Online activity management method and device

Abstract

The embodiment of the invention provides an online activity management method and device, and the method comprises the steps: constructing an activity full-link association data table which is used for linking all activity information; receiving application information of a user, performing logic detection on the application information according to a preset dependency syntax, determining corresponding real application information, and generating a corresponding user portrait label based on the real application information by adopting a three-dimensional attenuation factor and an integrated behavior prediction model; performing user group division on the user portrait label according to a clustering algorithm, determining a corresponding training plan, obtaining multi-dimensional learning behavior data of the user according to a computer vision technology, determining a comprehensive training score, receiving a final examination score of the user, and storing the comprehensive training score and the examination score into an activity full-link association data table; and based on the activity full-link association data table, performing one-key screening on the examination qualified personnel to generate the examination certificate. According to the invention, the efficiency and accuracy of activity full-process management can be improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to an online activity management method and device. Background Art

[0002] In the current technological landscape, event management, especially online event management involving registration, training, exams, and certification, faces numerous challenges. Traditional offline event management relies on manual processes, which are inefficient and prone to errors. Specifically, traditional methods require manual collection of registration information, offline exams or training, and the manual creation and distribution of certificates. This process is time-consuming and labor-intensive, and it is difficult to guarantee the accuracy of information and the timeliness of certificates.

[0003] With the development of internet technology, basic online event platforms have emerged. These platforms collect registration information through form-based tools and generate certificates in batches using templates after manually exporting the data. However, while this approach has improved efficiency to a certain extent, it still has significant shortcomings. First, registration, training, and certification often utilize different systems, requiring data to be imported and exported multiple times between systems. This not only increases operational complexity but can also lead to data loss or errors. Second, the certificate generation process still relies on manual operations, lacking automation and intelligentization, resulting in both the efficiency and quality of certificate generation being affected by human factors. Furthermore, such systems often fail to dynamically associate event rules, such as automatically adjusting the method and content of certificate generation based on event type, scale, or participants.

[0004] Therefore, there is an urgent need for an online activity management method that can improve the efficiency and accuracy of the entire activity management process. Summary of the Invention

[0005] In response to the problems in the prior art, the present application provides an online activity management method and device, which can improve the efficiency and accuracy of the entire activity management process.

[0006] In order to solve at least one of the above problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides an online activity management method, comprising:

[0008] Generate a corresponding activity unique identification code according to the preset activity management rules, construct an activity full-link association data table, and store the activity unique identification code in the activity full-link association data table, wherein the activity full-link association data table includes an activity attribution table, a user information table, a training result table, and an examination score table, and the activity attribution table is used to connect the user information table, the training result table, and the examination score table;

[0009] Receive user registration information, perform logic detection on the registration information according to a preset dependency syntax to determine the corresponding real registration information, construct a dynamic label generation algorithm using a three-dimensional attenuation factor and an integrated behavior prediction model, generate a corresponding user portrait label based on the dynamic label generation algorithm and the real registration information, and store the user portrait label in the user information table;

[0010] Divide the user portrait tags into user groups according to a clustering algorithm, determine a corresponding training plan, and output the training plan to the user terminal so that the user can learn according to the training plan; obtain the user's multi-dimensional learning behavior data according to computer vision technology, perform multimodal fusion on the multi-dimensional learning behavior data, determine a corresponding comprehensive training score, store the comprehensive training score in the training result table, receive the user's final test score, and store the test score in the test score table;

[0011] Obtain user full-link information based on the unique identification code of the activity, determine the qualified user and the corresponding qualified data based on the preset qualified rules, generate an examination certificate based on the qualified data and a preset template engine, and send the examination certificate to the qualified user terminal so that the qualified user can obtain the examination certificate

[0012] Furthermore, before performing a logical check on the registration information according to the preset dependency syntax to determine the corresponding real registration information, the process includes:

[0013] Performing entity extraction on the user's registration information based on a pre-trained language model;

[0014] The entity data after the entity extraction is compared with the preset information database, and the inconsistent data is deleted to determine the corresponding revised registration information.

[0015] Furthermore, before constructing the dynamic label generation algorithm using the three-dimensional attenuation factor and the integrated behavior prediction model, the method includes:

[0016] The attenuation of time, space, and behavior intensity is calculated simultaneously based on the preset spatiotemporal attenuation weight model to determine the corresponding three-dimensional attenuation factor. The spatial attenuation uses the Haversine formula to calculate the impact of geographic location changes.

[0017] Multimodal fusion is performed on the preset structured behavior data model, the preset unstructured behavior data model, and the preset time series data model to determine the corresponding integrated behavior prediction model.

[0018] Furthermore, the dynamic label generation algorithm constructed by using the three-dimensional attenuation factor and the integrated behavior prediction model includes:

[0019] Build a scenario matrix based on device type, network environment, and time period to determine the corresponding context-aware computing framework;

[0020] The corresponding three-dimensional attenuation factor is determined according to the context-aware computing framework and the spatiotemporal attenuation weight model, the integrated behavior prediction model is calibrated according to the three-dimensional attenuation factor, and the corresponding dynamic label generation algorithm is determined.

[0021] Furthermore, dividing the user portrait labels into user groups according to a clustering algorithm and determining corresponding training plans includes:

[0022] According to the clustering algorithm, the optimal learning path of historical students with similar tags to the user profile is matched to determine the initial training plan;

[0023] User learning data is acquired according to a preset time interval, the initial training plan is updated according to the user learning data and a set reinforcement learning network, and an updated training plan is determined.

[0024] Furthermore, before updating the initial training plan according to the user learning data and the set reinforcement learning network and determining the updated training plan, the method further includes:

[0025] Defining a state space according to a preset learning state, using a proximal policy optimization algorithm to adjust the execution strategy according to the learning state, and determining the corresponding action space;

[0026] A reward function is constructed based on a short-term reward and penalty mechanism, and a corresponding reinforcement learning network is determined according to the state space, the action space, and the reward function.

[0027] Furthermore, the method of obtaining multi-dimensional learning behavior data of the user based on computer vision technology, performing multimodal fusion on the multi-dimensional learning behavior data, and determining the corresponding comprehensive training score includes:

[0028] A lightweight convolutional neural network is used to capture user facial features in real time, user screen operations are captured based on the client SDK, and user eye tracking is achieved based on the Dlib library.

[0029] Multimodal fusion is performed according to the facial features, the screen operation, and the eye trajectory tracking to determine a corresponding comprehensive training score.

