A method, device, equipment, and storage medium for dynamic growth level management of users across multiple entertainment ecosystems.

Abstract

This invention provides a method, device, equipment, and storage medium for dynamic user growth level management based on a multi-ecosystem pan-entertainment model. It asynchronously collects user behavior data via message queues, constructs user behavior feature vectors containing characteristic values ​​of various behavior types, and calculates user scores based on these feature vectors through weighted calculations using configured weights and dynamic adjustment coefficients. Historical scores and monthly scores are combined using weighted coefficients to obtain a comprehensive score, achieving a balanced assessment of long-term user value and short-term activity. User levels are determined through a level mapping function based on the comprehensive score, and the level status is dynamically adjusted based on upgrade and level retention conditions. The dynamic adjustment coefficients are updated in real-time based on changes in user level and the contribution of each behavior type, forming an adaptive level management closed loop. This allows the system to automatically learn and optimize the importance of different behavior types, thereby achieving a truly fair and intelligent user value assessment system.

Description

Technical Field

[0001] This invention relates to the field of platform user management, and in particular to a method, apparatus, device, and storage medium for managing the dynamic growth levels of users across multiple entertainment ecosystems. Background Technology

[0002] With the booming development of the internet entertainment industry, diversified business scenarios such as live streaming, social networking, and gaming are constantly emerging. User level systems have become a core mechanism for enhancing user stickiness and promoting the prosperity of the platform ecosystem. Effective user level management can not only incentivize users to remain active but also create greater commercial value for the platform. However, existing user level management technologies have significant shortcomings in addressing multi-scenario and multi-dimensional user behavior assessments.

[0003] Existing user ranking systems generally employ a single-dimensional evaluation standard, primarily relying on users' payment amounts to calculate rankings, neglecting users' contributions in other dimensions such as content creation, social interaction, and platform activity. This mechanism not only easily leads to an unbalanced ecosystem where "paying equals being strong," but also discourages many valuable but low-paying users from progressing. Furthermore, traditional linear point accumulation methods lack dynamic adjustment mechanisms, failing to accurately reflect users' current activity levels and resulting in a rigid ranking system.

[0004] In view of the above, this application is hereby submitted. Summary of the Invention

[0005] This invention discloses a method, device, equipment, and storage medium for managing the dynamic growth level of users across multiple entertainment ecosystems, aiming to achieve a comprehensive quantification of the multi-dimensional behavioral value of users in order to truly reflect the current active status of users.

[0006] The first embodiment of the present invention provides a method for managing the dynamic growth level of users based on a pan-entertainment multi-ecosystem, including:

[0007] By asynchronously collecting user behavior data across multiple scenarios and services through message queues, a user behavior feature vector containing feature values ​​of various behavior types is constructed.

[0008] Based on the user behavior feature vector, the user score is calculated by weighting the configured weights and dynamic adjustment coefficients. The user score is then accumulated into the historical score and the monthly score, and the historical score and the monthly score are combined by the historical score weight coefficient to obtain the comprehensive score.

[0009] The user level is determined by a level mapping function based on the comprehensive score. Based on the current level and the monthly points, it is determined whether the user meets the conditions for upgrading or maintaining the level, and the user level status and level validity period are updated accordingly.

[0010] Based on the changes in user level and the contribution of each behavior type in the user behavior feature vector, the dynamic adjustment coefficients corresponding to each behavior type are updated through an iterative algorithm controlled by the learning rate. The updated dynamic adjustment coefficients are then applied to the integral calculation in the next cycle to form an adaptive level management closed loop.

[0011] Preferably, the calculation process of the user score involves multiplying each feature value in the user behavior feature vector by the corresponding configuration weight and dynamic adjustment coefficient, and summing all weighted feature values ​​to obtain the user score. The specific expression is as follows:

[0012] in, The configuration weight for the i-th behavior, Let m be the feature value of the i-th behavior, and m be the total number of behavior types. This is a dynamic adjustment coefficient;

[0013] The comprehensive score is calculated by weighting historical scores and monthly scores, and its expression is as follows:

[0014]

[0015] in, For comprehensive scoring, Historical integral weighting coefficients, For historical integration, Monthly points.

[0016] Preferably, the iterative algorithm controlling the learning rate updates the dynamic adjustment coefficients corresponding to each behavior type, specifically as follows:

[0017] The difference between the current period and the previous period is calculated as the change in level. The proportion of the feature value of each behavior type to the sum of all feature values ​​is calculated as the contribution of that behavior type.

