with the dynamic grouping of resources, and calculate the mean historical usage frequency of each group of resources for subsequent probability adjustment.

[0028] S1.4: Combine usage frequency, cumulative usage time, and resource grouping to calculate the comprehensive usage probability of each resource group , and record it in the grouping table and related parameters.

[0029] Among them, the comprehensive use probability The calculation formula is: ; in, Grouping tables for resources The comprehensive probability of use, For resources The historical frequency of use, For resources The cumulative usage time, Grouping tables for resources The historical average frequency of use, is the nonlinear adjustment parameter, is the offset parameter of the used probability distribution.

[0030] Output resource grouping table after dynamic grouping , including the content feature set and relevance information of each group of resources.

[0031] Integrate dynamic grouping results, content feature tags and usage probability information to generate a resource grouping table, which is then stored in the local database for subsequent recommendation processes.

[0032] S2: Extract students’ recent learning time period, learning frequency, learning resource preference type and usage progress characteristics to form a learning behavior feature set, where the feature set includes time tags, resource type weights and priority scores.

[0033] Specifically, according to the learning record association table, obtain the student's learning record timestamp set, and use the time grouping rules (such as dividing 24 hours a day into three time periods: morning, daytime, and evening) to count the learning frequency and cumulative learning time in each time period, and mark the active time period and learning habit characteristics to generate a time label .

[0034] Get resource usage records from learning records, combined with resource grouping table The content type attribute is used to count the usage frequency and cumulative usage time of each type of resource, and calculate the weight of each resource type. , generate resource type preference feature set .

[0035] The weight of each resource type is The calculation formula is: ; in, and They are the vectors of learning characteristics and resource characteristics, respectively, indicating the degree of match between students’ learning needs and resources.

[0036] Furthermore, by combining the learning records and resource grouping table, the cumulative usage time of each resource is obtained. and total resource duration , computing resource usage progress , and set progress status labels according to the learning stage (such as starting, halfway, and nearly completed); ; Furthermore, according to the time label , Resource Type Weight and learning progress characteristics , combined with the student's most recent learning period and resource type, the priority scoring formula is used to calculate the resource Priority score : ; in, ~ is the weight factor.

[0037] The extracted time labels, resource type weight sets, learning progress labels and priority scores are summarized as the learning behavior feature set , and record it in the feature data table.

[0038] S3: Based on the learning behavior feature set, combined with the resource index table, according to the learning resource preference type and capacity adaptability, select qualified educational resources from the local resource library, generate a recommended resource set based on time sequence, and determine the recommendation priority through a dynamic adjustment algorithm.

[0039] S3.1: Based on the learning behavior feature set The resource type weight in the resource index table is preliminarily screened according to the resource type label, the resource set matching the preferred type is extracted, and the unique identifier of the resource is recorded.

[0040] S3.2: Extract resource capacity data and remaining storage capacity of student terminals from the selected resource set of preferred type , and according to the capacity adaptation formula: ; in, For resources The capacity adaptability of For resources Capacity data.

[0041] Calculate the capacity fit of each resource , remove the fitness below the set threshold resources to generate a capacity adaptation resource set.

[0042] S3.3: Based on learning behavior feature set The learning progress label in the capacity adaptation resource set is used to filter resources in the capacity adaptation resource set according to their usage progress, give priority to resources with consistent progress status (for example, resources that are "near completion" are pushed first), and update the matching resource set.

[0043] S3.4: Update the resource collection according to the resource time tag Sorting, generating a preliminary recommended resource set, and combining it with the learning behavior feature set Priority scoring in , dynamically adjust the preliminary recommended resource set, calculate the final priority score of each resource according to the priority scoring formula, and generate the final recommended resource set after adjusting the sorting.

[0044] The priority scoring formula is: ; in, ~ is the weight coefficient.

[0045] Calculate the final priority score for each resource, adjust the ranking, and generate the final recommended resource set.

[0046] S4: Sort the final recommended resource set by time label, and provide a personalized resource list display on the student terminal interface, marking the resource priority score and estimated usage time to support students' independent selection.

[0047] According to the final recommended resource set generated in step S3 and the time tags of the resources, the resources are sorted in chronological order to ensure that the chronological order of the recommended resources is consistent with the students' learning progress and time requirements, and a sorted resource set is generated after sorting.

