Game user data management system based on big data

Abstract

The invention belongs to the technical field of game user management, and particularly discloses a game user data management system based on big data, which comprises a user performance data extraction module, a game expressive force scoring module, a social expressive force scoring module, a user label classification module, a game suitable-to-play tendency label analysis module and a game recommendation module. Game expressive force and social expressive force analysis is performed by calling game expression data and social expression data of users in a game platform, so that the users are classified and labeled, and meanwhile, a list of games played by the users with the same classification label is summarized; according to the game recommendation method and device, the classification labels of the users are selected from the game list, and then playing tendency label analysis is conducted on the games based on the classification labels of the users, the similar games highly matched with the classification labels of the users can be screened out from the game list for recommendation through the multi-dimensional data driving method, the game recommendation adaptation degree and accuracy are remarkably improved, and therefore the user experience is optimized, and the user experience is improved. And the user churn rate is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of game user management, and in particular relates to a game user classification recommendation technology, and specifically discloses a game user data management system based on big data. Background Art

[0002] The fast pace and multiple pressures of modern society have made games a choice for many people to relieve stress and seek pleasure. With the expansion of the user base, the game platform has developed rapidly. In order to improve user retention, the platform adopts a personalized recommendation system, which analyzes user preferences based on user data on the game platform to achieve targeted game recommendations and enhance user participation and satisfaction.

[0003] There are also technical solutions for recommending games through user classification in the prior art. For example, the Chinese invention patent with publication number CN118172131A discloses a method, device, equipment and storage medium for recommending information based on user behavior. By recording the detected operation behavior data and the associated scene information during the game operation, the corresponding user preference index under different scene information is determined; thereby determining the game recommendation information based on the user preference index, and recommending information. In this solution, the analysis of user preferences mainly depends on the user's operation behavior data during the game operation, that is, the user's performance in the game. Although this analysis method can capture the user's gameplay and interests, it ignores the user's social performance data, resulting in a certain one-sidedness in the preference analysis. Specifically, when the content of a game meets the user's preferences, but the game has a strong social demand, if the user's social performance is weak, recommending the game may result in a low degree of adaptation, thereby affecting the user experience and easily increasing the risk of user loss.

[0004] Another example is a Chinese invention patent with publication number CN109146627A, which discloses a game recommendation method. By obtaining user information (such as installed game lists, game usage time, address book friend lists, etc.), users are clustered into different clusters, and a corresponding game list is generated to recommend unregistered games to users. This solution can combine the user's game usage data and social data for cluster analysis, thereby improving the pertinence of recommendations. However, this method has a limitation when recommending unregistered games: it fails to fully consider the consistency between the game in the game list and the user's cluster. Specifically, when users register games, they cannot know the clusters of other users in the game, which will cause the game list to contain some games that are inconsistent with the characteristics of the cluster to which the user belongs. If these games are not further matched and screened, it will directly affect the adaptability and accuracy of the recommendation system. Summary of the invention

[0005] In view of the shortcomings of the prior art mentioned above, the present invention proposes a game user data management system based on big data, which comprehensively classifies the game performance data and social performance data of users in the game, and then matches and filters similar games from the game lists of classified users for recommendation, thereby effectively making up for the problems mentioned in the background technology.

[0006] The purpose of the present invention can be achieved through the following technical solutions: A game user data management system based on big data, including: a user performance data extraction module, which is used to count the total number of registered users on the game platform, and extract the user's game behavior records and social behavior records in each time period according to a set time gradient period, and extract game performance data and social performance data therefrom.

[0007] The game performance scoring module is used to form a game performance matrix with the game performance data of the user in each time period, thereby analyzing the user's game performance activity and game performance activity trend in each time period, and then scoring the user's game performance accordingly.

[0008] The social expression scoring module is used to form a social expression matrix with the social expression data of users in each time period, thereby analyzing the social expression activity and social expression activity trend of users in each time period, and then scoring the social expression of users accordingly.

[0009] The user tag classification module is used to classify users by tags based on their gaming performance and social performance.

