Multi-scene interaction control method and system in IPTV education platform

Abstract

The invention relates to the technical field of data processing, and discloses a multi-scene interaction control method and system in an IPTV education platform. The method comprises the steps that a user behavior data set is generated through educational behavior Latin hypercube sampling, a standardized educational scene type is obtained based on three-dimensional similarity calculation, an educational space-time hypergraph is constructed to extract scene feature vectors, a resource optimization model with the minimum delay cost as the target is established, and a standardized educational scene is obtained. And pre-allocation of network bandwidth and server resources and rapid scene switching scheduling are realized. The problem that multi-scene interaction control in an IPTV education platform lacks specialized sampling, modeling and optimization technologies is solved. According to the invention, the response speed of education scene switching and the intelligent level of resource allocation are improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a multi-scenario interactive control method and system in an IPTV education platform. Background Technology

[0002] With the rapid development of digital education, IPTV education platforms have been widely used as an important means of remote teaching. Existing IPTV education platforms typically employ a scene management approach based on channel switching, organizing different educational content into channels. Users can switch between different educational scenes using a remote control or a simple interface. Traditional scene control methods mainly rely on preset scene templates and fixed resource allocation strategies. They provide scene navigation through the EPG system, utilize CDN networks for content distribution, and employ QoS mechanisms to ensure basic network service quality. These technologies offer a certain degree of stability and reliability in single-scene applications.

[0003] However, existing technologies have significant shortcomings in handling concurrent and dynamic switching across multiple scenarios. First, the scenario switching response latency is a prominent issue. Traditional methods require re-establishing network connections and reallocating server resources, resulting in scenario switching times of 3-5 seconds, which seriously affects the continuity of teaching. Second, the resource scheduling algorithm lacks intelligence. Existing systems use a static average allocation strategy, which cannot be dynamically optimized according to the importance and real-time needs of different educational scenarios, resulting in key educational scenarios such as online examinations not receiving priority. Third, the scenario relationship modeling is overly simplified. Traditional graph structures can only represent pairwise connections between scenarios and cannot accurately describe the complex teaching logic relationships between multiple educational scenarios, leading to inaccurate scenario recommendation and prediction. Summary of the Invention

[0004] This application provides a multi-scene interactive control method and system for IPTV education platforms, addressing the lack of specialized sampling, modeling, and optimization techniques for multi-scene interactive control in IPTV education platforms. This application improves the response speed of educational scene switching and the level of intelligent resource allocation.

[0005] Firstly, this application provides a multi-scene interactive control method for an IPTV education platform, the multi-scene interactive control method for the IPTV education platform comprising: Step S1: Using Latin hypercube sampling of educational behavior, IPTV user behavior data is stratified and sampled according to educational scenario weight factors and learning progress correlation coefficients to generate an educational scenario sampling dataset. Step S2: Calculate the cluster centers of the educational scenario sampling dataset according to content similarity, time similarity, and user similarity to obtain a standardized set of educational scenario types; Step S3: Construct an educational spatiotemporal hypergraph using the standardized educational scenario type set as nodes, wherein hyperedges connect scenario combinations with teaching logic associations to obtain scenario feature vector representations; Step S4: Based on the scene feature vector representation, establish a resource optimization model with the objective function of minimizing the product of scene response latency and education priority, and determine the IPTV scene resource allocation scheme; Step S5: Pre-allocate network bandwidth and server resources according to the IPTV scenario resource allocation scheme, and perform resource scheduling when responding to user scenario switching requests.

[0006] Secondly, this application provides a multi-scene interactive control system for an IPTV education platform, the multi-scene interactive control system for the IPTV education platform comprising: The sampling module is used to perform Latin hypercube sampling of educational behavior, which stratifies IPTV user behavior data according to educational scenario weight factors and learning progress correlation coefficients to generate an educational scenario sampling dataset. The calculation module is used to calculate the cluster centers of the educational scenario sampling dataset according to content similarity, time similarity and user similarity, so as to obtain a standardized set of educational scenario types; The construction module is used to construct an educational spatiotemporal hypergraph with the standardized set of educational scenario types as nodes, wherein the hyperedges connect scenario combinations with teaching logic associations to obtain scenario feature vector representations; A module is established to build a resource optimization model based on the scene feature vector representation, with the objective function being to minimize the product of scene response delay and education priority, and to determine the IPTV scene resource allocation scheme; The response module is used to pre-allocate network bandwidth and server resources according to the IPTV scenario resource allocation scheme, and to perform resource scheduling when responding to user scenario switching requests.

[0007] Thirdly, a multi-scene interactive control device for an IPTV education platform is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the multi-scene interactive control device for the IPTV education platform to execute the aforementioned multi-scene interactive control method for the IPTV education platform.

[0008] Fourthly, a computer-readable storage medium is provided, wherein instructions are stored in the computer-readable storage medium, which, when executed on a computer, cause the computer to perform the aforementioned multi-scenario interactive control method in the IPTV education platform.

[0009] The technical solution provided in this application effectively solves the technical problem that traditional random sampling cannot distinguish the importance of educational scenarios by adopting the Latin hypercube sampling technology for educational behavior. This sampling method introduces educational scenario weight factors and learning progress correlation coefficients, enabling stratified sampling of user behavior data according to the specific application needs of the IPTV education platform. This ensures that high-importance scenarios, such as live classroom broadcasts, receive more sampling attention, while also taking into account users' personalized learning progress. The generated educational scenario sampling dataset more accurately reflects the actual characteristics of IPTV education applications. The three-dimensional clustering algorithm based on content similarity, time similarity, and user similarity overcomes the limitations of traditional clustering methods with their single dimension. By quantifying the similarity differences between the same knowledge point, related knowledge points, and different subject scenarios, and combining time window overlap analysis and user group characteristic comparison, it achieves accurate identification and classification of educational scenario types. The resulting standardized educational scenario type set lays a solid foundation for subsequent scenario relationship modeling. The construction of the educational spatiotemporal hypergraph structure breaks through the technical limitation of traditional graph structures that can only represent paired connections. By connecting multiple scene nodes with teaching logic through hyperedges, it accurately describes the complex multi-scene relationships in the IPTV education platform. Combined with the educational scene mask matrix and hypergraph neural network encoding, the generated scene feature vector representation contains multi-dimensional information such as scene content, time, space and association.