[0030] In a second aspect, the present application provides an online activity management device, comprising:

[0031] A full-link association relationship determination module is used to generate a corresponding activity unique identification code according to preset activity management rules, construct an activity full-link association data table, and store the activity unique identification code in the activity full-link association data table, wherein the activity full-link association data table includes an activity attribution table, a user information table, a training result table, and an examination score table, and the activity attribution table is used to connect the user information table, the training result table, and the examination score table;

[0032] A user information profile construction module is configured to receive user registration information, perform logic detection on the registration information according to a preset dependency syntax, determine the corresponding real registration information, construct a dynamic label generation algorithm using a three-dimensional attenuation factor and an integrated behavior prediction model, generate a corresponding user profile label based on the dynamic label generation algorithm and the real registration information, and store the user profile label in the user information table;

[0033] A training and test score determination module is used to divide the user portrait labels into user groups based on a clustering algorithm, determine a corresponding training plan, and output the training plan to the user terminal so that the user can learn according to the training plan; obtain the user's multi-dimensional learning behavior data based on computer vision technology, perform multimodal fusion on the multi-dimensional learning behavior data, determine the corresponding comprehensive training score, store the comprehensive training score in the training result table, receive the user's final test score, and store the test score in the test score table;

[0034] The full-link associated certificate generation module is used to obtain the user's full-link information based on the unique identification code of the activity, determine the qualified user and the corresponding qualified data according to the preset qualified rules, generate an examination certificate based on the qualified data and the preset template engine, and send the examination certificate to the qualified user terminal so that the qualified user can obtain the examination certificate.

[0035] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the online activity management method when executing the program.

[0036] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the online activity management method when executed by a processor.

[0037] In a fifth aspect, the present application provides a computer program product, comprising a computer program / instruction, which implements the steps of the online activity management method when executed by a processor.

[0038] It can be seen from the above technical solution that the present application provides an online activity management method and device, which constructs an activity full-link associated data table to link all activity information; receives the user's registration information, performs logical detection on the registration information according to the preset dependency syntax, determines the corresponding real registration information, and generates corresponding user portrait labels based on the real registration information using a three-dimensional attenuation factor and an integrated behavior prediction model; divides the user portrait labels into user groups according to a clustering algorithm, determines the corresponding training plan, obtains the user's multi-dimensional learning behavior data according to computer vision technology, determines the comprehensive training score, receives the user's final test results, and stores the comprehensive training score and test results in the activity full-link associated data table; based on the activity full-link associated data table, one-click screening of qualified candidates generates test certificates, thereby improving the efficiency and accuracy of the entire process management of the activity. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0040] Figure 1 This is one of the flow charts of the online activity management method in the embodiment of the present application;

[0041] Figure 2 This is a second flow chart of the online activity management method in an embodiment of the present application;

[0042] Figure 3 This is the third flow chart of the online activity management method in the embodiment of the present application;

[0043] Figure 4 This is a fourth flow chart of the online activity management method in an embodiment of the present application;

[0044] Figure 5 This is a fifth flow chart of the online activity management method in an embodiment of the present application;

[0045] Figure 6 This is the sixth flow chart of the online activity management method in the embodiment of the present application;

[0046] Figure 7 FIG7 is a flowchart of the online activity management method in an embodiment of the present application;

[0047] Figure 8 is a structural diagram of an online activity management device in an embodiment of the present application;

[0048] Figure 9 Schematic diagram of the structure of the electronic device in the embodiment of the present application.

[0049] Reference numerals:

[0050] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver program storage unit 9144, antenna 9111, speaker 9131, microphone 9132. DETAILED DESCRIPTION

[0051] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0052] The acquisition, storage, use, and processing of data in this application's technical solution comply with relevant national laws and regulations.

[0053] Considering that in the current technical environment, activity management, especially online activity management involving registration, training, examinations and certificate issuance, requires manual collection of registration information, offline examinations or training, and then manual production and sending of certificates one by one. This process is time-consuming and labor-intensive, and it is difficult to ensure the accuracy of information and the timeliness of certificates. The present application provides an online activity management method and device, which constructs an activity full-link associated data table to link all activity information; receives user registration information, performs logical detection on the registration information according to a preset dependency syntax, determines the corresponding real registration information, and generates corresponding user portrait labels based on the real registration information using a three-dimensional attenuation factor and an integrated behavior prediction model; divides user portrait labels into user groups according to a clustering algorithm, determines the corresponding training plan, obtains user multi-dimensional learning behavior data according to computer vision technology, determines the comprehensive training score, receives the user's final test score, and stores the comprehensive training score and test score in the activity full-link associated data table; based on the activity full-link associated data table, one-click screening of qualified test personnel generates test certificates, thereby improving the efficiency and accuracy of the entire process management of the activity.

[0054] In order to improve the efficiency and accuracy of the whole process management of activities, this application provides an embodiment of an online activity management method, see Figure 1 , the online activity management method specifically includes the following contents:

[0055] Step S101: Generate a corresponding activity unique identification code according to preset activity management rules, construct an activity full-link association data table, and store the activity unique identification code in the activity full-link association data table, wherein the activity full-link association data table includes an activity attribution table, a user information table, a training result table, and an examination score table, and the activity attribution table is used to connect the user information table, the training result table, and the examination score table;

[0056] Optionally, in this embodiment, the purpose of this step is to construct a full-link association data table for linking all subsequent activity information.

[0057] Specifically, the full-link associated data table includes:

[0058] activity_id (activity ownership table), used to ensure clear data ownership;

[0059] Participants (user information table) is used to store user profile information and is bound to the activity_id field;

[0060] training_progress (training result table) is used to store training information and is bound to the activity_id field;

[0061] exam_results (exam score table) is used to store exam score information and is bound to the activity_id field.

[0062] Each of the above tables corresponds to an activity stage of activity management.

[0063] The following examples illustrate the correspondence between each table in the full-link association data table and the activity stage. The activity process of this solution is divided into four stages. First, the activity creation stage automatically generates a unique registration link and QR code (bound to the activity ID) based on the exam activity rules (start registration time, training / exam duration, passing score, etc.). The system generates a globally unique ID (such as a UUID) for each activity and stores it in activity_id, which serves as the primary key for the full-link data association. The data tables of all stages of the activity contain the activity_id field to ensure clear data ownership.

[0064] Secondly, during the registration phase, users scan a QR code or click a link to fill in their information (name, company, etc.). After submitting the registration form, the system automatically verifies the authenticity of the submitted text using natural language processing (NLP), identifying false or abnormal input. Then, using artificial intelligence technology, it constructs a user profile. This user information (name, email address, etc.) is bound to the activity_id and stored in real time in the event-specific database (in the participants table).

[0065] During the training / exam phase, the system generates a training plan based on clustering technology and user profiles. Users then follow the training plan. Computer-based monitoring and scoring are used to monitor and score the training process. Scores of user learning behaviors (such as video viewing duration) are written to the training_progress table in real time, linking participant_id and activity_id. After the user completes training, they take the exam, record their scores, and store them in the exam_results table, again linking activity_id and participant_id.

[0066] Finally, during the certificate issuance phase, based on activity_id and preset rules (such as score ≥ 60 points), the qualified students and their related data are jointly queried from multiple tables, and the template engine Jinja2 is used to dynamically fill the query results into the certificate template. A hash value is generated for the key data of the certificate, and the certificate PDF metadata is embedded. An interface is provided for users to query the original activity data by certificate number.