[0010] The game playability tendency label parsing module is used to retrieve the game list of users with the same classification label and summarize it, and count the number of registered users of each game. At the same time, it parses the game playability tendency label based on the classification label of the registered users.

[0011] The game recommendation module is used to match the user's classification label with the playability label of each game in the summary game list, thereby making game recommendations.

[0012] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention retrieves the game performance data and social performance data of the user in the game platform to analyze the game performance and social performance, thereby classifying and labeling the users. At the same time, the present invention summarizes the game lists played by users with the same classification labels, and then analyzes the game suitability labels based on the user's classification labels. This multi-dimensional data-driven method can filter out similar games that are highly matched with the user's classification labels from the game list for recommendation, significantly improving the adaptability and accuracy of game recommendations, thereby optimizing the user experience and reducing user churn.

[0013] (2) Before selecting similar games from the list of games played by users with the same classification label for recommendation, the present invention adds the analysis of the distribution statistics of the new and old user groups of the selected games and the groups with a tendency to be lacking, and further selects games with a relatively lacking new user group for recommendation to users, which can effectively optimize the balance of the game ecology. By recommending games with few new users but potential, the platform can attract more new users to these games, promote the diversification and healthy development of the game community, and thus enhance the vitality of the entire game ecology. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0015] Figure 1 It is a schematic diagram of the connection of each module of the system of the present invention.

[0016] Figure 2 It is a schematic diagram of the game performance data composition in the present invention.

[0017] Figure 3 This is a schematic diagram of the social performance data composition in the present invention.

[0018] Figure 4 Several category diagrams are generated for cross-classifying the gaming performance level and the social performance level in the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Reference Figure 1 As shown, the present invention proposes a game user data management system based on big data, including a user performance data extraction module, a game performance scoring module, a social performance scoring module, a user label classification module, a game playability tendency label parsing module, and a game recommendation module, wherein the user performance data extraction module is connected to the game performance scoring module and the social performance scoring module respectively, the game performance scoring module and the social performance scoring module are both connected to the user label classification module, the user label classification module is connected to the game playability tendency label parsing module, and the user label classification module and the game playability tendency label parsing module are both connected to the game recommendation module.

[0021] The user performance data extraction module is used to count the total number of registered users on the game platform and obtain the user's registered account, and at the same time extract the user's game behavior records and social behavior records in each time period according to the set time gradient period, and extract game performance data from the game behavior records, and extract social performance data from the social behavior records.

[0022] In the example applied to the above scheme, the set time gradient period can be 30 days, 90 days, 6 months, 12 months, etc., and by setting the time gradient period, the dynamic changes of the user's performance behavior in the game platform can be captured. Specifically, by setting a shorter time gradient period (such as 30 days), the user's short-term performance behavior can be observed, by setting a medium-length time gradient period (such as 90 days, 6 months), the medium- and long-term performance behavior can be observed, and by setting a longer time gradient period (such as 12 months), the user's long-term performance behavior can be observed. In this way, the changing trend of the user's performance behavior can be captured through the time gradient performance behavior. Compared with simply analyzing the user's behavior performance based on a fixed time period, the setting of multiple time gradient periods can dynamically and reasonably analyze the user's game expressiveness and social expressiveness, which is conducive to improving the accuracy and real-time nature of the analysis, and also provides the platform with richer user behavior insights.

[0023] In the supplementary explanation of the above scheme, the game behavior record is the user's operation behavior in the game that the platform automatically generates and records every time the user logs in to the game platform. These behaviors include but are not limited to the user's game time, task progress, level completion, props use, etc., which constitute comprehensive game behavior data. The generated game behavior record will be stored by the platform and associated with the user's registered account.

[0024] Social behavior records are automatically generated and recorded by the platform every time a user logs into the game platform to conduct social activities. Specifically, the social behavior records can be extracted from the friend relationship list, instant messaging, team formation, participation in social events, and social evaluations. The generated social behavior records will be stored by the platform and associated with the user's registered account.

