A game guild activity allocation method and a terminal

Abstract

The application discloses a game guild activity distribution method and a terminal, and relates to the technical field of game guild activity distribution, and specifically discloses a game guild activity distribution method and a terminal. The method comprises the following steps: establishing a connection channel of a game database, collecting and preprocessing guild member data according to the connection channel, wherein the guild member data comprises basic data and real-time behavior data; constructing an online time model and a game habit model, inputting time data and operation records in the basic data into the online time model and the game habit model respectively, and obtaining game time characteristics and game habit characteristics of the guild members; if an instruction triggering game guild activity recommendation is received, distributing a guild activity to the guild members according to the game time characteristics and the game habit characteristics corresponding to the guild members, and starting the guild activity. The game time characteristics and the game habit characteristics of the guild members are used to distribute appropriate guild activities to the guild members, the accuracy of guild activity recommendation is improved, and the personalized needs of the guild members are met.

Description

Technical Field

[0001] This invention relates to the field of game development technology, and in particular to a method and terminal for allocating game guild activities. Background Technology

[0002] In traditional game guild activity organization, activity arrangements are usually fixed or rely on human experience and judgment. For example, the time of guild activities is often determined by guild managers based on their own experience or the opinions of a few core members. It is difficult to accurately match the online time distribution of all guild members, resulting in many members being unable to participate due to time conflicts, which reduces the participation rate and guild activity. Furthermore, the selection of activity types also lacks a scientific and systematic basis. They are usually carried out according to a fixed guild activity schedule or the suggestions of a few members, without fully considering the differences in members' gaming habits and skill levels. This may cause novice members to be frustrated in high-difficulty activities, while high-level members lack challenge in simple activities, thus affecting members' interest and enthusiasm for guild activities. This is not conducive to the long-term development of the guild and the improvement of cohesion. In addition, during the activities, there is a lack of effective real-time monitoring and intelligent guidance mechanisms. For the cooperation problems between members in multi-person cooperative activities and the tactical adjustments in combat activities, they often have to rely on the members' own communication and on-the-spot improvisation, which is inefficient and ineffective, and easily leads to the failure of activities or failure to achieve the expected goals. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a method and terminal for allocating game guild activities, which can ensure the accuracy of data mining and analysis, thereby improving the accuracy of guild activity recommendations, meeting the personalized needs of guild members, and significantly improving the organizational efficiency and quality of guild activities.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for allocating game guild activities, including the following steps: S1. Establish a connection channel for the game database, collect and preprocess guild member data according to the connection channel, the guild member data includes basic data and real-time behavior data; S2. Construct an online time model and a game habit model. Input the time data and operation records in the basic data into the online time model and the game habit model respectively to obtain the game time characteristics and game habit characteristics of the guild members. S3. If an instruction to trigger a game guild activity recommendation is received, then guild activities are assigned to the guild members according to their game time characteristics and game habit characteristics, and the guild activities are started.

[0005] To solve the above-mentioned technical problems, another technical solution adopted by the present invention is as follows: A game guild activity allocation terminal includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the various steps of the game guild activity allocation method described above.

[0006] The beneficial effects of this invention are as follows: This invention provides a method and terminal for allocating game guild activities. By establishing a connection channel to a game database, guild member data is collected and preprocessed through this connection channel. This guild member data includes basic data and real-time behavioral data, enabling a comprehensive and accurate understanding of each guild member's game characteristics and status, providing data support for subsequent activity recommendations and optimization. An online time model and a game habit model are constructed. The time data and operation records from the basic data are input into the online time model and the game habit model, respectively, to obtain the guild member's game time characteristics and game habit characteristics. If an instruction to trigger a game guild activity recommendation is received, guild activities are allocated to the guild member based on their corresponding game time characteristics and game habit characteristics, and the guild activities are initiated. By allocating suitable guild activities based on the guild member's online time characteristics and game habit characteristics, the accuracy of guild activity recommendations is improved, meeting the personalized needs of guild members and significantly improving the organizational efficiency and quality of guild activities. Attached Figure Description

[0007] Figure 1 This is a flowchart illustrating a method for allocating game guild activities according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a game guild activity allocation terminal according to an embodiment of the present invention; Label Explanation: 1. A game guild activity allocation terminal; 2. Memory; 3. Processor. Detailed Implementation

[0008] To explain in detail the technical content, objectives, and effects of the present invention, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0009] Please refer to Figure 1 This invention provides a method for allocating game guild activities, including the following steps: S1. Establish a connection channel for the game database, collect and preprocess guild member data according to the connection channel, the guild member data includes basic data and real-time behavior data; S2. Construct an online time model and a game habit model. Input the time data and operation records in the basic data into the online time model and the game habit model respectively to obtain the game time characteristics and game habit characteristics of the guild members. S3. If an instruction to trigger a game guild activity recommendation is received, then guild activities are assigned to the guild members according to their game time characteristics and game habit characteristics, and the guild activities are started.

[0010] As can be seen from the above description, the beneficial effects of the present invention are as follows: By establishing a connection channel for the game database, guild member data is collected and preprocessed according to the connection channel. The guild member data includes basic data and real-time behavior data, so as to comprehensively and accurately understand the game characteristics and status of each guild member, providing data basis for subsequent activity recommendations and optimizations; an online time model and a game habit model are constructed, and the time data and operation records in the basic data are respectively input into the online time model and the game habit model to obtain the game time characteristics and game habit characteristics of the guild members. If an instruction to trigger the recommendation of game guild activities is received, guild activities are assigned to the guild members according to the game time characteristics and game habit characteristics corresponding to the guild members, and the guild activities are started. By assigning appropriate guild activities to guild members based on their online time characteristics and game habit characteristics, the accuracy of guild activity recommendations is improved, the personalized needs of guild members are met, and the organizational efficiency and quality of guild activities are significantly improved.

