A data management method for a teaching service platform based on cloud services

Through edge computing and cloud distributed transmission, multi-dimensional correlation partitioning and deep learning prediction models, the problem of insufficient data management and personalized teaching capabilities of education platforms is solved, and efficient data management and personalized teaching experience are achieved.

CN118735749BActive Publication Date: 2024-11-26QINGGUO SOFTWARE GROUP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411193156.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-11-26
Estimated Expiration
2044-08-28

AI Technical Summary

Technical Problem

The existing educational platforms lack effective multi-dimensional data correlation and consistency verification mechanisms in data management, and cannot provide a personalized learning experience, resulting in low data management and personalized teaching capabilities.

Method used

By obtaining the initial multi-source teaching data stream for edge computing and cloud-based distributed transmission, deduplication and integration of data, perform multi-dimensional correlation partitioning and consistency verification, extract students' learning behavior and teaching content characteristics, identify potential teaching modes, and generate deep learning Predictive models, adaptive learning path planning, and real-time monitoring of learning progress through sensors.

Benefits of technology

It improves data quality and consistency, realizes refined teaching resource management, improves personalized teaching effect and efficiency, provides timely feedback to teachers, and optimizes teaching strategies.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118735749B_ABST
    Figure CN118735749B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of data management technology, and in particular to a data management method for a teaching service platform based on cloud services. The method comprises the following steps: obtaining a multi-source teaching initial data stream; performing multi-source data edge computing on the multi-source teaching initial data stream to generate a standard multi-source teaching initial data stream; performing cloud server distributed transmission on the standard multi-source teaching initial data stream, thereby generating cloud teaching transmission data; performing multi-source data deduplication and integration on the cloud teaching transmission data, thereby generating a teaching integration data packet; performing multi-dimensional correlation partitioning on the teaching integration data packet, and generating multi-dimensional partitioning data of teaching resources. The present invention improves the data management and personalized teaching capabilities of a teaching platform based on cloud services by integrating and deduplicating data, verifying consistency, deeply analyzing learning behavior and teaching content, generating personalized learning paths, and real-time monitoring progress.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a data management method for a teaching service platform based on cloud services. Background Art

[0002] The earliest educational platforms relied on local servers and data storage, which was not only costly, but also difficult to expand and complex to maintain. With the development of Internet technology, traditional data management methods have gradually shown their limitations, especially in terms of processing massive data, ensuring data security, and improving system flexibility. After entering the 21st century, the rise of cloud computing technology has brought revolutionary changes to the data management of educational platforms. Cloud service providers such as Amazon WebServices, Microsoft Azure, and Google Cloud provide high-performance, scalable computing resources and storage solutions. These cloud services enable educational platforms to become more efficient in processing big data, achieving real-time analysis, and dynamic expansion. At the same time, the elastic computing and storage capabilities of cloud platforms enable educational institutions to dynamically adjust resources according to demand and optimize cost structures. In recent years, cloud service platforms have also made significant progress in data security, privacy protection, and compliance. By adopting encryption technology, identity authentication, and access control, educational platforms can effectively protect user data from unauthorized access and attacks. However, the current data management of educational platforms usually lacks effective multi-dimensional data association and consistency verification mechanisms, and cannot provide personalized learning experiences, which leads to low data management and personalized teaching capabilities. Summary of the invention

[0003] Based on this, it is necessary to provide a data management method for a teaching service platform based on cloud services to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a data management method for a teaching service platform based on cloud services is provided, the method comprising the following steps:

[0005] Step S1: obtaining a multi-source teaching initial data stream; performing multi-source data edge computing on the multi-source teaching initial data stream to generate a standard multi-source teaching initial data stream; performing cloud server distributed transmission on the standard multi-source teaching initial data stream to generate cloud teaching transmission data; performing multi-source data deduplication and integration on the cloud teaching transmission data to generate a teaching integration data packet;

[0006] Step S2: perform multi-dimensional correlation partitioning on the teaching integrated data package to generate multi-dimensional partition data of teaching resources; perform partition consistency verification on the multi-dimensional partition data of teaching resources to generate consistent partition data of teaching resources; perform data encapsulation on the consistent partition data of teaching resources to generate partition teaching data sets;

[0007] Step S3: extract learning behavior and teaching content features from the partitioned teaching data set to obtain student learning behavior feature data and teaching content feature data; identify potential teaching patterns from the partitioned teaching data set using the student learning behavior feature data and teaching content feature data to generate teaching pattern recognition data;

[0008] Step S4: Perform model training on the teaching pattern recognition data to generate a deep learning prediction model; import the partitioned teaching data set into the deep learning prediction model to perform adaptive learning path planning, thereby generating personalized dynamic learning trajectory data;

[0009] Step S5: Set sensors for personalized dynamic learning trajectory data, and collect data at preset time intervals based on the set sensors to obtain real-time learning data streams; dynamically monitor the learning progress of the real-time learning data streams and generate a teaching resource feedback report.

[0010] The present invention standardizes the initial data stream of multi-source teaching through edge computing and cloud distributed transmission, removes redundant and duplicate data, and generates clear and accurate teaching integration data packets. This improves the quality and consistency of the data and provides a reliable data basis for subsequent analysis. Through multi-dimensional association partitioning and consistency verification, the teaching resources are managed in a refined manner to ensure the classification accuracy and consistency of the data. The generated partitioned teaching data set helps to organize and retrieve teaching resources more accurately and improves the structured level of the data. By extracting student learning behavior characteristics and teaching content characteristics and identifying potential teaching patterns, deep insights are provided for the teaching process. This helps to understand students' learning habits and the effects of teaching content, thereby optimizing teaching strategies and content. The teaching pattern recognition data is trained using a deep learning model to achieve personalized dynamic learning path planning. The generated personalized learning trajectory data can customize the most suitable learning path for each student and improve learning effects and efficiency. By setting sensors and real-time data collection, students' learning progress is dynamically monitored and a teaching resource feedback report is generated. This provides teachers with timely feedback information, helps adjust teaching strategies, and ensures the effectiveness and continuous improvement of teaching activities. Therefore, the present invention improves the data management and personalized teaching capabilities of the cloud service-based teaching platform by integrating and deduplicating data, verifying consistency, deeply analyzing learning behavior and teaching content, generating personalized learning paths, and monitoring progress in real time.

[0011] Preferably, step S1 comprises the following steps:

[0012] Step S11: using sensors to obtain multi-source teaching initial data streams;

[0013] Step S12: performing multi-source data edge computing on the multi-source teaching initial data stream to generate a standard multi-source teaching initial data stream;

[0014] Step S13: edge node confirmation is performed on the standard multi-source teaching initial data stream to obtain teaching edge node data; the standard multi-source teaching initial data stream is distributedly transmitted to the cloud server through the teaching edge node data through the distributed transmission protocol, thereby generating cloud teaching transmission data;

[0015] Step S14: deduplication and integration of multi-source data of cloud teaching transmission data to generate a teaching integration data packet.

[0016] The present invention obtains multi-source teaching initial data streams through sensors, improves the real-time and accuracy of data, and ensures that the collected information is comprehensive and detailed. Edge computing is performed on the multi-source teaching initial data streams, and data can be processed on-site at the data source, reducing the load of data transmission, improving processing efficiency and reducing latency. Confirming edge nodes can optimize data transmission paths, ensure that data is transmitted from the closest node to the cloud server, and further improve transmission efficiency and reliability. Using a distributed transmission protocol, data is transmitted to the cloud server through edge nodes to achieve rapid data transmission and efficient distribution, ensuring the timeliness of data. Multi-source data deduplication and integration are performed on the data transmitted from the cloud to eliminate redundant data, ensure that the final generated teaching integration data packet is of high quality and consistent, and facilitate subsequent teaching analysis and application.

[0017] Preferably, step S12 comprises the following steps:

[0018] Step S121: dividing the multi-source teaching initial data streams into data stream types to obtain teaching type divided data streams, wherein the teaching type divided data streams include video type data streams, audio type data streams and text type data streams;

[0019] Step S122: performing video type encoding on the video type data stream to generate a video type encoded data stream;

[0020] Step S123: performing audio noise reduction on the audio type data stream to generate an audio type noise reduction data stream;

[0021] Step S124: performing text parsing on the text type data stream to generate a text type parsed data stream;

[0022] Step S125: Standardize the behavior data of the video type encoding data stream, the audio type noise reduction data stream and the text type parsing data stream to generate a formatted teaching behavior data stream; synchronize the formatted teaching behavior data stream in time to generate a teaching behavior synchronization data stream; and encapsulate the teaching behavior synchronization data stream to generate a standard multi-source teaching initial data stream.