[0010] The resource optimization model employs an innovative approach where the objective function minimizes the product of scenario response latency and educational priority. This effectively balances the importance and performance requirements of different educational scenarios. By setting differentiated priority parameters for scenarios such as online exams, live classroom broadcasts, interactive Q&A, and resource downloads, it ensures that key educational activities receive priority resource guarantees. Simultaneously, the established bandwidth constraints guarantee the feasibility of the resource allocation scheme and the stable operation of the system. The education scenario-oriented optimization algorithm EDU-NGO demonstrates significant advantages in IPTV education platform applications. This algorithm is specifically optimized for the functional requirements of educational scenarios. Through an evolutionary mechanism similar to that of eagle populations and an elite retention strategy, it can efficiently search for optimal solutions in a complex multi-constraint optimization space. In particular, the key educational scenario detection and automatic weight adjustment mechanisms enable the algorithm to dynamically respond to real-time changes in educational activities, ensuring that key scenarios such as online exams always receive sufficient resource allocation. The pre-allocated resource scheduling mechanism completely changes the inefficient mode of traditional IPTV systems where scene switching requires re-applying for resources. By reserving network bandwidth according to priority and allocating computing resources according to scene type, the constructed scene resource mapping table realizes the direct mapping from scene identifier to network channel and server node. When a user initiates a scene switching request, the pre-allocated resources can be quickly activated through resource switching instructions, which significantly shortens the scene response time and improves the user experience and teaching effectiveness of the IPTV education platform. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of an embodiment of the multi-scene interactive control method in the IPTV education platform according to the present application. Figure 2 This is a schematic diagram of an embodiment of the multi-scene interactive control system in the IPTV education platform of this application; Figure 3 This is a schematic block diagram of the structure of the multi-scene interactive control device in the IPTV education platform in this embodiment of the invention. Detailed Implementation

[0013] This application provides a multi-scene interactive control method and system for an IPTV education platform. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0014] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the multi-scene interactive control method in the IPTV education platform in this application includes: Step S1: Using Latin hypercube sampling of educational behavior, IPTV user behavior data is stratified and sampled according to educational scenario weight factors and learning progress correlation coefficients to generate an educational scenario sampling dataset. Step S2: Calculate the cluster centers of the educational scenario sampling dataset based on content similarity, time similarity, and user similarity to obtain a standardized set of educational scenario types; Step S3: Construct an educational spatiotemporal hypergraph using a standardized set of educational scenario types as nodes, where hyperedges connect scenario combinations with teaching logic connections to obtain scenario feature vector representations; Step S4: Based on the scene feature vector representation, establish a resource optimization model with the objective function of minimizing the product of scene response latency and education priority, and determine the IPTV scene resource allocation scheme; Step S5: Pre-allocate network bandwidth and server resources according to the IPTV scenario resource allocation scheme, and perform resource scheduling when responding to user scenario switching requests.

[0015] It is understood that the executing entity of this application can be a multi-scene interactive control system in an IPTV education platform, or it can be a terminal or a server; the specific implementation is not limited here. This application's embodiment uses a server as an example for illustration.

[0016] Specifically, the Latin hypercube sampling method for educational behavior is an improved multidimensional data sampling technique optimized for user behavior characteristics of IPTV education platforms. IPTV user behavior data is divided into a five-dimensional hypercube space based on user ID, timestamp, scene type, network status, and device type, forming a user behavior space matrix. Then, each dimension is uniformly divided into k intervals to ensure an equal number of sampling points, generating a uniformly distributed sampling grid. The educational scene weight factor α and the learning progress correlation coefficient β are the core parameters of this invention. α is used to distinguish the importance of different educational scenes: the weight of a live classroom scene is set to 1.0 times α, the weight of an interactive quiz scene is set to 0.8 times α, and the weight of a resource download scene is set to 0.6 times α, forming a scene weight distribution matrix. The weight of each sampling point is calculated by multiplying the scene priority by α and the learning progress correlation by β, generating an educational scene sampling dataset.

[0017] The EDU-ISODATA clustering algorithm was used to identify scene types in the sampled data. This algorithm achieves accurate clustering through three similarity calculations. In the content similarity calculation, the similarity value for scenes with the same knowledge point was set to 1.0, for scenes with related knowledge points it was set to 0.7, and for scenes with different subjects it was set to 0.2, forming a content similarity matrix. Temporal similarity was determined by the ratio of the overlapping part of the time window to the union of the time windows, reflecting the user's usage patterns of educational scenes in different time periods, generating a temporal similarity matrix. User similarity was quantified by comprehensively considering factors such as grade, subject, and learning ability, reflecting the behavioral characteristics of different user groups, forming a user similarity matrix. Finally, the three similarity matrices were weighted and fused with weights of 0.5, 0.3, and 0.2, and the positions of 8 cluster centers were calculated through an iterative clustering algorithm, corresponding to standardized educational scene types such as live classroom teaching, interactive Q&A, resource download, online exams, group discussions, experimental demonstrations, course playback, and homework submission.

[0018] An educational spatiotemporal hypergraph structure is constructed, with eight educational scenarios set as hypergraph nodes. Hyperedges connect scenarios with logical connections in teaching. The advantage of hypergraphs over traditional graph structures lies in their ability to represent higher-order relationships between multiple nodes. Live classroom streaming, interactive Q&A, and course replay form the first hyperedge representing the teaching process, while online exams, experimental demonstrations, and assignment submission form the second hyperedge representing the evaluation and verification process. The processing of the educational scenario mask matrix is ​​a key technology. When two scenarios have a logical connection in teaching, the mask value is set to 0; otherwise, it is set to negative infinity, forming a scenario association mask matrix. The hypergraph neural network uses multi-layer graph convolutional encoding. Temporal convolution extracts temporal features of the scenarios, and spatial convolution extracts geographical distribution features, generating a 128-dimensional scenario feature vector representation that includes content features, temporal features, user features, and association features.

[0019] A resource optimization model is established, and priority parameters for educational scenarios are set based on scenario feature vector representations. The priority for online exams is set to 10, for live classroom sessions to 8, for interactive Q&A sessions to 6, and for resource downloads to 4, forming an educational priority parameter matrix. The priority parameter matrix is ​​multiplied by scenario response latency data to calculate the latency cost of each educational scenario, generating a scenario latency cost matrix. The constrained optimization model uses minimizing the sum of latency costs across all scenarios as its objective function, while setting a constraint that the total bandwidth must be greater than or equal to the sum of the products of bandwidth requirements for each scenario and the number of students. The Education Scenario-Oriented Optimization Algorithm (EDU-NGO) initializes an eagle population, with each eagle containing bandwidth allocation ratios and server allocation strategy parameters for eight educational scenarios. It searches the optimal solution space through fitness evaluation, position updates, and elite retention mechanisms, outputting an IPTV scenario resource allocation scheme after 200 iterations.

[0020] Based on the IPTV scenario resource allocation scheme, network bandwidth and server resources are pre-allocated. Network bandwidth pre-allocation prioritizes and reserves bandwidth requirements for each educational scenario to the corresponding network channels. Server resource pre-allocation distributes computing resources to edge nodes and cloud servers according to scenario type. A constructed resource scheduling mapping table records the network channel identifier and server node identifier corresponding to each educational scenario. When a user scenario switching request is received, the pre-allocated resources for the target scenario are found according to the mapping relationship, and the corresponding network channel and server node are activated through a resource switching command.