[0018] Multiplying the change in grade, contribution, and learning rate, and adding this product to the original dynamic adjustment coefficient yields a new dynamic adjustment coefficient, expressed as follows:

[0019]

[0020] in, The level of the current cycle. The level of the previous cycle, As for contribution level, This is the current dynamic adjustment coefficient. This is the updated dynamic adjustment coefficient. This represents the learning rate for the current period.

[0021] Preferably, the learning rate employs an adaptive adjustment strategy, increasing the learning rate when two consecutive level changes occur in the same direction and decreasing the learning rate when they occur in opposite directions, specifically:

[0022]

[0023] in, To adjust the factor, For symbolic functions, This represents the current cycle's trend of level changes. This represents the trend of level changes in the previous cycle.

[0024] Preferably, it further includes:

[0025] Calculate the user's points acquisition rate within a fixed time window. When the points acquisition rate exceeds the average rate plus n times the standard deviation, mark the user as an abnormal user. Multiply all dynamic adjustment coefficients of the abnormal user by a preset penalty factor. Also, periodically calculate the user retention rate and optimize the configuration weights of each behavior type based on the trend of retention rate changes.

[0026] Preferably, it further includes:

[0027] The CRC32 hash algorithm is used to calculate the user's unique identifier. The result is modulo 32 to determine the shard number to which the user data belongs. A historical points table and a monthly points table are created for each shard. The table names are generated by a formatting function. A Redis distributed lock mechanism is used to ensure the atomicity of user level updates. The lock key value is composed of the user's unique identifier, level value and update timestamp.

[0028] The second embodiment of the present invention provides a device for managing the dynamic growth level of users across multiple entertainment ecosystems, comprising:

[0029] The behavior data acquisition unit is used to asynchronously collect user behavior data in multiple scenarios and services through a message queue, and construct a user behavior feature vector containing feature values ​​of multiple behavior types.

[0030] The comprehensive score calculation unit is used to calculate user points based on the user behavior feature vector by weighting operations with configured weights and dynamic adjustment coefficients, accumulate the user points into historical points and monthly points respectively, and combine historical points and monthly points with historical point weight coefficients to calculate a comprehensive score.

[0031] The level management unit is used to determine the user level through the level mapping function based on the comprehensive score, determine whether the user meets the upgrade conditions or level retention conditions based on the current level and the monthly points, and update the user level status and level validity period accordingly.

[0032] The parameter adjustment unit is used to update the dynamic adjustment coefficients corresponding to each behavior type according to the changes in user level and the contribution of each behavior type in the user behavior feature vector through an iterative algorithm controlled by the learning rate. The updated dynamic adjustment coefficients are then applied to the integral calculation of the next cycle to form an adaptive level management closed loop.

[0033] The third embodiment of the present invention provides a device for managing the dynamic growth level of users based on a pan-entertainment multi-ecosystem, characterized in that it includes a memory and a processor, wherein the memory stores a computer program, and the computer program can be executed by the processor to implement the method for managing the dynamic growth level of users based on a pan-entertainment multi-ecosystem as described above.

[0034] The fourth embodiment of the present invention provides a computer-readable storage medium, characterized in that it stores a computer program, which can be executed by the processor of the device where the computer-readable storage medium is located, to implement the dynamic growth level management method for users based on a pan-entertainment multi-ecosystem as described in any of the above.

[0035] Based on the present invention, a method, device, equipment, and storage medium for dynamic growth level management of users across multiple entertainment ecosystems are provided. This invention asynchronously collects user behavior data via message queues, constructs user behavior feature vectors containing characteristic values ​​of various behavior types, and calculates user scores based on these feature vectors through weighted calculations using configured weights and dynamic adjustment coefficients. Historical scores and monthly scores are combined using weighted coefficients to obtain a comprehensive score, achieving a balanced assessment of long-term user value and short-term activity. User levels are determined through a level mapping function based on the comprehensive score, and the level status is dynamically adjusted based on upgrade and level retention conditions. The dynamic adjustment coefficients are updated in real time according to changes in user levels and the contribution of each behavior type, forming an adaptive level management closed loop. This allows the system to automatically learn and optimize the importance of different behavior types, thereby achieving a truly fair and intelligent user value assessment system. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating a method for managing the dynamic growth levels of users across multiple entertainment ecosystems, provided in the first embodiment of the present invention.

[0037] Figure 2 This is a schematic diagram of a module for a dynamic growth level management device for users in a pan-entertainment multi-ecosystem, provided in the second embodiment of the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] This invention discloses a method, device, equipment, and storage medium for managing the dynamic growth level of users across multiple entertainment ecosystems, aiming to achieve a comprehensive quantification of the multi-dimensional behavioral value of users in order to truly reflect the current active status of users.