[0011] Furthermore, the preprocessing of guild member data in step S1 includes: The data member data is cleaned using a data filtering function, and then the cleaned data is organized and formatted to make it uniform.

[0012] As described above, data filtering functions are used to clean guild member data, improving its quality and reliability, ensuring more accurate and reliable subsequent data analysis results. The cleaned guild member data is then organized and formatted to ensure it is in a uniform format, simplifying the data processing process, avoiding errors and confusion caused by inconsistent data formats, and improving the efficiency and accuracy of data processing.

[0013] Furthermore, in step S2, the time data from the basic data is input into the online time model to obtain the game time characteristics of the guild members, including: Extract the online time characteristics of each guild member from the time data, and then standardize the online time characteristics. According to the preset clustering algorithm, a preset number of data are randomly selected from the processed online time features to determine the initial cluster centers. The distance between the online time feature of each guild member and each initial cluster center is calculated. According to the principle of minimum distance, the online time feature of each guild member is assigned to the nearest initial cluster center. The center feature vector is calculated based on the online time features in each cluster. The initial cluster center is updated with the center feature vector. The initial cluster center is checked to see if it has converged. If it has not converged, the online time features of each guild member and the distance to each initial cluster center are recalculated. The online time features of each guild member are then redistributed to the nearest initial cluster center according to the principle of minimum distance, until the initial cluster center converges. The clustering results are stored in the game database to obtain the game time characteristics of each guild member. The clustering results include the cluster category to which each guild member belongs and the cluster center feature vector.

[0014] As described above, by extracting and standardizing online time features, the accuracy and completeness of the data are ensured, avoiding the impact of large differences in the absolute number of logins in different time periods on the clustering effect. A preset clustering algorithm is used to determine the initial cluster centers, avoiding the influence of subjective factors and improving the fairness and accuracy of clustering. The principle of minimum distance is used to allocate online time features, ensuring that guild members within the same cluster have similar online time characteristics, facilitating subsequent analysis and processing. By continuously updating the cluster centers and checking their convergence, the stability and reliability of the clustering results are ensured. The clustering results are stored in the game database for convenient subsequent querying and use, providing data support for game guild activity recommendations.

[0015] Furthermore, in step S2, the operation records in the basic data are input into the game habit model to obtain the game habit characteristics of the guild members, including: Natural language processing technology is used to extract keywords from the operation records of each guild member, and a corresponding game habit feature vector is constructed based on the keywords; The game habit model is trained using a machine learning algorithm, and the game habit feature vector is input into the trained game habit model to obtain the game habit features of each guild member.

[0016] As described above, by extracting keywords from the operation records of each guild member using natural language processing technology, the game preferences and operation characteristics of guild members can be accurately captured. Based on the keywords, a corresponding game habit feature vector is constructed to facilitate the training of a game habit model using machine learning algorithms. This allows for the automated analysis and evaluation of guild members' game habits, thereby accurately recommending guild activities that match the game habit characteristics of guild members, thus improving the accuracy of guild activity recommendations.

[0017] Furthermore, step S3 further includes: Identify the activity type of the guild activity, activate the corresponding guild activity monitoring mechanism according to the activity type, collect real-time data of the guild activity, generate and execute corresponding adjustment strategies based on the real-time data, and record the result information of the guild activity after it ends.

[0018] As described above, by identifying the type of guild activity and initiating corresponding monitoring mechanisms, the monitoring is ensured to be targeted and effective. Real-time data of guild activities is collected to understand the progress of guild activities in a timely manner. This allows for the generation and execution of corresponding adjustment strategies based on the real-time data, thereby improving the quality of guild activities. Furthermore, the recorded results of guild activities can be used to improve the planning and organization of guild activities, thus enhancing the overall level of guild activities.

[0019] Please refer to Figure 2 Another embodiment of the present invention provides a game guild activity allocation terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the various steps of the game guild activity allocation method described above.

[0020] The game guild activity allocation method and terminal described above are applicable to organizing game guild activities, ensuring the accuracy of data mining and analysis, thereby improving the accuracy of guild activity recommendations. The following is a detailed description of the implementation methods: Please refer to Figure 1 Embodiment 1 of the present invention is: a method for allocating game guild activities, comprising the following steps: S1. Establish a connection channel for the game database, collect and preprocess guild member data according to the connection channel, the guild member data includes basic data and real-time behavior data.

[0021] In this embodiment, a connection channel for the game database is established by creating a database connection object using a database connection library (such as JDBC) based on data block configuration information (such as database address, port, username, password, etc.). This ensures interaction with the database storing game-related data, laying the foundation for subsequent data read and write operations. Guild member data is collected and preprocessed through this connection channel. This guild member data includes basic data and real-time behavioral data to comprehensively and accurately understand the game characteristics and status of each guild member, providing data support for subsequent activity recommendations and optimizations. The processes for acquiring basic data and real-time behavioral data are as follows: 1. Obtain basic data: Retrieve data such as the guild member's role, ID, registration time, and level from the member information table in the database by executing SQL query statements, and store the retrieval results in a data structure (such as an array, list, or mapping) in memory for easy access and processing later; 2. Collect real-time behavioral data: Utilize the communication interface or hook technology between the game client and server to monitor guild members' in-game actions, such as logging in, logging out, accepting and completing quests, starting and ending battles, etc. Record these events and their related data (such as login time, quest type, battle results, etc.) in a temporary cache. The cache can be an in-memory data queue or a cache database. At predetermined time intervals (such as every 5 minutes), write the data in the cache to the behavior log table in the database in batches. Use database insert statements (such as INSERT INTO...VALUES...) to persist the data and ensure that the data is not lost due to program exceptions or system failures.

