Data processing method and device, equipment, medium and program product

Abstract

The invention discloses a data processing method and device, equipment, a medium and a program product, and belongs to the field of artificial intelligence. The method comprises the following steps: acquiring object data of a game object in a Tth time period; t is greater than or equal to 1; obtaining hyper-parameters corresponding to the strength calculation formula; the hyper-parameter is obtained by optimizing and solving an error function, and the error function is used for balancing an error between the estimated strength and the actual strength of the zone server where each game object is located; substituting the object data of the game object and the hyper-parameter into a strength calculation formula to obtain a strength value corresponding to the game object; accumulating the strength values corresponding to the game objects belonging to the same district clothes to obtain district clothes strength values corresponding to the district clothes respectively; and combining the zone clothes based on the zone clothes strength values to obtain a zone clothes combination result of the (T + 1) th time period. According to the scheme, the hyper-parameters in the strength calculation formula can reflect the correlation between the strength value of the game object and the strength value of the zone clothes, and the accuracy of the zone clothes combination result can be improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, and particularly to a data processing method, apparatus, device, medium and program product. Background Art

[0002] After a player logs in to a game, they usually need to select a server area. Different server areas usually have different game environments, gameplays, game contents, languages, versions, etc. At the end of a game season, multiple server areas are usually merged based on the relevant data of the players to ensure that the server areas in the next game season can provide a relatively fair game environment for different players.

[0003] In the related art, the strength of a player is evaluated based on a strength calculation formula, and the server area strength of the server area is determined according to the strengths of all players in the server area; the merger is performed based on the server area strength of each server area. Among them, the above strength calculation formula includes hyperparameters, and the hyperparameters are set by the artificial experience of game planners. This formula cannot reflect the correlation between the server area strength and the player strength. Summary of the Invention

[0004] This application provides a data processing method, apparatus, device, medium and program product. The technical solution is as follows:

[0005] According to one aspect of this application, a data processing method is provided. The method includes:

[0006] Obtain the object data of the game object in the T-th time period; T is greater than or equal to 1;

[0007] Obtain the hyperparameters corresponding to the strength calculation formula; the strength calculation formula is used to calculate the strength value corresponding to the game object, and the hyperparameters are obtained by optimizing and solving an error function, and the error function is determined based on the object data of the game object in the (T - 1)-th time period and the T-th time period, and is used to balance the error between the estimated strength and the actual strength of each server area where the game object is located;

[0008] Substitute the object data of the game object and the hyperparameters into the strength calculation formula to obtain the strength value corresponding to the game object;

[0009] Accumulate the strength values corresponding to the game objects belonging to the same server area to obtain the server area strength values corresponding to each server area;

[0010] Merge each server area based on the server area strength values to obtain the server area merger result in the (T + 1)-th time period.

[0011] According to another aspect of this application, a data processing apparatus is provided. The apparatus includes:

[0012] An acquisition module, configured to acquire object data of a game object in the T-th time period; where T is greater than or equal to 1;

[0013] The acquisition module is further configured to acquire hyperparameters corresponding to a strength calculation formula; the strength calculation formula is used to calculate a strength value corresponding to the game object, the hyperparameters are obtained by optimizing and solving an error function, and the error function is determined based on the object data of the game object in the (T - 1)-th time period and the T-th time period, and is used to balance the error between the estimated strength and the actual strength of each game object in the server area;

[0014] A processing module, configured to substitute the object data of the game object and the hyperparameters into the strength calculation formula to obtain a strength value corresponding to the game object;

[0015] The processing module is further configured to accumulate the strength values corresponding to the game objects belonging to the same server area to obtain a server area strength value corresponding to each server area;

[0016] A merging module, configured to merge each server area based on the server area strength value to obtain a server area merging result in the (T + 1)-th time period.

[0017] According to another aspect of the present application, there is provided a computer device, including: a processor and a memory, the memory stores a computer program, and the computer program is loaded and executed by the processor to implement the data processing method as described above.

[0018] According to another aspect of the present application, there is provided a computer-readable storage medium, which stores a computer program, and the computer program is loaded and executed by a processor to implement the data processing method as described above.

[0019] According to another aspect of the present application, there is provided a computer program product, which includes computer instructions, the computer instructions are stored in a computer-readable storage medium, and the processor obtains the computer instructions from the computer-readable storage medium, so that the processor loads and executes to implement the data processing method as described above.

[0020] The beneficial effects brought by the technical solution provided by the embodiments of the present application at least include:

[0021] The computer device obtains the object data of game objects in the T-th time period; T is greater than or equal to 1; obtains the hyperparameters corresponding to the strength calculation formula; substitutes the object data of the game objects and the hyperparameters into the strength calculation formula to obtain the strength value corresponding to the game objects; accumulates the strength values corresponding to the game objects belonging to the same server area to obtain the server area strength values corresponding to each server area; merges each server area based on the server area strength values to obtain the server area merging result in the (T + 1)-th time period. Since the hyperparameters are obtained by optimizing and solving the error function and are not manually set, and the error function is determined based on the object data of game objects in the (T - 1)-th time period and the T-th time period and can be used to balance the error between the estimated strength and the actual strength of each server area where the game objects are located, that is, the error function can reflect the strength of game objects in different time periods and the strength of the server areas where the game objects are located in different time periods, so the hyperparameters can reflect the correlation between the server area strength and the strength of game objects and have strong interpretability. Thus, it can improve the rationality and accuracy of the strength values of game objects and the server area strength values determined based on the hyperparameters. Since the accuracy of the server area strength values is improved, the accuracy of the server area merging result is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0023] Figure 1 Shows the structural block diagram of a computer system provided by an exemplary embodiment of the present application;

[0024] Figure 2 Shows the schematic diagram of a data processing method provided by an exemplary embodiment of the present application;

[0025] Figure 3 Shows the flowchart of a data processing method provided by an exemplary embodiment of the present application;

[0026] Figure 4 Shows the flowchart of a data processing method provided by an exemplary embodiment of the present application;

[0027] Figure 5 Shows the flowchart of a data processing method provided by an exemplary embodiment of the present application;

[0028] Figure 6 Shows the schematic diagram of the season-end title settlement interface provided by an exemplary embodiment of the present application;

[0029] Figure 7 Shows a schematic diagram of the ranking of the district service strength value provided by an exemplary embodiment of the present application;

[0030] Figure 8 Shows the overall architecture diagram of the data processing method provided by an exemplary embodiment of the present application;

[0031] Figure 9 Shows a schematic diagram of the optimized error function provided by an exemplary embodiment of the present application;

[0032] Figure 10 Shows a schematic diagram of the comparison of beneficial effects provided by an exemplary embodiment of the present application;

[0033] Figure 11 Shows a schematic diagram of the comparison of core indicators provided by an exemplary embodiment of the present application;

[0034] Figure 12 Shows the block diagram of the data processing device provided by an exemplary embodiment of the present application;

[0035] Figure 13 Shows the block diagram of the structure of the computer device provided by an exemplary embodiment of the present application. Detailed implementation manners

[0036] To make the objectives, technical solutions and advantages of the present application clearer, the embodiments of the present application will be further described in detail below with reference to the accompanying drawings.

[0037] Here, the exemplary embodiments will be described in detail, and the examples are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of the devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0038] The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. The singular forms of "a", "the" and "said" used in the present application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0039] It should be understood that although the terms first, second, etc. may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this application, the first parameter may also be referred to as the second parameter, and similarly, the second parameter may also be referred to as the first parameter. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0040] It should be noted that before collecting relevant data of the user (player / game object) (such as: object data of the game object in the (T-1)th time period and the Tth time period, data of each server area, data of each camp, etc.) and during the process of collecting relevant data of the user, a prompt interface, a pop-up window or voice prompt information can be displayed. The prompt interface, pop-up window or voice prompt information is used to prompt the user that their relevant data is being collected currently, so that this application only starts to execute the relevant steps of obtaining the relevant data of the user after obtaining the confirmation operation issued by the user for the prompt interface or the pop-up window. Otherwise (that is, when the confirmation operation issued by the user for the prompt interface or the pop-up window is not obtained), the relevant steps of obtaining the relevant data of the user are ended, that is, the relevant data of the user is not obtained. In other words, all user data collected by this application is collected with the consent and authorization of the user, and the collection, use and processing of the relevant user data need to comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0041] First, a brief introduction to the nouns involved in the embodiments of this application:

[0042] Server area: Refers to the server in the game. When a player logs in to the game, they need to select a server area. Different server areas are usually used to provide different game environments, gameplays, game contents, languages, versions, etc. At the end of a game season, usually multiple server areas are merged according to the relevant data of the players to ensure that the server areas in the next game season can provide a relatively fair game environment for different players. Among them, the game season (or simply called season) refers to the periodically occurring time period in the game. For example, a game season can be set to three months. After the end of this game season, rankings will be made according to the achievements, scores, etc. of the game object in this game season, and corresponding game rewards will be given to the game object. In this embodiment, a time period can correspond to one season.

[0043] Camp: Composed of several server areas. The game season server usually includes at least two camps, and there is a hostile relationship between at least two camps. Taking the game season server including two camps as an example, these two camps can be respectively called the first camp and the second camp, and the first camp and the second camp are in a hostile relationship with each other.

[0044] Server area strength value: A value used to represent the comprehensive strength of a server area. Since a server area includes multiple players, the server area strength value is also a value used to represent the comprehensive strength of all players or some players (e.g., active players with activity higher than a set value) in that server area. The larger the server area strength value corresponding to a server area, the stronger the comprehensive strength of the players in that server area. Generally, the server area strength value corresponding to a server area is the sum of the strength values corresponding to all players or some players belonging to that server area.

[0045] Faction strength value: A value used to represent the comprehensive strength of a faction. Since a faction includes multiple server areas, the faction strength value is also a value used to represent the comprehensive strength of all server areas or some server areas (e.g., the top N server areas with server area strength values higher than a set value) in that faction. The larger the faction strength value corresponding to a faction, the stronger the comprehensive strength of the server areas in that faction. Generally, the faction strength value corresponding to a faction is the sum of the server area strength values corresponding to all server areas or some server areas belonging to that faction.

[0046] Optimization algorithm: An algorithm used to solve hyperparameters. Optionally, the optimization algorithm can be implemented based on at least one of the Bayesian optimization algorithm, genetic algorithm, and simulated annealing algorithm. Exemplarily, the Bayesian optimization algorithm: is an optimization algorithm based on Bayes' theorem, used for optimization under high-dimensional, non-convex, noisy, and black-box objective functions. By continuously exploring and exploiting the information of the objective function, it seeks an approximate global optimal solution.

[0047] Gaussian Process (GP): A Bayesian model based on regression problems, also referred to as a Gaussian process model in this embodiment. The application fields of Gaussian processes include at least one of the following: regression analysis, classification problems, time series analysis, Bayesian optimization, and reinforcement learning.

[0048] Game object: Refers to a user account (player) participating in the game.

[0049] Figure 1 FIG. shows a structural block diagram of a computer system 100 provided by an exemplary embodiment of the present application. The computer system 100 can be implemented as the system architecture of a data processing method. The computer system 100 includes: a terminal 120 and a server 140.

[0050] The terminal 120 may be an electronic device such as a mobile phone, a tablet computer, a vehicle-mounted terminal (in-vehicle computer), a wearable device, a PC (Personal Computer), an unattended reservation terminal, etc. A client of a target application program may be installed and run in the terminal 120, and the target application program may be an application program that supports a virtual world. Exemplarily, the application program may be any one of a battle royale shooting game, a virtual reality (VR) client, an augmented reality (AR) program, a 3D map program, a virtual reality game, an augmented reality game, a first-person shooting game (FPS), a third-person shooting game (TPS), a multiplayer online battle arena game (MOBA), and a simulation game (SLG). The form of the target application program in this application is not limited, including but not limited to an App (Application) installed in the terminal 120, a mini-program, and may also be in the form of a web page.

[0051] The terminal 120 is connected to the server 140 through a wireless network or a wired network.

[0052] The server 140 may be an independent physical server, or a server cluster or a distributed system composed of multiple physical servers, or may also be a cloud server that provides cloud computing services, a cloud database, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, a content delivery network (CDN), and a cloud server that provides basic cloud computing services such as a big data and artificial intelligence platform. The server 140 includes at least one of a server, multiple servers, a cloud computing platform, and a virtualization center.

[0053] Among them, Cloud Technology refers to a hosting technology that unifies a series of resources such as hardware, software, and network within a wide area network or local area network to achieve data computing, storage, processing, and sharing. Cloud technology is the general term for network technology, information technology, integration technology, management platform technology, application technology, etc. based on the cloud computing business model. It can form a resource pool, be used as needed, and is flexible and convenient. Cloud computing technology will become an important support. The back-end services of the technical network system require a large amount of computing and storage resources, such as video websites, picture websites, and more portal websites. With the highly developed application of the Internet industry, in the future, each item may have its own identification mark and needs to be transmitted to the back-end system for logical processing. Data at different levels will be processed separately, and various industry data requires the support of a powerful system. This can only be achieved through cloud computing.