[0040] The first embodiment of the present invention provides a method for managing the dynamic growth level of users in a pan-entertainment multi-ecosystem, which can be executed by a pan-entertainment multi-ecosystem user dynamic growth level management device (hereinafter referred to as management device or system), specifically, by one or more processors within the management device or system, to at least implement the following steps:

[0041] S101 asynchronously collects user behavior data under multiple scenarios and services through message queues, and constructs user behavior feature vectors containing feature values ​​of multiple behavior types.

[0042] In this embodiment, the management device can be a terminal with data processing capabilities, such as a laptop, desktop computer, or server workstation. The management device can be equipped with a corresponding operating system and application software, and the functions required in this embodiment can be realized through the combination of the operating system and application software.

[0043] This embodiment uses the Pulsar message queue as the core component for data acquisition, constructing an asynchronous data processing architecture with multiple producers and consumers. In various business scenarios of the pan-entertainment platform, including live streaming, chat room, family, and social scenarios, when a user performs any action, the corresponding business system encapsulates the action into a message event and pushes it to the Pulsar queue. For example, when a user sends a virtual gift in a live stream, the live streaming system generates a message containing information such as user ID, gift type, gift value, and gifting time; when a user interacts on stage in a chat room, the chat room system records the user's start and end times on stage, calculates the stage duration, and generates a corresponding behavioral event message.

[0044] Each behavioral event is standardized into a unified data structure upon entering the Pulsar queue, containing three core attributes: event occurrence time, behavior type enumeration value, and behavior value. The behavior type enumeration value uses a predefined encoding system; for example, virtual gift giving is coded as 1, recharge amount as 2, item consumption as 3, PK win / loss result as 4, live chat duration as 5, viewing time as 6, leaderboard ranking as 7, and clan check-in as 8. The behavior value has different meanings depending on the behavior type. For example, the behavior value for virtual gift giving is the gift value (unit: yuan), the behavior value for live chat duration is the duration (unit: seconds), and the behavior value for PK win / loss result may include rules such as 1 point for a win, 0.5 points for a draw, and 0 points for a loss.

[0045] During data collection, the system filters behavioral events in real time using business filtering rules configured in Redis. For example, for virtual gift-giving, a minimum threshold of 0.1 yuan might be set; gift-giving below this threshold will be filtered out. For live-streaming duration statistics, a minimum valid duration of 30 seconds might be set; live-streaming sessions shorter than this duration will not be counted towards points. This filtering mechanism is implemented through an indicator function; the indicator function returns 1 when the behavior meets the configured business conditions, and 0 otherwise.

[0046] For the filtered behavioral data, the system will assign personal behavioral bonuses based on the user's special identity or the special items used. For example, when a user uses a 1.2x points card, the personal bonus coefficient for all their behaviors is 1.2; when a user uses a 2x experience item during a specific event, the bonus coefficient is 2.0. The system multiplies the filtered behavioral values ​​by the personal bonus coefficients to obtain the bonus behavioral value.

[0047] When constructing user behavior feature vectors, the system maintains an 8-dimensional vector for each user, with each dimension corresponding to a behavior type. Within a statistical period (e.g., one day), the system continuously accumulates the values ​​of the same type of behavior generated by the user in various scenarios. For example, if a user may send gifts in multiple live streams within a day, the system will sum up the value of all gifts as the feature value of the virtual gift-giving dimension; if a user may enter different chat rooms and speak multiple times, the system will sum up the duration of all speaking sessions as the feature value of the live stream duration dimension.

[0048] S102, Based on the user behavior feature vector, calculate the user score by weighting the configured weights and dynamic adjustment coefficients, accumulate the user score into the historical score and monthly score respectively, and combine the historical score and monthly score through the historical score weight coefficient to calculate the comprehensive score.

[0049] Once the system obtains user behavior feature vectors, it needs to convert this multi-dimensional behavioral data into a unified score. In this embodiment, the system uses a weighted aggregation method for score calculation. Specifically, the system maintains a weight configuration table in the etcd configuration center, setting corresponding weight coefficients for each behavior type. For example, in actual operation, the weight of virtual gift giving might be set to 0.3, indicating that this behavior accounts for 30% of the overall score calculation; the weight of recharge amount might be set to 0.2; and non-paid behaviors such as live streaming duration and viewing time might each account for a weight of 0.1-0.15. These weight configurations can be set according to the actual situation. In this implementation, both the user's paid contribution and the value of the user in content creation and social interaction are recognized.