[0022] Furthermore, in this embodiment, the preprocessing of guild member data in step S1 includes: using a data filtering function to clean the guild member data, and organizing and formatting the cleaned guild member data to make it uniform in format. Cleaning the guild member data through the data filtering function involves traversing the data in the behavior log table and deleting invalid data according to predefined rules (such as checking data integrity, removing duplicate records, filtering outliers, etc.). For example, if a login time record does not meet the time format requirements or the associated data of an operation record is missing, the record is marked as invalid and removed from the dataset, thereby improving efficiency. The quality and reliability of guild member data make it suitable for subsequent analysis and model building. The cleaned guild member data is converted into a unified format, meaning different types of data are processed accordingly. For example, time data is converted to a unified timestamp format using a date and time processing library for easier time series analysis. Operation record data is mapped to standardized event codes; for instance, "accepting a main quest" is mapped to code "MT1001," and "completing a combat mission and winning" is mapped to "BT2001," simplifying the data processing, avoiding errors and confusion caused by inconsistent data formats, and improving the efficiency and accuracy of data processing.

[0023] S2. Construct an online time model and a game habit model. Input the time data and operation records in the basic data into the online time model and the game habit model respectively to obtain the game time characteristics and game habit characteristics of the guild members.

[0024] In this embodiment, the game time characteristics and game habit characteristics of guild members are obtained through online time models and game habit models. This provides data support for subsequent guild activity recommendations, ensuring that the recommended guild activities match the online time and game habits of guild members and improving the accuracy of guild activity recommendations.

[0025] Furthermore, in this embodiment, step S2, inputting the time data from the basic data into the online time model to obtain the game time characteristics of the guild members, includes: extracting the online time characteristics of each guild member from the time data and standardizing the online time characteristics; randomly selecting a preset number of data from the processed online time characteristics according to a preset clustering algorithm to determine the initial cluster centers; calculating the distance between the online time characteristics of each guild member and each initial cluster center; assigning the online time characteristics of each guild member to the nearest initial cluster center according to the principle of minimum distance; calculating the center feature vector based on the online time characteristics in each cluster; updating the initial cluster centers with the center feature vector; and checking whether the initial cluster centers have converged. If they have not converged, recalculating the online time characteristics of each guild member and the distance to each initial cluster center, and assigning the online time characteristics of each guild member to the nearest initial cluster center according to the principle of minimum distance. The members are reassigned to the nearest initial cluster centers until the initial cluster centers converge. The clustering results are stored in the game database to obtain the game time characteristics of each guild member. The clustering results include the cluster category to which each guild member belongs and the cluster center feature vector. By extracting and standardizing online time characteristics, the accuracy and completeness of the data are ensured, avoiding the impact of large differences in the absolute number of logins in different time periods on the clustering effect. A preset clustering algorithm is used to determine the initial cluster centers to avoid the influence of subjective factors and improve the fairness and accuracy of clustering. The principle of minimum distance is used to assign online time characteristics so that guild members in the same cluster have similar online time characteristics, which is convenient for subsequent analysis and processing. By continuously updating the cluster centers and checking their convergence, the stability and reliability of the clustering results are ensured. The clustering results are stored in the game database for convenient subsequent querying and use, providing data support for game guild activity recommendations. The online time characteristics include: 1. Login Time Distribution: The number and frequency of guild members logging in at different times of the day (morning, afternoon, evening, etc.) reflects the guild members' activity preferences during the day. For example, some members may log in less during the day on weekdays and more frequently at night; while on weekends, their login frequency may be relatively even throughout the day. 2. First login time: The time when members log in to the game for the first time each day is important for understanding their gaming habits. For example, some members are used to logging in to check information or do some simple tasks after waking up in the morning, while others may log in for the first time after getting off work. 3. Last login time: The time when a member last logged into the game each day. Combined with the first login time, this can roughly estimate the range of a member's daily active time in the game, as well as whether there are instances of playing across different time periods (such as from evening to early morning). 4. Average online time: This is calculated by dividing the total online time of each login by the number of logins. It reflects the member's continued engagement after each login. For example, members with a longer average online time may be more inclined to participate in long-duration dungeon activities or continuous game tasks, while members with a shorter average online time may be more suited to short-duration mini-games or tasks that can be completed quickly.

[0026] 5. Consecutive Login Interval: Analyze the time interval between two consecutive logins of a member to determine the continuity and regularity of their gameplay. For example, some members may log in every day with short intervals, while others may only log in once every few days or even weeks. This difference has a significant impact on the timing and type of activities organized. For example, for members with short and stable consecutive login intervals, team activities can be organized regularly; while for members with longer intervals, some individual guidance or reward activities can be pushed when they return to help them reintegrate into the game.