[0025] In the preferred implementation of the above scheme, see Figure 2 As shown, the game performance data includes game duration data, game progress data and props usage data. In a further example, the game duration data includes but is not limited to single game duration, game login frequency, longest continuous game duration, etc., the game progress data includes but is not limited to level completion rate, clearance time, clearance rate, etc., and the props usage data includes but is not limited to the number of props held, frequency of props use, amount of props consumed, etc.

[0026] In another preferred implementation of the above scheme, see Figure 3 As shown, social performance data includes social relationship data, social interaction data and social influence data, among which social relationship data reflects the social network structure of users in the game. Specifically, social relationship data includes but is not limited to the number of friend lists, frequency of adding friends, duration of friend relationships, etc. Social interaction data reflects the specific social interaction behavior of users in the game. Specifically, social interaction data includes but is not limited to the length of interactive messages, frequency of interactive messages, number of consecutive interactive days, etc. Social influence data reflects the influence of users in the game community. Specifically, social influence data includes but is not limited to the growth rate of fans, frequency of content release, success rate of game recommendations, etc.

[0027] The data extracted by the present invention when retrieving the game behavior records and social behavior records of users on the game platform to extract game performance data and social performance data include multiple aspects, which can characterize the game performance and social performance in multiple dimensions, and provide solid data support for subsequent game performance and social performance scoring.

[0028] It should be noted that due to the inconsistency of the duration scales of different time gradient cycles, the number of game behavior records and social behavior records retrieved in each time period may be different. Therefore, when extracting game performance data and social performance data, some data in different game behavior records and social behavior records in the corresponding period can be averaged to eliminate the impact of time scale differences and improve the accuracy of data analysis. For example, when extracting a single game duration, each game behavior record has a game duration, and the game duration in each game behavior record can be averaged to obtain a single game duration. However, for certain specific types of data, such as clearance rate, friend adding frequency, longest continuous game duration, etc., there is no need to perform mean calculation, because taking the average may obscure its intrinsic meaning or lose key information, so it should be kept as it is.

[0029] The game performance scoring module is used to form a game performance matrix with the game performance data of the user in each time period, thereby analyzing the game performance activity and game performance activity trend of the user in each time period, and then scoring the game performance of the user accordingly.

[0030] Preferably, the game performance data of the user in each time period is used to form a game performance matrix as follows: define the time period as the row of the matrix, define the game performance data as the column of the matrix, construct the game performance matrix, and use the user's registered account as the matrix identifier.

[0031] In the above preferred implementation example, the game performance matrix is ​​constructed as In this matrix, T1, T2, T3, and T4 represent the time periods within the time gradient cycle, respectively.GD 、data GP 、data PU They represent game duration data, game progress data, and item usage data respectively.

[0032] The game performance data of users in each time period is extracted, normalized, and then filled into the constructed game performance matrix.

[0033] It should be noted that the normalization process mentioned above is to ensure that the numerical ranges of performance data of different types of games are consistent and to eliminate the dimensional differences between different indicators. Exemplarily, the normalization process can adopt a maximum-minimum normalization method.

[0034] Further preferably, the analysis of the user's gaming activity in each time period refers to the following process: weighting the game duration, game progress and prop usage.

[0035] In the above further preferred example, the weights of game duration, game progress and item usage can be assigned to 30%, 50% and 20%. The reason for such assignment is that the game progress is the core indicator to measure the actual progress and participation depth of users in the game. It reflects the key behaviors of users such as completing tasks, clearing levels and unlocking achievements, and directly reflects the user's skill level and game involvement. Game duration is an important indicator to measure user activity, reflecting the time and frequency spent by users in the game. Longer game duration usually means that users have a higher interest in the game, but there may also be inefficient games. Therefore, although game duration is an important indicator of activity, it cannot fully represent the actual quality of user participation. Item usage reflects the consumption and utilization of virtual items by users in the game, and is an important indicator to measure users' economic behavior and strategy selection. The frequency and type of item usage can reflect users' consumption habits and game strategies, but for all users, the frequency and amount of item usage may vary greatly. Not all users will use items frequently, so item usage can only be used as a supplementary indicator of game activity, which is much less important than game duration and game progress.