[0050] In addition to fixed configuration weights, the system also introduces a dynamic adjustment coefficient mechanism. Each behavior type has an independent dynamic adjustment coefficient, initially set to 1.0. This coefficient automatically adjusts based on changes in the user's level, achieving adaptive optimization of the points strategy. For example, if the system detects that a user primarily earns points by watching live streams but their level is increasing slowly, it may appropriately increase the dynamic adjustment coefficient for that user's viewing time to encourage continued activity.

[0051] When calculating user scores, the system iterates through each dimension of the user behavior feature vector, multiplies the feature value, configuration weight, and dynamic adjustment coefficient, and then sums the results for all dimensions (the specific expression is: For example, a user's behavioral feature vector within a statistical period is [600, 100, 50, 2, 180, 300, 5, 7], with corresponding configuration weights of [0.3, 0.2, 0.1, 0.05, 0.1, 0.1, 0.1, 0.05], and a dynamic adjustment coefficient of 1.0 for each weight. The user's score calculation process is: 600×0.3×1.0 + 100×0.2×1.0 + 50×0.1×1.0 + 2×0.05×1.0 + 180×0.1×1.0 + 300×0.1×1.0 + 5×0.1×1.0 + 7×0.05×1.0 = 253.45 points.

[0052] After earning points for a single transaction, the system needs to update the user's historical points and monthly points. Historical points are accumulated, meaning that points earned each time are added to the user's total historical points. This value accumulates continuously from the user's registration date, representing the user's long-term value contribution to the platform. Monthly points are reset to 0 on the first day of each calendar month, with all points earned within the month added to the monthly points, reflecting the user's short-term activity level.

[0053] To balance long-term value and short-term activity, the system combines the two using a historical integral weighting coefficient β to calculate a comprehensive score (the expression for which is: In this embodiment, the value of β is typically set between 0.3 and 0.7, with the specific value determined based on the platform's operational strategy. For example, when β = 0.4, it means that historical points account for 40% of the overall score, and monthly points account for 60%. This protects the vested interests of existing users while providing growth opportunities for new and reactivating users.

[0054] S103, determine the user level through the level mapping function based on the comprehensive score, determine whether the user meets the upgrade conditions or level retention conditions based on the current level and the monthly points, and update the user level status and level validity period accordingly;

[0055] The system uses a segmented mapping method to convert a user's overall score into a specific level value. A level mapping table is stored in the etcd configuration center, defining the score range corresponding to each level. For example, 0-999 points correspond to level 1, 1000-4999 points correspond to level 2, 5000-19999 points correspond to level 3, and so on, up to a preset level limit.

[0056] After determining a user's base level, the system further considers monthly points to determine whether to upgrade or downgrade. Historical points determine the user's base level, while monthly points determine whether the user can maintain or upgrade their current level. In practice, the system sets upgrade and retention thresholds for each level. The upgrade threshold is typically the same as or slightly higher than the maximum points required for that level, while the retention threshold is set to 60% of the previous level's maximum points. For example, a level 3 user might have an upgrade threshold of 20,000 points, while their retention threshold is 3,000 points (60% of the level 2 maximum of 5,000 points).

[0057] The system executes level determination logic after each points update. First, it checks if the user's monthly points have reached the upgrade threshold. If so, the user's level is upgraded by one level, and the level's validity period is set to the end of the following month. This relatively long protection period is designed to reward users who actively upgrade, giving them ample time to enjoy the benefits of the new level. If the monthly points have not reached the upgrade threshold but the level retention conditions are met, the system will maintain the user's current level, but the validity period will only be extended to the end of the following month. This protects the user's existing interests while maintaining a certain level of activity pressure.

[0058] For users who do not meet either the upgrade or retention requirements, the system will not immediately downgrade them. Instead, a scheduled task will be used to process these users at midnight on the 1st of each month. This batch processing method avoids the negative user experience caused by frequent level changes. During the downgrade process, the system will recalculate the user's correct level based on their actual monthly points, potentially lowering them by one or more levels until a level is found that allows them to retain their current level.

[0059] S104. Based on the changes in user level and the contribution of each behavior type in the user behavior feature vector, the dynamic adjustment coefficients corresponding to each behavior type are updated through an iterative algorithm controlled by the learning rate. The updated dynamic adjustment coefficients are then applied to the integral calculation in the next cycle to form an adaptive level management closed loop.