[0023] The present invention realizes specialized processing for different types of data by classifying the multi-source teaching initial data stream into types (video, audio, text), which helps to improve the processing quality and efficiency of various types of data. The video type data stream is encoded to generate a video type encoded data stream, which ensures the format unification and compression optimization of the video data, and helps to improve the transmission and storage efficiency of the video data. The audio type data stream is subjected to noise reduction processing to generate an audio type noise reduction data stream, which improves the clarity and quality of the audio data and provides a better audio resource for the teaching process. The text type data stream is parsed to generate a text type parsed data stream, and the text data is structured and standardized, which is convenient for subsequent text analysis and processing. Different types of data streams (video, audio, text) are standardized in behavior data to generate a formatted teaching behavior data stream to ensure the consistency and standardization of the data format. The formatted teaching behavior data stream is synchronized in time to generate a teaching behavior synchronization data stream to ensure the consistency of multi-source data in the time dimension, and provide a basis for multimodal data fusion and comprehensive analysis. The teaching behavior synchronization data stream is encapsulated to generate a standard multi-source teaching initial data stream to ensure the integrity and consistency of the data during transmission and storage, which is convenient for subsequent data processing and application.

[0024] Preferably, step S2 comprises the following steps:

[0025] Step S21: Perform data transmission security verification on the teaching integrated data package to generate security verification result data, wherein the data transmission security verification includes security verification and integrity verification;

[0026] Step S22: Based on the security verification result data, the teaching integrated data package is partitioned in multiple dimensions to generate teaching resource multi-dimensional partition data;

[0027] Step S23: performing partition consistency verification on the multi-dimensional partition data of teaching resources to generate consistent partition data of teaching resources;

[0028] Step S24: encapsulate the consistent partition data of teaching resources to generate a partitioned teaching data set.

[0029] The present invention verifies the security and integrity of the teaching integrated data packet, generates security verification result data, ensures that the data is not tampered with or damaged during the transmission process, and ensures the reliability and security of the data. Based on the security verification result data, the teaching integrated data packet is partitioned in multiple dimensions to generate multi-dimensional partition data of teaching resources, which helps to classify and manage the data in more detail and improve the efficiency of data organization and retrieval. The consistency of the multi-dimensional partition data of the teaching resources is verified, and the consistency partition data of the teaching resources is generated to ensure the consistency and integrity of each partition data, avoid data duplication and omission, and improve the accuracy of the data. The consistency partition data of the teaching resources is encapsulated to generate a partitioned teaching data set, which ensures the integrity and consistency of the data during storage and use, and is convenient for subsequent data analysis and teaching applications.

[0030] Preferably, step S22 includes the following steps:

[0031] Step S221: constructing a data relationship diagram for the teaching integrated data package based on the security verification result data to obtain a teaching resource data relationship diagram;

[0032] Step S222: identifying key nodes and main data clusters of the teaching resource data relationship diagram to obtain key node data of teaching resources and main data clusters of teaching resources;

[0033] Step S223: Formulate a partition strategy through key node data and teaching resource main data clusters to generate a teaching resource data partition strategy;

[0034] Step S224: Perform multi-dimensional associated partitioning on the teaching integrated data package according to the teaching resource data partitioning strategy to generate multi-dimensional partitioning data of teaching resources.

[0035] The present invention constructs a data relationship diagram for a teaching integrated data packet based on security verification result data to obtain a teaching resource data relationship diagram. This step helps to fully understand the relationship and structure between data, and provides a basis for subsequent key node identification and data partitioning. Key nodes and main data clusters are identified in the teaching resource data relationship diagram to obtain teaching resource key node data and teaching resource main data clusters. By identifying key nodes and main data clusters, the importance and relevance of data can be understood more accurately, and the effectiveness of data management can be improved. Partition strategy formulation is performed through key node data and teaching resource main data clusters to generate teaching resource data partition strategy. This step ensures the scientificity and rationality of the partition strategy, and helps to optimize the storage and retrieval efficiency of data. According to the teaching resource data partition strategy, the teaching integrated data packet is multi-dimensionally associated and partitioned to generate teaching resource multi-dimensional partition data. Through multi-dimensional associated partitions, data can be more effectively organized and managed, and the utilization rate and access efficiency of data can be improved.

[0036] Preferably, step S223 includes the following steps:

[0037] Step S2231: Perform historical teaching resource access analysis on the teaching integration data package through key node data and teaching resource main data clusters to generate historical teaching resource access data;

[0038] Step S2232: performing resource importance evaluation on the historical teaching resource access data to generate teaching resource importance evaluation data; performing usage frequency analysis on the historical teaching resource access data to generate historical teaching resource usage frequency data; performing first partition dimension confirmation based on the teaching resource importance evaluation data and the historical teaching resource usage frequency data to obtain first partition dimension data;

[0039] Step S2233: performing data correlation analysis on the teaching integrated data package to generate teaching resource strong correlation data; performing data similarity calculation on the teaching integrated data package to obtain teaching resource similarity distribution data; performing second partition dimension confirmation based on the teaching resource strong correlation data and the teaching resource similarity distribution data to obtain second partition dimension data;

[0040] Step S2234: performing sensitive data identification on the teaching integrated data package to obtain sensitive data of teaching resources; performing privacy protection demand analysis on the sensitive data of teaching resources to generate privacy protection demand data; performing third partition dimension confirmation through the sensitive data of teaching resources and the privacy protection demand data to obtain third partition dimension data;

[0041] Step S2235: Formulate a comprehensive partitioning strategy based on the first partitioning dimension data, the second partitioning dimension data and the third partitioning dimension data to generate a teaching resource data partitioning strategy.

[0042] The present invention analyzes historical teaching resource access through key node data and main data clusters, generates historical teaching resource access data, and understands resource usage and trends. The historical teaching resource access data is evaluated for resource importance and analyzed for usage frequency, and teaching resource importance evaluation data and historical teaching resource usage frequency data are generated. Through these evaluations and analyses, the importance and usage frequency of resources can be confirmed, which is helpful to optimize resource management and allocation. The teaching integrated data package is analyzed for data relevance, and teaching resource strong correlation data is generated to identify the close connection between data. The data similarity calculation is performed to obtain teaching resource similarity distribution data. The second partition dimension is confirmed based on the relevance and similarity data to ensure the scientificity and rationality of the partition strategy and improve the data organization and retrieval efficiency. The teaching integrated data package is identified for sensitive data to obtain teaching resource sensitive data to ensure special processing of important and sensitive data. The sensitive data is analyzed for privacy protection requirements to generate privacy protection requirements data. The third partition dimension is confirmed through sensitive data and privacy protection requirements data to ensure that sensitive data is properly protected and meets the privacy protection requirements. The first, second and third partition dimension data are integrated to formulate a comprehensive partition strategy to generate a teaching resource data partition strategy. Comprehensively consider resource importance, usage frequency, data correlation, similarity, and privacy protection requirements to ensure the comprehensiveness and effectiveness of the partitioning strategy.

[0043] Preferably, step S3 comprises the following steps:

[0044] Step S31: extracting learning behavior features from the partitioned teaching data set to obtain student learning behavior feature data, wherein the student learning behavior feature data includes interaction frequency data, interaction duration data, and homework submission status data;

[0045] Step S32: extracting teaching content features from the partitioned teaching data set to obtain teaching content feature data;

[0046] Step S33: extracting keywords from the teaching content feature data to obtain teaching content keyword data; performing sentiment analysis on the teaching content keyword data to generate teaching content sentiment analysis data;

[0047] Step S34: Evaluate the feedback positivity of the teaching content sentiment analysis data through the interaction frequency data, interaction duration data and homework submission status data to generate teaching feedback positivity data; identify potential teaching patterns of the partitioned teaching data set based on the teaching feedback positivity data to generate teaching pattern recognition data.

[0048] The present invention extracts learning behavior features from partitioned teaching data sets to obtain student learning behavior feature data (including interaction frequency, interaction duration, and homework submission status), which can fully understand students' learning behavior and participation, and is helpful for personalized teaching and student management. Extracting teaching content features from partitioned teaching data sets to obtain teaching content feature data can deeply analyze the specific characteristics of teaching content and provide data support for optimizing teaching design and content adjustment. Extracting keywords from teaching content feature data to obtain teaching content keyword data is helpful for quickly understanding and summarizing the core points of teaching content. Emotional analysis is performed on teaching content keyword data to generate teaching content emotional analysis data, which can understand students' emotional reactions to different teaching contents and provide feedback basis for teaching improvement. Feedback enthusiasm evaluation is performed on teaching content emotional analysis data through interaction frequency, interaction duration, and homework submission status data to generate teaching feedback enthusiasm data. This helps to understand students' acceptance and enthusiasm for teaching content and provides a reference for adjusting teaching methods. Potential teaching mode recognition is performed on the partitioned teaching data sets according to the teaching feedback enthusiasm data to generate teaching mode recognition data, which can discover and identify effective teaching modes and methods, and improve teaching effects and student learning experience.