[0067] By building a full-link associated data table for activities to integrate various links such as registration, training, examinations and certificate issuance, this solution realizes automated management of the entire process, screens qualified candidates for the activity with one click without manual control, extracts information based on the full-link associated data table, automatically fills it into the certificate template, and triggers the generation and delivery of certificates for all qualified students with one click, thereby improving the efficiency and accuracy of the entire process management of the activity.

[0068] It is understandable that this embodiment is illustrated by taking four activity stages as an example, but the increase or decrease in the number of activity stages does not affect the implementation of the embodiment of the present invention.

[0069] Step S102: Receive user registration information, perform logic checks on the registration information according to a preset dependency syntax to determine the corresponding real registration information, construct a dynamic label generation algorithm using a three-dimensional attenuation factor and an integrated behavior prediction model, generate a corresponding user profile label based on the dynamic label generation algorithm and the real registration information, and store the user profile label in the user information table;

[0070] Optionally, in this embodiment, this step is performed during the user registration stage, demonstrating how this application implements automated user registration information screening and user portrait construction based on artificial intelligence technology, greatly reducing labor costs, reducing review complexity, and improving process implementation efficiency.

[0071] Specifically, in the user information authenticity comparison stage, the system performs authenticity comparison on the collected user information based on NLP (natural language recognition) technology.

[0072] More specifically, this embodiment uses the pre-trained model RoBERTa to extract entities from user-submitted text (such as name, unit, and position) and compares it with public databases (such as business information databases and academic institution directories). User information that fails the comparison is marked as false registration and deleted. Dependency syntax analysis is used to detect logical contradictions and trigger a review process. Manual review is only required during the review process. After the authenticity comparison is passed, the remaining user registration information has filtered out untrue noise information, laying the foundation for subsequent user profile construction.

[0073] Specifically, during the user profile construction phase, the system uses an innovative dynamic label generation algorithm, combining a three-dimensional attenuation factor with an integrated behavior prediction model to achieve real-time dynamic updates of user profile labels. This is particularly suitable for scenarios such as training and examinations that require precise user segmentation.

[0074] More specifically, during the system's cold start phase, zero-shot profiles can be constructed based on user registration information and database data, without relying on historical data. This allows for identity-based user classification and initial user profiles. For example, a user can be labeled "engineer" or "teacher" based on their self-reported occupation and combined with database data.

[0075] For example: When user A registers, he fills in "Unit: Department of Cardiology, XX Hospital". The system automatically links his papers in the field of Cardiology published on PubMed and generates a label of "Cardiovascular Expert".

[0076] Subsequently, during the subsequent training phase, the user profile is dynamically updated based on user behavior data. Specifically, step S103 generates a training plan by clustering users based on the user identity tags in the initial user profile during the cold start phase. As users execute the training plan, the system provides feedback on their learning behavior data based on computer vision and multi-terminal interaction technology. At this point, this step updates the user profile based on the three-dimensional attenuation factor and the integrated behavior prediction model. The updated user profile is then fed back to step S103 for training plan updates, forming a complete system closed loop.

[0077] For example: Based on the initial user portrait, in the cold start phase of step S102, user C is labeled as "engineer", and the system executes the cold start mechanism of step S103, formulates an initial training plan based on the "engineer" label and provides feedback on the user's learning behavior; subsequently, step S102 updates the user portrait based on the user's learning behavior using a three-dimensional attenuation factor and an integrated behavior prediction model, and adds "not actively learning" and "frequently traveling to other places" labels for the user. At this time, it is speculated that the user may frequently travel for work reasons and therefore has insufficient time for learning. Based on the labels added in step S102, step S103 updates the training plan, changing the overall video learning to fragmented exercise practice to mobilize the user's learning enthusiasm.

[0078] Optionally, in this embodiment, this step updates the initial user portrait based on the three-dimensional attenuation factor and the integrated behavior prediction model.

[0079] Specifically, in the first step, after obtaining user learning behavior data based on the feedback from step S103, this step fuses the user-submitted information, the user information in the database, and the user learning behavior data to perform feature extraction, extracting the corresponding numerical features, text features, and graph data features. After the fusion feature extraction, the label coverage can be expanded from the traditional 3-5 categories to 20+ categories (such as industry experts, potential learners, and high-risk users).

[0080] For example:

[0081] Numerical features: such as "average score of previous exams" and "completion rate of training video viewing";

[0082] Text features: Extract keywords (such as "machine learning" and "clinical medicine") from user-submitted "personal profiles" through NLP;

[0083] Graph data features: Build a user-institution-industry association map to identify potential community relationships.

[0084] Next, in the second step, a context-aware computing framework is constructed to provide environmental context parameters to solve the "environmental adaptability" problem. Based on the context-aware computing framework, the attenuation weight model is regulated to obtain a three-dimensional attenuation factor. The three-dimensional attenuation factor is used to combine with the fusion features obtained after the above feature extraction to provide feature weights for the subsequent integrated behavior prediction model to reduce the impact of historical data on the current prediction.

[0085] Specifically, the three-dimensional decay factor is determined by a spatiotemporal decay weight model, which includes time decay, spatial decay, and behavior intensity weighting. Time decay calculates the timeliness of behavior using the formula W_t = e^(-0.02t), where t is in hours. Spatial decay calculates the impact of geographic location changes using the Haversine formula. Behavior intensity weighting assigns high weights based on the key types of operations.

[0086] For example, when a user takes a mock exam, the user completes the exam and gets a basic score of 100 points:

[0087] Time decay after 30 days: 100e^(-0.0272)≈23.2

[0088] Inter-provincial login spatial attenuation: 23.2*0.7(spatial attenuation)≈16.2

[0089] The behavior of submitting the test in one go: 16.2*1.2 (behavior intensity weighting coefficient) = 19.4

[0090] The weight calculation based on this three-dimensional attenuation factor has improved accuracy compared to the single time attenuation weight calculation. Based on the attenuation factor, the weight of the user's simulated test scores obtained 30 days ago (or longer ago) can be reduced, thereby generating a customized training plan based on the user's current simulated test score status.

[0091] Specifically, the context-aware computing framework is used to adapt to environmental parameters and construct a scenario matrix based on device type, network status, and time period detection. The context-aware computing framework is used to control the three-dimensional attenuation factors mentioned above.

[0092] For example, assuming that the user is taking a mock exam, the context-aware computing framework detects that the device type is a mobile terminal. In this case, the spatial attenuation coefficient is adjusted from 1 to 1.5. If the context-aware computing framework detects that the network status delay is too high, the behavior intensity weighted value is adjusted to reduce the behavior confidence.

[0093] Specifically, the three-dimensional attenuation factor regulated by the context-aware computing framework is fused with the multimodal features extracted in the first step and then input into the integrated behavior prediction model, so that the integrated behavior prediction model makes predictions based on the three-dimensional attenuation factor, reduces the user's too old historical training information, and increases the model's attention to the user's recent information, thereby generating a user portrait and improving the accuracy of the user portrait.