[0036] The user's game performance matrix is ​​focused on the columns corresponding to the time period of each row, and then the data in the corresponding columns are extracted and combined with the weight values ​​of game time, game progress and props usage to perform weighted average calculation to obtain the user's game performance activity in each time period.

[0037] It should be added that when using game time data, game progress data and item usage data to calculate the weighted average of game performance activity, since these data also contain different types of data, it is recommended to first normalize the internal data to ensure that data of different categories are comparable. Subsequently, a weighted average calculation is performed based on the normalized data to obtain the score values ​​of game time, game progress and item usage. The score values ​​of game time, game progress and item usage are then combined with the weight values ​​assigned to game time, game progress and item usage to perform a weighted average calculation to finally obtain a comprehensive score of game performance activity.

[0038] More preferably, the game performance activity trend is analyzed in the following process: the time periods in the time gradient cycle are arranged and numbered in order from far to near in time.

[0039] In the above further preferred example, the 30 days, 90 days, 6 months, and 12 months in the time gradient cycle are arranged in order from far to near as 12 months, 6 months, 90 days, and 30 days.

[0040] According to the arrangement number of the time periods, the game performance activity of each time period is subtracted from the game performance activity of the previous time period and then divided by the interval between the two time periods to obtain the game performance activity change rate of each time period.

[0041] It should be noted that since the first time period has no previous time period, it starts from the second time period.

[0042] The game performance activity change rate of each time period is calculated using the smoothing formula S t =α*S 0t +(1-α)*S t-1 Get the smoothed game activity change rate S for each time period t , where t represents the time period number, t=1,2,......,m, S 0t represents the game performance activity change rate in time period t, S t-1 It represents the smoothed game activity change rate in the t-1th time period, α represents the smoothing factor, and 0<α<1, which determines the weight of recent data. Generally, a larger α value will make the smoothed result closer to the recent data, while a smaller α will make the smoothed result more stable. For example, α=0.6.

[0043] It should be understood that after obtaining the game performance activity change rate in each time period, the present invention does not directly use these raw data to perform game performance activity trend analysis, but first smoothes the activity change rate in each time period to eliminate the impact of short-term fluctuations, thereby obtaining a more stable activity change rate. This helps to more accurately reflect the long-term behavioral trends of users. Specifically, the present invention adopts an exponentially weighted moving average method, which achieves smoothing by assigning higher weights to recent time periods, so that recent data has a greater impact on the smoothing result, while the impact of data in earlier time periods gradually weakens, so that short-term random fluctuations can be smoothed while retaining recent trends.

[0044] The smoothed game activity change rate in each time period is cumulatively averaged to obtain the user's corresponding game performance activity trend.

[0045] It should be added that the game performance activity trend value can be positive or negative. When the game performance activity trend is positive, it means that the user's overall game activity is increasing. When the game performance activity trend is negative, it means that the user's overall game activity is decreasing.

[0046] Further preferably, the scoring of the user's gaming performance is implemented as follows: the user's gaming performance activity in each time period is combined with the gaming performance activity trend into a scoring formula Get the user's game performance GP, where represents the average of the user's gaming performance activity in each time period, k represents the user's gaming performance activity trend, Q′ represents the user's gaming performance activity in the most recent time period, e represents a natural constant, d represents a scoring system, exemplarily, d can be a one-point system, a ten-point system, or a percentage system, when d is a one-point system, d=1, when d is a percentage system, d=100, η ​​represents the fluctuation index of the user's gaming performance activity in each time period, where max{Q t}、min{Q t} respectively represent the maximum game performance activity and the minimum game performance activity of the user in each time period, η0 represents the limited fluctuation index preset by the system, illustratively, η0=0.5.