[0060] After a user completes a period of behavioral data collection (e.g., one day), the system calculates the user's level change. By comparing the current level with the level of the previous period, the system obtains the level change amount. This change amount can be positive (upgrade), zero (maintain level), or negative (downgrade), directly reflecting the incentive effect of the current points strategy on the user. Simultaneously, the system analyzes the contribution of each dimension in the user's behavioral feature vector, that is, calculating the proportion of each behavior type's feature value to the total feature value. For example, if a user's virtual gift-giving feature value is 600, and the sum of all behavioral feature values ​​is 1000, then the contribution of virtual gift-giving is 0.6, indicating that 60% of the user's points come from gift-giving behavior.

[0061] Based on the change in user level and the contribution of behaviors, the system updates the dynamic adjustment coefficient using a gradient-up approach. If a user's level increases, it indicates that the current points-based strategy is effective, and the system will reinforce behaviors with high contributions, appropriately increasing their dynamic adjustment coefficients. Conversely, if a user's level decreases, the system will decrease the adjustment coefficients for major contributing behaviors, while potentially increasing the coefficients for other behavior types, guiding users to engage in more diverse platform activities. This update is achieved by multiplying the change in level, contribution, and learning rate, and then adding the result to the original adjustment coefficient. The expression is as follows: .

[0062] The learning rate controls the step size throughout the update process; its specific expression is: The system employs an adaptive learning rate strategy, dynamically adjusting the learning rate by analyzing the trend of level changes over two consecutive cycles. When the system detects that two consecutive level changes occur in the same direction (e.g., two consecutive upgrades or two consecutive downgrades), it indicates that the current adjustment direction is correct, and the system multiplies the learning rate by an adjustment factor of 1.2 to accelerate convergence. Conversely, if the level changes occur in opposite directions (e.g., an upgrade followed by a downgrade), it indicates that over-adjustment may have occurred, and the system multiplies the learning rate by 0.8 to slow down the adjustment and prevent system oscillations.

[0063] In practice, this adaptive mechanism has demonstrated good results. For example, for a user who primarily earns points through top-ups, if their level stagnates for an extended period, the system will gradually reduce the dynamic adjustment coefficient for top-up behavior while increasing the coefficients for other behaviors such as content creation and social interaction, guiding the user to participate more comprehensively in platform activities. For users with high activity levels and diverse behaviors, the system will maintain a relative balance across various coefficients to ensure they can grow through multiple avenues.

[0064] The updated dynamic adjustment coefficients are immediately applied to the next cycle's points calculation, creating a real-time feedback loop. This allows the system to automatically optimize the points strategy based on each user's behavioral characteristics and growth trajectory, avoiding both a one-size-fits-all approach and reducing the need for manual intervention. Through continuous learning and adjustment, the system will eventually find the most suitable combination of incentive parameters for each user, achieving a personalized level progression path.

[0065] In one possible implementation of the present invention, it further includes:

[0066] Calculate the user's points acquisition rate within a fixed time window. When the points acquisition rate exceeds the average rate plus n times the standard deviation, mark the user as an abnormal user. Multiply all dynamic adjustment coefficients of the abnormal user by a preset penalty factor. Also, periodically calculate the user retention rate and optimize the configuration weights of each behavior type based on the trend of retention rate changes.

[0067] It should be noted that the system continuously monitors each user's points accumulation through a sliding time window. In this embodiment, the time window is set to 24 hours, and the system calculates the user's points accumulation rate over the past 24 hours every hour. Specifically, the system reads the user's points at the start and end times of the time window from the database, and the difference between the two is the points accumulation rate divided by the time interval. For example, if a user's points increase from 1000 to 2500 in 24 hours, their points accumulation rate is 62.5 points / hour.

[0068] To identify abnormal behavior, the system maintains a global statistical model of points acquisition rate. This model calculates the average points acquisition rate and standard deviation for all active users in real time. In actual operation, it was found that the points acquisition rate of normal users roughly follows a normal distribution, while the rate of malicious point-farming users deviates significantly from the normal range. Based on this characteristic, the system uses the 3σ principle in statistics for anomaly detection. When a user's points acquisition rate exceeds the average plus three times the standard deviation, the system automatically marks that user as an abnormal user. This effectively identifies malicious behavior while avoiding misjudgment of normal, highly active users.

[0069] For users marked as abnormal, the system does not directly ban or reset their points. Instead, it employs a more lenient restriction strategy. Specifically, the dynamic adjustment coefficient for all of the user's behavior types is uniformly multiplied by a penalty factor of 0.5. This means that the points earned by the abnormal user for all subsequent behaviors will be halved, serving both a punishment purpose and giving the user an opportunity to correct their behavior. If the user returns to normal behavior patterns during the subsequent observation period, the system will gradually lift the restrictions, restoring the penalty factor from 0.5 to 1.0.