[0094] Finally, in the third step, an integrated behavior prediction model is constructed to generate user portraits and improve the accuracy of user portraits.

[0095] Specifically, the integrated behavior prediction model builds a multimodal integrated prediction framework based on the processing of multimodal data, including:

[0096] Structured data: XGBoost processes numerical features (such as learning time)

[0097] Unstructured data: Transformer processing text / image logs (such as forum posts)

[0098] Time series data: LSTM analyzes behavioral sequence patterns (such as sudden acceleration in answering questions)

[0099] When building user portrait labels, define the label logic through DSL (domain-specific language).

[0100] Combining the regulated three-dimensional attenuation factor, multimodal fusion features, and labels generated based on label logic, they are input into the multimodal integrated prediction framework for model training, thereby obtaining the final dynamic label generation model, which is used to generate user portraits, improve the accuracy of user portraits, and lay the foundation for subsequent targeted training plans.

[0101] Step S103: Divide the user portrait tags into user groups according to a clustering algorithm, determine a corresponding training plan, and output the training plan to the user terminal so that the user can learn according to the training plan. Obtain multidimensional learning behavior data of the user according to computer vision technology, perform multimodal fusion on the multidimensional learning behavior data, determine a corresponding comprehensive training score, store the comprehensive training score in the training result table, receive the user's final test score, and store the test score in the test score table;

[0102] Optionally, in this step, an adaptive training plan is generated based on the user portrait provided in step S102, and user learning behavior is fed back.

[0103] Specifically, the initial strategy for new students uses a clustering model (K-means) to match the optimal paths of students with similar history, solving the data sparsity problem in the early stages of RL and obtaining the initial training plan.

[0104] Next, the user learns based on the initial training plan. At this time, a multimodal behavior monitoring system driven by computer vision (CV) monitors the user's learning behavior, obtains the user's learning behavior data, and feeds it back to step S102 for updating the user portrait.

[0105] More specifically, user learning behavior monitoring involves using the lightweight MobileNetV3 convolutional neural network to capture students' facial features in real time, combined with liveness detection (such as blinking and micro-expression analysis) to ensure students' authenticity. Screen actions (such as focus switching and minimizing windows) are captured via browser plug-ins or client SDKs, and eye tracking is implemented using the Dlib library to detect non-learning-related behaviors (such as frequent switching to social media apps).

[0106] Facial recognition, screen operation, and eye tracking data are input into the LSTM time series model to comprehensively judge the student's concentration (output 0-100% score).

[0107] Next, based on user learning behavior, this step uses a reinforcement learning model to make real-time strategy adjustments to user learning behavior. For example, a short video summary is inserted for low-focus students to help them better digest and absorb the training content.

[0108] Specifically, reinforcement learning model construction:

[0109] Define the student status as a vector S = [historical accuracy, chapter stay time, distribution of incorrect knowledge points, and concentration score].

[0110] Use the PPO (Proximal Policy Optimization) algorithm to define the action space and dynamically adjust the learning content, such as:

[0111] Push special exercises for frequently incorrect knowledge points (Action A1);

[0112] Insert a short video summary for students with low concentration (Action A2).

[0113] Define the dual reward function and penalty mechanism:

[0114] Short-term reward: Improved accuracy of single questions (+R1); long-term reward: Passing the chapter test (+R2).

[0115] Penalty mechanism: 3 consecutive failed tests (-R3).

[0116] The reinforcement learning model updates its strategy every 30 minutes to adapt to the student's dynamic performance.

[0117] For example:

[0118] Student A had an error rate of 40% in the "Basics of Cybersecurity" section. The system automatically inserted a 3-minute micro-lecture and pushed 5 similar exercises, and the subsequent accuracy rate increased to 85%.

[0119] Student B's concentration level remains below 60%, and the system switches his learning mode to "segmented learning" (with a mandatory break of 2 minutes every 15 minutes).

[0120] It is understandable that although this step also updates the learning plan based on the reinforcement learning network, it is different from step S102 which updates the plan based on the three-dimensional attenuation factor and behavior prediction model. Step S102 is more about adjusting the user portrait, and then adjusting the overall training plan in the general direction based on the update of the user portrait. This step, on the basis of the overall training plan, adjusts the micro real-time training plan based on the user's real-time learning behavior.

[0121] The corresponding training scores obtained based on the user's performance during training are stored in the training result table and used to judge whether the trainees have qualified according to the qualification rules in the future.

[0122] Step S104: Obtain the user's full-link information based on the unique identification code of the activity, determine the qualified user and the corresponding qualified data according to the preset qualified rules, generate an examination certificate based on the qualified data and the preset template engine, and send the examination certificate to the qualified user terminal so that the qualified user can obtain the examination certificate.

[0123] Optionally, this step is the certificate issuance stage.

[0124] Specifically, this step can realize the joint query of qualified students and their related data from multiple tables based on activity_id and preset rules (such as score ≥ 60 points, training concentration ≥ 60), use the template engine (such as Jinja2) to dynamically fill the query results into the certificate template, generate a hash value for the key data of the certificate (such as participant_id + activity_id), embed the certificate PDF metadata, and provide an interface for users to query the original activity data by certificate number.

[0125] This step also has the following functions:

[0126] Batch processing: one-click triggering the generation and sending of certificates for all qualified students

[0127] Anti-counterfeiting measures: Each certificate is embedded with a unique number

[0128] This example demonstrates how this embodiment integrates various links such as registration, training, examinations, and certificate issuance by building a full-link association data table for activities. During the registration stage, false registration information is filtered out based on natural language recognition to reduce labor costs. During the training stage, user profiles are built based on the predictive model to adjust the overall training plan, and real-time training plans are adjusted based on the reinforcement model. During the certificate generation stage, qualified certificates are generated with one click based on certificate templates and qualification rules, realizing automated management of the entire process.

[0129] From the above description, it can be seen that the online activity management method provided in the embodiment of the present application can be used to link all activity information by constructing an activity full-link associated data table; receiving the user's registration information, performing logical detection on the registration information according to the preset dependency syntax, determining the corresponding real registration information, and generating corresponding user portrait labels based on the real registration information using a three-dimensional attenuation factor and an integrated behavior prediction model; dividing the user portrait labels into user groups according to a clustering algorithm, determining the corresponding training plan, obtaining the user's multi-dimensional learning behavior data according to computer vision technology, determining the comprehensive training score, receiving the user's final test results, and storing the comprehensive training score and test results in the activity full-link associated data table; based on the activity full-link associated data table, one-click screening of qualified candidates to generate test certificates can improve the efficiency and accuracy of the entire process management of the activity.

[0130] In one embodiment of the online activity management method of the present application, see Figure 2 , and can also include the following:

[0131] Step S201: extracting entities from the user's registration information based on a pre-trained language model;

[0132] Step S202: Compare the entity data after the entity extraction with the preset information database, delete the inconsistent data, and determine the corresponding revised registration information.