[0054] In some embodiments, the server 140 can also be implemented as a node in a blockchain system. Blockchain is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, and encryption algorithms. Blockchain, in essence, is a decentralized database, a string of data blocks generated by using cryptographic methods. Each data block contains information about a batch of network transactions, which is used to verify the validity of the information (anti-counterfeiting) and generate the next block. Blockchain can include the blockchain underlying platform, the platform product service layer, and the application service layer.

[0055] Exemplarily, the server 140 includes a processor 144 and a memory 142. The memory 142 further includes a receiving module 1421, a control module 1422, and a sending module 1423. The receiving module 1421 is used to receive the area server merging request sent by the terminal 120; the control module 1422 is used to generate the area server merging result; the sending module 1423 is used to send the area server merging result to the terminal 120. The server 140 is used to provide back-end services for the terminal 120.

[0056] Optionally, the server 140 undertakes the main computing work, and the terminal 120 undertakes the secondary computing work; or, the server 140 undertakes the secondary computing work, and the terminal 120 undertakes the main computing work; or, a distributed computing architecture is adopted between the server 140 and the terminal 120 for collaborative computing.

[0057] In the data processing method provided by the embodiments of the present application, the execution subject of each step can be a computer device. A computer device refers to an electronic device with the ability to calculate, process, and store data. Figure 1Taking the solution implementation environment shown as an example, the data processing method can be executed by the terminal 120 (for example, the target application installed and running in the terminal 120 executes the data processing method), or can be executed by the server 140, or can be executed by the interaction and cooperation of the terminal 120 and the server 140. This application does not make any limitations in this regard.

[0058] Those skilled in the art can know that the number of the above-mentioned terminals 120 can be more or less. For example, the above-mentioned terminal 120 can be only one, or the above-mentioned terminal 120 can be dozens or hundreds, or a larger number. The embodiments of this application do not limit the number and device type of the terminal 120.

[0059] In the related art, the strength of game objects is evaluated based on the strength calculation formula, and the server strength of the server area is determined according to the strength of all game objects in the server area; the server strength of each server area is merged. Among them, the above-mentioned strength calculation formula includes hyperparameters, and the hyperparameters are set by the manual experience of game planners. This method relies heavily on the historical experience of numerical planners and lacks a quantitative index to measure the rationality of server area strength evaluation. There are at least the following disadvantages:

[0060] 1. The hyperparameters in the strength calculation formula are all set by manual experience. Setting a large number of hyperparameters requires a lot of manpower, the setting efficiency is low, and the set hyperparameters remain unchanged and cannot dynamically reflect the correlation between the server area strength and the strength of game objects in a timely manner;

[0061] 2. There is a lack of an accurate quantitative index to measure the rationality of the server area strength calculated based on the hyperparameters set by manual experience, making the calculated server area strength inaccurate;

[0062] 3. When the server area strength cannot be accurately and efficiently quantified, it is impossible to assist game planners in obtaining an interpretable and highly balanced server merging list (server area merging result) before the start of a new season in a timely manner, resulting in a reduced gaming experience for game objects and affecting the overall activity and commercialization level of the game.

[0063] Based on this, the embodiments of this application provide a data processing method. Figure 2 The figure shows a schematic diagram of the data processing method provided by an exemplary embodiment of this application. Taking the application of this method to a computer device as an example, the computer device can be Figure 1 the server 140 shown. Taking the example that there are two camps in the game: Camp 1 and Camp 2, and four server areas: Server Area 1, Server Area 2, Server Area 3, and Server Area 4.

[0064] In the server merge result 142 for the current time period (the T-th time period, i.e., the T-th season), the faction strength value corresponding to Faction 1 is the same as or similar to the faction strength value corresponding to Faction 2. Among them, Faction 1 includes Server 1 and Server 2; Game Object 1, Game Object 2, and Game Object 3 belong to Server 1; Game Object 4, Game Object 5, and Game Object 6 belong to Server 2; Faction 2 includes Server 3 and Server 4; Game Object 7, Game Object 8, and Game Object 9 belong to Server 3; Game Object 10, Game Object 11, and Game Object 12 belong to Server 4. At the end of the T-th time period, the computer device needs to determine the server merge result 144 for the next time period (the (T + 1)-th time period, i.e., the (T + 1)-th season) based on the object data of the game objects in the T-th time period.

[0065] The steps of the data processing method executed by the computer device are briefly described as follows:

[0066] Step 1, the computer device obtains the object data of the game objects in the T-th time period; T is greater than or equal to 1.

[0067] Step 2, the computer device obtains the hyperparameters corresponding to the strength calculation formula.

[0068] The strength calculation formula is pre-set and stored in the computer device. This strength calculation formula is used to calculate the strength value corresponding to the game object. The strength calculation formula includes hyperparameters. The hyperparameters obtained by the computer device are obtained by optimizing and solving the error function. The error function is determined based on the object data of the game objects in the (T - 1)-th time period (i.e., the (T - 1)-th season) and the T-th time period, and is used to balance the error between the estimated strength and the actual strength of each game object in the server. It should be noted that this hyperparameter is calculated by the computer device before Step 2, or, this hyperparameter is calculated by another server 160 ( Figure 2 not shown in the figure), and the computer device is connected to the server 160 through a network. The computer device can obtain the hyperparameters corresponding to the strength calculation formula from the server 160.

[0069] Step 3, the computer device substitutes the object data of the game object and the hyperparameters into the strength calculation formula to obtain the strength value corresponding to the game object.

[0070] Step 4, accumulate the strength values corresponding to the game objects belonging to the same server to obtain the server strength value corresponding to each server.

[0071] Step 5, merge the servers based on the server strength values to obtain the server merge result 144 for the (T + 1)-th time period.

[0072] In the server merge result 144 in the (T + 1)-th time period, the faction strength value corresponding to Faction 1 is the same as or close to the faction strength value corresponding to Faction 2. Among them, Faction 1 includes Server 1 and Server 3; Game Object 1, Game Object 2, and Game Object 3 belong to Server 1; Game Object 7, Game Object 8, and Game Object 9 belong to Server 3; Faction 2 includes Server 2 and Server 4; Game Object 4, Game Object 5, and Game Object 6 belong to Server 2; Game Object 10, Game Object 11, and Game Object 12 belong to Server 4.

[0073] In summary, the data processing method provided by the embodiments of the present application has at least the following beneficial effects: 1. The hyperparameters are not set manually, but obtained by optimizing and solving the error function. Since the error function is determined based on the object data of the game objects in the (T - 1)-th time period and the T-th time period and is used to balance the error between the estimated strength and the actual strength of each game object in the server, the hyperparameters can reflect the correlation between the server strength and the strength of the game objects, and have strong interpretability. 2. The setting of the hyperparameters does not require a large amount of manpower, and the setting efficiency is improved. 3. Based on the above hyperparameters, the strength value of the game object and the server strength value can be calculated, which can improve the accuracy of the calculation result and make the calculation result more reasonable. 4. Since the accuracy of the server strength value is improved, the accuracy of the server merge result is improved after merging each server according to the server strength value. 5. On the basis of the above beneficial effects 1-4, the game experience can be improved, and the overall activity and commercialization degree of the game can be enhanced.

[0074] Figure 3 The flowchart of the data processing method provided by an exemplary embodiment of the present application is shown. Taking the application of this method to a computer device as an example, the computer device may be Figure 1 the server 140 shown, and this method includes steps 210, 220, 230, 240, and 250:

[0075] Step 210, obtain the object data of the game objects in the T-th time period; T is greater than or equal to 1.

[0076] The time period refers to a period of time. In this embodiment, a time period may correspond to a season. Among them, a season refers to a period of time divided according to certain systems or rules. During this period, at least one of new game characters, versions, equipment, dungeons, gameplay, and game content will be launched, and an activity or competition will be carried out during this period. After this period ends, various game objects, servers, and factions in the game will be ranked, and each server will be merged.

[0077] A game object refers to a user account participating in the game. Object data refers to the relevant data of the game object. In this embodiment, the object data of the game object includes at least one of the following: historical recharge and payment value, number of dead and injured soldiers in the current time period, average strength value of the confrontation lineup used. The average strength value of the confrontation lineup used refers to the total strength value corresponding to the confrontation lineup divided by the number of confrontations.

[0078] At the end of the T-th time period, the computer device obtains the object data of the game object in the T-th time period. Optionally, the computer device obtains the object data of all game objects in the T-th time period. Or, the computer device obtains the object data of the active game objects in the T-th time period. An active game object refers to a game object with an activity level higher than a set value, and the activity level can be characterized by at least one of the number of logins, login duration, number of confrontations, average strength value of the confrontation lineup used, number of historical recharge and payment times, historical recharge and payment value, and number of dead and injured soldiers in the current time period.

[0079] Step 220, obtain the hyperparameters corresponding to the strength calculation formula; the strength calculation formula is used to calculate the strength value corresponding to the game object, and the hyperparameters are obtained by optimizing and solving the error function, and the error function is determined based on the object data of the game objects in the (T - 1)-th time period and the T-th time period, and is used to balance the error between the estimated strength and the actual strength of each game object in the server area.

[0080] The strength calculation formula is preset and stored in the computer device. The strength calculation formula is used to calculate the strength value corresponding to the game object, and this strength value is used to represent the comprehensive strength of the game object in the T-th time period. The larger the strength value, the stronger the comprehensive strength of the game object.

[0081] Exemplarily, the strength calculation formula corresponds one-to-one with the type of the object data of the game object. The strength calculation formula includes hyperparameters, and different strength calculation formulas include different hyperparameters. In this embodiment, the hyperparameters obtained by the computer device are obtained by optimizing and solving the error function. The error function is determined based on the object data of the game objects in the (T - 1)-th time period and the T-th time period, and is used to balance the error between the estimated strength and the actual strength of each game object in the server area. Since a server area includes multiple game objects and a camp is composed of multiple server areas, this error function can also be understood as the error between the estimated strength and the actual strength of different camps. The error function in this embodiment can reflect the strength of the game object in different time periods and the strength of the server area where the game object is located in different time periods. Therefore, this hyperparameter can reflect the correlation between the strength value of the game object and the server area strength value corresponding to the server area, and can also reflect the correlation between the strength value of the game object, the server area strength value corresponding to the server area, and the camp strength value corresponding to the camp where the server area is located. The hyperparameters in this embodiment can also be called optimal hyperparameters. The calculation method of the hyperparameters will be introduced in subsequent embodiments.

[0082] In some embodiments, taking the object data of the game object including at least one of the historical recharge and payment value (HistoryPay), the number of dead and injured soldiers in the current time period (DeadHurtSoldierCount), and the average strength value of the confrontation used lineup (UsedLineupStrength) as an example. Then the hyperparameters in the strength calculation formula include at least one of the first hyperparameter P1, the second hyperparameter P2, the third hyperparameter P3, the fourth hyperparameter P4, the fifth hyperparameter P5, and the sixth hyperparameter P6. The strength calculation formula includes at least one of the first strength calculation formula P(HistoryPay) corresponding to the historical recharge and payment value, the second strength calculation formula D(DeadHurtSoldierCount) corresponding to the number of dead and injured soldiers in the current time period, and the third strength calculation formula S(UsedLineupStrength) corresponding to the average strength value of the confrontation used lineup.

[0083] The first strength calculation formula P(HistoryPay) is: the sum value between the product of the historical recharge and payment value (HistoryPay) and the first hyperparameter P1 and the first preset value, where the sum value is less than or equal to the second hyperparameter P2; or, the second hyperparameter P2, where the sum value is less than or equal to the second hyperparameter P2. Optionally, the first preset value is set to 1. The first strength calculation formula P(HistoryPay) is expressed as follows:

[0084]

[0085] The second strength calculation formula D(DeadHurtSoldierCount) is: the second preset value, where the number of dead and injured soldiers in the current time period (DeadHurtSoldierCount) is less than or equal to the third hyperparameter P3; or, the value corresponding to the base of the difference between the number of dead and injured soldiers in the current time period (DeadHurtSoldierCount) minus the third hyperparameter P3 and the fourth hyperparameter P4 as the exponent, where the number of dead and injured soldiers in the current time period (DeadHurtSoldierCount) is greater than the third hyperparameter P3. Optionally, the second preset value is set to 0. The second strength calculation formula D(DeadHurtSoldierCount) is expressed as follows:

[0086]

[0087] The third strength calculation formula S(UsedLineupStrength) is: the value corresponding to the base number which is the sum of the third preset value and the average strength value of the confrontation used lineup (UsedLineupStrength), and the fifth hyperparameter P5 as the exponent, and then subtract the sixth hyperparameter P6. Optionally, the third preset value is set to 1, and the third strength calculation formula S(UsedLineupStrength) is expressed as follows:

[0088] S(UsedLineupStrength) = (1 + UsedLineupStrength)^P5 - P6

[0089] It should be noted that the above three strength calculation formulas are set with the object data of three types of game objects as examples. In actual applications, at least one of the above three example calculation formulas can be used for subsequent operations according to actual technical needs, or more strength calculation formulas and more hyperparameters can be set according to more types of object data for subsequent operations.