[0021] In one specific embodiment, step S1 includes: IPTV user behavior data is divided into an n-dimensional hypercube space according to user ID, timestamp, scene type, network status, and device type to obtain the user behavior space matrix; The user behavior space matrix is ​​uniformly divided into k intervals in each dimension, ensuring that the number of sampling points in each interval is equal, thus obtaining a uniformly distributed sampling grid. Differential weights are assigned to different educational scenarios in a uniformly distributed sampling grid based on the educational scenario weight factor α. The weight of the classroom live broadcast scenario is 1.0 times that of α, the weight of the interactive Q&A scenario is 0.8 times that of α, and the weight of the resource download scenario is 0.6 times that of α, thus obtaining the scenario weight distribution matrix. By combining the learning progress correlation coefficient β with the scene weight distribution matrix, the weight of each sampling point is calculated by adding the product of scene priority and α to the product of progress correlation and β, thus obtaining the educational scene sampling dataset.

[0022] Specifically, IPTV user behavior data is spatially divided according to five dimensions: User ID, which records each student's unique identifier; Timestamp, which records the specific time point when the user accesses the educational scenario; Scenario Type, which identifies the category of the currently used educational scenario; Network Status, which reflects the user's current network connection quality and bandwidth; and Device Type, which distinguishes the user's terminal device, such as a set-top box, tablet, or mobile phone. These five dimensions constitute an n-dimensional hypercube space, where n equals 5. Each dimension represents a coordinate axis. Each user action is recorded as a data point in this five-dimensional space, and the set of all data points constitutes the user behavior space matrix.

[0023] When uniformly partitioning the user behavior space matrix, each dimension is divided into k equally spaced intervals to ensure that each interval contains an equal number of sampling points. Specifically, the user ID dimension is partitioned according to student number order, the timestamp dimension is partitioned according to the 24 hours of a day, the scenario type dimension is partitioned according to eight educational scenario types, the network status dimension is partitioned according to bandwidth range from low to high, and the device type dimension is partitioned according to different terminal devices. The boundary values ​​of each interval are determined by calculating the minimum and maximum values ​​of the data in that dimension, followed by equally spaced partitioning, forming a uniformly distributed sampling grid. This grid structure ensures that sampling bias is not caused by uneven data distribution in any dimension.

[0024] The differentiated weighting of the education scenario weight factor α is set based on the importance of different scenarios within the IPTV education platform. The live classroom scenario, as the core teaching activity, has the highest requirements for real-time performance and interactivity; therefore, its weight is set to 1.0 times α, indicating that this scenario enjoys the highest resource allocation priority. The interactive Q&A scenario requires real-time interaction between students and teachers but has a relatively high tolerance for network latency; its weight is set to 0.8 times α. The resource download scenario mainly involves the acquisition of teaching materials and has the lowest real-time requirements; its weight is set to 0.6 times α. During the weighting process, each sampling point in the uniformly distributed sampling grid is matched with a corresponding weight coefficient based on its scenario type dimension value. This weight coefficient is then multiplied by α to obtain the weight value of that sampling point in the scenario dimension. The weight values ​​of all sampling points form the scenario weight distribution matrix.

[0025] The calculation of the learning progress correlation coefficient β and the scene weight distribution matrix involves a complex weight fusion process. The learning progress correlation reflects the degree of matching between the user's current learning status and a specific educational scene, calculated by analyzing the user's historical learning records, current course progress, and knowledge mastery. In the weight calculation process for each sampling point, the scene priority value corresponding to that sampling point is extracted, and then multiplied by α to obtain the scene weight. Next, the progress correlation value of that sampling point is extracted, and multiplied by β to obtain the progress weight. Finally, the scene weight and the progress weight are added together to obtain the weight of that sampling point. This weight fusion mechanism ensures that the sampling process considers both the objective importance of the scene and the user's personalized learning needs.

[0026] In one specific embodiment, step S2 includes: Content similarity was calculated for scene data in the educational scene sampling dataset. The similarity value of scenes with the same knowledge point was 1.0, the similarity value of scenes with related knowledge points was 0.7, and the similarity value of scenes with different subjects was 0.2, resulting in a content similarity matrix. The time similarity is calculated based on the time window information in the educational scenario sampling dataset. The time similarity value is determined by the ratio of the overlapping part of the time window to the union of the time windows, and a time similarity matrix is ​​obtained. User similarity is calculated based on the user group characteristics in the educational scenario sampling dataset to obtain a user similarity matrix; The content similarity matrix, time similarity matrix, and user similarity matrix are weighted and fused with weights of 0.5, 0.3, and 0.2, respectively. The positions of the eight cluster centers are calculated by an iterative clustering algorithm to obtain a set of standardized educational scenario types.

[0027] Specifically, accurate scene type identification is achieved through three similarity calculation methods. Content similarity calculation quantifies the degree of association between knowledge points in educational scenes. Knowledge point label information for each scene is extracted from the educational scene sampling dataset, and a knowledge point association database is established, recording the logical relationships between different knowledge points. When two educational scenes contain completely identical knowledge points, such as an explanation of mathematical function concepts and function exercises, the content similarity value is set to 1.0, representing the highest similarity. When two scenes involve related knowledge points, such as basic algebra and introductory geometry both belonging to the category of basic mathematics, the content similarity value is set to 0.7, representing moderate similarity. When two scenes belong to completely different subjects, such as physics experiments and Chinese reading, the content similarity value is set to 0.2, representing low similarity. In constructing the content similarity matrix, all scenes are paired up, and the similarity value is determined by looking up a table based on their knowledge point association, forming a symmetric matrix structure. Each element in the matrix represents the degree of content similarity between the corresponding scene pairs.

[0028] Time similarity calculation is based on the analysis of the overlap of user time windows in educational scenarios. A time window refers to the range of time during which users concentrate on using a certain educational scenario. The overlap portion of the time window is calculated by comparing the active time periods of two scenarios. Specifically, the start and end times of each educational scenario are determined, and the intersection of the two time periods is identified as the overlap time. The union of time windows refers to the total coverage of the usage time periods of the two scenarios, calculated by merging the two time periods and removing the overlap. The time similarity value is equal to the overlap time length divided by the union time length; this ratio reflects the similarity of the two educational scenarios in terms of time usage patterns. In the process of generating the time similarity matrix, all scenario combinations in the educational scenario sampling dataset are traversed, the time window overlap ratio of each pair of scenarios is calculated, and the calculation results are filled into the corresponding positions in the matrix to form a data structure reflecting the similarity of time usage patterns.

[0029] User similarity calculation comprehensively considers multiple dimensions of user group characteristics, including grade level, subject, and learning ability. Grade level is categorized by educational stage, such as primary school, middle school, and high school. Subject level is divided based on the user's primary learning area, such as science-oriented, humanities-oriented, or integrated learning categories. Learning ability is assessed by analyzing the user's historical learning performance and test scores, including indicators such as learning speed, comprehension, and memory. In the user similarity calculation process, a weight coefficient is assigned to each user group characteristic. Then, the degree of difference between the main user groups in two educational scenarios across each characteristic dimension is calculated. Finally, the differences across dimensions are weighted and summed to obtain a comprehensive similarity value. The user similarity matrix records the similarity of user groups across all scenario pairs; higher matrix element values ​​indicate greater similarity between the target user groups of the two scenarios.