[0049] Preferably, step S4 comprises the following steps:

[0050] Step S41: performing principal component analysis on the teaching mode recognition data to generate a teaching mode feature matrix;

[0051] Step S42: dividing the teaching mode feature matrix into data sets to obtain a model training set and a model test set;

[0052] Step S43: Performing model training on the model training set by using a recurrent neural network algorithm to generate a deep learning training model; performing model iterative optimization on the deep learning training model by using a model test set to generate a deep learning prediction model;

[0053] Step S44: Import the partitioned teaching data set into the deep learning prediction model for adaptive learning path planning, thereby generating personalized dynamic learning trajectory data.

[0054] The present invention generates a teaching mode feature matrix by performing principal component analysis on teaching mode recognition data, which can effectively reduce data dimensions, retain the most important information, simplify the complexity of the model, and improve computational efficiency. The teaching mode feature matrix is ​​divided into a data set to obtain a model training set and a model test set, which is helpful for effective verification and evaluation during model training and testing, and ensures the generalization ability and prediction accuracy of the model. The model training set is trained by a recurrent neural network algorithm to generate a deep learning training model. The training model is iteratively optimized using the model test set to generate a deep learning prediction model, which can continuously improve the accuracy and stability of the model. The recurrent neural network can capture the complex time dependency between student learning behavior and teaching content, improve the accuracy of prediction and the adaptability of the model. The partitioned teaching data set is imported into the deep learning prediction model for adaptive learning path planning, and personalized dynamic learning trajectory data is generated. This helps to formulate personalized learning paths according to the learning behavior and characteristics of each student, and improve learning efficiency and effect. Adaptive learning path planning can dynamically adjust teaching content and progress, meet students' personalized needs, and provide a more flexible and efficient learning experience.

[0055] Preferably, step S44 includes the following steps:

[0056] Step S441: importing the partitioned teaching data set into the deep learning prediction model to perform teaching potential prediction and generate teaching potential prediction data;

[0057] Step S442: mapping the student learning behavior characteristic data to a learning progress according to the teaching potential prediction data to generate learning progress evaluation data;

[0058] Step S443: Matching teaching interest and ability with the teaching potential prediction data through the teaching mode recognition data to generate interest and ability matching data;

[0059] Step S444: Generate a dynamic learning path based on the learning progress evaluation data and the interest and ability matching data to obtain personalized dynamic learning trajectory data.

[0060] The present invention predicts teaching potential by importing partitioned teaching data sets into a deep learning prediction model, and generates teaching potential prediction data. By predicting the teaching potential of students, students' learning performance and potential learning problems can be understood in advance, which is helpful to formulate targeted teaching strategies. According to the teaching potential prediction data, the learning progress of students' learning behavior feature data is mapped to generate learning progress evaluation data. By evaluating the learning progress of students, the learning status of students can be monitored in real time, and the teaching progress can be discovered and adjusted in time to ensure that students can keep up with the course progress. The teaching potential prediction data is matched with teaching interest and ability through teaching mode recognition data to generate interest and ability matching data. By matching students' interests and abilities, more suitable learning content and methods can be recommended to students, thereby improving the initiative and effect of learning. Dynamic learning path generation is performed based on learning progress evaluation data and interest and ability matching data to obtain personalized dynamic learning trajectory data. By dynamically generating learning paths, teaching plans can be adjusted according to students' real-time learning progress and interests, providing a more personalized and efficient learning experience.

[0061] Preferably, step S5 comprises the following steps:

[0062] Step S51: Setting sensors for personalized dynamic learning trajectory data, and collecting data at preset time intervals based on the set sensors to obtain real-time learning data streams;

[0063] Step S52: dynamically monitor the learning progress of the real-time learning data stream to generate learning progress dynamic monitoring data;

[0064] Step S53: Visualize the learning progress dynamic monitoring data to generate a teaching resource feedback report.

[0065] The present invention obtains a real-time learning data stream by setting sensors for personalized dynamic learning trajectory data, and collecting data at preset time intervals based on the set sensors. By collecting data in real time through sensors, the student's immediate learning situation can be obtained to ensure the timeliness and accuracy of the data. The real-time learning data stream is dynamically monitored for learning progress, and learning progress dynamic monitoring data is generated. Dynamic monitoring of students' learning progress can track students' learning status in real time, timely discover and solve problems that arise in the learning process, and improve the flexibility and response speed of learning management. The learning progress dynamic monitoring data is visualized to generate a teaching resource feedback report. Through data visualization, students' learning progress and learning effects can be intuitively displayed, providing valuable feedback information for teachers, helping teachers adjust teaching strategies and content, and improving teaching quality.

[0066] The beneficial effect of the present invention is that by acquiring the multi-source teaching initial data stream and performing edge computing, a standardized multi-source teaching initial data stream is generated. This helps to improve data processing efficiency, reduce data transmission delay, and ensure data consistency and reliability. The standard multi-source teaching initial data stream is distributedly transmitted to the cloud server, and multi-source data is deduplicated and integrated to generate a teaching integration data packet. This can effectively manage and process large-scale, multi-source data, and improve the flexibility and efficiency of data processing. The teaching integration data packet is multi-dimensionally associated and partitioned to generate multi-dimensional partitioned data of teaching resources, and partition consistency verification is performed to ensure the accuracy and integrity of the data. This helps to finely manage teaching resources and improve resource utilization efficiency. Learning behavior and teaching content feature extraction are performed on the partitioned teaching data set to obtain student learning behavior feature data and teaching content feature data. This can fully understand the characteristics of students' learning behavior and teaching content, and provide data support for personalized teaching. Potential teaching mode recognition is performed on the partitioned teaching data set through student learning behavior feature data and teaching content feature data to generate teaching mode recognition data. Identifying effective teaching modes can improve teaching effects and student learning experience. Model training is performed on the teaching mode recognition data to generate a deep learning prediction model. The model is used for adaptive learning path planning to generate personalized dynamic learning trajectory data. This can adjust the learning path in real time according to the student's learning situation, improving learning efficiency and effectiveness. Sensors are set for personalized dynamic learning trajectory data, and data is collected at preset time intervals to obtain real-time learning data streams. Through real-time dynamic monitoring of learning progress, a teaching resource feedback report is generated. Real-time monitoring of students' learning progress can promptly discover and solve learning problems and provide a basis for teaching decisions. Therefore, the present invention improves the data management and personalized teaching capabilities of the cloud service-based teaching platform by integrating and deduplicating data, verifying consistency, deeply analyzing learning behavior and teaching content, generating personalized learning paths, and monitoring progress in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A schematic diagram of the steps of a data management method for a teaching service platform based on a cloud service;

[0068] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0069] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0070] Figure 4 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0071] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0072] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0073] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0074] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0075] To achieve this, please refer to Figures 1 to 4 , a data management method for a teaching service platform based on cloud services, the method comprising the following steps:

[0076] Step S1: obtaining a multi-source teaching initial data stream; performing multi-source data edge computing on the multi-source teaching initial data stream to generate a standard multi-source teaching initial data stream; performing cloud server distributed transmission on the standard multi-source teaching initial data stream to generate cloud teaching transmission data; performing multi-source data deduplication and integration on the cloud teaching transmission data to generate a teaching integration data packet;

[0077] Step S2: perform multi-dimensional correlation partitioning on the teaching integrated data package to generate multi-dimensional partition data of teaching resources; perform partition consistency verification on the multi-dimensional partition data of teaching resources to generate consistent partition data of teaching resources; perform data encapsulation on the consistent partition data of teaching resources to generate partition teaching data sets;

[0078] Step S3: extract learning behavior and teaching content features from the partitioned teaching data set to obtain student learning behavior feature data and teaching content feature data; identify potential teaching patterns from the partitioned teaching data set using the student learning behavior feature data and teaching content feature data to generate teaching pattern recognition data;

[0079] Step S4: Perform model training on the teaching pattern recognition data to generate a deep learning prediction model; import the partitioned teaching data set into the deep learning prediction model to perform adaptive learning path planning, thereby generating personalized dynamic learning trajectory data;

[0080] Step S5: Set sensors for personalized dynamic learning trajectory data, and collect data at preset time intervals based on the set sensors to obtain real-time learning data streams; dynamically monitor the learning progress of the real-time learning data streams and generate a teaching resource feedback report.