[0090] Step 230, substitute the object data and hyperparameters of the game object into the strength calculation formula to obtain the strength value corresponding to the game object.

[0091] Exemplarily, the computer device substitutes the object data and hyperparameters of the game object into the strength calculation formula matching the type of the object data to obtain the strength value corresponding to the game object.

[0092] In some embodiments, when the object data of the game object includes multiple types, each type of object data will correspond to a strength value, and the final strength value corresponding to the game object is the product of multiple strength values. For example, the object data includes the historical recharge and payment value, the number of dead and wounded soldiers in the current time period, and the average strength value of the confrontation used lineup. Then the game object includes a first sub-strength value, a second sub-strength value, and a third sub-strength value. Among them, the first sub-strength value is calculated based on the first strength calculation formula, the second sub-strength value is calculated based on the second strength calculation formula, and the third sub-strength value is calculated based on the third strength calculation formula. Then the strength value of the game object is the product of the first sub-strength value, the second sub-strength value, and the third sub-strength value.

[0093] Step 240, accumulate the strength values corresponding to the game objects belonging to the same server area to obtain the server area strength value corresponding to each server area.

[0094] Exemplarily, a game server includes multiple game objects. The computer device accumulates the strength values corresponding to the game objects belonging to the same game server to obtain the server strength values corresponding to each game server. Here, "accumulation" means addition. Or, in some other embodiments, the accumulation can also be a weighted sum, and the weight can be set according to the type of object data. The weights corresponding to the strength values calculated using different types of object data can be the same or different.

[0095] In some embodiments, the object data of the game object includes at least one of the historical recharge and payment value, the number of dead and wounded soldiers in the current time period, and the average strength value of the lineup used in the confrontation. The strength calculation formula includes at least one of the following as examples: the first strength calculation formula P(HistoryPay) corresponding to the historical recharge and payment value (HistoryPay), the second strength calculation formula D(DeadHurtSoldierCount) corresponding to the number of dead and wounded soldiers in the current time period (DeadHurtSoldierCount), and the third strength calculation formula S(UsedLineupStrength) corresponding to the average strength value of the lineup used in the confrontation (UsedLineupStrength). Then, the server strength value corresponding to a game server is the sum of the strength values corresponding to the game objects in this game server, and the strength value is the product of the first sub-strength value, the second sub-strength value, and the third sub-strength value; where the first sub-strength value is calculated based on the first strength calculation formula, the second sub-strength value is calculated based on the second strength calculation formula, and the third sub-strength value is calculated based on the third strength calculation formula. Representing the number of game objects in a game server as N, the server strength value (ZoneScore) corresponding to a game server is expressed as:

[0096]

[0097] Step 250: Merge each game server based on the server strength value to obtain the game server merging result in the (T + 1)-th time period.

[0098] The computer device merges each game server based on the server strength value to obtain the game server merging result in the (T + 1)-th time period. Exemplarily, the game server merging result includes at least two server sets, each server set in the at least two server sets includes at least one game server, the at least two server sets belong to at least two camps, and the at least two camps are hostile to each other.

[0099] Before the start of the (T + 1)-th time period (the (T + 1)-th season), the computer device also notifies the game objects of the game server merging result in the (T + 1)-th time period through the game promotion channels. Optionally, the game promotion channels include at least one of a promotional announcement, a splash screen advertisement, and a pop-up email.

[0100] In summary, in the data processing method provided by the embodiments of the present application, a computer device obtains object data of game objects in the T-th time period, where T is greater than or equal to 1; obtains hyperparameters corresponding to a strength calculation formula; substitutes the object data of the game objects and the hyperparameters into the strength calculation formula to obtain strength values corresponding to the game objects; accumulates the strength values corresponding to the game objects belonging to the same server area to obtain server area strength values corresponding to each server area; and merges each server area based on the server area strength values to obtain a server area merging result in the (T + 1)-th time period. Since the hyperparameters are obtained by optimizing and solving an error function, rather than being manually set, and the error function is determined based on the object data of the game objects in the (T - 1)-th time period and the T-th time period and is used to balance the error between the estimated strength and the actual strength of each game object in the server area, the hyperparameters can reflect the correlation between the server area strength and the strength of the game objects and have strong interpretability. Therefore, the rationality and accuracy of the strength values of the game objects and the server area strength values determined based on the hyperparameters can be improved. Since the accuracy of the server area strength values is improved, the accuracy of the server area merging result is improved when determining the server area merging result based on the server area strength values. In addition, since the accuracy of the server area merging result is improved, the matching of the game objects with our side's camp, our side's game objects, the enemy's camp, and the enemy's game objects can be more accurate in the (T + 1)-th time period, realizing a relatively fair game environment for different game objects. Therefore, this solution can also improve the user's gaming experience and enhance the overall activity and commercialization level of the game.

[0101] Next, the calculation method of the hyperparameters corresponding to the strength calculation formula will be introduced.

[0102] · Hyperparameters

[0103] In some embodiments, Figure 4 The flowchart of the data processing method provided by an exemplary embodiment of the present application is shown. Step 220 may optionally include steps 310, 320, 330, and 340:

[0104] Step 310, obtain the object data of the game objects in the (T - 1)-th time period and the T-th time period.

[0105] The (T - 1)-th time period is the previous time period of the T-th time period. In this embodiment, the computer device obtains the object data of the game objects in the consecutive (T - 1)-th time period and the T-th time period. Since the time of the (T - 1)-th time period and the T-th time period is consecutive, the correlation of the object data in these two time periods is higher, which can make the estimated camp strength value in the T-th time period closer to the actual camp strength value and make the calculated hyperparameters more interpretable.

[0106] In some other embodiments, the computer device obtains the object data of game objects during the X-th time period and the T-th time period. Herein, the X-th time period is an arbitrary random time period before the T-th time period. This embodiment is applicable to the case where the volume of object data is small and the degree of change in object data between any two time periods is less than a threshold value.

[0107] Step 320: Accumulate based on the object data of the (T - 1)-th time period to obtain the predicted faction strength value of the T-th time period; and, perform statistics based on the object data of the T-th time period to obtain the actual faction strength value of the T-th time period.

[0108] The predicted faction strength value (PredictFactionScore T ) of the T-th time period is obtained based on the object data of the (T - 1)-th time period, and the actual faction strength value of the (T - 1)-th time period is used as the predicted faction strength value of the T-th time period. The predicted faction strength value is used to represent the predicted strength of the faction corresponding to the server area where each game object is located during the T-th time period. The predicted faction strength value is calculated by substituting the object data of the (T - 1)-th time period and the initial hyperparameters into the strength calculation formula, and the calculation method can refer to Step 321-1, Step 322-1, and Step 323-1.

[0109] The actual faction strength value (FactionMilitaryExploit T ) of the T-th time period is obtained based on the object data of the T-th time period. The actual faction strength value is used to represent the actual strength of the faction corresponding to the server area where each game object is located during the T-th time period. The actual faction strength value is calculated by substituting the object data of the T-th time period and the initial hyperparameters into the strength calculation formula, and the calculation method can refer to Step 321-2, Step 322-2, and Step 323-2.

[0110] Exemplarily, the computer device accumulates based on the object data of the (T - 1)-th time period to obtain the predicted faction strength value of the T-th time period; and, performs statistics based on the object data of the T-th time period to obtain the actual faction strength value of the T-th time period. Herein, the predicted faction strength value and the actual faction strength value can be calculated simultaneously or separately, and the time sequence is not limited in this embodiment.

[0111] Step 330: Determine the error function between the predicted faction strength value and the actual faction strength value.

[0112] The error function is determined based on the object data of the game objects in the (T - 1)-th time period and the T-th time period, and is used to balance the error between the estimated strength and the actual strength of the camps corresponding to the regions where each game object is located. Exemplarily, the computer device determines the error function between the estimated camp strength value and the actual camp strength value. This error function can also be referred to as the objective function.

[0113] Step 340, taking minimizing the error function as the optimization objective, solve to obtain the hyperparameters corresponding to the strength calculation formula.

[0114] Exemplarily, the computer device takes minimizing the error function as the optimization objective, and solves to obtain the hyperparameters corresponding to the strength calculation formula. It should be noted that the computer device can calculate the hyperparameters only once. For example, when the game characters, versions, equipment, dungeons, gameplay, and game content in each time period change little, it is only necessary to calculate the hyperparameters once and apply them to the steps of determining the server merge result. Or, the computer device can also update the hyperparameters periodically, and this period can be determined based on at least one of game characters, versions, equipment, dungeons, gameplay, and game content. For example, compared with the (T - 1)-th time period, when the number of newly added dungeons in the T-th time period exceeds the set value, when performing server merge at the end of the T-th time period, first update the hyperparameters once, and then execute each step of determining the server merge result.

[0115] In this embodiment, the computer device obtains the object data of the game objects in the consecutive (T - 1)-th time period and the T-th time period, and determines the error function in the T-th time period based on the object data of the game objects in the (T - 1)-th time period and the T-th time period. Since the time of the (T - 1)-th time period and the T-th time period is consecutive, the correlation of the object data in these two time periods is higher, making the error function more accurate. When solving the hyperparameters corresponding to the strength calculation formula according to the error function, the obtained hyperparameters have stronger interpretability.

[0116] In some embodiments, taking the game including two camps as an example. The camps include the first camp and the second camp, then the estimated camp strength value includes the estimated first camp strength value and the estimated second camp strength value, and the actual camp strength value includes the actual first camp strength value and the actual second camp strength value.

[0117] Exemplarily, the cumulative calculation based on the object data in the (T - 1)-th time period in step 320 to obtain the estimated camp strength value in the T-th time period can be specifically implemented as steps 321-1, 322-1, and 323-1:

[0118] Step 321-1, substitute the object data of the game objects in the (T - 1)-th time period and the initial hyperparameters into the strength calculation formula to obtain the first strength value of the game objects in the (T - 1)-th time period.

[0119] The initial hyperparameter is the initial value of the hyperparameter. The first strength value refers to the strength value corresponding to the game object in the (T - 1)th time period. Exemplarily, the computer device substitutes the object data of the game object in the (T - 1)th time period and the initial hyperparameter into the strength calculation formula to obtain the first strength value of the game object in the (T - 1)th time period.

[0120] Step 322 - 1: Accumulate the first strength values corresponding to the game objects belonging to the same server area to obtain the first server - area strength values respectively corresponding to each server area in the (T - 1)th time period.

[0121] Accumulation means addition. The first server - area strength value refers to the server - area strength value corresponding to the server area in the (T - 1)th time period. Exemplarily, the computer device accumulates the first strength values corresponding to the game objects belonging to the same server area to obtain the first server - area strength values (ZoneScore T-1 ) respectively corresponding to each server area in the (T - 1)th time period.

[0122] Step 323 - 1: Accumulate the first server - area strength values corresponding to the server areas belonging to the same faction to obtain the predicted first - faction strength value and the predicted second - faction strength value in the Tth time period.

[0123] Accumulation means addition. The predicted first - faction strength value refers to the predicted faction strength value corresponding to the first faction. The predicted second - faction strength value refers to the predicted faction strength value corresponding to the second faction. Exemplarily, the computer device accumulates the first server - area strength values corresponding to the server areas belonging to the same faction to obtain the predicted first - faction strength value (PredictFactionScore of the first faction T ) and the predicted second - faction strength value (PredictFactionScore of the second faction T ) in the Tth time period.

[0124] In this embodiment, based on the object data of the game object in the (T - 1)th time period and the initial hyperparameter, the computer device can respectively obtain the first strength value of the game object in the (T - 1)th time period, the first server - area strength value of the server area, and the faction strength value. Since the (T - 1)th time period and the Tth time period are continuous in time, using the faction strength value in the (T - 1)th time period as the predicted faction strength value in the Tth time period can ensure the prediction accuracy of the faction strength value in the Tth time period.