[0030] The weighted fusion process linearly combines three similarity matrices according to preset weights: content similarity weight is set to 0.5, time similarity weight to 0.3, and user similarity weight to 0.2. The weight allocation is determined based on the influence of different factors on scene clustering within the IPTV education platform. In the fusion calculation, matrix elements at corresponding positions are multiplied by their respective weights, and the three products are then summed to obtain the comprehensive similarity matrix. The iterative clustering algorithm classifies scenes based on the comprehensive similarity matrix. During algorithm initialization, eight cluster center locations are randomly selected, corresponding to eight main educational scene types: live classroom broadcasts, interactive Q&A, resource downloads, online exams, group discussions, experimental demonstrations, course playback, and homework submission. During iteration, the comprehensive similarity distance from each scene data point to each cluster center is calculated, and the data point is assigned to the cluster center with the highest similarity. The cluster center locations are then recalculated based on the new clustering results, and this process is repeated until the cluster center locations stabilize or the maximum number of iterations is reached.

[0031] In one specific embodiment, step S3 includes: Eight educational scenarios from the standardized educational scenario type set are set as hypergraph nodes, including classroom live streaming, interactive Q&A, resource download, online exam, group discussion, experiment demonstration, course playback, and homework submission nodes, resulting in a set of educational scenario nodes; Hyperedge connections are constructed based on the set of educational scenario nodes. Each hyperedge connection has multiple scenario nodes with teaching logic. Classroom live broadcast, interactive Q&A, and course replay form the first hyperedge, and online exams, experimental demonstrations, and homework submission form the second hyperedge, resulting in an educational spatiotemporal hypergraph structure. The scene nodes in the education spatiotemporal hypergraph structure are processed by an education scene mask matrix. When two scenes have a teaching logic relationship, the mask value is set to 0, and when there is no relationship, the mask value is set to negative infinity, thus obtaining the scene relationship mask matrix. The scene association mask matrix is ​​input into a hypergraph neural network for multi-layer graph convolutional encoding. Temporal features of the scene are extracted through temporal convolution and geographical distribution features of the scene are extracted through spatial convolution, resulting in a scene feature vector representation.

[0032] Specifically, the construction process of the educational spatiotemporal hypergraph structure in this invention is specifically designed for the multi-scene association characteristics of IPTV education platforms, solving the technical problem that traditional graph structures cannot represent high-order relationships between multiple scenes. The establishment of the educational scene node set is based on eight core scene types identified from a standardized set of educational scene types. Each scene type is abstracted as a node in the hypergraph: classroom live streaming nodes represent real-time teaching activities, interactive Q&A nodes represent teacher-student interaction, resource download nodes represent learning material acquisition, online exam nodes represent learning assessment activities, group discussion nodes represent collaborative learning, experimental demonstration nodes represent practical teaching activities, course replay nodes represent review and consolidation, and homework submission nodes represent the display of learning outcomes. During node setup, each node is assigned a unique identifier and attribute information. The attribute information includes characteristic data such as the scene's functional type, resource requirements, and user interaction patterns. The initial connections between nodes are temporarily empty, awaiting the establishment of subsequent hyperedge connections.

[0033] The construction of hyperedge connections is based on grouping according to the logical relevance of teaching. Hyperedges are a core concept in hypergraph structures. Unlike edges in traditional graphs that can only connect two nodes, hyperedges can connect multiple nodes simultaneously, indicating complex multi-faceted relationships between them. The first hyperedge connects three nodes: live classroom streaming, interactive quizzes, and course replays. This connection reflects the logical teaching process: live classroom streaming provides the main teaching content, interactive quizzes verify learning outcomes, and course replays support review and consolidation. These three scenarios form a closed-loop structure in the teaching process. The second hyperedge connects three nodes: online exams, experimental demonstrations, and assignment submissions. This reflects the logical relationship of learning assessment and verification. Experimental demonstrations showcase the practical application of theoretical knowledge, online exams assess knowledge mastery, and assignment submissions record learning outcomes. These three scenarios together constitute the learning assessment system. The weight of the hyperedge is determined by calculating the strength of the teaching relevance between the connected nodes. The relevance strength is calculated based on factors such as the frequency of use, temporal order, and content dependencies between scenarios. A higher weight indicates a tighter logical connection between the nodes within the hyperedge.

[0034] The educational scene masking matrix is ​​a key data preprocessing step in hypergraph neural networks. The masking matrix controls information propagation between different scene nodes in the attention mechanism. In constructing the masking matrix, an 8×8 matrix structure is established, with rows and columns corresponding to eight different educational scene nodes. Masking values ​​are then filled in according to the teaching logic relationships between the nodes. When two scene nodes are located within the same hyperedge, it indicates a direct teaching logic relationship, and the corresponding masking value is set to 0. This value allows the attention mechanism to exchange information normally between these two nodes. When two scene nodes are not located within any shared hyperedge, it indicates a lack of direct teaching logic relationship, and the corresponding masking value is set to negative infinity. This value is converted to a near-zero probability in the softmax calculation, thus preventing information interference between unrelated scenes. All diagonal elements of the masking matrix are set to 0, representing the relationship between each scene node and itself. The scene association masking matrix is ​​a symmetric matrix, reflecting the bidirectional logical relationships between scenes.

[0035] The multi-layer graph convolutional encoding process of a hypergraph neural network uses a scene association mask matrix as input for feature extraction. Graph convolution is a deep learning technique specifically designed for processing graph-structured data, capable of updating the feature representation of the current node by aggregating information from neighboring nodes. Temporal convolution specifically extracts the temporal features of the scene, generating feature vectors reflecting the scene's temporal behavior by analyzing time-related attributes such as usage patterns, switching frequency, and duration of the scene across different time periods. Spatial convolution specifically extracts the scene's geographical distribution features, generating feature vectors reflecting the scene's spatial distribution by analyzing spatial attributes such as differences in user preferences for scenes in different regions, the impact of network conditions on scene performance, and the requirements of device type for scene adaptation. In the multi-layer convolution process, the first convolutional layer extracts basic node features and neighbor information; the second convolutional layer further aggregates neighbor information from a wider range based on the first layer; and the third convolutional layer forms global scene association features, outputting a 128-dimensional scene feature vector representation. This vector contains information from four dimensions: content features, temporal features, spatial features, and association features of the scene.