[0133] Optionally, in this embodiment, during the user information authenticity comparison stage, the system performs an authenticity comparison on the collected user information based on NLP (natural language recognition) technology.

[0134] This embodiment uses the pre-trained model RoBERTa to extract entities from user-submitted text (such as name, unit, and position) and compares it with public databases (such as business information databases and academic institution directories). User information that fails the comparison is marked as false registration and deleted. Dependency syntax analysis is used to detect logical contradictions and trigger a review process. Human participation is only required during the review process. After the authenticity comparison passes, the remaining user registration information has filtered out untrue noise information, laying the foundation for subsequent user profile construction.

[0135] Through step S202, this embodiment successfully filters the registration information, reduces labor costs, and increases the data authenticity of the registration information, laying the foundation for subsequent user portrait construction.

[0136] In one embodiment of the online activity management method of the present application, see Figure 3 , and can also include the following:

[0137] Step S301: Calculate the attenuation of time, space, and behavior intensity simultaneously according to a preset spatiotemporal attenuation weight model, and determine the corresponding three-dimensional attenuation factor. The spatial attenuation uses the Haversine formula to calculate the impact of geographic location changes.

[0138] Step S302: performing multimodal fusion on the preset structured behavior data model, the preset unstructured behavior data model, and the preset time series data model to determine a corresponding integrated behavior prediction model.

[0139] Optionally, in this embodiment, the three-dimensional attenuation model includes time attenuation, space attenuation, and behavior intensity weighting.

[0140] Among them, the time decay is calculated according to the formula W_t=e^(-0.02t), and the unit of t is hours;

[0141] Spatial attenuation calculates the impact of geographic location changes based on the Haversine formula;

[0142] Behavior intensity weighting assigns high weights based on the key types of operations.

[0143] Optionally, in this embodiment, the behavior prediction model is integrated:

[0144] Structured data: XGBoost processes numerical features (such as learning time)

[0145] Unstructured data: Transformer processing text / image logs (such as forum posts)

[0146] Time series data: LSTM analyzes behavioral sequence patterns (such as sudden acceleration in answering questions)

[0147] When building user portrait labels, define the label logic through DSL (domain-specific language).

[0148] Through step S302, this embodiment successfully constructs a three-dimensional attenuation model for obtaining a three-dimensional attenuation factor and constructs an integrated behavior prediction model, laying the foundation for subsequent model fusion to construct a user portrait.

[0149] In one embodiment of the online activity management method of the present application, see Figure 4 , and can also include the following:

[0150] Step S401: constructing a scenario matrix based on device type, network environment, and time period to determine the corresponding context-aware computing framework;

[0151] Step S402: determining a corresponding three-dimensional attenuation factor according to the context-aware computing framework and the spatiotemporal attenuation weight model, calibrating the integrated behavior prediction model according to the three-dimensional attenuation factor, and determining a corresponding dynamic label generation algorithm.

[0152] Optionally, in this embodiment, a context-aware computing framework is used to adapt to environmental parameters and construct a scenario matrix based on device type, network status, and time period detection. The context-aware computing framework is used to control the three-dimensional attenuation factor.

[0153] For example, assuming that the user is taking a mock exam, the context-aware computing framework detects that the device type is a mobile terminal. In this case, the spatial attenuation coefficient is adjusted from 1 to 1.5. If the context-aware computing framework detects that the network status delay is too high, the behavior intensity weighted value is adjusted to reduce the behavior confidence.

[0154] Specifically, the three-dimensional attenuation factor regulated by the context-aware computing framework is fused with the multimodal features and then input into the integrated behavior prediction model, so that the integrated behavior prediction model makes predictions based on the three-dimensional attenuation factor, reduces the user's too old historical training information, and increases the model's attention to the user's recent information, thereby generating a user portrait and improving the accuracy of the user portrait.

[0155] Through step S402, this embodiment successfully constructs a dynamic tag generation algorithm for generating and updating user portraits.

[0156] In one embodiment of the online activity management method of the present application, see Figure 5 , and can also include the following:

[0157] Step S501: Match the optimal learning paths of historical students with similar tags to the user portrait according to a clustering algorithm to determine an initial training plan;

[0158] Step S502: acquiring user learning data at a preset time interval, updating the initial training plan according to the user learning data and the set reinforcement learning network, and determining an updated training plan.

[0159] Optionally, the initial strategy for new students uses a clustering model (K-means) to match the optimal paths of students with similar history to solve the data sparsity problem in the early stages of RL and obtain the initial training plan.

[0160] Optionally, user learning data is captured in real time using a lightweight convolutional neural network, MobileNetV3, to capture student facial features. This is combined with liveness detection (e.g., blinking and micro-expression analysis) to ensure the student's true identity. Screen operations (e.g., focus switching, minimizing windows) are captured via browser plug-ins or client SDKs, and eye tracking is implemented in conjunction with the Dlib library to detect non-learning-related behaviors (e.g., frequent switching to social media apps). Facial recognition, screen operation, and eye tracking data are input into an LSTM time series model to comprehensively judge the student's concentration (outputting a 0-100% score).

[0161] Optionally, based on user learning behavior, a reinforcement learning model can be used to make real-time strategy adjustments. For example, a short video summary can be inserted for low-focus trainees to help them better digest and absorb the training content.

[0162] Through step S502 , this embodiment successfully performs adaptive micro-adjustments and real-time updates of the training program based on the user's real-time learning behavior.

[0163] In one embodiment of the online activity management method of the present application, see Figure 6 , and can also include the following:

[0164] Step S601: defining a state space according to a preset learning state, using a proximal policy optimization algorithm to adjust the execution strategy according to the learning state, and determining a corresponding action space;

[0165] Step S602: constructing a reward function based on a short-term reward and penalty mechanism, and determining a corresponding reinforcement learning network according to the state space, the action space, and the reward function.

[0166] This step is optional and is the reinforcement learning network construction process.

[0167] Define the student status as a vector S = [historical accuracy, chapter stay time, distribution of incorrect knowledge points, and concentration score].

[0168] Use the PPO (Proximal Policy Optimization) algorithm to define the action space and dynamically adjust the learning content, such as:

[0169] Push special exercises for frequently incorrect knowledge points (Action A1);

[0170] Insert a short video summary for students with low concentration (Action A2).

[0171] Define the dual reward function and penalty mechanism:

[0172] Short-term reward: Improved accuracy of single questions (+R1); long-term reward: Passing the chapter test (+R2).

[0173] Penalty mechanism: 3 consecutive failed tests (-R3).

[0174] The reinforcement learning model updates its strategy every 30 minutes to adapt to the student's dynamic performance.

[0175] For example:

[0176] Student A had an error rate of 40% in the "Basics of Cybersecurity" section. The system automatically inserted a 3-minute micro-lecture and pushed 5 similar exercises, and the subsequent accuracy rate increased to 85%.