[0081] The present invention standardizes the initial data stream of multi-source teaching through edge computing and cloud distributed transmission, removes redundant and duplicate data, and generates clear and accurate teaching integration data packets. This improves the quality and consistency of the data and provides a reliable data basis for subsequent analysis. Through multi-dimensional association partitioning and consistency verification, the teaching resources are managed in a refined manner to ensure the classification accuracy and consistency of the data. The generated partitioned teaching data set helps to organize and retrieve teaching resources more accurately and improves the structured level of the data. By extracting student learning behavior characteristics and teaching content characteristics and identifying potential teaching patterns, deep insights are provided for the teaching process. This helps to understand students' learning habits and the effects of teaching content, thereby optimizing teaching strategies and content. The teaching pattern recognition data is trained using a deep learning model to achieve personalized dynamic learning path planning. The generated personalized learning trajectory data can customize the most suitable learning path for each student and improve learning effects and efficiency. By setting sensors and real-time data collection, students' learning progress is dynamically monitored and a teaching resource feedback report is generated. This provides teachers with timely feedback information, helps adjust teaching strategies, and ensures the effectiveness and continuous improvement of teaching activities. Therefore, the present invention improves the data management and personalized teaching capabilities of the cloud service-based teaching platform by integrating and deduplicating data, verifying consistency, deeply analyzing learning behavior and teaching content, generating personalized learning paths, and monitoring progress in real time.

[0082] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a data management method for a teaching service platform based on a cloud service of the present invention. In this example, the data management method for a teaching service platform based on a cloud service includes the following steps:

[0083] Step S1: obtaining a multi-source teaching initial data stream; performing multi-source data edge computing on the multi-source teaching initial data stream to generate a standard multi-source teaching initial data stream; performing cloud server distributed transmission on the standard multi-source teaching initial data stream to generate cloud teaching transmission data; performing multi-source data deduplication and integration on the cloud teaching transmission data to generate a teaching integration data packet;

[0084] In an embodiment of the present invention, various data sources such as video, audio, text, online course platform data, interactive data, etc. are identified and confirmed. Data is collected from the above data sources by means of API, sensors, network crawlers, etc. Edge computing devices, such as edge servers, routers, etc., are configured and deployed close to the data source. Preprocessing work such as data cleaning, format conversion, and data filtering is performed on the edge device to reduce noise and redundant data. Preliminary data analysis and calculation, such as real-time video analysis, speech recognition, and text processing, are performed on the edge device. The preprocessed data is formatted and standardized to ensure consistent data formats and uniform data types. Metadata (such as data source, timestamp, data type, etc.) is added to the standardized data for subsequent processing and analysis. A suitable distributed transmission protocol (such as HTTP, MQTT, Kafka, etc.) is selected for data transmission. Compression, encryption, load balancing and other technologies are used to optimize the data transmission process to ensure transmission efficiency and security. A server is configured and deployed in the cloud to receive and store the transmitted data stream. After receiving the data stream, the cloud server performs data decompression and decryption processing. The integrity and accuracy of the received data are verified to ensure that the data is not lost or damaged. Store the verified data in a cloud database or file system for subsequent processing and analysis. Use hash, MD5 and other algorithms to deduplicate the data and delete duplicate data. Merge and integrate data from different data sources to generate a unified teaching integration data package. Verify the integrated data to ensure data consistency and integrity.

[0085] Step S2: perform multi-dimensional correlation partitioning on the teaching integrated data package to generate multi-dimensional partition data of teaching resources; perform partition consistency verification on the multi-dimensional partition data of teaching resources to generate consistent partition data of teaching resources; perform data encapsulation on the consistent partition data of teaching resources to generate partition teaching data sets;

[0086] In an embodiment of the present invention, partition dimensions such as courses, chapters, topics, teachers, student interactions, and resource types (video, audio, text, etc.) are determined according to the characteristics of the teaching content. An association rule algorithm (such as an Apriori algorithm and an FP-growth algorithm) is used to perform multi-dimensional partitioning of the data and establish association relationships between the dimensions. Based on the results of the association partitioning algorithm, multi-dimensional partitioning data of teaching resources is generated. Consistency verification rules are defined, such as data integrity, consistency, accuracy, uniqueness, etc. The partitioning data is verified for consistency using methods such as SQL queries and script verification. The verified data is marked as consistent data, and consistent partitioning data of teaching resources is generated. A suitable data encapsulation format is selected, such as JSON, XML, CSV, etc. A data encapsulation script is written to encapsulate the consistent partitioning data into a partitioned teaching data set. The encapsulated data set is compressed (such as using Gzip) and encrypted (such as using AES) to ensure efficient and secure data transmission and storage.

[0087] Step S3: extract learning behavior and teaching content features from the partitioned teaching data set to obtain student learning behavior feature data and teaching content feature data; identify potential teaching patterns from the partitioned teaching data set using the student learning behavior feature data and teaching content feature data to generate teaching pattern recognition data;

[0088] In an embodiment of the present invention, by determining the feature dimensions to be extracted, such as student behavior, interaction frequency, resource access times, teaching content type, etc., a suitable feature extraction algorithm is selected, such as time series analysis, NLP (natural language processing) algorithm, etc. Student learning behavior feature data and teaching content feature data are extracted from the partitioned teaching data set. The extracted feature data is standardized, and a suitable pattern recognition algorithm is selected, such as a clustering algorithm (K-means, DBSCAN), a classification algorithm (decision tree, random forest), etc., and the pattern recognition algorithm is used to identify potential teaching patterns of the partitioned teaching data set to generate teaching pattern recognition data.

[0089] Step S4: Perform model training on the teaching pattern recognition data to generate a deep learning prediction model; import the partitioned teaching data set into the deep learning prediction model to perform adaptive learning path planning, thereby generating personalized dynamic learning trajectory data;

[0090] In an embodiment of the present invention, by collecting teaching mode recognition data, including interaction frequency data, interaction duration data, homework submission data, etc., the teaching mode recognition data is standardized to eliminate dimensional differences. The main features are extracted by the PCA method to generate a teaching mode feature matrix. The teaching mode feature matrix is ​​randomly divided into a training set and a test set, usually divided in a ratio of 8:2 or 7:3. Ensure that the data in the training set and the test set are representative and balanced. Select a suitable recurrent neural network structure, such as LSTM or GRU, and set the hyperparameters of the model, including the number of layers, the number of neurons, the learning rate, etc. The model training set is input into the RNN model for training, the model parameters are adjusted by the back propagation algorithm, and a suitable loss function (such as mean square error) is selected for model optimization. The trained model is verified and optimized using the model test set, the hyperparameters and structure are continuously adjusted to improve the model performance, the model performance is monitored on the verification set, overfitting is prevented, and the early stopping rule is used to control the training process. The preprocessed partitioned teaching data set is input into the trained deep learning prediction model, and the learning potential of the students is predicted by the model to generate teaching potential prediction data. According to the teaching potential prediction data, the learning progress is mapped to the student learning behavior feature data to generate learning progress evaluation data. The teaching potential prediction data is matched with teaching interest and ability through the teaching pattern recognition data to generate interest and ability matching data. Based on the learning progress evaluation data and interest and ability matching data, dynamic learning paths are generated to obtain personalized dynamic learning trajectory data.

[0091] Step S5: Set sensors for personalized dynamic learning trajectory data, and collect data at preset time intervals based on the set sensors to obtain real-time learning data streams; dynamically monitor the learning progress of the real-time learning data streams and generate a teaching resource feedback report.

[0092] In an embodiment of the present invention, the required sensor type (such as heart rate monitoring, eye tracking, keystroke recording, etc.) is determined according to the personalized dynamic learning trajectory data. The sensor is installed and configured in the student's learning environment to ensure that the sensor can accurately collect relevant data. The data collection frequency and time interval of the sensor are set according to the learning process and data requirements. The sensor continuously collects the student's learning behavior data within a preset time interval to form a real-time learning data stream. The real-time learning data stream is denoised and cleaned to ensure the accuracy and consistency of the data. Key learning behavior features such as heart rate changes, eye movement patterns, input speed, etc. are extracted from the real-time data. Combined with the personalized dynamic learning trajectory data, the real-time learning data is dynamically analyzed to evaluate the student's learning progress. Abnormal behaviors such as decreased attention and fatigue in the learning process are monitored to generate corresponding monitoring data. The dynamic monitoring data of learning progress is visualized in the form of charts, such as line charts, bar charts, pie charts, etc., to intuitively display the student's learning status and progress. The trend and changes in the learning process are displayed through visualization tools to help teachers and students better understand the learning situation. The report content contains key information such as students' learning progress, behavioral characteristics, abnormal situations, etc. Provide students with personalized learning suggestions and improvement measures based on feedback data to help them optimize their learning outcomes. Send feedback reports to relevant students and teachers through email, learning management systems and other channels to ensure timely communication and feedback.