[0027] In this embodiment, the specific process of obtaining the game time characteristics of guild members by analyzing the time data in the basic data through clustering algorithms is as follows: 1. Data preparation: Extract the online time characteristics of the above members from the preprocessed login time data to ensure the accuracy and completeness of the data. These data will be used as input for cluster analysis. Standardize the data to make different characteristics comparable. For example, for login time distribution data, it can be converted into a relative proportion or probability distribution form to avoid the clustering effect being affected by the large difference in the absolute number of logins in different time periods. 2. Clustering Algorithm Selection and Parameter Initialization: The K-Means clustering algorithm is selected. This algorithm divides data points into K clusters based on distance metrics, minimizing the sum of distances from each data point in a cluster to its cluster center. K data points are randomly selected as initial cluster centers, or representative points are selected as initial centers based on prior knowledge or data distribution characteristics. For example, if there is a known clear morning and evening peak login pattern, member data points with typical login characteristics during the morning and evening peak periods can be selected as initial cluster centers. 3. Assign members to clusters: Calculate the distance between each member's online time feature and each cluster center. Here, appropriate distance metrics such as Euclidean distance and Manhattan distance can be used. For example, for a member's login time distribution feature vector and a cluster center's feature vector, calculate the Euclidean distance between them. Based on the principle of minimum distance, assign each member to the cluster to which the nearest cluster center belongs. 4. Update cluster centers: For each cluster, recalculate its center feature vector by taking the average of the online time features of all members in the cluster. For example, for the login frequency feature, calculate the average login frequency of all members in the cluster in each time period, and use it as the login frequency feature value of the new cluster center. 5. Convergence Judgment: Check whether the cluster centers have converged, that is, determine whether the difference between the current cluster center and the cluster center calculated in the previous iteration is less than a preset threshold. If it is less than the threshold, the clustering process is considered to have converged and the iteration stops; otherwise, return to step 3 to continue the iteration. 6. Determining the Number of Clusters (Application of the Elbow Rule): To determine the appropriate number of clusters K, the elbow rule is used. This involves calculating the sum of squared cluster errors (SSE) for different K values ​​(usually starting from 2 and gradually increasing). SSE measures the sum of squared distances from each data point to its cluster center. A curve is plotted showing the relationship between K and SSE. Observing the trend of the curve, the K value corresponding to the "elbow" position of the curve, i.e., the point where the downward trend of SSE begins to slow down, is generally considered a suitable number of clusters. For example, when K is small, increasing K will significantly reduce SSE; however, when K increases to a certain extent, the rate of decrease in SSE significantly decreases. The K value at this point is the appropriate number of clusters. 7. Result Storage: The final clustering results, including the cluster category to which each member belongs and the cluster center feature vector, are stored in the member model table in the database. Database update statements (such as UPDATE... SET... WHERE...) are used to accurately write the relevant information into the database according to the member ID, so that these model data can be easily queried and called in the subsequent activity organization and recommendation process. This provides a basis for accurately matching activity types and times. For example, when it is necessary to recommend an activity to a member, the typical online time pattern of the cluster category to which the member belongs can be queried, and appropriate activities within the active time period can be selected for recommendation.

[0028] Furthermore, in this embodiment, step S2, inputting the operation records from the basic data into the game habit model to obtain the game habit characteristics of the guild members, includes: extracting keywords from the operation records of each guild member using natural language processing technology, constructing a corresponding game habit feature vector based on the keywords; training the game habit model using machine learning algorithms, and inputting the game habit feature vector into the trained game habit model to obtain the game habit characteristics of each guild member. By extracting keywords from the operation records of each guild member using natural language processing technology, the game preferences and operation characteristics of guild members are accurately captured, and a corresponding game habit feature vector is constructed based on the keywords. This facilitates the training of the game habit model using machine learning algorithms, thereby enabling automated analysis and evaluation of guild members' game habits through the game habit model, and accurately recommending guild activities that match the game habit characteristics of guild members. The specific process of training the game habit model using machine learning algorithms includes: 1. Data preparation stage Feature Vector Construction: From members' game operation records (such as task participation records, game-related content in chat logs, etc.), natural language processing techniques (such as word frequency statistics, keyword extraction, topic modeling, etc.) are used to extract features such as keywords of frequently participated game task types (such as "combat task", "puzzle task", "exploration task", etc.), preferred game mode keywords (such as "PVP mode", "PVE mode", etc.), and commonly used skill combination keywords (such as "healing + control skill combination", "output + movement skill combination", etc.). These features are then converted into numerical feature vectors. For example, using the bag-of-words model or TF-IDF vector representation, each keyword is mapped to a feature dimension, and its weight in the text is calculated, thereby constructing a feature vector of each member's game habits. Data partitioning: The constructed game habit feature vector dataset is divided into training set, validation set and test set. Usually, it is randomly partitioned according to a certain ratio (e.g., 70% training set, 20% validation set and 10% test set) to ensure the accuracy of the model's generalization ability assessment on different data subsets. 2. Model Selection and Initialization Choose an appropriate machine learning algorithm based on actual needs and data characteristics, such as decision tree algorithm or neural network algorithm. The initialization process of the two machine learning algorithms is as follows: Decision tree algorithm initialization: For the decision tree algorithm, determine the hyperparameters such as the maximum depth of the decision tree, the minimum number of sample splits, and the minimum number of sample leaf nodes. These hyperparameters will affect the growth and complexity of the decision tree, thereby affecting the performance of the model. For example, a small maximum depth may lead to underfitting of the model, while an excessively large maximum depth may lead to overfitting. Appropriate hyperparameter values ​​can be selected based on experience or through cross-validation. Neural network algorithm initialization: For the neural network algorithm, determine the network structure, including the number of input layer nodes (consistent with the dimension of the game habit feature vector), the number of hidden layers and the number of nodes in each hidden layer, and the number of output layer nodes (related to the number of predicted activity types). At the same time, randomly initialize the weights and bias parameters of the neural network, and continuously adjust and optimize the initial values ​​during the training process. 3. Model Training Phase Decision tree training: Using training data, the decision tree algorithm constructs a decision tree structure based on feature vectors. Starting from the root node, the algorithm selects the best feature (evaluated based on metrics such as information gain, information gain ratio, or Gini index) for splitting, dividing the dataset into different subsets. This process is repeated on each subset until a stopping condition is met (such as reaching the maximum depth, the number of node samples being less than the minimum number of splits, or the number of leaf node samples being less than the minimum number of leaf nodes). By learning from the training data, decision rules between different features are determined to predict members' preferences for different activity types. Neural network training: Input the training set data into the neural network, calculate the prediction result through forward propagation, calculate the loss function (such as mean squared error, cross-entropy loss, etc.) based on the difference between the prediction result and the true label (the member's actual preference for activity type), and use the backpropagation algorithm to propagate the error back from the output layer to the input layer. Adjust the weights and bias parameters in the network according to the error to minimize the loss function. Iterate this process repeatedly on the training set until the preset stopping conditions are met (such as reaching the maximum number of iterations, the loss function value converging, etc.). 4. Model Evaluation and Optimization Model evaluation metric selection: During training, the model is evaluated using validation set data. Appropriate evaluation metrics are selected, such as accuracy, recall, F1 score, and mean squared error, depending on the model type and the nature of the prediction task. For example, for classification problems (predicting which category a member's preference for different activity types belongs to), accuracy and F1 score can measure the accuracy of the model's predictions; for regression problems (such as predicting a member's participation rating for a certain activity), mean squared error can assess the degree of difference between the predicted value and the true value. Hyperparameter tuning and model optimization: Based on the evaluation results on the validation set, if the model performance does not meet expectations, the model's hyperparameters (such as the maximum depth of the decision tree, the learning rate of the neural network, etc.) are adjusted, and the model is retrained. The optimal combination of hyperparameters is found through grid search, random search, or more advanced optimization algorithms (such as genetic algorithms, particle swarm optimization algorithms, etc.). At the same time, regularization techniques (such as L1 regularization, L2 regularization, etc.) are used to prevent the model from overfitting and improve the model's generalization ability. 5. Model Saving and Application Once the model has been trained, evaluated, and optimized, its parameters (such as node information of the decision tree, weights and biases of the neural network, etc.) and feature weights are stored in the database. The model parameters can be serialized into binary data or text format, and then saved to the database using the database's stored procedures or file storage functions, so that the model can be quickly loaded and used for prediction during activity recommendation. In practical applications, when it is necessary to recommend activities to members, a pre-trained model is loaded from the database, and the member's game habit feature vector is input into the model. The model predicts the member's preference for different types of activities based on the learned decision rules (decision tree) or mapping relationship (neural network), thereby recommending the most suitable guild activities for members, improving activity participation and member satisfaction.