[0048] For each educational resource in the sorted resource set, a corresponding priority label is added to each resource in combination with the priority score generated in step S3. In the recommendation display interface, each resource is labeled with a priority score and an estimated usage time.

[0049] Estimated usage time It can be estimated by the following formula: ; in, is the learning completion coefficient of the resource, The resource capacity.

[0050] The sorted recommended resources and their corresponding priority scores and estimated usage time are displayed through the student learning terminal interface. The displayed content includes the resource name, priority score, estimated usage time and description information of the learning resource.

[0051] Through personalized display methods, students can clearly understand the priority of recommended resources and their usage time, thereby supporting students' independent choices.

[0052] S5: When the network connection is restored, the latest learning record information is extracted and compared with the learning behavior feature set generated offline, the recommendation strategy is updated and synchronized to the resource index table to ensure that the recommendation strategy is consistent with the student's real-time learning status.

[0053] Specifically, the latest learning record information will be extracted Compare with the offline generated learning behavior feature set. Calculate the difference between the offline behavior feature set and the real-time behavior feature set through similarity metrics (such as cosine similarity or Euclidean distance): ; in, is the number of dimensions of the feature set; and They are the first and second feature values.

[0054] if If the set threshold is exceeded, the learning behavior is considered to have changed, and the learning behavior feature set must be updated to reflect the student's real-time learning status.

[0055] According to the comparison result, the learning behavior feature set is updated and the resource recommendation priority based on the new feature set is recalculated. The formula can refer to the dynamic adjustment formula in step S3.

[0056] After updating the recommendation strategy, the new learning behavior feature set needs to be synchronized to the resource index table and updated together with the learning record association table to ensure that the recommendation strategy is consistent with the student's real-time learning status and can provide an accurate personalized resource list at the next recommendation.

[0057] Furthermore, this embodiment also provides a remote education resource recommendation system, including: The learning resource data collection module is used to obtain the downloaded educational resource data and related learning record information in the student learning terminal, dynamically group the resources according to the content type and content relevance of the learning resources, establish a resource grouping table for offline analysis, and record the corresponding usage probability in each group of resources; The learning behavior analysis module is used to extract students’ recent learning time period, learning frequency, learning resource preference type and usage progress characteristics to form a learning behavior feature set; A screening and recommendation module is used to screen qualified educational resources from the local resource library according to the learning behavior feature set, combined with the resource index table, the learning resource preference type and capacity adaptability, generate a recommended resource set based on time sequence, and determine the recommendation priority through a dynamic adjustment algorithm; Personalized resource display module, used to provide personalized resource list display on the student terminal interface, marking resource priority rating and estimated usage time to support students' independent selection; The network recovery and synchronization module is used to extract the student's latest learning record information when the network connection is restored, compare it with the learning behavior feature set generated offline, update the recommendation strategy and synchronize it to the resource index table to ensure that the recommendation strategy is consistent with the student's real-time learning status.

[0058] This embodiment also provides a computer device suitable for the remote education resource recommendation method, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the remote education resource recommendation method proposed in the above embodiment.

[0059] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0060] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for recommending remote education resources as proposed in the above embodiment is implemented.

[0061] In summary, the present invention introduces a learning behavior feature set in the recommendation process, and combines the resource index table for dynamic resource grouping and screening, so as to generate a personalized recommendation list based on the student's learning time period, frequency, resource preference and other characteristics. At the same time, the recommendation algorithm proposed in the present invention can be dynamically adjusted based on the student's learning progress and priority score, further improving the recommendation accuracy and user satisfaction. In addition, the present invention solves the problem that traditional technologies are difficult to update in an offline state by synchronizing the latest learning records when the network is restored. Therefore, the method of the present invention can significantly improve the recommendation quality of distance education resources and provide students with more personalized and accurate learning support.

[0062] Example 2, referring to Table 1, is the second example of the present invention. This example provides a method for recommending remote education resources. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0063] In order to verify the effectiveness of the distance education resource recommendation method, the present invention designs an experiment based on simulated data, and the experimental object is the learning terminal data of a student. It is assumed that the student has downloaded several educational resources and recorded relevant learning behavior data in the past period of time. All learning resources are divided into multiple different categories, such as mathematics, physics, English, etc. The experiment is carried out through the following steps: The downloaded educational resource data was extracted from the students' learning terminals, and the content feature vector was extracted for each resource. For each resource, its metadata (such as subject tags, keywords, and document structure) was first parsed to generate a content feature set. Then, the content relevance between resources was calculated, and the cosine similarity of the resource content feature vector was used for calculation, and the resources were dynamically grouped. The resource relevance within each group is higher than the set threshold, and the comprehensive usage probability of each group of resources is calculated based on historical usage frequency, cumulative usage time and other data. This step records the usage probability of each group of resources by establishing a resource grouping table for subsequent offline analysis and recommendation.