[0125] Exemplarily, the statistics based on the object data in the Tth time period in step 320 to obtain the actual faction strength value in the Tth time period can be specifically implemented as steps 321 - 2, 322 - 2, and 323 - 2:

[0126] Step 321-2: Substitute the object data of the game object in the T-th time period and the initial hyperparameters into the strength calculation formula to obtain the second strength value of the game object in the T-th time period.

[0127] The initial hyperparameters are the initial values of the hyperparameters. The second strength value refers to the strength value corresponding to the game object in the T-th time period. Exemplarily, the computer device substitutes the object data of the game object in the T-th time period and the initial hyperparameters into the strength calculation formula to obtain the second strength value of the game object in the T-th time period.

[0128] Step 322-2: Accumulate the second strength values corresponding to the game objects belonging to the same server area to obtain the second server area strength values respectively corresponding to each server area in the T-th time period.

[0129] Accumulation means summing up. The second server area strength value refers to the server area strength value corresponding to the server area in the T-th time period. Exemplarily, the computer device accumulates the second strength values corresponding to the game objects belonging to the same server area to obtain the second server area strength values (ZoneScore T ) respectively corresponding to each server area in the T-th time period.

[0130] Step 323-2: Accumulate the second server area strength values corresponding to the server areas belonging to the same camp to obtain the actual first camp strength value and the actual second camp strength value in the T-th time period.

[0131] Accumulation means summing up. The actual first camp strength value refers to the actual camp strength value corresponding to the first camp. The actual second camp strength value refers to the actual camp strength value corresponding to the second camp. Exemplarily, the computer device accumulates the second server area strength values corresponding to the server areas belonging to the same camp to obtain the actual first camp strength value (FactionMilitaryExploit of the first camp T ) and the actual second camp strength value (FactionMilitaryExploit of the second camp T ) in the T-th time period.

[0132] In this embodiment, based on the object data of the game object in the T-th time period and the initial hyperparameters, the computer device can respectively obtain the second strength value of the game object in the T-th time period, the second server area strength value of the server area, and the actual camp strength value, and can accurately determine the camp strength value in the T-th time period.

[0133] In some embodiments, step 330 can be optionally implemented as steps 331, 332, and 333:

[0134] Step 331: Determine the first ratio of the estimated first camp strength value to the estimated second camp strength value.

[0135] The first ratio refers to the value obtained by dividing the estimated strength value of the first camp by the estimated strength value of the second camp. Exemplarily, the computer device determines the first ratio of the estimated strength value of the first camp to the estimated strength value of the second camp.

[0136] Step 332: Determine the second ratio of the actual strength value of the first camp to the actual strength value of the second camp.

[0137] The second ratio refers to the value obtained by dividing the actual strength value of the first camp by the actual strength value of the second camp. Exemplarily, the computer device determines the second ratio of the actual strength value of the first camp to the actual strength value of the second camp.

[0138] Step 333: Subtract the second ratio from the first ratio to obtain an error function.

[0139] Exemplarily, the computer device subtracts the second ratio from the first ratio to obtain an error function (Loss). Optionally, the absolute value corresponding to the difference between the first ratio and the second ratio is determined as the error function. This error function is the objective function for solving the hyperparameters in the strength calculation formula, which is expressed as follows:

[0140]

[0141] The error function in this embodiment can minimize the relative ratio error between the estimated camp strength and the actual camp strength of different camps in the server area. Thus, after determining the hyperparameters based on this error function, when calculating the strength of the server area and camps, the error between the estimated strength and the actual strength of the camp is smaller, which is beneficial to accurately estimating the server area strength value and further improving the accuracy of the server area merging result.

[0142] In some embodiments, an optimization algorithm is used to solve the hyperparameters. In the following embodiments, it is exemplified that the object data of the game object includes the historical recharge and payment value, the number of dead and wounded soldiers in the current time period, and the average strength value of the confrontation lineup. Then, the hyperparameters in the strength calculation formula include the first hyperparameter P1, the second hyperparameter P2, the third hyperparameter P3, the fourth hyperparameter P4, the fifth hyperparameter P5, and the sixth hyperparameter P6 for illustration.

[0143] Figure 5 The flowchart of the data processing method provided by an exemplary embodiment of the present application is shown. Step 340 is specifically implemented as steps 410, 420, 430, 440, and 450:

[0144] Step 410: Obtain an initialization data set; the initialization data set includes N sample data, and the N sample data are determined based on the hyperparameters to be solved in the strength calculation formula and the error function, where N is greater than or equal to 1.

[0145] Initialize the dataset for fitting to obtain the first Gaussian process model. Exemplarily, the computer device obtains the initialization dataset D; the initialization dataset includes N sample data, and the N sample data are determined based on the hyperparameters to be solved and the error function in the strength calculation formula, where N is greater than or equal to 1. Among them, the sample point x is determined based on the hyperparameters to be solved, and the error Y corresponding to the sample point is determined based on the error function. The sample point x and the corresponding error Y together form a sample data. Denote the i-th sample point x i as:

[0146] Step 420, use the initialization dataset to fit and obtain the first Gaussian process model; the first Gaussian process model includes a Gaussian kernel function, and the Gaussian kernel function is used to describe the similarity between any two sample data in the initialization dataset.

[0147] The first Gaussian process model is a Gaussian process model obtained or trained based on the initialization dataset. Exemplarily, the computer device uses the initialization dataset D to fit and obtain the first Gaussian process model. The first Gaussian process model includes a Gaussian Kernel, and the Gaussian kernel function is used to describe the similarity between any two sample data in the initialization dataset.

[0148] Specifically, the first Gaussian process model is expressed as f(x) ∼ GP(μ(x), k(x, x′)). Among them, f(x) is the first Gaussian process model, GP represents the Gaussian process, μ(x) is the mean function of f(x), which can usually be set to 0; k(x, x′) is the Gaussian kernel function, and x and x′ respectively represent the sample points corresponding to any two sample data in the initialization dataset D. The Gaussian kernel function x(x, x′) is specifically expressed as follows:

[0149]

[0150] where, ‖x - x′‖ 2 represents the Euclidean distance between x and x′, and σ is used to control the range of action of the Gaussian kernel function. The Gaussian kernel function k(x, x′) describes the covariance and is used to describe the similarity between x and x′. Since the sample points are determined based on the hyperparameters, the Gaussian kernel function can also be used to describe the similarity between different hyperparameters.

[0151] Step 430, iteratively update the initialization dataset, and use the iteratively updated initialization dataset to update the first Gaussian process model until the iteration stop condition is satisfied to obtain the second Gaussian process model; the iteratively updated initialization dataset includes N + i sample data, where i is the number of updates of the initialization dataset, and i is greater than or equal to 1.

[0152] The first Gaussian process model is continuously iteratively updated based on the iteratively updated initial dataset, and the second Gaussian process model is the final Gaussian process model obtained when the iterative update stops. The iteratively updated initial dataset refers to adding new sample data to the initial dataset continuously during the iterative process, and each iterative update adds a new sample data to the initial dataset. When the update count of the initial dataset is i, the obtained iteratively updated initial dataset will include N + i sample data.

[0153] The iteration stop condition is a pre-set stop condition. In some embodiments, the iteration stop condition includes at least one of the following: reaching a pre-set number of iterations, the degree of change between the error functions corresponding to the second Gaussian process model and the first Gaussian process model is less than a pre-set error threshold.

[0154] Exemplarily, the computer device iteratively updates the initial dataset D, and uses the iteratively updated initial dataset to update the first Gaussian process model until the iteration stop condition is satisfied to obtain the second Gaussian process model; the iteratively updated initial dataset includes N + i sample data, where i is the update count of the initial dataset and i is greater than or equal to 1.

[0155] Step 440, determine the minimum value of the error function based on the second Gaussian process model.

[0156] Exemplarily, the computer device determines the minimum value of the error function based on the second Gaussian process model.

[0157] Step 450, determine the hyperparameters corresponding to the strength calculation formula as the hyperparameters corresponding to the minimum value of the error function.

[0158] Exemplarily, the computer device determines the hyperparameters corresponding to the minimum value of the error function as the hyperparameters corresponding to the strength calculation formula. Among them, the minimum value of the error function is also the minimum value of the error Y corresponding to the sample point x in the iteratively updated initial dataset, and the sample point x is determined based on the hyperparameters. Therefore, the hyperparameters corresponding to the minimum value of the error function can be determined. The computer device represents the hyperparameters x corresponding to the strength calculation formula as: best It is expressed as:

[0159] Since the calculation formula of the error function is relatively complex and there are no first-order and second-order derivatives, it is impossible to directly use algorithms such as gradient descent to solve the extreme value. The computer device in this embodiment trains a Gaussian process model using the initial dataset, and solves the minimum value of the error function by iteratively updating the initial dataset and the Gaussian process model, which can improve the data processing efficiency and quickly solve the hyperparameters corresponding to the strength calculation formula.

[0160] In some embodiments, step 410 specifically includes step 411, step 412, and step 413:

[0161] Step 411, randomly set N groups of hyperparameters within the value range of the hyperparameters in the strength calculation formula.

[0162] Exemplarily, the computer device randomly sets N groups of hyperparameters within the value range of the hyperparameters in the strength calculation formula. Among them, the value range of each hyperparameter is a specified value range.

[0163] Specifically, the computer device randomly sets N groups of hyperparameters within the respective specified value ranges of each of the hyperparameters P1, P2, P3, P4, P5, and P6. Each group of hyperparameters in the N groups of hyperparameters includes a randomly set value of the above 6 hyperparameters.

[0164] Step 412, based on the error function, determine the errors corresponding to each of the N groups of hyperparameters.

[0165] Exemplarily, the computer device determines the errors corresponding to each of the N groups of hyperparameters based on the error function. Among them, the expression of the error function (Loss) can be referred to in the embodiment of step 333. For each group of hyperparameters in the N groups of hyperparameters, the computer device calculates the corresponding error using this expression.

[0166] Specifically, represent the errors corresponding to the N groups of hyperparameters as loss = {Y1, Y2, …, Y N}. Among them, Y1 represents the error corresponding to the first group of hyperparameters corresponding error.

[0167] Step 413, based on the N groups of hyperparameters and their respective corresponding errors, obtain an initial dataset; among them, each sample data in the initial dataset respectively includes a sample point and the error corresponding to the sample point, the error is determined based on the error function, and each sample point is a group of hyperparameters in the N groups of hyperparameters.

[0168] Exemplarily, the computer device obtains an initial dataset D based on the N groups of hyperparameters and their respective corresponding errors; among them, the initial dataset D includes N sample data, and each sample data respectively includes a sample point x and the error Y corresponding to the sample point x, the error is determined based on the error function, and each sample point is a group of hyperparameters in the N groups of hyperparameters.

[0169] Specifically, the initial dataset D is represented as: D = {(x1, Y1), (x2, Y2), …, (x N , Y N )}. Among them, a sample point x1 and the error Y1 corresponding to the sample point x1 are called a sample data (x1, Y1). The i-th sample point xi Expressed as:

[0170] In this embodiment, the computer device randomly sets N groups of hyperparameters within the value ranges of the respective hyperparameters in the strength calculation formula, and obtains an initial dataset according to the error corresponding to each group of hyperparameters. This initial dataset can be used to represent each group of hyperparameters and can also represent the error corresponding to each group of hyperparameters, which is beneficial for subsequent solving and determining the minimum value of the error function.

[0171] In some embodiments, step 420 specifically includes step 421 and step 422:

[0172] Step 421, determine the joint Gaussian distribution corresponding to the N sample data in the initial dataset.

[0173] Step 422, based on the joint Gaussian distribution, fit to obtain the first Gaussian process model.

[0174] For each sample data in the initial dataset, there is a corresponding Gaussian distribution. Then, there is a joint Gaussian distribution for all the sample data in the initial data. This joint Gaussian distribution is the first Gaussian process model trained using the initial dataset.

[0175] Exemplarily, the computer device determines the joint Gaussian distribution corresponding to all the sample points {x1, x2, …, x N} in the N sample data in the initial dataset D. The mean of the joint Gaussian distribution is 0, and the covariance is expressed as:

[0176]

[0177] Among them, k(x1, x1) represents the covariance between the sample point x1 and the sample point x1, which is obtained by substituting x1 into the Gaussian kernel function k(x, x′); k(x N , x2) represents the covariance between the sample point x N and the sample point x2, which is obtained by substituting x N and x2 into the Gaussian kernel function k(x, x′), and so on. The computer device fits to obtain the first Gaussian process model f(x 1:N ) based on this joint Gaussian distribution. The first Gaussian process model f(x 1:N ) follows an N-dimensional Gaussian distribution.

[0178] In this embodiment, the computer device can train a Gaussian process model using the initial dataset, which is beneficial for subsequently accurately finding the global optimal point that makes the error function reach the minimum value.