[0029] S3. If an instruction to trigger a game guild activity recommendation is received, then guild activities are assigned to the guild members according to their game time characteristics and game habit characteristics, and the guild activities are started.

[0030] In this embodiment, suitable guild activities are assigned to guild members based on their online time and gaming habits, improving the accuracy of guild activity recommendations, meeting the personalized needs of guild members, and significantly improving the organizational efficiency and quality of guild activities. The game guild activity recommendation triggering methods include timed triggering and manual triggering. Scheduled triggering: Use the scheduled task function provided by the operating system (such as the cron task of the Linux system) or a programming language to set a timer to check whether the current time meets the recommended time conditions of the activity at a predetermined time interval (such as 6 pm every day). If it does, start the activity type matching process. Manual Trigger: Provides guild administrators with an interface or command-line interface that allows them to manually trigger the activity recommendation process when needed. When the administrator performs the corresponding operation (such as clicking the "Recommend Activity" button or entering a specific command), the program receives the trigger signal and starts the activity type matching process.

[0031] Furthermore, in this embodiment, the allocation of guild activities includes online time matching, fragmented time matching, and skill level matching, specifically as follows: 1. Online Time Matching: Using SQL query statements (such as SELECT... FROM... WHERE...), the online time model data of members is retrieved based on the guild ID to obtain the distribution of members' online time, such as the proportion of online members in each time period. Based on the retrieved online time model data, suitable activity types are selected from the activity library. The activity library is a database table that stores various activity information (such as activity name, activity type, suitable number of participants, activity time requirements, etc.). The activity filtering function traverses the activity records in the activity library and matches them according to the time requirements of the activity and the current online time distribution of the members. For example, if it is found that most members have a high proportion of online time between 7 pm and 10 pm, and there is a large guild war activity in the activity library that requires to be held during this time period and is suitable for multiple participants, then this activity is selected as a candidate activity. The information of the selected candidate activities is stored in the activity list data structure in memory for subsequent recommendation to guild members. 2. Scattered Time Matching: Member filtering is achieved using SQL queries combined with conditional functions. Specifically, it retrieves information on members with scattered online times from the member model table in the database. Based on predefined criteria for scattered online time (e.g., short average daily online time and dispersed login times), it filters out members who meet these criteria. For these members, it queries their game habit model to obtain their preferred single-player activity types. It then finds corresponding asynchronous activities such as single-player exploration tasks from the activity library and recommends these activities to relevant members via the in-game messaging system (e.g., pop-up messages, emails). Simultaneously, it creates a task achievement accumulation rule table in the database, using database table creation statements (e.g., CREATE TABLE...) to define the conversion relationship between single-player task achievements and overall guild resources. For example, completing a specific single-player exploration task can add a certain number of resource points to the guild or unlock specific guild functions. After a member completes a task, the task achievement processing function updates the guild's resource data and the member's personal task completion record according to the task achievement accumulation rules. Finally, it uses database update statements to write the updated data to the corresponding database tables. 3. Skill Level Matching: Retrieve skill level data of guild members from the member information table in the database, and count the number and proportion of members with different skill levels. The counting method is to use SQL query statements (such as SELECT COUNT(), skill_level FROM member_info GROUP BY skill_level) to query the number of members for each skill level and calculate its proportion in the total number of guild members. Based on the skill level proportion information, select appropriate activity types from the activity library. If the proportion of novice members is high (e.g., more than 50%), select basic tutorial activities. These activities are associated with specific novice skill level ranges in the activity library. Use the activity selection function to query the activity library based on skill level conditions and filter out activities that meet the requirements. If there are many high-level members, select high-difficulty challenge activities. Push the recommended activity information to the message queue of guild members. Members can view the recommended activities in the message center when they log in to the game. The message queue is implemented by message middleware (such as RabbitMQ, Kafka, etc.).