[0064] Based on the students' learning records, extract their recent learning time period, learning frequency, learning resource preference type and learning progress characteristics. By analyzing the learning timestamp, count the learning frequency and duration in each time period to form a time label. At the same time, count the frequency and cumulative usage time of students on different types of resources, calculate the resource type preference weight, and generate a progress label based on the learning progress. Summarize these data into the student's learning behavior feature set and record it in the feature data table.

[0065] Combine the learning behavior feature set with the resource index table, and select resources according to resource type, capacity adaptability, learning progress and other conditions. According to the student's learning progress and resource capacity adaptability, select suitable resources from the local resource library, and generate the final recommended resource set through the priority scoring algorithm. Resources are arranged in order of recommendation priority to ensure the rationality of the recommendation order and match the current needs of students.

[0066] According to the generated final recommended resource set, the resources are arranged in chronological order and displayed through the student learning terminal interface. In the display interface, each recommended resource is marked with a priority score and an estimated usage time to help students better choose and arrange their studies independently.

[0067] When the network connection is restored, the student's latest learning record information is extracted and compared with the previously generated offline learning behavior feature set to calculate the difference. If the student's learning behavior has changed significantly, the learning behavior feature set is updated and the recommendation priority is recalculated. Finally, the updated recommendation strategy is synchronized to the resource index table to ensure the accuracy of the next recommendation. The following table records the key parameters and data involved in the simulation experiment. Through this table, the difference in effect between the present invention and the traditional recommendation method can be intuitively compared: Table 1 Experimental data table

[0068] From the above table, we can see that the number and priority rating of recommended resources are closely related to students' learning behavior characteristics (such as learning time period, learning frequency, and resource type preference). For example, student S001's learning behavior in the morning, daytime, and evening are fully considered, and the number and priority rating of recommended resources reflect his learning habits and resource demand preferences. Based on his historical learning records and preferences, the recommendation system can dynamically adjust the recommended order and quantity of resources to ensure personalized and targeted recommendations. Compared with traditional static recommendation systems, this dynamic adjustment based on learning behavior characteristics and progress significantly improves the accuracy and adaptability of recommendations.

[0069] Specifically, the dynamic grouping and priority scoring algorithms in the invention content have significant advantages. By introducing multi-dimensional features such as historical usage frequency, study time period, resource type preference, etc., the personalized needs of students can be accurately reflected when recommending resources. For example, in the case of student S001, the resource recommendation priority score for the morning study period is 95, which is much higher than other time periods, showing the system's advantage in automatically adjusting the recommendation priority according to the student's active time period. At the same time, the resource type preference weight of student S002 is relatively low, and the system can reasonably reduce the number of recommendations and adjust the priority to ensure that the recommendation results are not overly redundant.

[0070] Different from traditional recommendation algorithms based on collaborative filtering or simple content-based recommendation algorithms, the dynamic adjustment and capacity adaptation strategies proposed in this paper can be continuously optimized according to the real-time learning status of students. This optimization process is not only reflected in the personalization of recommendations, but also in real-time synchronous updates. When the network is restored, the system can accurately update the student's learning behavior feature set and adjust the resource recommendation strategy accordingly. This mechanism greatly improves the real-time and accuracy of the recommendation system.

[0071] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A distance education resource recommendation method, characterized in that: include: S1: Obtain the downloaded educational resource data and associated learning record information in the student learning terminal, dynamically group the resources according to the content type and content relevance of the learning resources, establish a resource grouping table for offline analysis, and record the corresponding usage probability in each group of resources; The dynamic grouping of resources includes: calculating the content relevance between resources, presetting multiple groups of relevance thresholds, and dynamically grouping the resources into a plurality of groups according to the content relevance; S2: Extract students’ recent learning time period, learning frequency, learning resource preference type and usage progress characteristics to form a learning behavior feature set, where the feature set includes time tags, resource type weights and priority scores; S3: Based on the learning behavior feature set, combined with the resource index table, and according to the learning resource preference type and capacity adaptability, qualified educational resources are screened from the local resource library, a recommended resource set based on time sequence is generated, and a recommendation priority is determined through a dynamic adjustment algorithm; S4: Provide a personalized resource list display on the student terminal interface, mark the resource priority score and estimated usage time to support students' independent selection; S5: When the network connection is restored, the student's latest learning record information is extracted and compared with the learning behavior feature set generated offline, the recommendation strategy is updated and synchronized to the resource index table to ensure that the recommendation strategy is consistent with the student's real-time learning status.