[0177] Student B's concentration level remains below 60%, and the system switches his learning mode to "segmented learning" (with a mandatory break of 2 minutes every 15 minutes).

[0178] Through step S602 , this embodiment successfully constructs a reinforcement learning model for real-time adjustment of micro-learning plans.

[0179] In one embodiment of the online activity management method of the present application, see Figure 7 , and can also include the following:

[0180] Step S701: Use a lightweight convolutional neural network to capture user facial features in real time, capture user screen operations based on the client SDK, and implement user eye trajectory tracking based on the Dlib library;

[0181] Step S702: performing multimodal fusion according to the facial features, the screen operation, and the eye trajectory tracking to determine a corresponding comprehensive training score.

[0182] Optionally, in this embodiment, a lightweight convolutional neural network, MobileNetV3, is used to capture students' facial features in real time, combined with liveness detection (such as blinking and micro-expression analysis) to ensure the student's true identity. Screen operations (such as focus switching and minimizing windows) are captured through a browser plug-in or client SDK, and eye tracking is implemented in conjunction with the Dlib library to detect non-learning-related behaviors (such as frequent switching to social media apps).

[0183] Facial recognition, screen operation, and eye tracking data are fed into an LSTM time series model to comprehensively determine the trainee's concentration (outputting a score from 0-100%). The training score is generated based on the user's performance (behavior + concentration) during training and stored in the training results table. This score is then used to determine if a trainee has qualified based on the qualification criteria.

[0184] Through step S702, this embodiment successfully obtains user training behavior data, laying a solid foundation for subsequent adjustment of user portraits.

[0185] In order to improve the efficiency and accuracy of the whole process management of activities, the present application provides an embodiment of an online activity management device for implementing all or part of the contents of the online activity management method, see Figure 8 The online activity management device specifically includes the following contents:

[0186] The full-link association relationship determination module 10 is used to generate a corresponding activity unique identification code according to the preset activity management rules, construct an activity full-link association data table, and store the activity unique identification code in the activity full-link association data table, wherein the activity full-link association data table includes an activity attribution table, a user information table, a training result table, and an examination score table, and the activity attribution table is used to connect the user information table, the training result table, and the examination score table;

[0187] A user information profile construction module 20 is configured to receive user registration information, perform logic checks on the registration information according to a preset dependency syntax, determine the corresponding real registration information, construct a dynamic label generation algorithm using a three-dimensional attenuation factor and an integrated behavior prediction model, generate a corresponding user profile label based on the dynamic label generation algorithm and the real registration information, and store the user profile label in the user information table;

[0188] The training and test score determination module 30 is used to divide the user portrait labels into user groups according to a clustering algorithm, determine the corresponding training plan, and output the training plan to the user terminal so that the user can learn according to the training plan; obtain the user's multi-dimensional learning behavior data based on computer vision technology, perform multimodal fusion on the multi-dimensional learning behavior data, determine the corresponding comprehensive training score, store the comprehensive training score in the training result table, receive the user's final test score, and store the test score in the test score table;

[0189] The full-link associated certificate generation module 40 is used to obtain the user's full-link information based on the unique identification code of the activity, determine the qualified user and the corresponding qualified data according to the preset qualified rules, generate an examination certificate based on the qualified data and the preset template engine, and send the examination certificate to the qualified user terminal so that the qualified user can obtain the examination certificate.

[0190] From the above description, it can be seen that the online activity management device provided by the embodiment of the present application can be used to link all activity information by constructing an activity full-link associated data table; receive the user's registration information, perform logical detection on the registration information according to the preset dependency syntax, determine the corresponding real registration information, and generate corresponding user portrait labels based on the real registration information using a three-dimensional attenuation factor and an integrated behavior prediction model; divide the user portrait labels into user groups according to a clustering algorithm, determine the corresponding training plan, obtain the user's multi-dimensional learning behavior data according to computer vision technology, determine the comprehensive training score, receive the user's final test results, and store the comprehensive training score and test results in the activity full-link associated data table; based on the activity full-link associated data table, one-click screening of qualified candidates generates test certificates, thereby improving the efficiency and accuracy of the entire process management of the activity.

[0191] From a hardware perspective, in order to improve the efficiency and accuracy of full-process activity management, the present application provides an embodiment of an electronic device for implementing all or part of the online activity management method. The electronic device specifically includes the following:

[0192] A processor, memory, a communications interface, and a bus; wherein the processor, memory, and communications interface communicate with each other via the bus; the communications interface is used to transmit information between the online activity management method and related devices such as the core business system, user terminals, and related databases; the logic controller can be a desktop computer, a tablet computer, a mobile terminal, etc., but this embodiment is not limited thereto. In this embodiment, the logic controller can be implemented with reference to the embodiments of the online activity management method and the embodiments of the online activity management method in the embodiment, and their contents are incorporated herein, and repeated parts are not repeated.

[0193] It is understandable that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0194] In practical applications, portions of the online activity management method may be executed on the electronic device as described above, or all operations may be performed on the client device. The specific method may be selected based on the processing capabilities of the client device and the limitations of the user's usage scenario. This application does not impose any restrictions on this. If all operations are performed on the client device, the client device may also include a processor.

[0195] The client device may include a communication module (i.e., a communication unit) that can establish a communication connection with a remote server to implement data transmission with the server. The server may include a server on the task scheduling center side, and in other implementation scenarios, may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, a server cluster consisting of multiple servers, or a server structure of a distributed device.

[0196] Figure 9 Schematic block diagram of the system structure of the electronic device 9600 according to an embodiment of the present application. Figure 9 As shown, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It is worth noting that the Figure 9 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0197] In one embodiment, the online activity management method function may be integrated into the central processing unit 9100. The central processing unit 9100 may be configured to perform the following control:

[0198] Step S101: Generate a corresponding activity unique identification code according to preset activity management rules, construct an activity full-link association data table, and store the activity unique identification code in the activity full-link association data table, wherein the activity full-link association data table includes an activity attribution table, a user information table, a training result table, and an examination score table, and the activity attribution table is used to connect the user information table, the training result table, and the examination score table;

[0199] Step S102: Receive user registration information, perform logic checks on the registration information according to a preset dependency syntax to determine the corresponding real registration information, construct a dynamic label generation algorithm using a three-dimensional attenuation factor and an integrated behavior prediction model, generate a corresponding user profile label based on the dynamic label generation algorithm and the real registration information, and store the user profile label in the user information table;

[0200] Step S103: Divide the user portrait tags into user groups according to a clustering algorithm, determine a corresponding training plan, and output the training plan to the user terminal so that the user can learn according to the training plan. Obtain multidimensional learning behavior data of the user according to computer vision technology, perform multimodal fusion on the multidimensional learning behavior data, determine a corresponding comprehensive training score, store the comprehensive training score in the training result table, receive the user's final test score, and store the test score in the test score table;

[0201] Step S104: Obtain the user's full-link information based on the unique identification code of the activity, determine the qualified user and the corresponding qualified data according to the preset qualified rules, generate an examination certificate based on the qualified data and the preset template engine, and send the examination certificate to the qualified user terminal so that the qualified user can obtain the examination certificate.