[0032] Furthermore, in this embodiment, step S3 further includes: identifying the activity type of the guild activity, activating the corresponding guild activity monitoring mechanism according to the activity type, collecting real-time data of the guild activity, generating and executing corresponding adjustment strategies based on the real-time data, and recording the result information of the guild activity after its conclusion. The activity types include guild wars and team puzzles. By identifying the activity type of the guild activity, the corresponding monitoring strategy and data interface are determined, thereby activating the corresponding monitoring thread. Multi-threaded programming techniques (such as Java's Thread class and Python's threading module) are used to create the monitoring thread, with each thread running independently to avoid delays or blockages during data collection affecting the execution of other program functions, ensuring the targeting and effectiveness of monitoring. Each monitoring thread is responsible for communicating with a specific data interface of the game server to collect real-time data of the guild activity, promptly understanding the progress of the guild activity, so as to generate and execute corresponding adjustment strategies based on the real-time data, improving the quality of the guild activity, and using database insert statements (such as INSERT INTO). The activity_history…VALUES… table stores the results of guild activities (such as the outcome of guild wars, the completion time of team puzzles, etc.) in the activity history table of the database. This allows for improvements in the planning and organization of guild activities based on the results, thereby enhancing the overall quality of guild activities.

[0033] In this embodiment, corresponding adjustment strategies are generated and executed based on real-time monitored data to optimize guild activities. Taking guild battles and team puzzle-solving as examples, the optimization process includes: 1. Guild Wars (1) Data collection Combat Strength Data Acquisition: Every second, the system communicates with the game server's combat data interface, sending data requests using specific communication protocols (such as HTTP requests or custom binary protocols) to obtain combat strength data for both sides' members, including member levels, equipment attributes, skill proficiency, etc. Upon receiving the request, the server queries the game database for member information and combat-related data, returning the data to the monitoring program in JSON or binary format. After receiving the data, the monitoring program uses a data parsing library (such as a JSON parsing library or a custom binary data parser) to parse the data into in-memory data structures (such as object arrays or structure arrays) for subsequent analysis. Member positioning data acquisition: Obtain the coordinate position information of members in the battle scene from the scene data interface of the game server, and use data parsing technology to convert it into a data structure in memory, such as a two-dimensional array to represent the battle scene, and each element in the array records the member information or null value at the corresponding position; Combat operation data collection: By listening to the game server's combat log interface, the combat operation data of members, such as attack frequency and skill release order, is received. This data is stored in a data queue in memory. The data in the queue is summarized and analyzed at certain time intervals (such as every 10 seconds) to calculate statistical indicators such as average attack frequency and number of skill uses. (2) Analysis and decision making Combat Situation Analysis: The collected combat data is input into a pre-trained combat situation analysis model. This model can be a neural network model built based on deep learning algorithms. The model receives combat strength data, member positioning data, and combat operation data as input features. Through multi-layer computation of the neural network, it outputs information such as the current advantageous and disadvantageous areas of the battle, as well as the strength comparison between the two sides. For example, the model can calculate the combat power value of each member based on their level and equipment attributes, analyze the distribution of combat power in different areas based on member positioning information, and predict the development trend of the battle based on combat operation data. If it is detected that one of the guilds is at a disadvantage in a certain combat area (e.g., the area advantage score output by the model is lower than a set threshold), the tactical suggestion module is activated. The multi-layer computation process of the neural network includes: 1) Data preprocessing and input layer Data normalization: Normalizing the input combat strength data (such as member level, equipment attributes, skill proficiency, etc.), member positioning data (such as coordinate position information), and combat operation data (such as attack frequency, skill release order, etc.) makes the data have better training effect in the neural network, speeds up the model convergence speed, and improves the stability of the model. Common normalization methods include min-max normalization (mapping the data to a specific interval, such as [0, 1]) or standardization (making the data have zero mean and unit variance). Input layer construction: The normalized combat data will be passed as input to the input layer of the neural network. The number of neurons in the input layer depends on the number of features in the input data. For example, if the combat strength data includes multiple dimensions of member level, equipment attributes (such as attack power, defense power, health points, etc.), skill proficiency, etc., the member positioning data includes two-dimensional or three-dimensional coordinate information, and the combat operation data includes attack frequency, skill release order, etc., then the number of neurons in the input layer will be the sum of the number of these features. Each input data point will correspond to one neuron in the input layer, and the data value will be used as the input signal of the neuron. 2) Hidden layer calculation Hidden layer structure: Neural networks typically contain multiple hidden layers, each consisting of several neurons. The number of hidden layers and the number of neurons in each hidden layer are hyperparameters that need to be selected and adjusted according to the specific problem and data characteristics. In combat situation analysis models, the design of hidden layers aims to learn complex relationships and patterns in combat data. For example, through the calculation of multiple hidden layers, deeper information such as the collaborative relationships between members and the effectiveness of tactical layouts can be gradually extracted. Neuron computation: In each hidden layer, neurons receive output signals from neurons in the previous layer (for the input layer, the signal is the preprocessed combat data) and calculate the weighted sum. Each neuron has a set of associated weights, which are continuously adjusted during training. The weighted sum is calculated as follows: Let the input of the j-th neuron in the l-th layer be xl / j (for the first hidden layer, xl / j is the output of the input layer), and the weight be wl / ij (representing the connection weight from the i-th neuron in the (l-1)-th layer to the j-th neuron in the l-th layer), then the weighted input of this neuron is... The weighted input is processed through an activation function. Common activation functions include ReLU (corrected linear unit, f(x) = max(0, x)) and Sigmoid (f(x) = 1 / (1+ex)). The role of the activation function is to introduce nonlinear factors so that the neural network can learn the nonlinear relationships in the data. The result after processing by the activation function is used as the output of the neuron and passed to the next layer of neurons. 3) Multilayer computation and feature extraction Multi-layer information transmission and integration: As data is transmitted between hidden layers, the neural network gradually learns the feature representations of different levels in the combat data by continuously adjusting the weights. For example, in the lower hidden layers, it may learn the combat ability characteristics of individual members; in the higher hidden layers, it can integrate more complex features such as the cooperative relationship and tactical strategies between members. This multi-layer computing method enables the model to extract deeper information from the raw combat data, thereby accurately analyzing the combat situation. Feature abstraction and representation learning: Through the computation of multiple hidden layers, the neural network gradually abstracts the raw combat data into feature representations that are meaningful to the combat situation. For example, the model may learn the correlation between specific member positioning patterns and combat advantage areas, or the impact of the combination of attack frequency and skill release order on the combat outcome. These learned feature representations will be used in the output layer to generate judgments on the combat situation. 4) Output layer and result generation Output layer design: The number of neurons in the output layer is determined by the combat situation information to be predicted. In this model, the output layer may have multiple neurons, which are used to represent the advantageous area, disadvantageous area and the strength comparison between the two sides in the current battle. For example, one or more neurons can be used to represent the location of the advantageous area (such as the coordinate range), the location of the disadvantageous area, and some neurons can be used to represent the quantitative indicators of the strength comparison between the two sides (such as the difference in combat power, the prediction of the win rate, etc.). Result generation: After multiple layers of computation in the hidden layers, the neurons in the output layer calculate the final output results based on the input signals received from the last hidden layer, through weighted summation and activation functions (if necessary). These output results are the model's analysis of the current battle situation, including information such as advantageous and disadvantageous areas and the comparison of strength between the two sides. During training, the model will adjust the weights in the neural network through the backpropagation algorithm based on the difference between the actual battle results (such as the outcome, actual advantageous areas, etc.) and the predicted results, continuously optimizing the model's performance so that it can analyze the battle situation more accurately. Tactical Suggestions: The tactical suggestion module uses path planning algorithms to determine the most suitable support members, their support routes, and timing based on the current battle situation, member skill status, and battlefield map information. The path planning algorithm can use either the A algorithm or Dijkstra's algorithm. Based on terrain information (such as obstacle locations and passable areas) and member position information, it calculates the shortest or optimal path from the support member's current position to the target battle area and estimates the support time based on the member's movement speed. Simultaneously, a skill recommendation algorithm selects the best strategy for changing the attack direction or recommends specific team skill combinations. This algorithm can select the most suitable skill combination or attack direction change plan from a predefined skill strategy library based on the current battle situation (such as the enemy's defensive layout and our attack effectiveness) and member skill characteristics. These tactical suggestions are pushed to the guild commander or relevant members' interfaces via the in-game messaging system. The tactical suggestion information is also sent to the target members using in-game messaging interfaces (such as sending chat messages or pop-up notifications) so they can adjust their tactics promptly. 2. Team puzzle solving (1) Data collection Interactive data acquisition: By listening to the chat data interface of the game server, the chat history between members is obtained, and communication information related to puzzle solving is extracted, such as sharing of puzzle solving ideas and discussion of problems. Text processing technology (such as string matching, keyword extraction, etc.) is used to analyze the chat history, and the information related to puzzle solving is stored in the data structure in memory (such as string list or text database) for subsequent analysis of members' puzzle solving ideas and difficulties. Puzzle progress data collection: Obtain the puzzle-solving steps completed by each member and the current puzzle status information from the puzzle-solving task interface of the game server, and store this data in a data structure (such as a dictionary or array) in memory. The key can be the member ID, and the value can be the member's puzzle-solving progress information, such as the number of completed steps, the key parameters of the current puzzle, etc. (2) Assistance and guidance Difficulty Detection: Periodically (e.g., every 30 seconds), check the members' puzzle-solving progress data to determine if any member is encountering difficulties at any stage of the puzzle. Based on predefined difficulty judgment criteria (e.g., the number of attempts for a certain puzzle step exceeds a set threshold without progress), use a loop to traverse the puzzle progress data structure, count the number of attempts for each member at each puzzle step, and compare it with the set threshold. If a member is found to be encountering difficulties, an assistance and guidance mechanism is activated. Member Matching and Assistance Request: The member skill matching algorithm searches for members who are good at this type of puzzle-solving among guild members. The member skill matching algorithm can calculate the similarity of puzzle-solving skills between members based on the puzzle-solving skill feature information in the members' game habit model, select the member with the highest similarity as the assistant, send an assistance request message to the assistant, and use the in-game messaging system (such as sending private messages or team notification messages) to send the assistance request information to the assistant's client, informing them that there is a member who needs puzzle-solving assistance, and providing relevant puzzle information and information about the member who is having difficulties; Clue Generation and Push: The system queries the puzzle knowledge base and, based on the current puzzle status and team members' problem-solving approaches, uses a clue generation algorithm to generate targeted puzzle clues. The puzzle knowledge base is a database storing a large number of puzzle types, problem-solving approaches, and clues. Database queries can be used to retrieve relevant clues and approaches based on puzzle type and current status. The clue generation algorithm can filter, combine, and optimize clues based on information in the knowledge base and the team members' progress, generating suitable clues for the current team member. Clues are then pushed to the interface of members encountering difficulties. The in-game message push interface sends clue information to members in the form of prompts, chat messages, etc., guiding them to continue solving the puzzle and promoting cooperation and communication among team members.