[0070] In addition to anomaly detection, the system also optimizes the overall points strategy by analyzing user retention rates. At the beginning of each month, the system calculates the percentage of users who are still active among those who registered or were active 30 days ago; this percentage is the monthly retention rate. By comparing retention rate data over several consecutive months, the system can evaluate the effectiveness of the current points strategy. If the retention rate shows a downward trend, it indicates that the current incentive mechanism may have problems, and the configuration weights of each behavior type need to be adjusted.

[0071] The weight optimization is implemented using A / B testing combined with machine learning. The system randomly divides users into multiple experimental groups, assigning slightly different weight configurations to each group. For example, group A receives a higher weight for non-paying behaviors, while group B receives a higher weight for social interactions. After a testing period, the system analyzes key metrics such as retention rate, activity level, and paid conversion rate for each group, selecting the best-performing weight configuration to roll out to all users. This ensures that the ranking system remains aligned with user needs and platform development goals, enabling continuous system evolution and optimization.

[0072] In one possible implementation of the present invention, it further includes:

[0073] The CRC32 hash algorithm is used to calculate the user's unique identifier. The result is modulo 32 to determine the shard number to which the user data belongs. A historical points table and a monthly points table are created for each shard. The table names are generated by a formatting function. A Redis distributed lock mechanism is used to ensure the atomicity of user level updates. The lock key value is composed of the user's unique identifier, level value and update timestamp.

[0074] When user data needs to be stored or retrieved, the system first performs a CRC32 hash operation on the user's UUID. The UUID is a 32-bit unique identifier generated by the platform for each user, consisting of numbers, letters, and underscores. The CRC32 algorithm converts this string into a 32-bit integer value, then performs a modulo operation on 32 to obtain a shard number between 0 and 31. For example, if the user's UUID is "a1b2c3d4_e5f6g7h8_i9j0k1l2_m3n4o5p6", the CRC32 operation yields a hash value of 2147483647, and the modulo operation gives shard number 31. All of this user's data will be stored in shard number 31.

[0075] The system creates two types of data tables for each shard: a historical points table and a monthly points table. Table names are generated using the `sprintf` formatting function, following the naming convention of "radiant_" followed by a 4-digit shard number. For example, the historical points table for shard 0 is named "radiant_0000", and the monthly points table for shard 31 is named "radiant_0031". This ensures standardized table names and facilitates quick data location for maintenance personnel. Each table contains core fields such as user UUID, points value, and update time. The historical points table records the user's accumulated points since registration, while the monthly points table is reset to zero at the beginning of each month.

[0076] In the critical operation of user level updates, the system must ensure data consistency and atomicity. Since multiple service instances may simultaneously process the same user's request in a distributed environment, without proper concurrency control, duplicate level upgrades or data inconsistencies may occur. Therefore, the system employs a Redis distributed lock mechanism for protection.

[0077] When a user's level needs to be updated, the system first constructs a lock key value. The lock key value consists of three parts: the user's UUID, the target level value, and the current timestamp, separated by colons. For example, to upgrade user "a1b2c3d4_e5f6g7h8_i9j0k1l2_m3n4o5p6" to level 5,

[0078] At timestamp 1625097600, the lock key value is:

[0079] "a1b2c3d4_e5f6g7h8_i9j0k1l2_m3n4o5p6:5:1625097600". This ensures that each level change has a unique lock identifier, avoiding duplicate processing.

[0080] The system uses the Redis SETNX command to attempt to acquire the lock, which will only succeed if the key does not exist. If the lock is acquired successfully, the system performs a level update operation, including updating the user level table in MySQL, logging the level change, and sending a level change notification. The lock is released after the operation is complete. If acquiring the lock fails, it means another process is processing the user's level update, and the current request will be rejected or wait for a retry. The lock expiration time is set to 30 seconds, ensuring sufficient time for normal operations to complete while preventing deadlocks due to exceptions.

[0081] The second embodiment of the present invention provides a device for managing the dynamic growth level of users across multiple entertainment ecosystems, comprising:

[0082] The behavior data acquisition unit 201 is used to asynchronously collect user behavior data in multiple scenarios and services through a message queue, and construct a user behavior feature vector containing feature values ​​of multiple behavior types.

[0083] The comprehensive score calculation unit 202 is used to calculate user points based on the user behavior feature vector by weighting the configured weights and dynamic adjustment coefficients, accumulate the user points into historical points and monthly points respectively, and combine the historical points and monthly points through the historical point weight coefficient to calculate the comprehensive score.