[0093] Preferably, step S1 comprises the following steps:

[0094] Step S11: using sensors to obtain multi-source teaching initial data streams;

[0095] Step S12: performing multi-source data edge computing on the multi-source teaching initial data stream to generate a standard multi-source teaching initial data stream;

[0096] Step S13: edge node confirmation is performed on the standard multi-source teaching initial data stream to obtain teaching edge node data; the standard multi-source teaching initial data stream is distributedly transmitted to the cloud server through the teaching edge node data through the distributed transmission protocol, thereby generating cloud teaching transmission data;

[0097] Step S14: deduplication and integration of multi-source data of cloud teaching transmission data to generate a teaching integration data packet.

[0098] In an embodiment of the present invention, multiple sensors are deployed in the classroom and learning environment, including video cameras, microphones, touch screen sensors, RFID tags, etc. Ensure that the sensors can cover the entire teaching area and capture multi-source data such as student-teacher interaction and classroom activities. The sensors collect audio, video, touch and RFID data in the classroom in real time. Use a local storage device (such as an edge computing device) to temporarily store the collected data. The collected multi-source data is transmitted to the edge computing device through the local network for preliminary processing. The initial data is pre-processed on the edge computing device, including denoising, deduplication, format conversion, etc. For example, denoising audio data and adjusting the frame rate of video data. The initial data is pre-processed on the edge computing device, including denoising, deduplication, format conversion, etc. For example, denoising audio data and adjusting the frame rate of video data. Perform preliminary analysis and calculation on the edge computing device, such as behavior recognition, sentiment analysis, interaction detection, etc. Generate preliminary analysis results and store them together with the original data. Determine the location and role of the edge computing device, such as the edge computing node in the classroom, the school server, etc. According to the direction and load of the data flow, the computing and transmission tasks of each edge node are allocated. Distributed transmission protocols (such as MQTT, CoAP, DDS, etc.) are used for data transmission between edge nodes. To ensure the integrity and reliability of data during transmission, data encryption and verification technology are used. The processed standard multi-source teaching initial data stream is transmitted from the edge node to the cloud server. The distributed transmission protocol is used to ensure the efficiency and stability of data transmission. The received data is deduplicated on the cloud server to delete duplicate data records. The hash algorithm or other deduplication technology is used to ensure the uniqueness of the data. The deduplicated multi-source data is integrated to form a complete teaching data set. Data from different sources are matched and merged, such as aligning the timestamps of audio, video and touch data to generate a synchronized multi-source data set. The integrated data is packaged into a standardized teaching integration data package to facilitate subsequent analysis and application. The data package is stored and the corresponding metadata description file is generated to record the content, format and source of the data package.

[0099] Preferably, step S12 comprises the following steps:

[0100] Step S121: dividing the multi-source teaching initial data streams into data stream types to obtain teaching type divided data streams, wherein the teaching type divided data streams include video type data streams, audio type data streams and text type data streams;

[0101] Step S122: performing video type encoding on the video type data stream to generate a video type encoded data stream;

[0102] Step S123: performing audio noise reduction on the audio type data stream to generate an audio type noise reduction data stream;

[0103] Step S124: performing text parsing on the text type data stream to generate a text type parsed data stream;

[0104] Step S125: Standardize the behavior data of the video type encoding data stream, the audio type noise reduction data stream and the text type parsing data stream to generate a formatted teaching behavior data stream; synchronize the formatted teaching behavior data stream in time to generate a teaching behavior synchronization data stream; and encapsulate the teaching behavior synchronization data stream to generate a standard multi-source teaching initial data stream.

[0105] In an embodiment of the present invention, the type of data stream, such as video, audio and text, is identified from the initial data stream of multi-source teaching. Data stream classification is performed using predefined classification rules or machine learning models. According to the recognition result, the data stream is divided into a video type data stream, an audio type data stream and a text type data stream. Each type of data stream is stored for subsequent processing. An independent data stream is created for each data type to ensure clear and accurate classification of the data type. A suitable video encoding format, such as H.264, H.265, etc., is selected to ensure efficient compression and transmission of video data, and a video encoding tool or library (such as FFmpeg) is used to encode the video data stream, compress the video stream, reduce the amount of data, optimize storage and transmission efficiency, and store the encoded video data stream in a standardized video encoding data stream format. An audio denoising algorithm, such as Wiener filtering, adaptive filtering, etc., is selected to perform denoising on the audio data stream to remove background noise and interference, and an audio processing tool or library (such as Audacity, librosa) is used to perform denoising operations, and the processed audio data stream is stored in a denoised audio data stream format. Select appropriate text parsing tools or libraries (such as NLTK, spaCy) for text analysis, perform parsing operations such as word segmentation, part-of-speech tagging, and entity recognition on the text data stream, extract key information and structured data, generate a structured text data stream, and store the parsed text data stream in a standardized text parsing data stream format. Unify and standardize the formats of video, audio, and text data streams. Ensure that all data streams conform to unified format specifications for subsequent processing. Merge the standardized video, audio, and text data streams to generate a formatted teaching behavior data stream. Time synchronize the formatted teaching behavior data stream to ensure the temporal consistency of different types of data streams. Use timestamp alignment technology to ensure the consistency of the time axis of the data stream. Encapsulate the teaching behavior synchronization data stream to create a comprehensive data package containing all standardized data streams. Use data encapsulation technology (such as JSON, Protobuf) to integrate all data streams into a standardized multi-source teaching data package, and store the encapsulated data package as a standard multi-source teaching initial data stream for subsequent processing and analysis.

[0106] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0107] Step S21: Perform data transmission security verification on the teaching integrated data package to generate security verification result data, wherein the data transmission security verification includes security verification and integrity verification;

[0108] Step S22: Based on the security verification result data, the teaching integrated data package is partitioned in multiple dimensions to generate teaching resource multi-dimensional partition data;

[0109] Step S23: performing partition consistency verification on the multi-dimensional partition data of teaching resources to generate consistent partition data of teaching resources;

[0110] Step S24: encapsulate the consistent partition data of teaching resources to generate a partitioned teaching data set.

[0111] In an embodiment of the present invention, by selecting a suitable data security verification tool or technology, such as SSL / TLS protocol, data encryption algorithm (AES, RSA, etc.). Encrypt the teaching integration data packet to ensure the confidentiality of the data during transmission. Use a security verification tool to check the encryption standard of the data packet and the security during transmission. Ensure that the data packet is not unauthorizedly accessed or tampered with during transmission. Record the security verification results, including encryption standards, authentication status, etc. Select a data integrity verification tool or algorithm, such as the SHA-256 hash algorithm, calculate the hash value of the teaching integration data packet, compare the hash values ​​before and after the data packet transmission, ensure that the data is not tampered with, compare the hash values ​​before and after the transmission, record the integrity verification results, and ensure the integrity of the data. According to the security verification result data, analyze the data resources in the teaching integration data packet, including data source, data type and data use. Determine the associated dimensions of the data, such as the time dimension, space dimension, content dimension, etc. Formulate a partitioning strategy and partition according to the associated dimensions of the data. Determine the boundaries and data content of each partition to ensure the rationality and effectiveness of the partition. Partition the teaching integration data packet according to the formulated partitioning strategy. Record the data content and associated information of each partition to generate multi-dimensional partition data. Select appropriate consistency verification tools or technologies, such as consistent hashing algorithms or data integrity verification tools. Perform consistency checks on the data of each partition to ensure that the data within the partition is complete and consistent. Verify the data association and consistency between partitions, and check whether there is data missing or inconsistency. According to the consistency verification results, record the verification status of each partition. Generate consistency verification result data to ensure the data quality and consistency of each partition. According to the consistency verification results, encapsulate the consistent partition data, select a suitable encapsulation format, such as JSON, XML or database format, and package the partition consistency data into a partition teaching data set in a standard format, including the data content, verification results and metadata description of each partition. Store the generated partition teaching data set to ensure that it can be effectively read and used by subsequent systems, and record detailed information on data encapsulation, including encapsulation standards, version information, etc.

[0112] Preferably, step S22 includes the following steps:

[0113] Step S221: constructing a data relationship diagram for the teaching integrated data package based on the security verification result data to obtain a teaching resource data relationship diagram;

[0114] Step S222: identifying key nodes and main data clusters of the teaching resource data relationship diagram to obtain key node data of teaching resources and main data clusters of teaching resources;

[0115] Step S223: Formulate a partition strategy through key node data and teaching resource main data clusters to generate a teaching resource data partition strategy;

[0116] Step S224: Perform multi-dimensional associated partitioning on the teaching integrated data package according to the teaching resource data partitioning strategy to generate multi-dimensional partitioning data of teaching resources.