[0179] In some embodiments, step 430 specifically includes steps 431, 432, 433, 434, 425, and 436:

[0180] Step 431, determine the update count i of the updated initialization dataset;

[0181] Step 432, determine the (N + i)-th sample point;

[0182] Step 433, based on the error function, determine the error corresponding to the (N + i)-th sample point;

[0183] Step 434, based on the (N + i)-th sample point and the corresponding error, determine the (N + i)-th sample data;

[0184] Step 435, add the (N + i)-th sample data to the initialization dataset to iteratively update the initialization dataset, and use the iteratively updated initialization dataset to update the first Gaussian process model;

[0185] Step 436, update i to i + 1, and repeat the step of determining the (N + i)-th sample point until the iteration stop condition is met to obtain the second Gaussian process model.

[0186] The computer device iteratively updates the initialization dataset, and each iterative update adds a new sample data to the initialization dataset. And the first Gaussian process model is iteratively updated using the iteratively updated initialization dataset. Among them, the new sample data consists of a new sample point (a set of new hyperparameters) and its error.

[0187] Exemplarily, the computer device determines the update count i of the updated initialization dataset. Determine the (N + i)-th sample point, denoted as Based on the error function, determine the error Y corresponding to the (N + i)-th sample point N+i . Among them, for the expression of the error function (Loss), please refer to the example shown in step 333. The error corresponding to the (N + i)-th sample point is obtained by solving using this expression. Combine the (N + i)-th sample point and the corresponding error into the (N + i)-th sample data, denoted as (x N+i , Y N+i ). Add the (N + i)-th sample data to the initialization dataset to iteratively update the initialization dataset, and use the iteratively updated initialization dataset to update the first Gaussian process model. After this iterative update is completed, then update i to i + 1, and repeat the step of determining the (N + i)-th sample point until the iteration stop condition is met to obtain the second Gaussian process model.

[0188] Taking i = 1 and the first iterative update as an example, the computer device determines the update count 1 of the updated initialization dataset. Determine the (N + 1)-th sample point, denoted as Based on the error function, determine the error Y corresponding to the (N + 1)-th sample point N+1 . Combine the (N + 1)-th sample point and the corresponding error into the (N + 1)-th sample data, denoted as (x N+1 , Y N+1 ). Add the (N + 1)-th sample data to the initialized data set to iteratively update the initialized data set, and use the iteratively updated initialized data set to update the first Gaussian process model.

[0189] In this embodiment, the computer device realizes the iterative update of the initialized data set by determining the next new sample point and adding the sample point and its corresponding error to the initialized data set. By using the iteratively updated initialized data set to iteratively update the Gaussian process model, the iterative update of the Gaussian process model is realized. For each iterative update, the region where the set of hyperparameters that minimizes the error function is located becomes clearer. Through multiple iterative updates, the sample point corresponding to the minimum value of the error function can be found with fewer steps, so that a set of hyperparameters that minimizes the error function can be quickly determined, improving the calculation efficiency of the hyperparameters.

[0190] In some embodiments, the determination method of the (N + i)-th sample point is introduced. In this embodiment, the (N + i)-th sample point is determined based on the Upper Confidence Bound (UCB) strategy.

[0191] Specifically, step 432 can be implemented as step 520 and step 540:

[0192] Step 520, determine the mean and variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point.

[0193] Step 540, based on the mean and variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point, determine the (N + i)-th sample point that minimizes the error function.

[0194] When a given data set is provided, a corresponding multi-dimensional Gaussian process model can be fitted. For example, using the initialized data set, the first Gaussian process model is fitted, and this first Gaussian process model follows an N-dimensional Gaussian distribution. On this basis, for any given new sample point, according to the properties of the multi-dimensional Gaussian process model, the mean and variance of the one-dimensional Gaussian distribution corresponding to this sample point are calculated. On this basis, based on the UCB strategy, the new (N + i)-th sample point that minimizes the error function is determined.

[0195] Specifically, the computer device determines the mean μ(x) and variance σ(x) of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point. Subtract the product wσ(x) of the standard deviation σ(x) corresponding to the variance and the weight w from the mean μ(x) of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point to determine the (N + i)-th sample point that minimizes the error function. Among them, the weight is a preset fixed value used to balance the expected mean and variance.

[0196] Taking i = 1 and the first iterative update as an example, the computer device needs to determine the next sample point x that minimizes the error function N+1 . The calculation formula is: x N+1 = argmin x UCB = μ(x) - wσ(x), where the argmin function is used to determine the value of the independent variable when the error function reaches the minimum value. w is the weight used to balance the expected mean and variance. It can be seen from the above calculation formula that the UCB strategy tends to select the area with a low mean and a large standard deviation. Through the UCB strategy, the next sample point x N+ 1 that minimizes the error function can be calculated.

[0197] In this embodiment, the computer device can quickly determine the next new sample point that minimizes the error function, which is beneficial to realizing the iterative update of the initialization dataset.

[0198] In some embodiments, step 520 can be specifically implemented as step 521, step 522, and step 523:

[0199] Step 521, based on the N-dimensional Gaussian distribution that the first Gaussian process model follows on the initialization dataset and the one-dimensional Gaussian distribution that the first Gaussian process model follows at the (N + i)-th sample point, perform partitioning to determine the covariance matrix and the mean vector; the covariance matrix is calculated based on the initialization dataset and the Gaussian kernel function.

[0200] After adding the (N + i)-th sample point to the initialization dataset, at this time, the first Gaussian process model f(x 1：N+i ) still follows an (N + 1)-dimensional Gaussian distribution. Then, the computer device performs partitioning on the mean vector and the covariance matrix based on the N-dimensional Gaussian distribution that the first Gaussian process model follows on the initialization dataset and the one-dimensional Gaussian distribution that the first Gaussian process model follows at the (N + i)-th sample point to determine the covariance matrix and the mean vector. The first Gaussian process model at this time can be expressed as:

[0201]

[0202] Among them, f(x 1:N) represents the N-dimensional Gaussian distribution that the first Gaussian process model follows on the initialization dataset; μ(x 1：N ) is the mean function of f(x 1：N ); the capital K represents the covariance matrix, which is calculated based on the initialization dataset D and the Gaussian kernel function k(x, x′); the lowercase k represents the covariance, which is determined based on the covariance between any two sample points from x1 to x N+j , and is expressed as k = [k(x N+i , x1), k(x N+i , x2),..., k(x N+i , x N ); k T represents the transpose of k; f(x N+i ) represents the one-dimensional Gaussian distribution that the first Gaussian process model follows on the (N + i)-th sample point.

[0203] After partitioning, the mean vector is: The covariance matrix is:

[0204] Step 522, determine the one-dimensional Gaussian distribution that the first Gaussian process model follows on the (N + i)-th sample point, given that the first Gaussian process model follows an N-dimensional Gaussian distribution on the initialization dataset.

[0205] Since the conditional distribution of a multi-dimensional Gaussian distribution is still a Gaussian distribution, the computer device can determine the conditional distribution that the first Gaussian process model follows on the (N + i)-th sample point, given that the first Gaussian process model follows an N-dimensional Gaussian distribution on the initialization dataset. According to the properties of the multi-dimensional Gaussian distribution, this conditional distribution follows a one-dimensional Gaussian distribution.

[0206] Specifically, the computer device calculates the conditional distribution that f(x 1：N ) follows given f(x N+i ). This conditional distribution follows a one-dimensional Gaussian distribution, and the calculation formula for this conditional distribution is: f(x N+i )|f(x 1：N ) ~ GP(μ, σ 2 ), where μ is the mean and σ 2 is the variance.

[0207] Step 523, based on the covariance matrix, the mean vector, and the one-dimensional Gaussian distribution, determine the mean and variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point.

[0208] Exemplarily, the computer device substitutes the covariance matrix and the mean vector obtained by partitioning into the calculation formula of the one-dimensional Gaussian distribution, and can obtain the mean μ and variance σ 2. They are respectively represented as follows:

[0209] μ = k T K -1 f(x 1：N )

[0210] σ 2 = k(x N+i , x N+i ) - k T K -1 k

[0211] Then for a new set of hyperparameters x N+i , it can be considered that the error function follows a one-dimensional Gaussian distribution with a mean of μ and a variance of σ N+i on x. 2

[0212] Taking i = 1, the first iteration update as an example, the computer device needs to determine the mean and variance of the one-dimensional Gaussian distribution corresponding to the (N + 1)-th sample point. After adding the (N + 1)-th sample point to the initial dataset, the first Gaussian process model f(x 1：N+1 ) still follows an (N + 1)-dimensional Gaussian distribution. Then, based on the N-dimensional Gaussian distribution that the first Gaussian process model follows on the initial dataset and the one-dimensional Gaussian distribution that the first Gaussian process model follows at the (N + 1)-th sample point, the computer device partitions the mean vector and covariance matrix to determine the covariance matrix and mean vector. The first Gaussian process model at this time can be expressed as:

[0213]

[0214] Among them, f(x 1：N ) represents the N-dimensional Gaussian distribution that the first Gaussian process model follows on the initial dataset; μ(x 1：N ) is the mean function of f(x 1：N ); the capital K represents the covariance matrix, which is calculated based on the initial dataset D and the Gaussian kernel function k(x, x'); the lowercase k represents the covariance, which is determined based on the covariance between any two sample points from x1 to x N+1 , and is expressed as k = [k(x N+1 , x1), k(x N+1 , x2),..., k(x N+1 , x N )]; k T represents the transpose of k; f(x N+1 ) represents the one-dimensional Gaussian distribution that the first Gaussian process model follows at the (N + 1)-th sample point.

[0215] Since the conditional distribution of the multi-dimensional Gaussian distribution is still a Gaussian distribution, the computer device can calculate the conditional distribution that f(x 1：N ) follows given that f(x N+1 ) is known. This conditional distribution follows a one-dimensional Gaussian distribution. The calculation formula for this conditional distribution is: f(x N+1 )|f(x 1：N )~GP(μ, σ 2 ). According to the block scheme of the mean vector and covariance matrix and the calculation formula of the conditional distribution, the mean μ and variance σ 2 of this conditional distribution can be obtained, expressed as: μ = k T K -1 f(x 1：N ), σ 2 = k(x N+1 , x N+1 ) - k T K -1 k. For a new set of hyperparameters x N+1 , it can be considered that the error function follows a one-dimensional Gaussian distribution with a mean of μ and a variance of σ N+1 on x 2 .

[0216] In this embodiment, the computer device can calculate the one-dimensional Gaussian distribution that the Gaussian process model follows at the next new sample point by using the N-dimensional Gaussian distribution that the known Gaussian process model follows on the initialization dataset, which is beneficial to determining the next new sample point that minimizes the error function.

[0217] Next, the strategy for server area merging will be introduced.

[0218] · Server Area Merging

[0219] In some embodiments, taking a game with two camps, namely the first camp and the second camp, as an example. Step 250 can be optionally implemented as steps 251, 252, 253, and 254:

[0220] Step 251, divide the server area strength value into intervals to obtain at least two intervals.

[0221] The computer device divides the server area strength value into intervals to obtain at least two intervals. Exemplarily, taking the division into 5 intervals as an example, the server area strength values corresponding to the 5 intervals are expressed as follows: Very strong: [10000, +∞), Relatively strong: [60000, 100000), Medium strength: [40000, 60000), Relatively weak: [20000, 40000), Very weak: (0, 20000].

[0222] Step 252: Randomly select at least one zone and server from at least one interval of the at least two intervals to obtain a first zone and server set.

[0223] The first zone and server set is a set of randomly selected zones and servers.

[0224] The computer device randomly selects at least one zone and server from at least one interval of the at least two intervals to obtain a first zone and server set. Exemplarily, the computer device can randomly select at least one zone and server from each interval of the at least two intervals to obtain a first zone and server set, and the first zone and server set includes at least 5 zones and servers. The first zone and server set can be recorded as a first camp.

[0225] Step 253: Based on the server strength values ​​and the number of game objects corresponding to the first server set, at least one server is selected from at least one interval of the at least two intervals to obtain a second server set.

[0226] Since the zone server includes multiple game objects, the computer device selects at least one zone server from at least one of the at least two intervals based on the zone server strength value and the number of game objects corresponding to the first zone server set, and obtains a second zone server set. The second zone server set can be recorded as the second camp.

[0227] Exemplarily, when determining the second server set, the computer device makes the difference in server strength between the second server set and the first server set less than the first threshold, and the difference in the number of game objects between the second server set and the first server set less than the second threshold, so as to ensure that the number of game objects in the first server set and the second server set are the same or similar, and the overall server strength values ​​are the same or similar.

[0228] Step 254, determine the first server set and the second server set as the server merger result of the T+1th time period; wherein the difference in server strength value between the second server set and the first server set is less than the first threshold, and the difference in the number of game objects between the second server set and the first server set is less than the second threshold; the first server set and the second server set belong to different camps respectively.