[0202] From the above description, it can be seen that the electronic device provided in the embodiment of the present application is used to link all activity information by constructing an activity full-link associated data table; receiving the user's registration information, performing logical detection on the registration information according to the preset dependency syntax, determining the corresponding real registration information, and generating corresponding user portrait labels based on the real registration information using a three-dimensional attenuation factor and an integrated behavior prediction model; dividing the user portrait labels into user groups according to a clustering algorithm, determining the corresponding training plan, obtaining the user's multi-dimensional learning behavior data according to computer vision technology, determining the comprehensive training score, receiving the user's final test results, and storing the comprehensive training score and test results in the activity full-link associated data table; based on the activity full-link associated data table, one-click screening of qualified candidates generates test certificates, thereby improving the efficiency and accuracy of the entire activity process management.

[0203] In another embodiment, the online activity management method can be configured separately from the central processor 9100. For example, the online activity management method can be configured as a chip connected to the central processor 9100, and the functions of the online activity management method are implemented under the control of the central processor.

[0204] like Figure 9 As shown, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily have to include Figure 9 In addition, the electronic device 9600 may also include all components shown in Figure 9 For components not shown, reference may be made to the prior art.

[0205] like Figure 9 As shown, the central processing unit 9100 is sometimes also referred to as a controller or operation control, and may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.

[0206] Memory 9140 can be, for example, one or more of a cache, flash memory, hard drive, removable media, volatile memory, non-volatile memory, or other suitable devices. It can store the aforementioned failure-related information and also store programs that execute the relevant information. The CPU 9100 can execute the programs stored in memory 9140 to implement information storage or processing.

[0207] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 may be, for example, a keypad or touch input device. The power supply 9170 is used to provide power to the electronic device 9600. The display 9160 is used to display objects such as images and text. The display may be, for example, an LCD display, but is not limited thereto.

[0208] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), or a SIM card. Alternatively, it may be a memory that retains information even when power is off, can be selectively erased, and is provided with more data. Examples of such memory are sometimes referred to as EPROMs. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142 for storing application programs and function programs or processes for executing the operation of the electronic device 9600 by the central processing unit 9100.

[0209] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various driver programs for communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0210] The communication module 9110 is a transmitter / receiver that transmits and receives signals via the antenna 9111. The communication module 9110 is coupled to the central processing unit 9100 to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal.

[0211] Based on different communication technologies, multiple communication modules 9110 can be provided in the same electronic device, such as a cellular network module, a Bluetooth module, and / or a wireless local area network module. The communication module 9110 is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide audio output via the speaker 9131 and receive audio input from the microphone 9132, thereby implementing common telecommunication functions. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. Furthermore, the audio processor 9130 is also coupled to the central processing unit 9100, enabling local recording via the microphone 9132 and playback of stored audio via the speaker 9131.

[0212] Embodiments of the present application also provide a computer-readable storage medium capable of implementing all steps of the online activity management method in the above-mentioned embodiment, where the execution subject is a server or a client. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the computer program implements all steps of the online activity management method in the above-mentioned embodiment, where the execution subject is a server or a client. For example, when the processor executes the computer program, the following steps are implemented:

[0213] Step S101: Generate a corresponding activity unique identification code according to preset activity management rules, construct an activity full-link association data table, and store the activity unique identification code in the activity full-link association data table, wherein the activity full-link association data table includes an activity attribution table, a user information table, a training result table, and an examination score table, and the activity attribution table is used to connect the user information table, the training result table, and the examination score table;

[0214] Step S102: Receive user registration information, perform logic checks on the registration information according to a preset dependency syntax to determine the corresponding real registration information, construct a dynamic label generation algorithm using a three-dimensional attenuation factor and an integrated behavior prediction model, generate a corresponding user profile label based on the dynamic label generation algorithm and the real registration information, and store the user profile label in the user information table;

[0215] Step S103: Divide the user portrait tags into user groups according to a clustering algorithm, determine a corresponding training plan, and output the training plan to the user terminal so that the user can learn according to the training plan. Obtain multidimensional learning behavior data of the user according to computer vision technology, perform multimodal fusion on the multidimensional learning behavior data, determine a corresponding comprehensive training score, store the comprehensive training score in the training result table, receive the user's final test score, and store the test score in the test score table;

[0216] Step S104: Obtain the user's full-link information based on the unique identification code of the activity, determine the qualified user and the corresponding qualified data according to the preset qualified rules, generate an examination certificate based on the qualified data and the preset template engine, and send the examination certificate to the qualified user terminal so that the qualified user can obtain the examination certificate.

[0217] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present application is used to link all activity information by constructing an activity full-link associated data table; receiving the user's registration information, performing logical detection on the registration information according to the preset dependency syntax, determining the corresponding real registration information, and generating corresponding user portrait labels based on the real registration information using a three-dimensional attenuation factor and an integrated behavior prediction model; dividing the user portrait labels into user groups according to a clustering algorithm, determining the corresponding training plan, obtaining the user's multi-dimensional learning behavior data according to computer vision technology, determining the comprehensive training score, receiving the user's final test results, and storing the comprehensive training score and test results in the activity full-link associated data table; based on the activity full-link associated data table, one-click screening of qualified candidates generates test certificates, thereby improving the efficiency and accuracy of the entire process management of the activity.

[0218] The present application also provides a computer program product capable of implementing all steps of the online activity management method described in the above embodiment, where the execution subject is a server or a client. When the computer program / instructions are executed by a processor, the computer program / instructions implement the steps of the online activity management method. For example, the computer program / instructions implement the following steps:

[0219] Step S101: Generate a corresponding activity unique identification code according to preset activity management rules, construct an activity full-link association data table, and store the activity unique identification code in the activity full-link association data table, wherein the activity full-link association data table includes an activity attribution table, a user information table, a training result table, and an examination score table, and the activity attribution table is used to connect the user information table, the training result table, and the examination score table;

[0220] Step S102: Receive user registration information, perform logic checks on the registration information according to a preset dependency syntax to determine the corresponding real registration information, construct a dynamic label generation algorithm using a three-dimensional attenuation factor and an integrated behavior prediction model, generate a corresponding user profile label based on the dynamic label generation algorithm and the real registration information, and store the user profile label in the user information table;

[0221] Step S103: Divide the user portrait tags into user groups according to a clustering algorithm, determine a corresponding training plan, and output the training plan to the user terminal so that the user can learn according to the training plan. Obtain multidimensional learning behavior data of the user according to computer vision technology, perform multimodal fusion on the multidimensional learning behavior data, determine a corresponding comprehensive training score, store the comprehensive training score in the training result table, receive the user's final test score, and store the test score in the test score table;

[0222] Step S104: Obtain the user's full-link information based on the unique identification code of the activity, determine the qualified user and the corresponding qualified data according to the preset qualified rules, generate an examination certificate based on the qualified data and the preset template engine, and send the examination certificate to the qualified user terminal so that the qualified user can obtain the examination certificate.