[0117] In an embodiment of the present invention, a suitable data relationship graph construction tool is selected, such as a graph database (Neo4j), a data visualization tool (Gephi, Cytoscape) or a custom tool. Relationship information between data is extracted from the teaching integration data package, including data source, data type, data interaction, etc. A data relationship graph is created to represent the relationship between data sources and data sources, and between data sources and data streams. The relationship graph includes nodes (data sources or data sets) and edges (data relationships), such as the relationship between sensors in a classroom and the collected data, and the structure and node relationships of the data relationship graph are recorded to generate a teaching resource data relationship graph. Graph analysis algorithms (such as PageRank, centrality analysis) are used to identify key nodes in the data relationship graph. Key nodes are usually the core of information flow or the main hub of data interaction, such as the main camera, the main audio source, etc. The main data clusters in the data relationship graph are identified by cluster analysis algorithms (such as community detection algorithms, K-means). Data clusters are sets of data with strong correlations, such as data clusters generated by the same teaching activities. The identified key nodes and data clusters are recorded, and descriptions of key node data and main data clusters are generated. According to the identified key nodes and main data clusters, a data partitioning strategy is formulated. The partitioning strategy should take into account factors such as data access frequency, processing requirements, and storage requirements. For example, key node data and major data clusters are allocated to different storage areas or processing units. Optimize the partitioning strategy to balance data load, reduce data transmission delay, and improve processing efficiency. Perform strategy simulation and evaluation, and adjust the partitioning strategy to meet actual needs. Record the details of the partitioning strategy, including partitioning criteria, data allocation rules, etc. According to the formulated partitioning strategy, perform multi-dimensional correlation partitioning on the teaching integration data package. Divide the data package into different areas or modules, each of which contains specific key node data and data clusters. Process and verify the partitioned data to ensure the integrity and consistency of each partition. Use partition processing tools or custom scripts to perform data partitioning operations. Record the data content and partition information of each partition to generate complete multi-dimensional partitioned data. Store the partitioned data and generate corresponding metadata description files to record the structure and content of the partitioned data.

[0118] Preferably, step S223 includes the following steps:

[0119] Step S2231: Perform historical teaching resource access analysis on the teaching integration data package through key node data and teaching resource main data clusters to generate historical teaching resource access data;

[0120] Step S2232: performing resource importance evaluation on the historical teaching resource access data to generate teaching resource importance evaluation data; performing usage frequency analysis on the historical teaching resource access data to generate historical teaching resource usage frequency data; performing first partition dimension confirmation based on the teaching resource importance evaluation data and the historical teaching resource usage frequency data to obtain first partition dimension data;

[0121] Step S2233: performing data correlation analysis on the teaching integrated data package to generate teaching resource strong correlation data; performing data similarity calculation on the teaching integrated data package to obtain teaching resource similarity distribution data; performing second partition dimension confirmation based on the teaching resource strong correlation data and the teaching resource similarity distribution data to obtain second partition dimension data;

[0122] Step S2234: performing sensitive data identification on the teaching integrated data package to obtain sensitive data of teaching resources; performing privacy protection demand analysis on the sensitive data of teaching resources to generate privacy protection demand data; performing third partition dimension confirmation through the sensitive data of teaching resources and the privacy protection demand data to obtain third partition dimension data;

[0123] Step S2235: Formulate a comprehensive partitioning strategy based on the first partitioning dimension data, the second partitioning dimension data and the third partitioning dimension data to generate a teaching resource data partitioning strategy.

[0124] In an embodiment of the present invention, historical access records of teaching resources are collected, including access time, access users, access types, etc. By counting and analyzing historical access records, teaching resources with high access frequency and importance are identified. Historical teaching resource access data is generated using data analysis tools (such as database query, analysis software). Detailed information of access data is recorded, including resource access frequency, user access behavior, resource access time, etc. A suitable resource importance evaluation model is selected, such as a comprehensive scoring model based on access frequency, user feedback, and resource content. The historical teaching resource access data is analyzed using the evaluation model to evaluate the importance of each resource. The evaluation results are recorded to generate teaching resource importance evaluation data. The number and frequency of access to each teaching resource are counted. Detailed information of access frequency is recorded, including high-frequency access resources and low-frequency access resources. The analysis results are organized into usage frequency data, and the access frequency level of the resource is marked. The importance and usage frequency of the resource are comprehensively considered to confirm the partition dimension. The data of the first partition dimension is determined, such as important resources with high frequency of use and secondary resources with low frequency of use. The detailed information of the first partition dimension is recorded, including resource partition standards and dimension data. The association relationship between data is analyzed, such as the collaborative use relationship between resources and the data flow relationship. Generate strong correlation data, including the connection strength and relationship type of related resources, record the detailed information of strong correlation data, and identify the strong correlation relationship between resources. Select a data similarity calculation method, such as cosine similarity, Euclidean distance, etc. Perform similarity calculation on the resources in the teaching integration data package, generate similarity distribution data, record the similarity level and distribution between resources, and record the results of similarity calculation, including the similarity distribution and association information of resources. Determine the second partition dimension based on the strong correlation data and similarity distribution data of the resources, for example, divide highly correlated resources and resources with high similarity into the same partition, and record the detailed information of the second partition dimension, including resource partition standards and dimension data. Select appropriate sensitive data identification tools, such as data leakage detection tools and sensitive information identification algorithms, scan the teaching integration data package, identify data containing sensitive information, such as personal information, confidential content, etc., and record the detailed information of sensitive data, including the type and location of sensitive data. According to the type and content of sensitive data, analyze its privacy protection requirements, such as encryption, access control, etc., evaluate the privacy protection requirements and compliance requirements of the data, and record the detailed information of privacy protection requirements, including protection measures and requirement descriptions. Determine the third partition dimension based on sensitive data and privacy protection requirements. For example, assign sensitive data and data that require privacy protection to partitions with high security levels, and record detailed information about the third partition dimension, including resource partition standards and dimension data.Comprehensively consider the first, second and third partition dimension data, formulate the final data partitioning strategy, ensure that the strategy covers the importance, correlation, similarity and privacy protection requirements of resources, record the detailed information of the comprehensive partitioning strategy, including partitioning standards, data allocation rules, etc., and generate a partitioning strategy document as a guide for partitioning implementation.

[0125] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0126] Step S31: extracting learning behavior features from the partitioned teaching data set to obtain student learning behavior feature data, wherein the student learning behavior feature data includes interaction frequency data, interaction duration data, and homework submission status data;

[0127] Step S32: extracting teaching content features from the partitioned teaching data set to obtain teaching content feature data;

[0128] Step S33: extracting keywords from the teaching content feature data to obtain teaching content keyword data; performing sentiment analysis on the teaching content keyword data to generate teaching content sentiment analysis data;

[0129] Step S34: Evaluate the feedback positivity of the teaching content sentiment analysis data through the interaction frequency data, interaction duration data and homework submission status data to generate teaching feedback positivity data; identify potential teaching patterns of the partitioned teaching data set based on the teaching feedback positivity data to generate teaching pattern recognition data.

[0130] In an embodiment of the present invention, data such as student interaction records and homework submission status are collected from a partitioned teaching data set. The interaction frequency between students and teaching content or systems is calculated. For example, the number of interactions per week, the number of interactions per month, etc. Data processing tools (such as SQL, Python data analysis library) are used to extract and count interaction frequency data. The duration of each student interaction is counted, and the total interaction duration is calculated. The duration of each interaction is calculated using a timestamp data analysis tool. The student's homework submission status is recorded, including submission time, submission frequency, completion status, etc. The homework submission data is obtained using the data export function of the homework management system, and the interaction frequency data, interaction duration data, and homework submission status data are summarized to generate student learning behavior feature data. Teaching content-related data, including teaching materials, courseware content, handouts, etc., are extracted from the partitioned teaching data set. Text analysis and natural language processing tools (such as NLTK, spaCy) are used to extract features of the teaching content, such as themes, key concepts, structures, etc., and the extracted feature data, including content themes, key concepts, structural descriptions, etc., are recorded. Select a keyword extraction tool or algorithm, such as TF-IDF, TextRank, BERT, etc., use the selected tool to extract keywords from the teaching content feature data, record the frequency, importance and other information of the keywords, summarize the keyword extraction results, and generate teaching content keyword data. Select a sentiment analysis tool or algorithm, such as VADER, TextBlob, BERT, etc., perform sentiment analysis on the teaching content keyword data, determine the sentiment tendency (such as positive, negative, neutral), and record the sentiment analysis results, including sentiment tendency and intensity. Select a suitable feedback positivity evaluation model, such as a machine learning model (regression model, classification model) or a statistical model. Combine the interaction frequency data, interaction duration data, and homework submission data with the teaching content sentiment analysis data. Use the evaluation model to analyze the students' feedback positivity, score the sentiment analysis data based on the interaction frequency, interaction duration and homework submission, and record the feedback positivity evaluation results, including the positivity score and related data. Select appropriate pattern recognition methods, such as cluster analysis, association rule mining, and pattern matching algorithms, use teaching feedback enthusiasm data to identify potential teaching patterns, identify different teaching patterns and student learning behavior patterns, and record the identified potential teaching patterns, including pattern characteristics, pattern categories, etc.