[0047] It should be understood that the above-mentioned game performance score is scored in two aspects, static dimension and dynamic dimension, by combining the game performance activity and game performance activity trend of the user in each time period, wherein the effective representative selected from the game performance activity of each time period reflects the user's static game performance, while the game performance activity trend reflects the user's dynamic game performance. Through this dynamic and static combination, the obtained game performance score is more comprehensive and accurate. In addition, when screening static effective representatives, the average game performance activity of the user is not directly selected, but is determined according to the fluctuation of the game performance activity of the user in each time period: when the game performance activity of the user in each time period fluctuates less, indicating that its activity is relatively stable, the average game performance activity of the user is selected as the static effective representative. This can reflect the overall activity level of the user in the time period and avoid deviations caused by individual outliers. When the game performance activity of the user in each time period fluctuates greatly, it indicates that there is a great instability in its activity. At this time, the game performance activity of the most recent time period is selected as the static effective representative. This can better capture the user's current real active state and avoid the impact of abnormal fluctuations in historical data on the evaluation results.

[0048] The social expressiveness scoring module is used to form a social expressiveness matrix from the social expressiveness data of the user in each time period, thereby analyzing the social expressiveness and social expressiveness trend of the user in each time period, and then scoring the social expressiveness of the user accordingly.

[0049] The above-mentioned process of constructing a social performance matrix, analyzing the social performance activity and social performance activity trend of users in each time period, and scoring the social expression of users can be referred to the analysis process of game performance.

[0050] The user label classification module is used to label and classify users according to their gaming performance and social performance. The specific classification is as follows: different gaming performance levels and social performance levels are defined according to the value ranges of gaming performance and social performance, and the gaming performance division intervals corresponding to each gaming performance level and the social performance division intervals corresponding to each social performance level are recorded.

[0051] In the above exemplary implementation, in the expressiveness scoring, when the scoring system is a one-point system, the value range of the expressiveness is 0 to 1, when the scoring system is a ten-point system, the value range of the expressiveness is 0 to 10, and when the scoring system is a hundred-point system, the value range of the expressiveness is 0 to 100.

[0052] Further, in the above exemplary implementation, different game performance levels and social performance levels can be divided into two levels, three levels, four levels, five levels, etc. The game performance level and its corresponding game performance division interval obtained under the two-level level of expressiveness ranging from 0 to 100 are high game performance: 60-100, low game performance: 0-59. The social performance level and its corresponding social expressiveness division interval are the same as above.

[0053] The game performance and social performance of each user on the game platform are compared with the game performance division intervals corresponding to each game performance level and the social performance division intervals corresponding to each social performance level to obtain the game performance level and social performance level corresponding to each user.

[0054] The degree of gaming performance and the degree of social performance were cross-classified to generate several categories.

[0055] In a further exemplary implementation, the cross-classification of the game performance level and the social performance level into two levels is divided into four categories, see Figure 4 As shown, one category can represent high gaming performance and high social performance, the second category can represent high gaming performance and low social performance, the third category can represent low gaming performance and high social performance, and the fourth category can represent low gaming performance and low social performance.

[0056] Users are matched to corresponding categories according to their gaming performance and social performance, thereby assigning a classification label to each user.

[0057] The game playability tendency tag parsing module is used to retrieve and summarize the game lists of users with the same classification tag, and count the number of registered users of each game, and at the same time parse the game playability tendency tags based on the classification tags of the registered users;

[0058] It should be pointed out that deduplication is required when aggregating game lists of users with the same category label.

[0059] Preferably, parsing the game's playability tendency tags is implemented as follows: classifying the classification tags of all registered users in each game according to the same classification tag, thereby summarizing the occurrence ratio of each classification tag among the game's registered users.

[0060] The cumulative game time of users registered under each category tag is counted, and the contribution factor of each category tag is calculated based on the cumulative game time.

[0061] It should be emphasized that the cumulative game time mentioned above does not only refer to the time users spend on the game, but includes the time invested in the game platform with social elements. It reflects the user's level of investment in the game platform and can reflect the contribution of different classification tags. Specifically, the contribution factor of each classification tag can be calculated by dividing the cumulative game time of registered users under each classification tag by the total cumulative game time of registered users under all classification tags.