[0084] The level management unit 203 is used to determine the user level through the level mapping function based on the comprehensive score, determine whether the user meets the upgrade conditions or level retention conditions based on the current level and the monthly points, and update the user level status and level validity period accordingly.

[0085] The parameter adjustment unit 204 is used to update the dynamic adjustment coefficients corresponding to each behavior type according to the changes in user level and the contribution of each behavior type in the user behavior feature vector through an iterative algorithm controlled by the learning rate, and apply the updated dynamic adjustment coefficients to the integral calculation of the next cycle to form an adaptive level management closed loop.

[0086] The third embodiment of the present invention provides a device for managing the dynamic growth level of users based on a pan-entertainment multi-ecosystem, characterized in that it includes a memory and a processor, wherein the memory stores a computer program, and the computer program can be executed by the processor to implement the method for managing the dynamic growth level of users based on a pan-entertainment multi-ecosystem as described above.

[0087] The fourth embodiment of the present invention provides a computer-readable storage medium, characterized in that it stores a computer program, which can be executed by the processor of the device where the computer-readable storage medium is located, to implement the dynamic growth level management method for users based on a pan-entertainment multi-ecosystem as described in any of the above.

[0088] Based on the present invention, a method, device, equipment, and storage medium for dynamic growth level management of users across multiple entertainment ecosystems are provided. This invention asynchronously collects user behavior data via message queues, constructs user behavior feature vectors containing feature values ​​of various behavior types, and calculates user scores based on these feature vectors through weighted calculations using configured weights and dynamic adjustment coefficients. Historical scores and monthly scores are combined using weighted coefficients to obtain a comprehensive score, achieving a balanced assessment of long-term user value and short-term activity. User levels are determined using a level mapping function based on the comprehensive score, and the level status is dynamically adjusted based on upgrade and level retention conditions. The dynamic adjustment coefficients are updated in real-time based on changes in user level and the contribution of each behavior type, forming an adaptive level management closed loop. This allows the system to automatically learn and optimize the importance of different behavior types, thereby achieving a truly fair and intelligent user value assessment system.

[0089] Exemplary examples show that the computer program described in the third and fourth embodiments of the present invention can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in implementing a device for dynamic growth level management of users based on a pan-entertainment multi-ecosystem model. For example, the apparatus described in the second embodiment of the present invention.

[0090] The processor referred to can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the aforementioned dynamic growth level management method for users across a multi-ecosystem of pan-entertainment, utilizing various interfaces and lines to connect all parts of the method and implement its control.

[0091] The memory can be used to store the computer programs and / or modules. The processor, by running or executing the computer programs and / or modules stored in the memory, and by calling the data stored in the memory, implements various functions based on a pan-entertainment multi-ecosystem user dynamic growth level management method. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, text conversion function, etc.), etc.; the data storage area may store data created based on the use of the mobile phone (such as audio data, text message data, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0092] If the implemented module is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0093] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0094] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for managing the dynamic growth levels of users across multiple entertainment ecosystems, characterized in that: include: By asynchronously collecting user behavior data across multiple scenarios and services through message queues, a user behavior feature vector containing feature values ​​of various behavior types is constructed. Based on the user behavior feature vector, the user score is calculated by weighting the configured weights and dynamic adjustment coefficients. The user score is then accumulated into the historical score and the monthly score, and the historical score and the monthly score are combined by the historical score weight coefficient to obtain the comprehensive score. The user level is determined by a level mapping function based on the comprehensive score. Based on the current level and the monthly points, it is determined whether the user meets the conditions for upgrading or maintaining the level, and the user level status and level validity period are updated accordingly. Based on the changes in user levels and the contribution of each behavior type in the user behavior feature vector, the dynamic adjustment coefficients corresponding to each behavior type are updated using an iterative algorithm controlled by the learning rate. Specifically: the level difference between the current period and the previous period is calculated as the level change; the proportion of each behavior type's feature value to the sum of all feature values ​​is calculated as the contribution of that behavior type; the level change, contribution, and learning rate are multiplied together and added to the original dynamic adjustment coefficient to obtain a new dynamic adjustment coefficient, the expression of which is: ;in, The level of the current cycle. The level of the previous cycle, As for contribution level, This is the current dynamic adjustment coefficient. This is the updated dynamic adjustment coefficient. The learning rate for the current period uses an adaptive adjustment strategy: the learning rate increases when two consecutive level changes are in the same direction and decreases when they are in opposite directions. Specifically: in, To adjust the factor, For symbolic functions, This represents the current cycle's trend of level changes. To reflect the level change trend of the previous period, the updated dynamic adjustment coefficient is applied to the score calculation of the next period to form an adaptive level management closed loop. Specifically, the CRC32 hash algorithm is used to calculate the user's unique identifier, and the result is modulo 32 to determine the shard number to which the user data belongs. Historical score tables and monthly score tables are created for each shard, with table names generated using a formatting function. A Redis distributed lock mechanism is used to ensure the atomicity of user level updates; the lock key value is composed of the user's unique identifier, level value, and update timestamp.