[0131] As an example of the present invention, refer to Figure 4 As shown, in this example, step S4 includes:

[0132] Step S41: performing principal component analysis on the teaching mode recognition data to generate a teaching mode feature matrix;

[0133] Step S42: dividing the teaching mode feature matrix into data sets to obtain a model training set and a model test set;

[0134] Step S43: Performing model training on the model training set by using a recurrent neural network algorithm to generate a deep learning training model; performing model iterative optimization on the deep learning training model by using a model test set to generate a deep learning prediction model;

[0135] Step S44: Import the partitioned teaching data set into the deep learning prediction model for adaptive learning path planning, thereby generating personalized dynamic learning trajectory data.

[0136] In an embodiment of the present invention, features are extracted from teaching mode recognition data for principal component analysis (PCA), and the data is standardized (for example, the mean is 0 and the variance is 1) to ensure that the contribution of each feature to the principal component is fair, the covariance matrix of the feature data is calculated, the correlation between the features is analyzed, and the covariance matrix is ​​subjected to eigenvalue decomposition to obtain eigenvalues ​​and eigenvectors. Several main principal components are selected according to the size of the eigenvalues, and the eigenvectors corresponding to the first k eigenvalues ​​are usually selected. These principal components explain most of the variability of the data. The original data is mapped to a new feature space using the selected principal components to generate a teaching mode feature matrix. The teaching mode feature matrix is ​​divided into a model training set and a model test set. Generally, 70%-80% of the data set can be used for the training set, and 20%-30% can be used for the test set. Ensure that the training set and the test set are representative, that is, they contain various types of teaching mode data to avoid data imbalance problems. Randomization is performed when dividing the data set to avoid deviations caused by data set division. The divided training set and test set are saved as files for subsequent model training and testing. Choose a suitable deep learning framework, such as TensorFlow, Keras, or PyTorch, to build and train a recurrent neural network (RNN). Define the RNN structure, including the number of layers, the number of neurons in each layer, activation functions, etc. Use LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit) to enhance the performance of the model. Select an optimizer (such as Adam, SGD), a loss function (such as mean squared error, cross entropy), and an evaluation metric (such as accuracy, F1 score). Use the model training set to train the RNN model and adjust the hyperparameters to optimize the model performance. Use the model test set to evaluate the trained model and check its performance. Iteratively optimize the model based on the test results, adjust the model parameters, and improve the accuracy and generalization ability of the model. After training and optimization, save the deep learning prediction model for subsequent learning path planning. Import the partitioned teaching dataset into the deep learning prediction model to preprocess and convert the input data. Use the deep learning prediction model for adaptive learning path planning. Generate personalized learning paths and dynamic learning trajectories based on students' learning behavior characteristics and teaching content characteristics. The generated learning path and trajectory data are saved as personalized dynamic learning trajectory data and provided to students and education administrators.

[0137] Preferably, step S44 includes the following steps:

[0138] Step S441: importing the partitioned teaching data set into the deep learning prediction model to perform teaching potential prediction and generate teaching potential prediction data;

[0139] Step S442: mapping the student learning behavior characteristic data to a learning progress according to the teaching potential prediction data to generate learning progress evaluation data;

[0140] Step S443: Matching teaching interest and ability with the teaching potential prediction data through the teaching mode recognition data to generate interest and ability matching data;

[0141] Step S444: Generate a dynamic learning path based on the learning progress evaluation data and the interest and ability matching data to obtain personalized dynamic learning trajectory data.

[0142] In an embodiment of the present invention, the partitioned teaching data set is preprocessed, including data cleaning, standardization and feature selection, to ensure the quality and consistency of the data. The data format is converted to meet the input requirements of the deep learning model. The preprocessed data is input into the deep learning prediction model to predict the teaching potential. The results output by the prediction model are used to generate the teaching potential prediction data of the students, and the potential ability and development space of the students in various teaching fields are recorded. The teaching potential prediction data is compared and mapped with the learning behavior characteristic data of the students to evaluate the degree of matching between the students' actual learning progress and the predicted potential. The mapping results are used to generate learning progress evaluation data, including the progress, grades and potential matching of the students at different learning stages. The learning progress evaluation results are summarized to provide a detailed report on the students' learning progress and help formulate personalized learning plans. The teaching mode recognition data is used to analyze the students' learning interests and abilities, and the students' interest levels and abilities in different teaching contents are evaluated. The teaching potential prediction data is matched with the students' interest and ability data to evaluate the degree of matching between the students' interests and abilities in various teaching fields. The matching results of interest and ability are recorded to generate interest and ability matching data to provide a basis for personalized learning path planning. Integrate learning progress assessment data and interest and ability matching data to develop strategies for personalized learning paths. Generate dynamic learning paths based on integrated data, and adjust learning content and progress to suit students' actual needs and potential abilities. Use dynamic programming algorithms to generate personalized learning trajectories to ensure that the paths can be adjusted as students progress. Record and save personalized learning trajectory data, including recommended learning paths, key content, learning goals, and other information, and provide them to students and education administrators.

[0143] Preferably, step S5 comprises the following steps:

[0144] Step S51: Setting sensors for personalized dynamic learning trajectory data, and collecting data at preset time intervals based on the set sensors to obtain real-time learning data streams;

[0145] Step S52: dynamically monitor the learning progress of the real-time learning data stream to generate learning progress dynamic monitoring data;

[0146] Step S53: Visualize the learning progress dynamic monitoring data to generate a teaching resource feedback report.

[0147] In an embodiment of the present invention, various data (such as students' interaction frequency, usage time, etc.) in the learning process are monitored by selecting appropriate sensor devices (such as smart watches, learning tracking devices, environmental sensors, etc.). Set the parameters of the sensor, including sampling frequency, data type, transmission mode, etc., to ensure that it can accurately capture the key data in the learning process. Define the time interval for data collection (such as every minute, every hour, etc.), and select the appropriate interval according to the characteristics of the learning activity. Start the sensor for data collection and record various data streams in the learning process (such as learning time, number of interactions, activity content, etc.) in real time. Integrate and process the data collected by the sensor to generate a real-time learning data stream for subsequent analysis and monitoring. Perform real-time analysis on the real-time learning data stream to monitor the students' learning progress, including the frequency, duration and changes in learning content of learning activities. According to the preset learning progress standards, evaluate the actual progress of students on the personalized learning trajectory, and identify potential problems and improvement points. Organize the analysis results into dynamic monitoring data of learning progress, including students' real-time learning status, progress trends and possible deviations. If necessary, set an automatic alarm mechanism to promptly remind relevant personnel to intervene when abnormal learning progress is detected. Use data visualization tools (such as Tableau, Power BI, D3.js, etc.) to graphically display the dynamic monitoring data of learning progress. Generate various charts (such as line charts, bar charts, heat maps, etc.) to display students' learning progress, interaction status, learning time distribution and other information. Generate a teaching resource feedback report based on the data visualization results. The report should include an overall assessment of learning progress, major problems found, suggestions for improvement, and future learning plans. Send the feedback report to relevant education administrators, teachers, and students so that they can adjust and optimize their learning strategies based on the information in the report.

[0148] The beneficial effect of the present invention is that by acquiring the multi-source teaching initial data stream and performing edge computing, a standardized multi-source teaching initial data stream is generated. This helps to improve data processing efficiency, reduce data transmission delay, and ensure data consistency and reliability. The standard multi-source teaching initial data stream is distributedly transmitted to the cloud server, and multi-source data is deduplicated and integrated to generate a teaching integration data packet. This can effectively manage and process large-scale, multi-source data, and improve the flexibility and efficiency of data processing. The teaching integration data packet is multi-dimensionally associated and partitioned to generate multi-dimensional partitioned data of teaching resources, and partition consistency verification is performed to ensure the accuracy and integrity of the data. This helps to finely manage teaching resources and improve resource utilization efficiency. Learning behavior and teaching content feature extraction are performed on the partitioned teaching data set to obtain student learning behavior feature data and teaching content feature data. This can fully understand the characteristics of students' learning behavior and teaching content, and provide data support for personalized teaching. Potential teaching mode recognition is performed on the partitioned teaching data set through student learning behavior feature data and teaching content feature data to generate teaching mode recognition data. Identifying effective teaching modes can improve teaching effects and student learning experience. Model training is performed on the teaching mode recognition data to generate a deep learning prediction model. The model is used for adaptive learning path planning to generate personalized dynamic learning trajectory data. This can adjust the learning path in real time according to the student's learning situation, improving learning efficiency and effectiveness. Sensors are set for personalized dynamic learning trajectory data, and data is collected at preset time intervals to obtain real-time learning data streams. Through real-time dynamic monitoring of learning progress, a teaching resource feedback report is generated. Real-time monitoring of students' learning progress can promptly discover and solve learning problems and provide a basis for teaching decisions. Therefore, the present invention improves the data management and personalized teaching capabilities of the cloud service-based teaching platform by integrating and deduplicating data, verifying consistency, deeply analyzing learning behavior and teaching content, generating personalized learning paths, and monitoring progress in real time.