[0062] The value of each category label is obtained by multiplying the occurrence ratio of each category label corresponding to the game by the contribution factor, and then the category label corresponding to the maximum value is taken as the game's playability tendency label.

[0063] It should be explained that the playability label reflects the classification label that best represents the characteristics of the game's user group.

[0064] The game recommendation module is used to match the user's classification label with the playability tendency label of each game in the summary game list, thereby making game recommendations. The specific implementation is as follows: the user's classification label is matched with the playability tendency label of each game in the summary game list, and the successfully matched games constitute the matching game set corresponding to the user.

[0065] Obtain the game names registered by each registered user in the game platform, remove the registered game names from the matching game set corresponding to the corresponding user, and record the matching game set after removal as the valid game set.

[0066] The distribution statistics of the games in the effective game set between new and old user groups are collected to identify the groups with lack of tendency, and then the games with lack of tendency as new users are selected to recommend to users.

[0067] Preferably, in the above scheme, the identification of the group lacking tendency refers to the following process: the online time of each game in the valid game set is obtained respectively, and the statistical time range is formed in combination with the current time.

[0068] Determine the demarcation time within the statistical time range, and then compare the registration time of each registered user of each game with the demarcation time, thereby determining the new and old user classification corresponding to each registered user, where users before the demarcation time are old users, and users after the demarcation time are new users.

[0069] Exemplarily, the demarcation time may be one year after the game is launched.

[0070] The ratio of new to old users corresponding to each game is counted, and the ratios of new and old users are compared to obtain the game's lacking group. Specifically, when the old user ratio is higher than the new user ratio, the game's lacking group is new users; and when the new user ratio is higher than the old user ratio, the game's lacking group is old users.

[0071] It should be added that the group with a lack of tendency obtained based on the ratio of new and old users corresponding to the game reflects the user development direction of the game. By recommending games with user development direction of new users to users, the platform can help these games attract more new users, improve user composition, and achieve ecological balance. In addition, the continuous influx of new users helps to extend the life cycle of the game and avoid game decline due to user loss. By recommending games with a relatively lack of new user groups, the platform can inject new vitality into these games and maintain their long-term appeal and development potential.

[0072] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they shall all fall within the protection scope of the present invention.

Claims

1. A game user data management system based on big data, characterized in that ,include: The user performance data extraction module is used to count the total number of registered users on the game platform, and extract the game behavior records and social behavior records of users in each time period according to the set time gradient period, and extract the game performance data and social performance data from them; The game performance scoring module is used to form a game performance matrix with the game performance data of the user in each time period, thereby analyzing the game performance activity and game performance activity trend of the user in each time period, and then scoring the game performance of the user accordingly; A social expression scoring module is used to form a social expression matrix with the social expression data of users in each time period, thereby analyzing the social expression activity and social expression activity trend of users in each time period, and then scoring the social expression of users accordingly; User label classification module, used to classify users by label according to their gaming performance and social performance; The game playability tendency tag parsing module is used to retrieve the game list of users with the same classification tag and summarize it, and count the number of registered users of each game. At the same time, it parses the game playability tendency tag based on the classification tag of the registered users; The game recommendation module is used to match the user's classification label with the playability label of each game in the summary game list, thereby making game recommendations.

2. A game user data management system based on big data as claimed in claim 1, characterized in that: The game performance data includes game duration data, game process data and prop usage data, and the social performance data includes social relationship data, social interaction data and social influence data.

3. A game user data management system based on big data as claimed in claim 2, characterized in that: The process of forming a game performance matrix from the game performance data of the user in each time period is as follows: Define time periods as matrix rows, define game performance data as matrix columns, construct a game performance matrix, and use the user's registered account as the matrix identifier; The game performance data of users in each time period is extracted, normalized, and then filled into the constructed game performance matrix.

4. A game user data management system based on big data as claimed in claim 3, characterized in that: The analysis of the user's gaming activity in each time period refers to the following process: Assign weights to game duration, game progress, and item usage; The user's game performance matrix is ​​focused on the columns corresponding to the time period of each row, and then the data in the corresponding columns are extracted and combined with the weight values ​​of game time, game progress and props usage to perform weighted average calculation to obtain the user's game performance activity in each time period.