[0229] Exemplarily, the computer device repeatedly executes step 252 and step 253 until all the regions and servers are merged. The computer device determines the first region and server set and the second region and server set as the region and server merge result of the T+1th time period.

[0230] For example, take the case where a computer device divides into 5 intervals. The computer device randomly selects a server area from each interval, and the first server area set composed of these 5 server areas is denoted as the first camp. According to the server area strength value and the number of game objects in each server area in the first camp, 4 to 6 server areas are then selected from these 5 intervals, and these server areas form the second server area set, constituting the second camp. It is necessary to ensure that the server area strength values and the number of game objects in the first camp and the second camp are as close as possible. After the first camp and the second camp are matched, they can be used as the opposing camps in the (T + 1)-th time period. For the remaining server areas, the selection operation is repeated, and finally the server area merger in the (T + 1)-th time period is completed. Before the (T + 1)-th time period starts, the game objects are informed through the game promotion channels.

[0231] It should be noted that the above embodiments are described by taking the game including two camps as an example. In the case where the game includes more than two camps, more server area sets can also be determined. For example, if the game also includes a third camp, the computer device can also determine the third server area set in the same way, so that the difference in the server area strength values among the first server area set, the second server area set, and the third server area set is less than the third threshold, and the difference in the number of game objects among the first server area set, the second server area set, and the third server area set is less than the fourth threshold.

[0232] In this embodiment, through the above server area merger strategy, the computer device can merge each server area based on the server area strength value. In the server area merger result, the server area strength values between the server area sets of different camps are the same or close, and the number of game objects between the server area sets of different camps is the same or close. Since the accuracy of the server area strength value is relatively high, the accuracy of the server area merger result is also improved. On this basis, in the next time period, a relatively fair game environment can be provided for the game objects in different camps and different server areas, improving the game experience.

[0233] To more clearly understand the data processing method provided by the embodiments of the present application, the following embodiments will describe the application scenario, overall architecture, and processing flow of the data processing method in combination with schematic diagrams.

[0234] · Application Scenario Example

[0235] When the game is in the settlement stage at the end of the season, the strength evaluation module of the computer device will be started regularly. The computer device will load the hyperparameters calculated by the optimization algorithm, substitute the object data of the game objects and the hyperparameters into the strength calculation formula, calculate the strength values of the game objects, and accumulate the strength values of the game objects to obtain the server area strength values of each server area.

[0236] 1) Figure 6The figure shows a schematic diagram of the end-of-season title settlement interface 700 provided by an exemplary embodiment of the present application. For a game object, the end-of-season title settlement interface 700 is displayed on the computer device of the game object, and the title obtained by the game object can be determined according to the strength value of the game object. Optionally, the title can be at least one of "Ascend to the Throne", "Separate Regime", and "Heroes of the Time".

[0237] 2) Figure 7 The figure shows a schematic diagram of the server strength value ranking provided by an exemplary embodiment of the present application. For each server, the servers can be ranked according to the server strength value. Figure 7 For the servers ranked in the top 20 in terms of server strength value, the horizontal axis represents the server number, and the vertical axis represents the server strength value.

[0238] 3) By combining each server according to the server strength value, the server combination result, also known as the server merger list, can be obtained. Before the new season starts, the server merger list for the new season is announced to each game object.

[0239] An example of a server merger list is shown in Table 1 below:

[0240] Table 1 Example of Server Merger List

[0241]

[0242] In the following embodiments, the game includes two camps; the object data of the game object includes at least one of the following: historical recharge and payment value (HistoryPay), the number of dead and injured soldiers in the current time period (DeadHurtSoldierCount), and the average strength value of the lineup used in confrontation (UsedLineupStrength); the strength calculation formula includes at least one of the first strength calculation formula, the second strength calculation formula, and the third strength calculation formula.

[0243] The first strength calculation formula corresponding to the historical recharge and payment value is expressed as:

[0244]

[0245] The second strength calculation formula corresponding to the number of dead and injured soldiers in the current time period is expressed as:

[0246]

[0247] The third strength calculation formula corresponding to the average strength value of the lineup used in confrontation is expressed as:

[0248] S(UsedLineupStrength) = (1 + UsedLineupStrength)^P5 - P6

[0249] The hyperparameters in the strength calculation formula include the first hyperparameter P1, the second hyperparameter P2, the third hyperparameter P3, the fourth hyperparameter P4, the fifth hyperparameter P5, and the sixth hyperparameter P6.

[0250] The zone strength value (ZoneScore) corresponding to a zone is expressed as:

[0251]

[0252] · Overall architecture

[0253] Figure 8 The overall architecture diagram of the data processing method provided by an exemplary embodiment of the present application is shown.

[0254] This data processing method mainly includes three parts: The first part: Solving hyperparameters based on an optimization algorithm, including Figure 8 Steps 1 to 12 in; The second part: Estimating the strength value of game objects and the zone strength value, including Figure 8 Steps 13 to 16 in; The third part: Zone merging, including Figure 8 Steps 17 and 18 in. The following will explain the three parts separately:

[0255] · The first part: Solving hyperparameters based on an optimization algorithm

[0256] 1. Obtain the object data of game objects in the consecutive (T - 1)-th time period and the T-th time period: historical recharge and payment value, the number of dead and wounded soldiers in the current time period, and the average strength value of the confrontation lineup used.

[0257] 2. Obtain a set of set initial hyperparameters P1, P2, P3, P4, P5, P6.

[0258] 3. Obtain the strength calculation formula.

[0259] 4. Substitute the object data of game objects in the (T - 1)-th time period and the initial hyperparameters into the strength calculation formula to obtain the first strength value of game objects in the (T - 1)-th time period; Substitute the object data of game objects in the T-th time period and the initial hyperparameters into the strength calculation formula to obtain the second strength value of game objects in the T-th time period.

[0260] 5. Add up the first strength values corresponding to all game objects belonging to the same zone to obtain the first zone strength value (ZoneScore T-1 ) corresponding to each zone in the (T - 1)-th time period; Add up the second strength values corresponding to all game objects belonging to the same zone to obtain the second zone strength value (ZoneScore T ) corresponding to each zone in the T-th time period.

[0261] 6. Add up the first server strength values corresponding to all server regions belonging to the same camp to obtain the predicted first camp strength value (PredictFactionScore of the first camp) T ) and the predicted second camp strength value (PredictFactionScore of the second camp) T ) for the T - th time period.

[0262] 7. Determine the first ratio of the predicted first camp strength value to the predicted second camp strength value.

[0263] 8. Add up the second server strength values corresponding to the server regions belonging to the same camp to obtain the actual first camp strength value (FactionMilitaryExploit of the first camp) T ) and the actual second camp strength value (FactionMiltaryExploit of the second camp) T ) for the T - th time period.

[0264] 9. Determine the second ratio of the actual first camp strength value to the actual second camp strength value.

[0265] 10. Determine the error function; specifically, take the difference between the first ratio and the second ratio, and use the absolute value of this difference as the error function. Use this error function as the objective function for optimizing the hyperparameters P1, P2, P3, P4, P5, P6 in the strength calculation formula in step 11. The error function is expressed as:

[0266]

[0267] 11. Solve for the hyperparameters in the strength calculation formula using an optimization algorithm. The specific steps are as follows:

[0268] 11.1) Initialization

[0269] Randomly set N groups of hyperparameters within the value ranges of each hyperparameter P1, P2, P3, P4, P5, P6 in the strength calculation formula, denoted as: According to the expression of the error function, determine the errors corresponding to each of the N groups of hyperparameters, denoted as: loss = {Y1, Y2, …, Y N}. Based on the N groups of hyperparameters and their corresponding errors, obtain the initialization dataset, denoted as D = {(x1, Y1), (Y2, Y2), …, (x N , Y N )},

[0270] 11.2) Gaussian process training and prediction

[0271] Since the expression of the above error function is relatively complex and there are no first-order and second-order derivatives, it is impossible to directly use algorithms such as gradient descent to solve for the extreme value. Therefore, in this embodiment, an initialized dataset D is selected for training. It describes the covariance between sample points x and x′, and is used to characterize the similarity between any two sample points in the initialized dataset D.

[0272] The Gaussian process is a Bayesian method based on regression problems. For each sample point x, there is a corresponding Gaussian distribution, and for all sample points {x1, x2, …, x N} in the initialized dataset D, there exists a joint Gaussian distribution, and the mean of this joint Gaussian distribution is 0, and the covariance is expressed as:

[0273]

[0274] This joint Gaussian distribution is the Gaussian process model f(x 1:N ) trained using the initialized dataset D. For a new sample point After adding this sample point x N+1 to the initialized dataset D, the Gaussian process model f(x 1：N+1 ) follows an N + 1-dimensional Gaussian distribution. Then, according to the Gaussian process models f(x 1：N ) and f(x 1：N+1 ), by partitioning the mean vector and the covariance matrix, the Gaussian process model can be expressed as:

[0275]

[0276] In the above expression of the Gaussian process model, the covariance matrix K can be calculated based on the initialized dataset D and k(x, x′), and the covariance k = [k(x N+1 , x1), k(x N+1 , x2),..., k(x N+1 , x N )].

[0277] Since the conditional distribution of the multi-dimensional Gaussian distribution is still a Gaussian distribution, it is possible to calculate the conditional distribution that f(x 1：N ) follows when f(x N+1 ) is known. According to the properties of the multi-dimensional Gaussian distribution, this conditional distribution follows a one-dimensional Gaussian distribution, and the calculation formula of the conditional distribution is: f(x N+1 )|f(x 1：N ) ∼ GP(μ, σ 2 ).

[0278] Based on the block scheme of the mean vector and covariance matrix and the calculation formula of the conditional distribution, the mean and variance of this conditional distribution can be obtained, which are respectively expressed as: μ = k T K -1 f(x 1：N ); σ 2 = k(x N+1 , x N+1 ) - k T K -1 k. Therefore, for a new sample point (a new set of hyperparameters) x N+1 , it can be considered that the error function follows a one-dimensional Gaussian distribution with a mean of μ and a variance of σ N+1 on x 2 .

[0279] 11_3) Determine the new sample point x N+1

[0280] According to 11.2), given the initial dataset, the corresponding Gaussian process model can be fitted using this initial dataset. On this basis, for any given new sample point, based on the properties of the multi-dimensional Gaussian distribution, the mean and variance of the one-dimensional Gaussian distribution corresponding to the Gaussian process model at this sample point can be calculated. On this basis, the UCB acquisition strategy can be used to find the next sample point x N+1 that minimizes the error function (the value is the minimum). The calculation formula is: x N+1 = argmin x UCB = μ(x) - wσ(x). Where w is the weight used to balance the expected mean and variance. It can be seen from the above calculation formula that the UCB strategy tends to select regions with low mean and large standard deviation. Through the UCB acquisition strategy, the next new sample point x N+1 that minimizes the error function can be calculated, and the error Y N+1 corresponding to the sample point x N+1 can be obtained using the calculation formula of the error function. Add (x N+1 , Y N+1 ) to the initial dataset D.

[0281] 11.4) Iterative update

[0282] According to the updated initial dataset, update the Gaussian process model in the manner of 11.2) until the iteration stop condition is satisfied. The iteration stop condition includes at least one of the following: reaching the preset number of iterations, the degree of change between the error functions corresponding to the Gaussian process models before and after iteration is less than the preset error threshold.

[0283] 11.5) Generate the optimal hyperparameter combination

[0284] After reaching the iteration stop condition, a sample point that minimizes the error function can be obtained. This sample point corresponds to a set of hyperparameters, denoted as

[0285] Figure 9 FIG. shows a schematic diagram of optimizing an error function provided by an exemplary embodiment of the present application. The processing flow of step 11 above can be briefly described as follows: 1. Start; 2. Set the error function; 3. Determine the initial dataset: by determining the sample point x and determining the error Y according to the error function, combine them into the initial dataset D; 4. Update the Gaussian process model; 5. Determine the next new sample point, and each sample point is a set of hyperparameters: the i-th sample point is denoted as x i = argmax(μ(x|D)); 6. Add the new sample point x i and its corresponding error Y i to the initial dataset to iteratively update the initial dataset: D ← D ∪ {x i , Y i}, i ← i + 1; 7. Determine whether the number of updates t is greater than the preset number of iterations T. Otherwise, re-execute 4. If yes, execute 8; 8. Return the optimization result; 9. End.

[0286] 12. Determine the x in 11.5) best as the hyperparameter (optimal hyperparameter) in the strength calculation formula.