[0149] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0150] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A data management method for a teaching service platform based on cloud services, It is characterized in that The following steps are involved: Step S1: Obtaining multi-source teaching initial data stream; Perform multi-source data edge computing on the multi-source teaching initial data stream to generate a standard multi-source teaching initial data stream; Performing distributed transmission of the standard multi-source teaching initial data stream on the cloud server to generate cloud teaching transmission data; performing multi-source data deduplication and integration on the cloud teaching transmission data to generate a teaching integration data packet; Step S2: performing multi-dimensional correlation partitioning on the teaching integrated data package to generate multi-dimensional partitioning data of teaching resources; Perform partition consistency verification on multi-dimensional partition data of teaching resources to generate consistent partition data of teaching resources; Encapsulate the consistent partition data of teaching resources to generate partition teaching data sets; Step S2 includes the following steps: Step S21: Perform data transmission security verification on the teaching integrated data package to generate security verification result data, wherein the data transmission security verification includes security verification and integrity verification; Step S22: Based on the security verification result data, the teaching integrated data package is partitioned in multiple dimensions to generate multi-dimensional partition data of teaching resources; Step S22 includes the following steps: Step S221: constructing a data relationship diagram for the teaching integrated data package based on the security verification result data to obtain a teaching resource data relationship diagram; Step S222: identifying key nodes and main data clusters of the teaching resource data relationship diagram to obtain key node data of teaching resources and main data clusters of teaching resources; Step S223: Formulate a partition strategy through key node data and teaching resource main data clusters to generate a teaching resource data partition strategy; Step S224: performing multi-dimensional associated partitioning on the teaching integrated data package according to the teaching resource data partitioning strategy to generate multi-dimensional partitioning data of teaching resources; Step S23: performing partition consistency verification on the multi-dimensional partition data of teaching resources to generate consistent partition data of teaching resources; Step S24: encapsulating the consistent partition data of teaching resources to generate a partition teaching data set; Step S3: extract learning behavior and teaching content features from the partitioned teaching data set to obtain student learning behavior feature data and teaching content feature data; identify potential teaching patterns from the partitioned teaching data set through the student learning behavior feature data and teaching content feature data to generate teaching pattern recognition data; Step S3 includes the following steps: Step S31: extracting learning behavior features from the partitioned teaching data set to obtain student learning behavior feature data, wherein the student learning behavior feature data includes interaction frequency data, interaction duration data, and homework submission status data; Step S32: extracting teaching content features from the partitioned teaching data set to obtain teaching content feature data; Step S33: extracting keywords from the teaching content feature data to obtain teaching content keyword data; performing sentiment analysis on the teaching content keyword data to generate teaching content sentiment analysis data; Step S34: evaluating the feedback positivity of the teaching content sentiment analysis data through the interaction frequency data, interaction duration data, and homework submission status data to generate teaching feedback positivity data; identifying potential teaching patterns of the partitioned teaching data set according to the teaching feedback positivity data to generate teaching pattern recognition data; Step S4: Perform model training on the teaching pattern recognition data to generate a deep learning prediction model; import the partitioned teaching data set into the deep learning prediction model for adaptive learning path planning, thereby generating personalized dynamic learning trajectory data; Step S4 includes the following steps: Step S41: performing principal component analysis on the teaching mode recognition data to generate a teaching mode feature matrix; Step S42: dividing the teaching mode feature matrix into data sets to obtain a model training set and a model test set; Step S43: Performing model training on the model training set by using a recurrent neural network algorithm to generate a deep learning training model; performing model iterative optimization on the deep learning training model by using a model test set to generate a deep learning prediction model; Step S44: importing the partitioned teaching data set into the deep learning prediction model to perform adaptive learning path planning, thereby generating personalized dynamic learning trajectory data; Step S44 includes the following steps: Step S441: importing the partitioned teaching data set into the deep learning prediction model to perform teaching potential prediction and generate teaching potential prediction data; Step S442: mapping the student learning behavior characteristic data to a learning progress according to the teaching potential prediction data to generate learning progress evaluation data; Step S443: Matching teaching interest and ability with the teaching potential prediction data through the teaching mode recognition data to generate interest and ability matching data; Step S444: Generate a dynamic learning path based on the learning progress evaluation data and the interest and ability matching data to obtain personalized dynamic learning trajectory data; Step S5: Set sensors for personalized dynamic learning trajectory data, and collect data at preset time intervals based on the set sensors to obtain real-time learning data streams; dynamically monitor the learning progress of the real-time learning data streams and generate a teaching resource feedback report.

2. The data management method of the cloud-based teaching service platform according to claim 1, It is characterized in that Step S1 includes the following steps: Step S11: using sensors to obtain multi-source teaching initial data streams; Step S12: performing multi-source data edge computing on the multi-source teaching initial data stream to generate a standard multi-source teaching initial data stream; Step S13: edge node confirmation is performed on the standard multi-source teaching initial data stream to obtain teaching edge node data; the standard multi-source teaching initial data stream is distributedly transmitted to the cloud server through the teaching edge node data using a distributed transmission protocol, thereby generating cloud teaching transmission data; Step S14: deduplication and integration of multi-source data of cloud teaching transmission data to generate a teaching integration data packet.

3. The data management method of the cloud-based teaching service platform according to claim 2, It is characterized in that Step S12 includes the following steps: Step S121: dividing the multi-source teaching initial data streams into data stream types to obtain teaching type divided data streams, wherein the teaching type divided data streams include video type data streams, audio type data streams and text type data streams; Step S122: performing video type encoding on the video type data stream to generate a video type encoded data stream; Step S123: performing audio noise reduction on the audio type data stream to generate an audio type noise reduction data stream; Step S124: performing text parsing on the text type data stream to generate a text type parsed data stream; Step S125: Standardize the behavior data of the video type encoding data stream, the audio type noise reduction data stream and the text type parsing data stream to generate a formatted teaching behavior data stream; synchronize the formatted teaching behavior data stream in time to generate a teaching behavior synchronization data stream; and encapsulate the teaching behavior synchronization data stream to generate a standard multi-source teaching initial data stream.

4. The data management method of the cloud-based teaching service platform according to claim 1, It is characterized in that Step S223 includes the following steps: Step S2231: Perform historical teaching resource access analysis on the teaching integration data package through key node data and teaching resource main data clusters to generate historical teaching resource access data; Step S2232: performing resource importance evaluation on the historical teaching resource access data to generate teaching resource importance evaluation data; performing usage frequency analysis on the historical teaching resource access data to generate historical teaching resource usage frequency data; performing first partition dimension confirmation based on the teaching resource importance evaluation data and the historical teaching resource usage frequency data to obtain first partition dimension data; Step S2233: performing data correlation analysis on the teaching integrated data package to generate teaching resource strong correlation data; performing data similarity calculation on the teaching integrated data package to obtain teaching resource similarity distribution data; performing second partition dimension confirmation based on the teaching resource strong correlation data and the teaching resource similarity distribution data to obtain second partition dimension data; Step S2234: performing sensitive data identification on the teaching integrated data package to obtain sensitive data of teaching resources; performing privacy protection demand analysis on the sensitive data of teaching resources to generate privacy protection demand data; performing third partition dimension confirmation through the sensitive data of teaching resources and the privacy protection demand data to obtain third partition dimension data; Step S2235: Formulate a comprehensive partitioning strategy based on the first partitioning dimension data, the second partitioning dimension data and the third partitioning dimension data to generate a teaching resource data partitioning strategy.

5. The data management method of the teaching service platform based on cloud service according to claim 1, It is characterized in that Step S5 includes the following steps: Step S51: setting sensors for personalized dynamic learning trajectory data, and collecting data at preset time intervals based on the set sensors to obtain real-time learning data streams; Step S52: dynamically monitor the learning progress of the real-time learning data stream to generate learning progress dynamic monitoring data; Step S53: Visualize the learning progress dynamic monitoring data to generate a teaching resource feedback report.