[0034] Please refer to Figure 2 The second embodiment of the present invention is: a game guild activity allocation terminal 1, including a memory 2, a processor 3, and a computer program stored on the memory 2 and executable on the processor 3. When the processor 3 executes the computer program, it implements the various steps of the game guild activity allocation method of the first embodiment.

[0035] In summary, the present invention provides a method and terminal for allocating game guild activities. By establishing a connection channel to a game database, the method collects and preprocesses guild member data through this connection channel. This includes: cleaning the guild member data using a data filtering function, and then organizing and formatting the cleaned data to ensure a uniform format, simplifying the data processing, avoiding errors and confusion caused by inconsistent data formats, and improving the efficiency and accuracy of data processing; constructing an online time model and a game habit model, inputting the time data and operation records from the basic data into the online time model and the game habit model respectively, to obtain the game time characteristics and game habit characteristics of the guild members. This includes: extracting and standardizing the online time characteristics of guild members, and using a preset clustering algorithm to perform cluster analysis on the online time characteristics to obtain the game time characteristics of each guild member; extracting keywords from the operation records of each guild member, and constructing a corresponding game habit feature vector, and then using the game habit feature vector to perform cluster analysis on a pre-trained game... The system employs a game habit model to identify the game habit characteristics of each guild member, accurately capturing their game preferences and operational features. Upon receiving a command to trigger a recommended guild activity, it assigns and initiates guild activities based on the member's online time and game habit characteristics. By allocating suitable activities based on these characteristics, the system improves the accuracy of activity recommendations, meets the personalized needs of guild members, and significantly enhances the efficiency and quality of guild activity organization. The system also identifies the activity type, activates a corresponding monitoring mechanism, collects real-time data, generates and executes corresponding adjustment strategies based on this data, and records the results after each activity. By identifying the activity type and activating the corresponding monitoring mechanism, the system ensures targeted and effective monitoring, enabling the generation and execution of adjustment strategies based on real-time data and ultimately improving the overall level of guild activities.