[0287] It should be noted that the above optimization algorithm is implemented based on the Bayesian optimization algorithm. In actual applications, other optimization algorithms can be used according to actual technical needs, such as: genetic algorithm, simulated annealing algorithm. The above first part can be executed only once, or can be executed periodically. Optionally, when the computer device executes the data processing method each time, it can execute the first part, the second part, and the third part in sequence. Or, the computer device can also pre-execute the first part and store the hyperparameters. When executing the data processing method each time, only the second part and the third part are executed in sequence. This embodiment does not make a limitation on this.

[0288] · The second part: Estimate the strength value of the game object and the server strength value

[0289] 13. Obtain the object data of the game object in the T-th time period: historical recharge and payment value, the number of dead and wounded soldiers in the current time period, and the average strength value of the confrontation lineup used.

[0290] 14. Obtain the strength calculation formula.

[0291] 15. Substitute the object data of the game object in the T-th time period and the optimal hyperparameters into the strength calculation formula to obtain the strength value corresponding to the game object in the T-th time period.

[0292] 16. Add up the strength values corresponding to all game objects belonging to the same server area to obtain the server area strength values corresponding to each server area in the T-th time period.

[0293] · Part Three: Server Area Merging

[0294] 17. The server area merging strategy is as follows: Divide the server area strength values into intervals to obtain 5 intervals. The server area strength values corresponding to the 5 intervals are represented as follows: Very strong in strength: [10000, +∞), Relatively strong in strength: [60000, 100000), Medium in strength: [40000, 60000), Relatively weak in strength: [20000, 40000), Very weak in strength: (0, 20000]. Randomly select one server area from each interval. The first server area set composed of these 5 server areas is recorded as the first camp. According to the server area strength value and the number of game objects of each server area in the first camp, then select 4 to 6 server areas from these 5 intervals. These server areas form the second server area set and constitute the second camp, and it is necessary to ensure that the server area strength values and the number of game objects of the first camp and the second camp are as close as possible. After the first camp and the second camp are matched, they can be used as the opposing camps in the (T + 1)-th time period. For the remaining server areas, repeat the selection operation to finally complete the server area merging in the (T + 1)-th time period.

[0295] 18. Generate the server area merging list in the (T + 1)-th time period. Before the (T + 1)-th time period starts, inform the game objects through the game promotion channels.

[0296] · Beneficial Effects

[0297] Figure 10 Shows a schematic diagram of the comparison of beneficial effects provided by an exemplary embodiment of the present application. The calculation efficiency of the hyperparameters of the method in the embodiment of the present application is higher and the solution effect is better. As can be seen from Figure 10 it, compared with the related technology method, the time length for solving the hyperparameters of the method in the embodiment of the present application is shortened by 539 minutes, with a relative reduction of 96.3%; the method in the embodiment of the present application can accurately depict the strength of game objects and the server area strength, making the error between the predicted camp strength and the actual camp strength corresponding to the current time period smaller, and the cumulative error of the relative ratio of the predicted camp strength and the actual camp strength is relatively reduced by 23.6%.

[0298] Figure 11 Shows a schematic diagram of the comparison of core indicators provided by an exemplary embodiment of the present application. Use the method in the embodiment of the present application and the related technology method to determine the server area merging results respectively, and count the two indicators of the average daily active number and the average daily killing number in the (T + 1)-th time period. As can be seen from Figure 11It can be seen that, compared with the related art methods, the average daily active quantity using the method of the embodiment of the present application is relatively increased by 11.6%, and the average daily enemy-killing quantity is relatively increased by 20.2%.

[0299] The above embodiments effectively verify the positive value brought by the method of the embodiment of the present application to the game. The method of the embodiment of the present application can more accurately and efficiently evaluate the strength of game objects and the strength of server regions, so that game planners can design more reasonable server region merging strategies for new time periods and new seasons, improve the effect of camp matching in the game in new time periods and new seasons, and promote the improvement of the commercial value of the game.

[0300] Figure 12 The block diagram of a data processing device 800 provided by an exemplary embodiment of the present application is shown. The data processing device 800 includes:

[0301] An obtaining module 810, configured to obtain object data of game objects in the T-th time period; the T is greater than or equal to 1;

[0302] The obtaining module 810 is further configured to obtain hyperparameters corresponding to a strength calculation formula; the strength calculation formula is used to calculate the strength value corresponding to the game object, the hyperparameters are obtained by optimizing and solving an error function, and the error function is determined based on the object data of the game objects in the (T - 1)-th time period and the T-th time period, and is used to balance the error between the estimated strength and the actual strength of each server region where the game objects are located;

[0303] A processing module 820, configured to substitute the object data of the game object and the hyperparameters into the strength calculation formula to obtain the strength value corresponding to the game object;

[0304] The processing module 820 is further configured to accumulate the strength values corresponding to the game objects belonging to the same server region to obtain the server region strength values corresponding to each server region;

[0305] A merging module 830, configured to merge each server region based on the server region strength values to obtain a server region merging result in the (T + 1)-th time period.

[0306] In some embodiments, the obtaining module 810 is configured to:

[0307] Obtain the object data of the game objects in the (T - 1)-th time period and the T-th time period;

[0308] Accumulate based on the object data in the (T - 1)-th time period to obtain the estimated camp strength value in the T-th time period; the estimated camp strength value is used to represent the estimated strength of the camp corresponding to the server area where each of the game objects is located in the T-th time period, and the estimated camp strength value is calculated by substituting the object data in the (T - 1)-th time period and the initial hyperparameters into the strength calculation formula;

[0309] Also, perform statistics based on the object data in the T-th time period to obtain the actual camp strength value in the T-th time period; the actual camp strength value is used to represent the actual strength of the camp corresponding to the server area where each of the game objects is located in the T-th time period, and the actual camp strength value is calculated by substituting the object data in the T-th time period and the initial hyperparameters into the strength calculation formula;

[0310] Determine the error function between the estimated camp strength value and the actual camp strength value;

[0311] Taking the minimization of the error function as the optimization goal, solve to obtain the hyperparameters corresponding to the strength calculation formula.

[0312] In some embodiments, the obtaining module 810 is configured to:

[0313] Obtain an initialization data set; the initialization data set includes N sample data, the N sample data are determined based on the hyperparameters to be solved in the strength calculation formula and the error function, and N is greater than or equal to 1;

[0314] Use the initialization data set to fit to obtain a first Gaussian process model; the first Gaussian process model includes a Gaussian kernel function, and the Gaussian kernel function is used to describe the similarity between any two sample data in the initialization data set;

[0315] Iteratively update the initialization data set, and use the iteratively updated initialization data set to update the first Gaussian process model until the iteration stop condition is satisfied to obtain a second Gaussian process model; the iteratively updated initialization data set includes N + i sample data, where i is the update times of the initialization data set, and i is greater than or equal to 1;

[0316] Based on the second Gaussian process model, determine the minimum value of the error function;

[0317] Determine the hyperparameters corresponding to the strength calculation formula as the hyperparameters corresponding to the minimum value of the error function.

[0318] In some embodiments, the obtaining module 810 is configured to:

[0319] Randomly set N groups of hyperparameters within the value range of the hyperparameters in the strength calculation formula;

[0320] Based on the error function, determine the error corresponding to each of the N groups of hyperparameters;

[0321] Based on the N groups of hyperparameters and their respective corresponding errors, obtain the initialization dataset;

[0322] Among them, each sample data in the initialization dataset includes a sample point and the error corresponding to the sample point. The error is determined based on the error function, and each sample point is a group of hyperparameters in the N groups of hyperparameters.

[0323] In some embodiments, the obtaining module 810 is configured to:

[0324] Determine the joint Gaussian distribution corresponding to the N sample data in the initialization dataset;

[0325] Based on the joint Gaussian distribution, fit to obtain the first Gaussian process model.

[0326] In some embodiments, the obtaining module 810 is configured to:

[0327] Determine the update times i for updating the initialization dataset;

[0328] Determine the (N + i)-th sample point;

[0329] Based on the error function, determine the error corresponding to the (N + i)-th sample point;

[0330] Based on the (N + i)-th sample point and the corresponding error, determine the (N + i)-th sample data;

[0331] Add the (N + i)-th sample data to the initialization dataset to iteratively update the initialization dataset, and use the iteratively updated initialization dataset to update the first Gaussian process model;

[0332] Update i to i + 1, and repeat the step of determining the (N + i)-th sample point until the iteration stop condition is satisfied to obtain the second Gaussian process model.

[0333] In some embodiments, the iteration stop condition includes at least one of the following: reaching a preset number of iterations, and the degree of change between the error functions corresponding to the second Gaussian process model and the first Gaussian process model is less than a preset error threshold.

[0334] In some embodiments, the obtaining module 810 is configured to:

[0335] Determine the mean and variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point;

[0336] Based on the mean and the variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point, determine the (N + i)-th sample point that minimizes the error function.

[0337] In some embodiments, the obtaining module 810 is configured to:

[0338] Perform block partitioning based on the N-dimensional Gaussian distribution that the first Gaussian process model follows on the initialization dataset and the one-dimensional Gaussian distribution that the first Gaussian process model follows at the (N + i)-th sample point, and determine the covariance matrix and the mean vector; the covariance matrix is calculated based on the initialization dataset and the Gaussian kernel function;

[0339] Determine the one-dimensional Gaussian distribution that the first Gaussian process model follows at the (N + i)-th sample point given that the first Gaussian process model follows an N-dimensional Gaussian distribution on the initialization dataset;

[0340] Based on the covariance matrix, the mean vector, and the one-dimensional Gaussian distribution, determine the mean and the variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point.

[0341] In some embodiments, the camp includes a first camp and a second camp, the estimated camp strength value includes an estimated first camp strength value and an estimated second camp strength value, and the actual camp strength value includes an actual first camp strength value and an actual second camp strength value.

[0342] In some embodiments, the obtaining module 810 is configured to:

[0343] Determine a first ratio of the estimated first camp strength value to the estimated second camp strength value;

[0344] Determine a second ratio of the actual first camp strength value to the actual second camp strength value;

[0345] Subtract the second ratio from the first ratio to obtain the error function.

[0346] In some embodiments, the obtaining module 810 is configured to:

[0347] Substitute the object data of the game object in the (T - 1)-th time period and the initial hyperparameters into the strength calculation formula to obtain the first strength value of the game object in the (T - 1)-th time period;

[0348] Accumulate the first strength values corresponding to the game objects belonging to the same server area to obtain the first server area strength values respectively corresponding to each server area in the (T - 1)-th time period;

[0349] Accumulate the first server area strength values corresponding to the server areas belonging to the same camp to obtain the estimated first camp strength value and the estimated second camp strength value in the T-th time period.

[0350] In some embodiments, the obtaining module 810 is configured to:

[0351] Substitute the object data of the game object and the initial hyperparameters in the strength calculation formula in the T-th time period to obtain the second strength value of the game object in the T-th time period;

[0352] Accumulate the second strength values corresponding to the game objects belonging to the same server area to obtain the second server area strength values respectively corresponding to each server area in the T-th time period;

[0353] Accumulate the second server area strength values corresponding to the server areas belonging to the same camp to obtain the actual first camp strength value and the actual second camp strength value in the T-th time period.

[0354] In some embodiments, the merging module 830 is configured to:

[0355] Perform interval division on the server area strength values to obtain at least two intervals;

[0356] Randomly select at least one server area from at least one of the at least two intervals to obtain a first server area set;

[0357] Based on the server area strength values and the number of game objects corresponding to the first server area set, select at least one server area from at least one of the at least two intervals to obtain a second server area set;

[0358] Determine the first server area set and the second server area set as the server area merging result in the (T + 1)-th time period;

[0359] Wherein, the difference in the server area strength values between the second server area set and the first server area set is less than a first threshold, and the difference in the number of game objects between the second server area set and the first server area set is less than a second threshold; the first server area set and the second server area set belong to different camps respectively.

[0360] In some embodiments, the hyperparameters in the strength calculation formula include at least one of: a first hyperparameter, a second hyperparameter, a third hyperparameter, a fourth hyperparameter, a fifth hyperparameter, and a sixth hyperparameter;

[0361] The described strength calculation formula includes at least one of: a first strength calculation formula corresponding to the historical recharge and payment value, a second strength calculation formula corresponding to the number of dead and wounded soldiers in the current time period, and a third strength calculation formula corresponding to the average strength value of the confrontation usage lineup;

[0362] The first strength calculation formula is: the sum value between the product of the historical recharge and payment value and the first hyperparameter and the first preset value, where the sum value is less than or equal to the second hyperparameter; or, the second hyperparameter, where the sum value is less than or equal to the second hyperparameter;

[0363] The second strength calculation formula is: the second preset value, where the number of dead and wounded soldiers in the current time period is less than or equal to the third hyperparameter; or, the value corresponding to taking the difference between the number of dead and wounded soldiers in the current time period minus the third hyperparameter as the base and the fourth hyperparameter as the exponent, where the number of dead and wounded soldiers in the current time period is greater than the third hyperparameter;

[0364] The third strength calculation formula is: the value corresponding to taking the sum value of the third preset value and the average strength value of the confrontation usage lineup as the base and the fifth hyperparameter as the exponent, and subtracting the sixth hyperparameter.