2. The remote education resource recommendation method according to claim 1, characterized in that: The S1 specifically includes the following steps: Extract the downloaded educational resource data from the student learning terminal and extract the content feature vector of each resource ,in, For resources No. Content feature values ​​are generated by parsing the metadata of resources to generate a content feature set; Calculate the content relevance between resources based on the extracted content feature set ,in, Representation Resources and resources relevance; Extract historical usage frequency and cumulative usage time from students’ learning record information, combine with dynamic resource grouping, and calculate the average historical usage frequency of each resource group; Combine the average frequency of use, cumulative usage time, and dynamic resource grouping to determine the comprehensive usage probability of each group of resources .

3. The remote education resource recommendation method according to claim 2, characterized in that: The content relevance between the resources The calculation formula is: ； in, For resources and resources The relevance of the content, is the weight of the content feature, used to represent the The contribution of each feature to the content relevance. For resources In the The value of content features, n is the total number of content features, For resources In the The value of a content feature.

4. The remote education resource recommendation method according to claim 3, characterized in that: The comprehensive utilization probability of each group of resources The calculation formula is: ； in, Grouping tables for resources The comprehensive probability of use, For resources The historical frequency of use, For resources The cumulative usage time, Grouping tables for resources The historical average frequency of use, is the nonlinear adjustment parameter, is the offset parameter of the used probability distribution.

5. The remote education resource recommendation method according to claim 4, characterized in that: Determining the recommendation priority order by dynamically adjusting the algorithm includes: The updated resource collection is sorted by resource time tag Sorting, generating a preliminary recommended resource set, and combining it with the learning behavior feature set Priority scoring in , dynamically adjust the preliminary recommended resource set, and calculate the final priority score of each resource according to the priority score formula; The priority scoring formula is: ； in, ~ is the weight coefficient, For resources Capacity adaptability.

6. The remote education resource recommendation method according to claim 5, characterized in that: The comparing with the learning behavior feature set generated offline comprises: The latest learning record information will be extracted Compare with the offline generated learning behavior feature set, and calculate the difference between the offline behavior feature set and the real-time behavior feature set through similarity measurement. ; if If the set threshold is exceeded, the learning behavior is considered to have changed, and the learning behavior feature set must be updated to reflect the student's real-time learning status.

7. The remote education resource recommendation method according to claim 6, characterized in that: The difference The calculation formula is as follows: ； in, is the number of dimensions of the feature set; and They are the first and second feature values.

8. A distance education resource recommendation system, based on the distance education resource recommendation method according to any one of claims 1 to 7, characterized in that: include: The learning resource data collection module is used to obtain the downloaded educational resource data and related learning record information in the student learning terminal, dynamically group the resources according to the content type and content relevance of the learning resources, establish a resource grouping table for offline analysis, and record the corresponding usage probability in each group of resources; The learning behavior analysis module is used to extract students’ recent learning time period, learning frequency, learning resource preference type and usage progress characteristics to form a learning behavior feature set; A screening and recommendation module is used to screen qualified educational resources from the local resource library according to the learning behavior feature set, combined with the resource index table, the learning resource preference type and capacity adaptability, generate a recommended resource set based on time sequence, and determine the recommendation priority through a dynamic adjustment algorithm; The personalized resource display module is used to provide a personalized resource list display on the student terminal interface, marking the resource priority score and estimated usage time to support students' independent selection; The network recovery and synchronization module is used to extract the student's latest learning record information when the network connection is restored, compare it with the learning behavior feature set generated offline, update the recommendation strategy and synchronize it to the resource index table to ensure that the recommendation strategy is consistent with the student's real-time learning status.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the distance education resource recommendation method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the distance education resource recommendation method described in any one of claims 1 to 7 are implemented.

Images