5. A game user data management system based on big data as claimed in claim 1, characterized in that: The game performance trend is analyzed as follows: Arrange and number each time period in the time gradient cycle in order from far to near; According to the arrangement number of the time periods, the game performance activity of each time period is subtracted from the game performance activity of the previous time period and then divided by the interval between the two time periods to obtain the game performance activity change rate of each time period; The game performance activity change rate of each time period is calculated using the smoothing formula S t =α*S 0t +(1-α)*S t-1 Get the smoothed game activity change rate S for each time period t , where t represents the time period number, t=1,2,......,m, S 0t represents the game performance activity change rate in time period t, S t-1 represents the smoothed game activity change rate in the t-1th time period, α represents the smoothing factor, and 0<α<1; The smoothed game activity change rate in each time period is cumulatively averaged to obtain the user's corresponding game performance activity trend.

6. A game user data management system based on big data as claimed in claim 1, characterized in that: The scoring of the user's gaming performance is implemented as follows: The user's gaming activity in each time period is combined with the gaming activity trend to enter the scoring formula Get the user's game performance GP, where represents the average of the user's gaming performance activity in each time period, k represents the user's gaming performance activity trend, Q′ represents the user's gaming performance activity in the most recent time period, e represents a natural constant, d represents a scoring system, and η represents the fluctuation index of the user's gaming performance activity in each time period, where max{Q t }、min{Q t } respectively represent the maximum game performance activity and the minimum game performance activity of the user in each time period, and η0 represents the limited fluctuation index preset by the system.

7. A game user data management system based on big data as claimed in claim 6, characterized in that: The process of classifying users by labeling them based on their gaming and social performance is as follows: Define different levels of game performance and social performance according to the scoring range of game performance and social performance, and record the game performance division interval corresponding to each game performance level and the social performance division interval corresponding to each social performance level; Compare the game performance and social performance of each user on the game platform with the game performance division intervals corresponding to each game performance degree and the social performance division intervals corresponding to each social performance degree to obtain the game performance degree and social performance degree corresponding to each user; The degree of gaming performance and the degree of social performance were cross-classified to generate several categories; Users are matched to corresponding categories according to their gaming performance and social performance, thereby assigning a classification label to each user.

8. A game user data management system based on big data as claimed in claim 1, characterized in that: The process of parsing the game playability labels based on the classification labels of registered users is as follows: Classify the category tags of all registered users in each game according to the same category tag, and summarize the occurrence ratio of each category tag among the registered users of the game; Count the cumulative game time of users registered under each category tag, and calculate the contribution factor of each category tag based on the cumulative game time; The value of each category label is obtained by multiplying the occurrence ratio of each category label corresponding to the game by the contribution factor, and then the category label corresponding to the maximum value is taken as the game's playability tendency label.

9. A game user data management system based on big data as claimed in claim 1, characterized in that: The game recommends the following process: Match the user's classification label with the playability label of each game in the summary game list, and obtain the successfully matched games to form the matching game set corresponding to the user; Obtain the game names registered by each registered user in the game platform, remove the registered game names from the matching game set corresponding to the corresponding user, and record the matching game set after removal as the valid game set; The distribution statistics of the games in the effective game set between new and old user groups are collected to identify the groups with lack of tendency, and then the games with lack of tendency as new users are selected to recommend to users.

10. A game user data management system based on big data as claimed in claim 9, characterized in that: The process of collecting statistics on the distribution of new and old user groups of games in the effective game collection and identifying the groups lacking in tendency is as follows: Get the launch time of each game in the valid game set respectively, and combine it with the current time to form the statistical time range; Determine the demarcation time from the statistical time range, and then compare the registration time of each registered user of each game with the time, thereby determining the new and old user classification corresponding to each registered user; Count the ratio of new to old users for each game, and compare the two to get the group that tends to lack the game.

Images