[0223] From the above description, it can be seen that the computer program product provided by the embodiment of the present application is used to link all activity information by constructing an activity full-link associated data table; receiving the user's registration information, performing logical detection on the registration information according to the preset dependency syntax, determining the corresponding real registration information, and generating corresponding user portrait labels based on the real registration information using a three-dimensional attenuation factor and an integrated behavior prediction model; dividing the user portrait labels into user groups according to a clustering algorithm, determining the corresponding training plan, obtaining the user's multi-dimensional learning behavior data according to computer vision technology, determining the comprehensive training score, receiving the user's final test results, and storing the comprehensive training score and test results in the activity full-link associated data table; based on the activity full-link associated data table, one-click screening of qualified candidates generates test certificates, thereby improving the efficiency and accuracy of the entire process management of the activity.

[0224] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0225] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0226] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0227] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0228] Specific embodiments are used in the present invention to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A method for managing online activities, characterized in that: The method comprises: Generate a corresponding activity unique identification code according to the preset activity management rules, construct an activity full-link association data table, and store the activity unique identification code in the activity full-link association data table, wherein the activity full-link association data table includes an activity attribution table, a user information table, a training result table, and an examination score table, and the activity attribution table is used to connect the user information table, the training result table, and the examination score table; Receive user registration information, perform logic detection on the registration information according to a preset dependency syntax to determine the corresponding real registration information, construct a dynamic label generation algorithm using a three-dimensional attenuation factor and an integrated behavior prediction model, generate a corresponding user portrait label based on the dynamic label generation algorithm and the real registration information, and store the user portrait label in the user information table; Divide the user portrait tags into user groups according to a clustering algorithm, determine a corresponding training plan, and output the training plan to the user terminal so that the user can learn according to the training plan; obtain the user's multi-dimensional learning behavior data according to computer vision technology, perform multimodal fusion on the multi-dimensional learning behavior data, determine a corresponding comprehensive training score, store the comprehensive training score in the training result table, receive the user's final test score, and store the test score in the test score table; The user's full-link information is obtained based on the unique identification code of the activity, the qualified user and the corresponding qualified data are determined according to the preset qualified rules, the examination certificate is generated based on the qualified data and the preset template engine, and the examination certificate is sent to the qualified user terminal so that the qualified user can obtain the examination certificate.

2. The online activity management method according to claim 1, characterized in that: Before performing a logical check on the registration information according to the preset dependency syntax to determine the corresponding real registration information, the method includes: Performing entity extraction on the user's registration information based on a pre-trained language model; The entity data after the entity extraction is compared with the preset information database, and the inconsistent data is deleted to determine the corresponding revised registration information.

3. The online activity management method according to claim 1, characterized in that: Before constructing the dynamic label generation algorithm using the three-dimensional attenuation factor and the integrated behavior prediction model, the method includes: The attenuation of time, space, and behavior intensity is calculated simultaneously based on the preset spatiotemporal attenuation weight model to determine the corresponding three-dimensional attenuation factor. The spatial attenuation uses the Haversine formula to calculate the impact of geographic location changes. Multimodal fusion is performed on the preset structured behavior data model, the preset unstructured behavior data model, and the preset time series data model to determine the corresponding integrated behavior prediction model.

4. The online activity management method according to claim 3, characterized in that: The dynamic label generation algorithm constructed by using the three-dimensional attenuation factor and the integrated behavior prediction model includes: Build a scenario matrix based on device type, network environment, and time period to determine the corresponding context-aware computing framework; The corresponding three-dimensional attenuation factor is determined according to the context-aware computing framework and the spatiotemporal attenuation weight model, the integrated behavior prediction model is calibrated according to the three-dimensional attenuation factor, and the corresponding dynamic label generation algorithm is determined.

5. The online activity management method according to claim 1, wherein: The step of dividing the user portrait labels into user groups according to a clustering algorithm and determining corresponding training plans includes: According to the clustering algorithm, the optimal learning path of historical students with similar tags to the user profile is matched to determine the initial training plan; User learning data is acquired according to a preset time interval, the initial training plan is updated according to the user learning data and a set reinforcement learning network, and an updated training plan is determined.

6. The online activity management method according to claim 5, characterized in that: Before updating the initial training plan according to the user learning data and setting the reinforcement learning network and determining the updated training plan, the method includes: Defining a state space according to a preset learning state, using a proximal policy optimization algorithm to adjust the execution strategy according to the learning state, and determining the corresponding action space; A reward function is constructed based on a short-term reward and penalty mechanism, and a corresponding reinforcement learning network is determined according to the state space, the action space, and the reward function.

7. The online activity management method according to claim 1, wherein: The method of obtaining multi-dimensional learning behavior data of the user based on computer vision technology, performing multimodal fusion on the multi-dimensional learning behavior data, and determining a corresponding comprehensive training score includes: A lightweight convolutional neural network is used to capture user facial features in real time, user screen operations are captured based on the client SDK, and user eye tracking is achieved based on the Dlib library. Multimodal fusion is performed according to the facial features, the screen operation, and the eye trajectory tracking to determine a corresponding comprehensive training score.

8. An online activity management device, characterized in that: The device comprises: A full-link association relationship determination module is used to generate a corresponding activity unique identification code according to preset activity management rules, construct an activity full-link association data table, and store the activity unique identification code in the activity full-link association data table, wherein the activity full-link association data table includes an activity attribution table, a user information table, a training result table, and an examination score table, and the activity attribution table is used to connect the user information table, the training result table, and the examination score table; A user information profile construction module is configured to receive user registration information, perform logic detection on the registration information according to a preset dependency syntax, determine the corresponding real registration information, construct a dynamic label generation algorithm using a three-dimensional attenuation factor and an integrated behavior prediction model, generate a corresponding user profile label based on the dynamic label generation algorithm and the real registration information, and store the user profile label in the user information table; A training and test score determination module is used to divide the user portrait labels into user groups based on a clustering algorithm, determine a corresponding training plan, and output the training plan to the user terminal so that the user can learn according to the training plan; obtain the user's multi-dimensional learning behavior data based on computer vision technology, perform multimodal fusion on the multi-dimensional learning behavior data, determine the corresponding comprehensive training score, store the comprehensive training score in the training result table, receive the user's final test score, and store the test score in the test score table; The full-link associated certificate generation module is used to obtain the user's full-link information based on the unique identification code of the activity, determine the qualified user and the corresponding qualified data according to the preset qualified rules, generate an examination certificate based on the qualified data and the preset template engine, and send the examination certificate to the qualified user terminal so that the qualified user can obtain the examination certificate.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the online activity management method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the online activity management method according to any one of claims 1 to 7 are implemented.

Images