[0036] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for allocating game guild activities, characterized in that, Including the following steps: S1. Establish a connection channel for the game database, collect and preprocess guild member data according to the connection channel, the guild member data includes basic data and real-time behavior data; S2. Construct an online time model and a game habit model. Input the time data and operation records in the basic data into the online time model and the game habit model respectively to obtain the game time characteristics and game habit characteristics of the guild members. S3. If an instruction to trigger a game guild activity recommendation is received, then guild activities are assigned to the guild members according to their game time characteristics and game habit characteristics, and the guild activities are started.

2. The method for allocating game guild activities according to claim 1, characterized in that, The preprocessing of guild member data in step S1 includes: The data member data is cleaned using a data filtering function, and then the cleaned data is organized and formatted to make it uniform.

3. The method for allocating game guild activities according to claim 1, characterized in that, In step S2, the time data from the basic data is input into the online time model to obtain the game time characteristics of the guild members, including: Extract the online time characteristics of each guild member from the time data, and then standardize the online time characteristics. According to the preset clustering algorithm, a preset number of data are randomly selected from the processed online time features to determine the initial cluster centers. The distance between the online time feature of each guild member and each initial cluster center is calculated. According to the principle of minimum distance, the online time feature of each guild member is assigned to the nearest initial cluster center. The center feature vector is calculated based on the online time features in each cluster. The initial cluster center is updated with the center feature vector. The initial cluster center is checked to see if it has converged. If it has not converged, the online time features of each guild member and the distance to each initial cluster center are recalculated. The online time features of each guild member are then redistributed to the nearest initial cluster center according to the principle of minimum distance, until the initial cluster center converges. The clustering results are stored in the game database to obtain the game time characteristics of each guild member. The clustering results include the cluster category to which each guild member belongs and the cluster center feature vector.

4. The method for allocating game guild activities according to claim 1, characterized in that, In step S2, the operation records in the basic data are input into the game habit model to obtain the game habit characteristics of the guild members, including: Natural language processing technology is used to extract keywords from the operation records of each guild member, and a corresponding game habit feature vector is constructed based on the keywords; The game habit model is trained using a machine learning algorithm, and the game habit feature vector is input into the trained game habit model to obtain the game habit features of each guild member.

5. The method for allocating game guild activities according to claim 1, characterized in that, Step S3 is followed by: Identify the activity type of the guild activity, activate the corresponding guild activity monitoring mechanism according to the activity type, collect real-time data of the guild activity, generate and execute corresponding adjustment strategies based on the real-time data, and record the result information of the guild activity after it ends.

6. A game guild activity allocation terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it performs the following steps: S1. Establish a connection channel for the game database, collect and preprocess guild member data according to the connection channel, the guild member data includes basic data and real-time behavior data; S2. Construct an online time model and a game habit model. Input the time data and operation records in the basic data into the online time model and the game habit model respectively to obtain the game time characteristics and game habit characteristics of the guild members. S3. If an instruction to trigger a game guild activity recommendation is received, then guild activities are assigned to the guild members according to their game time characteristics and game habit characteristics, and the guild activities are started.

7. A game guild activity allocation terminal according to claim 6, characterized in that, The preprocessing of guild member data in step S1 includes: The data member data is cleaned using a data filtering function, and then the cleaned data is organized and formatted to make it uniform.

8. A game guild activity allocation terminal according to claim 6, characterized in that, In step S2, the time data from the basic data is input into the online time model to obtain the game time characteristics of the guild members, including: Extract the online time characteristics of each guild member from the time data, and then standardize the online time characteristics. According to the preset clustering algorithm, a preset number of data are randomly selected from the processed online time features to determine the initial cluster centers. The distance between the online time feature of each guild member and each initial cluster center is calculated. According to the principle of minimum distance, the online time feature of each guild member is assigned to the nearest initial cluster center. The center feature vector is calculated based on the online time features in each cluster. The initial cluster center is updated with the center feature vector. The initial cluster center is checked to see if it has converged. If it has not converged, the online time features of each guild member and the distance to each initial cluster center are recalculated. The online time features of each guild member are then redistributed to the nearest initial cluster center according to the principle of minimum distance, until the initial cluster center converges. The clustering results are stored in the game database to obtain the game time characteristics of each guild member. The clustering results include the cluster category to which each guild member belongs and the cluster center feature vector.

9. A game guild activity allocation terminal according to claim 6, characterized in that, In step S2, the operation records in the basic data are input into the game habit model to obtain the game habit characteristics of the guild members, including: Natural language processing technology is used to extract keywords from the operation records of each guild member, and a corresponding game habit feature vector is constructed based on the keywords; The game habit model is trained using a machine learning algorithm, and the game habit feature vector is input into the trained game habit model to obtain the game habit features of each guild member.

10. A game guild activity allocation terminal according to claim 6, characterized in that, Step S3 is followed by: Identify the activity type of the guild activity, activate the corresponding guild activity monitoring mechanism according to the activity type, collect real-time data of the guild activity, generate and execute corresponding adjustment strategies based on the real-time data, and record the result information of the guild activity after it ends.

Images