[0036] Taking the construction of a hypergraph for a high school chemistry course IPTV education platform as an example, the standardized educational scene type set includes four core scenes: chemistry theory explanation, experimental demonstration, exercise practice, and mock exams. When establishing the educational scene node set, the attribute information of the chemistry theory explanation node includes theoretical knowledge transmission function, high bandwidth requirement, and one-way interaction mode; the experimental demonstration node includes practical operation demonstration function, ultra-high-definition video requirement, and observation and interaction mode. In the construction of hyperedge connections, the first hyperedge connects the chemistry theory explanation and exercise practice nodes, reflecting the logical relationship between theoretical learning and practical application; the second hyperedge connects the experimental demonstration and mock exam nodes, reflecting the correlation between experimental observation and ability testing. When processing the educational scene mask matrix, the chemistry theory explanation and exercise practice nodes are located within the same hyperedge, so the mask value is set to 0; while the chemistry theory explanation and mock exam nodes are not located within the same hyperedge, so the mask value is set to negative infinity. In the convolutional encoding process of hypergraph neural networks, the temporal dimension convolution reveals that chemical theory explanations are mainly used during morning classes, while experimental demonstrations are concentrated during afternoon lab sessions. The spatial dimension convolution analysis reveals that the network requirements for experimental demonstration scenarios differ between laboratory environments and ordinary classroom environments. The generated scene feature vectors accurately reflect the multi-dimensional features and complex relationships of chemical education scenarios, solving the technical problem that traditional methods cannot handle high-order dependencies between multiple scenarios.

[0037] In one specific embodiment, step S4 includes: Based on the scene feature vector representation, the priority parameters of the education scene are set, where the priority of the online examination scene is set to 10, the priority of the live classroom scene is set to 8, the priority of the interactive Q&A scene is set to 6, and the priority of the resource download scene is set to 4, thus obtaining the education priority parameter matrix. The delay cost matrix of each educational scenario is calculated by multiplying the education priority parameter matrix with the scenario response delay data. Based on the scenario delay cost matrix, a resource optimization objective function is constructed, with the goal of minimizing the sum of the delay costs of all scenarios. At the same time, a constraint condition is set that the total bandwidth is greater than or equal to the sum of the products of the bandwidth requirements of each scenario and the number of students, thus obtaining a constrained optimization model. The constrained optimization model is input into the education scenario-oriented optimization algorithm for iterative solution. The optimal solution space is searched through population evolution and elite retention strategies to obtain the IPTV scenario resource allocation scheme.

[0038] Specifically, the priority parameters for educational scenarios are quantified based on importance indicators extracted from the scenario feature vector representation. The priority values ​​reflect the criticality of different educational scenarios in the teaching process and their urgent need for network resources. The priority of online examination scenarios is set to 10. This value is determined based on the real-time requirements and non-repeatable nature of examinations; any delay or interruption during the examination will directly affect student performance evaluation, thus requiring the highest level of resource support. The priority of live-streaming classroom scenarios is set to 8. This value considers the real-time requirements of live-streaming and the continuity of teacher-student interaction. Live-streaming delays will affect teaching effectiveness but have a certain degree of tolerance compared to examinations. The priority of interactive quiz scenarios is set to 6. Although quiz activities require real-time response, the impact of a single delay is relatively small, and students have time to wait for the network to recover. The priority of resource download scenarios is set to 4. Download activities have the lowest real-time requirements; delays mainly affect user experience without affecting core teaching functions. In constructing the educational priority parameter matrix, the priority values ​​of the eight educational scenarios are arranged into a one-dimensional vector according to scenario order, with each element in the vector corresponding to the priority weight of a scenario.

[0039] Scene response latency data originates from the real-time monitoring module of the IPTV education platform. This module continuously records the time interval from when a user initiates a request to when the scene is successfully loaded for each educational scene. Latency data includes multiple components such as network transmission latency, server processing latency, and content loading latency. The average response latency value for each scene is derived through historical data statistical analysis. In the product operation between the education priority parameter matrix and the scene response latency data, each element in the priority parameter vector is multiplied by the latency data of the corresponding scene. The formula is: latency cost value equals scene priority multiplied by scene response latency. The product result reflects the negative impact of the scene latency on the overall teaching quality. A higher latency cost value indicates a more severe latency problem in that scene, requiring more resource investment for optimization. The formation of the scene latency cost matrix involves organizing the latency costs of all scenes into a matrix form. Each element in the matrix represents the latency cost value of the corresponding scene. This matrix serves as the core data input for the resource optimization objective function.

[0040] The constrained optimization model is constructed based on a mathematical optimization framework established using the scenario delay cost matrix. The objective function is set to minimize the sum of the delay costs of all scenarios. This objective function is derived by summing all elements in the scenario delay cost matrix to obtain the total cost. The task of the optimization algorithm is to find the resource allocation strategy that minimizes the total cost. Constraints ensure the feasibility and rationality of the resource allocation scheme. The bandwidth constraint requires that the total bandwidth be greater than or equal to the sum of the products of the bandwidth demand of each scenario and the number of students. This constraint is calculated by statistically analyzing the bandwidth demand of a single user in each scenario and the number of students currently using that scenario, thus calculating the total bandwidth demand for each scenario. The bandwidth demands of all scenarios are then summed to obtain the total system demand, ensuring that the allocated bandwidth resources meet actual usage needs. Server load constraints limit CPU utilization to a preset threshold to prevent server overload and system performance degradation. The constrained optimization model combines the objective function and constraints into a mathematical model, providing a clear solution framework for the subsequent optimization algorithm.

[0041] The Education Scenario-Oriented Optimization Algorithm (EDU-NGO) is an intelligent optimization technology improved from the Northern Eagle Optimization Algorithm, specifically designed for the resource allocation characteristics of IPTV education platforms. The algorithm's population evolution process simulates the hunting behavior of eagles, with each individual eagle representing a possible resource allocation scheme. The individual's genetic code contains bandwidth allocation ratios and server allocation strategy parameters for eight education scenarios. During algorithm initialization, an initial population containing multiple eagle individuals is randomly generated, with each individual's genetic values ​​randomly set within a reasonable range to ensure the diversity of initial solutions. The fitness evaluation process substitutes each individual into the constrained optimization model for calculation. The fitness level of an individual is determined by a comprehensive score based on the objective function value and the degree of constraint violation; higher fitness indicates a better resource allocation scheme. The position update operation guides the movement direction of other individuals based on the current optimal individual's position information. The update formula is: the new position equals the current position plus the product of the random factor and the difference between the current and optimal positions. The random factor controls the update step size, ensuring the algorithm has both exploratory capability and convergence. The elite retention strategy selects the top 20% of individuals with the highest fitness after each iteration to directly enter the next generation of the population, thus preventing the loss of excellent solutions during the evolutionary process and providing learning targets for the remaining individuals.

[0042] In one specific embodiment, the process of inputting the constrained optimization model into the educational scenario-oriented optimization algorithm for iterative solution may specifically include the following steps: The constrained optimization model is used as the fitness function to initialize the goshawk population, where each goshawk individual contains bandwidth allocation ratios and server allocation strategy parameters for 8 educational scenarios, thus obtaining the initial optimized population. The fitness of each individual goshawk in the initial optimized population is evaluated. The fitness value of the individual is determined by calculating the weighted combination of the sum of the scene delay cost matrix and the degree of constraint violation, and the fitness distribution of the population is obtained. Based on the population fitness distribution, the position update operation of the eagle is performed. The new position coordinates are calculated by adding a random factor to the current position and the product of the difference between the current position and the optimal position. At the same time, the resource allocation weight is automatically increased when key educational scenarios such as online exams are detected, and the updated population is obtained. An elite retention mechanism is applied to the updated population, selecting the top 20% of individuals with the highest fitness values ​​as elite solutions to be retained in the next generation of the population. After 200 iterations, the globally optimal solution is output, resulting in the IPTV scenario resource allocation scheme.