2. The method for managing the dynamic growth level of users based on a multi-ecosystem pan-entertainment model according to claim 1, characterized in that, The calculation process for the user score involves multiplying each feature value in the user behavior feature vector by its corresponding configuration weight and dynamic adjustment coefficient, and then summing all the weighted feature values ​​to obtain the user score. The specific expression is as follows: in, The configuration weight for the i-th behavior, Let m be the feature value of the i-th behavior, and m be the total number of behavior types. This is a dynamic adjustment coefficient; The comprehensive score is calculated by weighting historical scores and monthly scores, and its expression is as follows: in, For comprehensive scoring, Historical integral weighting coefficients, For historical integration, Monthly points.

3. The method for managing the dynamic growth level of users based on a multi-ecosystem pan-entertainment model according to claim 1, characterized in that, Also includes: Calculate the user's points acquisition rate within a fixed time window. When the points acquisition rate exceeds the average rate plus n times the standard deviation, mark the user as an abnormal user. Multiply all dynamic adjustment coefficients of the abnormal user by a preset penalty factor. Also, periodically calculate the user retention rate and optimize the configuration weights of each behavior type based on the trend of retention rate changes.

4. A device for managing the dynamic growth level of users across multiple entertainment ecosystems, characterized in that: include: The behavior data acquisition unit is used to asynchronously collect user behavior data in multiple scenarios and services through a message queue, and construct a user behavior feature vector containing feature values ​​of multiple behavior types. The comprehensive score calculation unit is used to calculate user points based on the user behavior feature vector by weighting operations with configured weights and dynamic adjustment coefficients, accumulate the user points into historical points and monthly points respectively, and combine historical points and monthly points with historical point weight coefficients to calculate a comprehensive score. The level management unit is used to determine the user level through the level mapping function based on the comprehensive score, determine whether the user meets the upgrade conditions or level retention conditions based on the current level and the monthly points, and update the user level status and level validity period accordingly. The parameter adjustment unit is used to update the dynamic adjustment coefficients corresponding to each behavior type based on changes in user level and the contribution of each behavior type in the user behavior feature vector using an iterative algorithm controlled by the learning rate. Specifically, it calculates the level difference between the current period and the previous period as the level change, calculates the proportion of each behavior type's feature value to the sum of all feature values ​​as the contribution of that behavior type, and multiplies the level change, contribution, and learning rate together, adding the result to the original dynamic adjustment coefficient to obtain a new dynamic adjustment coefficient, the expression of which is: ;in, The level of the current cycle. The level of the previous cycle, As for contribution level, This is the current dynamic adjustment coefficient. This is the updated dynamic adjustment coefficient. The learning rate for the current period uses an adaptive adjustment strategy: the learning rate increases when two consecutive level changes are in the same direction and decreases when they are in opposite directions. Specifically: in, To adjust the factor, For symbolic functions, This represents the current cycle's trend of level changes. To reflect the level change trend of the previous period, the updated dynamic adjustment coefficient is applied to the score calculation of the next period to form an adaptive level management closed loop. Specifically, the CRC32 hash algorithm is used to calculate the user's unique identifier, and the result is modulo 32 to determine the shard number to which the user data belongs. Historical score tables and monthly score tables are created for each shard, with table names generated using a formatting function. A Redis distributed lock mechanism is used to ensure the atomicity of user level updates; the lock key value is composed of the user's unique identifier, level value, and update timestamp.

5. A device for managing the dynamic growth level of users across multiple entertainment ecosystems, characterized in that: It includes a memory and a processor, wherein the memory stores a computer program that can be executed by the processor to implement a method for managing the dynamic growth level of users based on a pan-entertainment multi-ecosystem as described in any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, The device contains a computer program that can be executed by a processor of the device where the computer-readable storage medium is located, to implement the method for managing the dynamic growth level of users based on a pan-entertainment multi-ecosystem as described in any one of claims 1 to 3.

Images