[0043] Specifically, the initialization process of the goshawk population uses a constrained optimization model as the fitness evaluation standard. Each individual goshawk represents a resource allocation scheme. The individual's genetic code contains bandwidth allocation ratios and server allocation strategy parameters for eight educational scenarios. The bandwidth allocation ratio data structure is an 8-dimensional vector, where each element represents the proportion of the corresponding educational scenario in the total bandwidth. The sum of all elements equals 1 to ensure the integrity of the bandwidth allocation. The server allocation strategy parameters contain the load distribution of each scenario across different server nodes, controlling the deployment location of scenarios through the load ratio of edge servers and cloud servers. During the initial optimization population generation process, a random number generator assigns values ​​to the genetic parameters of each individual within a reasonable range. The random range of the bandwidth allocation ratio is determined based on the historical usage frequency of each educational scenario. The upper limit of the initial allocation ratio for high-frequency usage scenarios is relatively high, while the upper limit of the allocation ratio for low-frequency usage scenarios is relatively low, ensuring the rationality and feasibility of the initial solution.

[0044] The fitness evaluation process comprehensively evaluates each individual eagle in the initial optimization population. The fitness function is designed based on a weighted combination of the sum of the scenario delay cost matrix and the degree of constraint violation. The sum of the scenario delay cost matrix is ​​obtained by substituting the individual's resource allocation scheme into the delay calculation model. In the specific calculation process, the actual bandwidth obtained by each educational scenario is calculated according to the individual's bandwidth allocation ratio. Then, based on the inverse relationship between bandwidth and latency, the expected response latency of the scenario is estimated. The expected latency is multiplied by the scenario priority to obtain the delay cost of the scenario. Finally, the delay cost of all scenarios is added together to obtain the total delay cost. The calculation of the degree of constraint violation involves checking bandwidth constraints and server load constraints. When an individual's resource allocation scheme causes the bandwidth demand of some scenarios to be unmet, the degree of violation is quantified according to the size of the gap. When the server load exceeds the limit, a violation penalty is also incurred. The fitness value is calculated by the difference between the reciprocal of the total delay cost and the constraint violation penalty. The higher the fitness value, the better the individual's resource allocation scheme. The population fitness distribution records the fitness ranking and statistical information of all individuals.

[0045] The goshawk position update operation is guided by the optimal individual position in the population fitness distribution. The calculation process simulates the goshawk's behavior of approaching prey. The current position data comes from the individual's gene coding value in the previous iteration, while the optimal position data comes from the gene coding value of the individual with the highest fitness in the population. A random factor is generated between zero and one using a pseudo-random number generator to control the step size of the position update. The calculation of the new position coordinates is to add the current position to the product of the random factor and the difference between the current position and the optimal position. This calculation process ensures that individuals can move towards the optimal solution while maintaining a certain degree of randomness to avoid premature convergence to a local optimum. A key educational scenario detection mechanism plays a dynamic adjustment role in the position update process. When an online examination scenario is detected, the algorithm automatically increases the resource allocation weight of that scenario. Specifically, this is achieved by adding a fixed increment to the bandwidth allocation ratio of the online examination scenario in the position update calculation, while proportionally reducing the allocation ratio of other non-key scenarios to ensure that key scenarios receive sufficient resource guarantees. After the update, the population contains the new gene coding values ​​of all individuals after the position update, preparing data for the next round of fitness evaluation.

[0046] The elite retention mechanism selects and preserves superior individuals in the updated population. The selection criteria for elite individuals are determined based on the ranking of fitness values. The algorithm sorts all individuals in the updated population in descending order of fitness value, and then selects the top 20% of individuals as elite solutions. The number of elite individuals is equal to the total population multiplied by the retention ratio and rounded up. The elite retention process directly copies the selected superior individuals to the next generation population without further mutation or crossover operations, ensuring that superior solutions are not lost during evolution. The remaining population positions are filled by genetic operations such as selection, crossover, and mutation on other individuals in the current population to form the next generation population. The iteration control mechanism sets a maximum of 200 iterations. Each iteration includes a complete operation process such as fitness evaluation, position update, and elite retention. The iteration terminates when the maximum number of iterations is reached or the optimal solution in the population does not significantly improve over several generations. The output of the global optimal solution selects the individual with the highest fitness from the iteration results. The genetic code of this individual is the IPTV scene resource allocation scheme, which includes the optimal bandwidth allocation ratio and server deployment strategy for each educational scene.

[0047] In one specific embodiment, step S5 includes: Based on the IPTV scenario resource allocation scheme, the network bandwidth is pre-allocated, and the bandwidth requirements of each education scenario are reserved to the corresponding network channels according to priority, thus obtaining the bandwidth pre-allocation configuration. Based on the IPTV scenario resource allocation scheme, server resources are pre-allocated, and computing resources are allocated to edge nodes and cloud servers according to scenario type to obtain the server pre-allocation configuration. The bandwidth pre-allocation configuration and the server pre-allocation configuration are merged to construct a resource scheduling mapping table, which records the network channel identifier and server node identifier corresponding to each education scenario, thus obtaining the scenario resource mapping relationship. When a user scene switching request is received, the pre-allocated resources of the target scene are found according to the scene resource mapping relationship. The corresponding network channel and server node are activated through the resource switching command to obtain the scene switching execution result.

[0048] Specifically, the pre-allocation of network bandwidth is implemented based on the bandwidth allocation ratio in the IPTV scenario resource allocation scheme. A tiered reservation mechanism divides network bandwidth into different priority channels according to the priority parameters of the education scenarios. The data processing for bandwidth pre-allocation extracts the bandwidth percentage data for each education scenario in the resource allocation scheme, then calculates the specific bandwidth allocation for each scenario based on the percentage of the total bandwidth capacity. Next, bandwidth is allocated to the corresponding network channels according to scenario priority: high-priority scenarios such as online exams are allocated to the first priority channel, medium-priority scenarios such as live classroom broadcasts are allocated to the second priority channel, and low-priority scenarios such as resource downloads are allocated to the third priority channel. Network channels are identified using a combination of channel ID and priority label. Each channel is assigned a unique numerical identifier and priority code. Channels are isolated and prioritized through a QoS mechanism to ensure that the bandwidth requirements of high-priority scenarios are met first. The data structure for bandwidth pre-allocation configuration includes key information such as scenario identifier, channel identifier, allocated bandwidth, and priority level. The configuration data is stored in tabular form and updated in real time.

[0049] The pre-allocation of server resources follows the server allocation strategy in the IPTV scenario resource allocation scheme, distributing computing resources in a distributed manner. The selection of edge nodes and cloud servers is based on the computing requirements of the scenario type and the geographical distribution of users. The data processing of computing resource allocation involves the quantitative allocation of resources across multiple dimensions, such as the number of CPU cores, memory capacity, and storage space. Each educational scenario determines the required scale of computing resources based on its computational complexity and the number of concurrent users. Edge nodes primarily handle latency-sensitive scenarios such as live classroom broadcasts and interactive Q&A sessions, which require localized computing to reduce network transmission latency. Cloud servers primarily handle computationally intensive scenarios such as large file downloads and data analysis, fully utilizing the powerful computing capabilities of the cloud. The server pre-allocation decision algorithm calculates the optimal server deployment scheme based on factors such as the geographical distribution hotspots of the scenario, user access patterns, and network topology. A load balancing algorithm distributes computing tasks for the same scenario across multiple server nodes, avoiding single-point overload. The data structure for server pre-allocation configuration records detailed information such as scenario identifier, server node identifier, allocated computing resource amount, and load distribution ratio. The configuration data is stored and synchronized through a distributed database.

[0050] The resource scheduling mapping table construction process integrates bandwidth pre-allocation configuration and server pre-allocation configuration data. The mapping table employs a multi-level index structure to quickly locate resource configuration information for specific scenarios. The data processing logic of the mapping table establishes an index structure with the scenario identifier as the primary key. Then, it stores the network channel identifier and server node identifier corresponding to each scenario as associated data, while also recording detailed resource configuration parameters such as bandwidth capacity, number of CPU cores, and memory size. The mapping relationship establishment process uses a hash algorithm for fast lookup. The scenario identifier is hashed to obtain the storage address, directly accessing the corresponding resource configuration information; the lookup time complexity is constant. The data synchronization mechanism of the mapping table ensures that the mapping information in multiple network nodes and server clusters remains consistent. When resource configuration changes, the update information is broadcast to all relevant nodes through a message queue mechanism. The data structure of the scenario resource mapping relationship adopts a key-value pair format, where the key is the scenario identifier, and the value is a composite data structure containing network channel information and server node information, supporting fast read and update operations.

[0051] The data processing for scene switching responds to user scene switching requests and rapidly activates network channels and server nodes through resource switching commands. User scene switching requests include basic information such as user identifier, source scene identifier, target scene identifier, and switching timestamp. The request processing module verifies user permissions and the legitimacy of the scene switching, then searches for corresponding pre-allocated resource information in the scene resource mapping relationship based on the target scene identifier. The resource lookup process directly locates the resource configuration record of the target scene using the hash index of the mapping table, extracts key information such as network channel identifier and server node identifier, and checks the current status and availability of the resources. The generation of resource switching commands includes two parts: commands to activate the network channel and commands to start the server node. The network channel activation command includes configuration information such as channel identifier, QoS parameters, and traffic control policies. The server node start command includes detailed parameters such as node identifier, computing resource allocation, and storage resource allocation. Command execution is automated through network management protocols and server management interfaces. Network channel activation is achieved through flow table rules issued by the SDN controller to implement traffic routing and QoS guarantees. Server node startup utilizes containerization technology to quickly deploy the computing environment required for the scene. The scene switching execution result includes feedback information such as switching status, switching time, and resource allocation status. The result data is used to monitor switching performance and optimize subsequent resource allocation strategies.

[0052] The above describes the multi-scene interactive control method in the IPTV education platform according to the embodiments of this application. The following describes the multi-scene interactive control system in the IPTV education platform according to the embodiments of this application. Please refer to [link / reference]. Figure 2 One embodiment of the multi-scene interactive control system in the IPTV education platform in this application includes: The sampling module is used to perform Latin hypercube sampling of educational behavior, which stratifies IPTV user behavior data according to educational scenario weight factors and learning progress correlation coefficients to generate an educational scenario sampling dataset. The calculation module is used to calculate the cluster centers of the educational scenario sampling dataset according to content similarity, time similarity and user similarity, so as to obtain a standardized set of educational scenario types; The construction module is used to construct an educational spatiotemporal hypergraph with the standardized set of educational scenario types as nodes, wherein the hyperedges connect scenario combinations with teaching logic associations to obtain scenario feature vector representations; A module is established to build a resource optimization model based on the scene feature vector representation, with the objective function being to minimize the product of scene response delay and education priority, and to determine the IPTV scene resource allocation scheme; The response module is used to pre-allocate network bandwidth and server resources according to the IPTV scenario resource allocation scheme, and to perform resource scheduling when responding to user scenario switching requests.

[0053] above Figure 2 The multi-scene interactive control system in the IPTV education platform of this invention will be described in detail from the perspective of modular functional entities. The multi-scene interactive control device in the IPTV education platform of this invention will be described in detail from the perspective of hardware processing.

[0054] Reference Figure 3 This invention also provides a multi-scene interactive control device for an IPTV education platform. This multi-scene interactive control device can be a server, and its internal structure can be as follows: Figure 3 As shown, the multi-scene interactive control device in this IPTV education platform includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor in this computer design provides computing and control capabilities. The memory of the multi-scene interactive control device in the IPTV education platform includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the multi-scene interactive control device in the IPTV education platform is used to store the data corresponding to this embodiment. The network interface of the multi-scene interactive control device in the IPTV education platform is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.

[0055] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the multi-scenario interactive control device in the IPTV education platform to which the present invention is applied.

[0056] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the multi-scene interactive control method in the IPTV education platform.

[0057] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0058] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a multi-scene interactive control device (which may be a personal computer, server, or network device, etc.) in an IPTV education platform to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0059] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-scene interactive control method in an IPTV education platform, characterized in that, The method includes: Step S1: Using Latin hypercube sampling of educational behavior, IPTV user behavior data is stratified and sampled according to educational scenario weight factors and learning progress correlation coefficients to generate an educational scenario sampling dataset. Step S2: Calculate the cluster centers of the educational scenario sampling dataset according to content similarity, time similarity, and user similarity to obtain a standardized set of educational scenario types; Step S3: Construct an educational spatiotemporal hypergraph using the standardized educational scenario type set as nodes, wherein hyperedges connect scenario combinations with teaching logic associations to obtain scenario feature vector representations; Step S4: Based on the scene feature vector representation, establish a resource optimization model with the objective function of minimizing the product of scene response latency and education priority, and determine the IPTV scene resource allocation scheme; Step S5: Pre-allocate network bandwidth and server resources according to the IPTV scenario resource allocation scheme, and perform resource scheduling when responding to user scenario switching requests.

2. The multi-scene interactive control method in the IPTV education platform according to claim 1, characterized in that, Step S1 includes: IPTV user behavior data is divided into an n-dimensional hypercube space according to user ID, timestamp, scene type, network status, and device type to obtain the user behavior space matrix; The user behavior space matrix is ​​uniformly divided into k intervals to ensure that the number of sampling points in each interval is equal, resulting in a uniformly distributed sampling grid. Differential weights are assigned to different educational scenarios in the uniformly distributed sampling grid based on the educational scenario weight factor α. The weight of the classroom live broadcast scenario is 1.0 times that of α, the weight of the interactive Q&A scenario is 0.8 times that of α, and the weight of the resource download scenario is 0.6 times that of α, thus obtaining the scenario weight distribution matrix. The learning progress correlation coefficient β is combined with the scene weight distribution matrix, and the weight of each sampling point is calculated by adding the product of scene priority and α to the product of progress correlation and β, thus obtaining the educational scene sampling dataset.

3. The multi-scene interactive control method in the IPTV education platform according to claim 1, characterized in that, Step S2 includes: Content similarity is calculated for the scene data in the educational scene sampling dataset, where the similarity value of scenes with the same knowledge point is 1.0, the similarity value of scenes with related knowledge points is 0.7, and the similarity value of scenes with different subjects is 0.2, thus obtaining a content similarity matrix; The time similarity is calculated based on the time window information in the educational scenario sampling dataset. The time similarity value is determined by the ratio of the overlapping part of the time window to the union of the time windows, and a time similarity matrix is ​​obtained. User similarity is calculated based on the user group characteristics in the educational scenario sampling dataset to obtain a user similarity matrix; The content similarity matrix, time similarity matrix, and user similarity matrix are weighted and fused with weights of 0.5, 0.3, and 0.2, respectively. The positions of the eight cluster centers are calculated by an iterative clustering algorithm to obtain a set of standardized educational scenario types.

4. The multi-scene interactive control method in the IPTV education platform according to claim 1, characterized in that, Step S3 includes: Eight educational scenarios from the standardized educational scenario type set are set as hypergraph nodes, including classroom live streaming, interactive Q&A, resource download, online exam, group discussion, experiment demonstration, course playback, and homework submission nodes, to obtain a set of educational scenario nodes; Based on the set of educational scenario nodes, a hyperedge connection is constructed, wherein each hyperedge connection has multiple scenario nodes with teaching logic association. Classroom live broadcast, interactive Q&A, and course replay are combined into the first hyperedge, and online exams, experimental demonstrations, and homework submissions are combined into the second hyperedge, resulting in an educational spatiotemporal hypergraph structure. The scene nodes in the educational spatiotemporal hypergraph structure are processed by an educational scene mask matrix. When two scenes have a teaching logic relationship, the mask value is set to 0, and when there is no relationship, the mask value is set to negative infinity, thus obtaining the scene relationship mask matrix. The scene association mask matrix is ​​input into a hypergraph neural network for multi-layer graph convolutional encoding. Temporal features of the scene are extracted through temporal convolution and geographical distribution features of the scene are extracted through spatial convolution, resulting in a scene feature vector representation.

5. The multi-scene interactive control method in the IPTV education platform according to claim 1, characterized in that, Step S4 includes: Based on the scene feature vector representation, educational scene priority parameters are set, wherein the priority of online examination scene is set to 10, the priority of classroom live broadcast scene is set to 8, the priority of interactive Q&A scene is set to 6, and the priority of resource download scene is set to 4, thus obtaining the educational priority parameter matrix. The education priority parameter matrix is ​​multiplied with the scenario response delay data to calculate the delay cost of each education scenario, thus obtaining the scenario delay cost matrix. Based on the scenario delay cost matrix, a resource optimization objective function is constructed with the goal of minimizing the sum of delay costs across all scenarios. At the same time, a constraint condition is set that the total bandwidth is greater than or equal to the sum of the products of the bandwidth requirements of each scenario and the number of students, thus obtaining a constrained optimization model. The constrained optimization model is input into the education scenario-oriented optimization algorithm for iterative solution. The optimal solution space is searched through population evolution and elite retention strategies to obtain the IPTV scenario resource allocation scheme.

6. The multi-scene interactive control method in the IPTV education platform according to claim 5, characterized in that, The process involves inputting the constrained optimization model into an education-oriented optimization algorithm for iterative solution, searching the optimal solution space through population evolution and elite retention strategies, and obtaining an IPTV scenario resource allocation scheme, including: The constrained optimization model is used as the fitness function to initialize the goshawk population, where each goshawk individual contains bandwidth allocation ratios and server allocation strategy parameters for 8 educational scenarios, thus obtaining the initial optimized population. The fitness of each individual goshawk in the initial optimized population is evaluated. The individual fitness value is determined by calculating the weighted combination of the sum of the scene delay cost matrix and the degree of constraint violation, thus obtaining the population fitness distribution. Based on the population fitness distribution, perform a goshawk position update operation. Calculate the new position coordinates by adding a random factor to the current position and the product of the difference between the current position and the optimal position. At the same time, detect key educational scenarios such as online exams and automatically increase resource allocation weights to obtain the updated population. An elite retention mechanism is applied to the updated population, selecting the top 20% of individuals with the highest fitness values ​​as elite solutions to be retained in the next generation of the population. After 200 iterations, the globally optimal solution is output, resulting in the IPTV scenario resource allocation scheme.

7. The multi-scene interactive control method in the IPTV education platform according to claim 1, characterized in that, Step S5 includes: Based on the IPTV scenario resource allocation scheme, the network bandwidth is pre-allocated, and the bandwidth requirements of each education scenario are reserved to the corresponding network channels according to priority, thus obtaining the bandwidth pre-allocation configuration. According to the IPTV scene resource allocation scheme, server resources are pre-allocated, and computing resources are allocated to edge nodes and cloud servers according to scene type to obtain the server pre-allocation configuration. The bandwidth pre-allocation configuration and the server pre-allocation configuration are merged to construct a resource scheduling mapping table, which records the network channel identifier and server node identifier corresponding to each education scenario, thus obtaining the scenario resource mapping relationship. When a user scene switching request is received, the pre-allocated resources of the target scene are found according to the scene resource mapping relationship. The corresponding network channel and server node are activated through the resource switching command to obtain the scene switching execution result.

8. A multi-scene interactive control system for an IPTV education platform, characterized in that, For implementing the multi-scene interactive control method in the IPTV education platform as described in any one of claims 1-7, the multi-scene interactive control system of the IPTV education platform includes: The sampling module is used to perform Latin hypercube sampling of educational behavior, which stratifies IPTV user behavior data according to educational scenario weight factors and learning progress correlation coefficients to generate an educational scenario sampling dataset. The calculation module is used to calculate the cluster centers of the educational scenario sampling dataset according to content similarity, time similarity and user similarity, so as to obtain a standardized set of educational scenario types; The construction module is used to construct an educational spatiotemporal hypergraph with the standardized set of educational scenario types as nodes, wherein the hyperedges connect scenario combinations with teaching logic associations to obtain scenario feature vector representations; A module is established to build a resource optimization model based on the scene feature vector representation, with the objective function being to minimize the product of scene response delay and education priority, and to determine the IPTV scene resource allocation scheme; The response module is used to pre-allocate network bandwidth and server resources according to the IPTV scenario resource allocation scheme, and to perform resource scheduling when responding to user scenario switching requests.

9. A multi-scene interactive control device for an IPTV education platform, characterized in that, The system includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the multi-scene interactive control method in the IPTV education platform according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the multi-scene interactive control method in the IPTV education platform as described in any one of claims 1 to 7.

Images