[0365] In some embodiments, the server region corresponding server region strength value is the sum of the strength values corresponding to the game objects within the server region, and the strength value is the product of a first sub-strength value, a second sub-strength value, and a third sub-strength value;

[0366] Wherein, the first sub-strength value is calculated based on the first strength calculation formula, the second sub-strength value is calculated based on the second strength calculation formula, and the third sub-strength value is calculated based on the third strength calculation formula.

[0367] It should be noted that the specific limitations in the above-described embodiments of one or more data processing devices 800 can be referred to the limitations on the data processing method in the foregoing text, and will not be elaborated herein. All or part of the modules of the above device can be implemented through software, hardware, and their combination. Each module can be embedded in the processor of the computer device in a hardware form or be independent of it, or can be stored in the memory of the computer device in a software form, so that the processor can call and execute the operations corresponding to each module.

[0368] An embodiment of the present application further provides a computer device, which includes: a processor and a memory, and a computer program is stored in the memory; the processor is configured to execute the computer program in the memory to implement the data processing method provided by the above method embodiments.

[0369] Exemplarily, Figure 13It is a structural block diagram of a computer device 1000 provided by an exemplary embodiment of the present application. Optionally, the computer device 1000 is a server 1000.

[0370] Generally, the server 1000 includes: a processor 1001 and a memory 1002.

[0371] The processor 1001 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 1001 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), or programmable logic array (PLA). The processor 1001 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the wake state, also known as the central processing unit (CPU); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 1001 may be integrated with a graphics processing unit (GPU), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 1001 may further include an artificial intelligence (AI) processor, and the AI processor is used to process computational operations related to machine learning.

[0372] The memory 1002 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 1002 may further include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In some embodiments, the non-transitory computer-readable storage media in the memory 1002 is used to store at least one instruction, and the at least one instruction is used to be executed by the processor 1001 to implement the data processing method provided by the method embodiment of the present application.

[0373] In some embodiments, the server 1000 may further optionally include: an input interface 1003 and an output interface 1004. The processor 1001, the memory 1002, the input interface 1003, and the output interface 1004 may be connected through a bus or signal lines. Each peripheral device may be connected to the input interface 1003 and the output interface 1004 through a bus, signal lines, or a circuit board. The input interface 1003 and the output interface 1004 may be used to connect at least one peripheral device related to input / output (I / O) to the processor 1001 and the memory 1002. In some embodiments, the processor 1001, the memory 1002, the input interface 1003, and the output interface 1004 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 1001, the memory 1002, the input interface 1003, and the output interface 1004 may be implemented on a separate chip or circuit board, and the embodiments of the present application do not limit this.

[0374] Those skilled in the art can understand that Figure 13 the structure shown in does not constitute a limitation on the computer device 1000, and it may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component layout.

[0375] In an exemplary embodiment, the present application provides a chip, which includes a programmable logic circuit and / or program instructions, and when the chip runs on a computer device, it is used to implement the data processing method provided in the above method embodiment.

[0376] The present application provides a computer-readable storage medium, which stores a computer program, and the computer program is loaded and executed by a processor to implement the data processing method provided in the above method embodiment.

[0377] The present application provides a computer program product or a computer program, which includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the processor of the computer device is loaded and executed to implement the data processing method provided in the above method embodiment.

[0378] The serial numbers of the above embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments.

[0379] Those of ordinary skill in the art can understand that all or part of the steps to implement the above embodiments can be completed by hardware, or can be completed by instructing relevant hardware through a program. The program can be stored in a computer-readable storage medium. The computer-readable storage medium mentioned above can be a read-only memory, a magnetic disk, an optical disc, or the like.

[0380] Those skilled in the art should be able to realize that in one or more of the above examples, the functions described in the embodiments of the present application can be implemented by hardware, software, firmware, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. The computer-readable medium includes computer storage media and communication media, where the communication media includes any medium that facilitates the transfer of a computer program from one place to another. The storage media can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0381] The foregoing are only alternative embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A data processing method, characterized in that, The method includes: Obtaining object data of game objects in the T-th time period; T is greater than or equal to 1; Obtaining hyperparameters corresponding to the strength calculation formula; the strength calculation formula is used to calculate the strength value corresponding to the game object, and the hyperparameters are obtained by optimizing and solving the error function, and the error function is determined based on the object data of the game objects in the (T - 1)-th time period and the T-th time period, and is used to balance the error between the estimated strength and the actual strength of each game object in the server area; Substituting the object data of the game object and the hyperparameters into the strength calculation formula to obtain the strength value corresponding to the game object; Accumulating the strength values corresponding to the game objects belonging to the same server area to obtain the server area strength values respectively corresponding to each server area; Merging each server area based on the server area strength values to obtain the server area merging result in the (T + 1)-th time period.

2. The method according to claim 1, wherein The obtaining of the hyperparameters corresponding to the strength calculation formula includes: Obtaining the object data of the game objects in the (T - 1)-th time period and the T-th time period; Accumulating based on the object data in the (T - 1)-th time period to obtain the estimated camp strength value in the T-th time period; the estimated camp strength value is used to represent the estimated strength of the camp corresponding to each server area of the game objects in the T-th time period, and the estimated camp strength value is calculated by substituting the object data in the (T - 1)-th time period and the initial hyperparameters into the strength calculation formula; And, performing statistics based on the object data in the T-th time period to obtain the actual camp strength value in the T-th time period; the actual camp strength value is used to represent the actual strength of the camp corresponding to each server area of the game objects in the T-th time period, and the actual camp strength value is calculated by substituting the object data in the T-th time period and the initial hyperparameters into the strength calculation formula; Determining the error function between the estimated camp strength value and the actual camp strength value; Taking minimizing the error function as the optimization goal to solve and obtain the hyperparameters corresponding to the strength calculation formula.

3. The method according to claim 2, wherein The taking minimizing the error function as the optimization goal to solve and obtain the hyperparameters corresponding to the strength calculation formula includes: Obtaining an initialization data set; the initialization data set includes N sample data, and the N sample data are determined based on the hyperparameters to be solved in the strength calculation formula and the error function, and N is greater than or equal to 1; Using the initialization data set to fit and obtain a first Gaussian process model; the first Gaussian process model includes a Gaussian kernel function, and the Gaussian kernel function is used to describe the similarity between any two sample data in the initialization data set; Iteratively updating the initialization data set, and using the iteratively updated initialization data set to update the first Gaussian process model until the iteration stop condition is satisfied to obtain a second Gaussian process model; the iteratively updated initialization data set includes N + i sample data, where i is the update times of the initialization data set, and i is greater than or equal to 1; Based on the second Gaussian process model, determine the minimum value of the error function; Determine the hyperparameters corresponding to the strength calculation formula as the hyperparameters corresponding to the minimum value of the error function.

4. The method according to claim 3, wherein The obtaining of the initial dataset includes: Randomly set N groups of hyperparameters within the value range of the hyperparameters in the strength calculation formula; Based on the error function, determine the errors corresponding to each of the N groups of hyperparameters; Based on the N groups of hyperparameters and their corresponding errors, obtain the initial dataset; Wherein, each sample data in the initial dataset respectively includes a sample point and the error corresponding to the sample point, the error is determined based on the error function, and each sample point is a group of hyperparameters among the N groups of hyperparameters.

5. The method according to claim 3, wherein The fitting of the first Gaussian process model using the initial dataset includes: Determine the joint Gaussian distribution corresponding to the N sample data in the initial dataset; Based on the joint Gaussian distribution, fit the first Gaussian process model.

6. The method according to claim 3, wherein The iterative updating of the initial dataset and the use of the iteratively updated initial dataset to update the first Gaussian process model until the iteration stop condition is met to obtain the second Gaussian process model includes: Determine the update times i for updating the initial dataset; Determine the (N + i)-th sample point; Based on the error function, determine the error corresponding to the (N + i)-th sample point; Based on the (N + i)-th sample point and the corresponding error, determine the (N + i)-th sample data; Add the (N + i)-th sample data to the initial dataset to iteratively update the initial dataset, and use the iteratively updated initial dataset to update the first Gaussian process model; Update i to i + 1, and repeat the step of determining the (N + i)-th sample point until the iteration stop condition is met to obtain the second Gaussian process model.

7. The method according to claim 6, characterized in that, The iteration stop condition includes at least one of the following: reaching a preset number of iterations, the degree of change between the error functions corresponding to the second Gaussian process model and the first Gaussian process model is less than a preset error threshold.

8. The method according to claim 6, wherein The determining of the (N + i)-th sample point includes: Determine the mean and variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point; Based on the mean and the variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point, determine the (N + i)-th sample point that minimizes the error function.

9. The method according to claim 8, wherein The determining of the mean and variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point includes: Based on the N-dimensional Gaussian distribution that the first Gaussian process model follows on the initial dataset and the one-dimensional Gaussian distribution that the first Gaussian process model follows on the (N + i)-th sample point, perform block partitioning to determine the covariance matrix and the mean vector; the covariance matrix is calculated based on the initial dataset and the Gaussian kernel function; Determine the one-dimensional Gaussian distribution that the first Gaussian process model follows at the (N + i)-th sample point given that the first Gaussian process model follows an N-dimensional Gaussian distribution on the initialization dataset; Based on the covariance matrix, the mean vector, and the one-dimensional Gaussian distribution, determine the mean and the variance of the one-dimensional Gaussian distribution corresponding to the (N + i)-th sample point.

10. The method according to any one of claims 2 to 9, characterized in that The camps include a first camp and a second camp. The predicted camp strength values include the predicted first camp strength value and the predicted second camp strength value. The actual camp strength values include the actual first camp strength value and the actual second camp strength value; Determining the error function between the predicted camp strength values and the actual camp strength values includes: Determine a first ratio of the predicted first camp strength value to the predicted second camp strength value; Determine a second ratio of the actual first camp strength value to the actual second camp strength value; Subtract the second ratio from the first ratio to obtain the error function.

11. The method according to claim 10, wherein The accumulating the object data in the (T - 1)-th time period to obtain the predicted camp strength value in the T-th time period includes: Substitute the object data of the game object in the (T - 1)-th time period and the initial hyperparameters into the strength calculation formula to obtain the first strength value of the game object in the (T - 1)-th time period; Accumulate the first strength values corresponding to the game objects belonging to the same server area to obtain the first server area strength values corresponding to the respective server areas in the (T - 1)-th time period; Accumulate the first server area strength values corresponding to the server areas belonging to the same camp to obtain the predicted first camp strength value and the predicted second camp strength value in the T-th time period.

12. The method according to claim 10, wherein The statistics based on the object data in the T-th time period to obtain the actual camp strength value in the T-th time period includes: Substitute the object data of the game object in the T-th time period and the initial hyperparameters into the strength calculation formula to obtain the second strength value of the game object in the T-th time period; Accumulate the second strength values corresponding to the game objects belonging to the same server area to obtain the second server area strength values corresponding to the respective server areas in the T-th time period; Accumulate the second server area strength values corresponding to the server areas belonging to the same camp to obtain the actual first camp strength value and the actual second camp strength value in the T-th time period.

13. The method according to any one of claims 1 to 12, characterized in that The merging of the respective server areas based on the server area strength values to obtain the server area merging result in the (T + 1)-th time period includes: Perform interval partitioning on the server area strength values to obtain at least two intervals; Randomly select at least one server area from at least one of the at least two intervals to obtain a first server area set; Based on the server area strength values and the number of game objects corresponding to the first server area set, select at least one server area from at least one of the at least two intervals to obtain a second server area set; Determine the first server area set and the second server area set as the server area merging result in the (T + 1)-th time period; Among them, the difference in the server strength value between the second server set and the first server set is less than a first threshold, and the difference in the number of game objects between the second server set and the first server set is less than a second threshold; the first server set and the second server set belong to different camps respectively.

14. The method according to any one of claims 1 to 12, characterized in that, The hyperparameters in the strength calculation formula include at least one of: a first hyperparameter, a second hyperparameter, a third hyperparameter, a fourth hyperparameter, a fifth hyperparameter, and a sixth hyperparameter; The strength calculation formula includes at least one of: a first strength calculation formula corresponding to the historical recharge and payment value, a second strength calculation formula corresponding to the number of dead and wounded soldiers in the current time period, and a third strength calculation formula corresponding to the average strength value of the confrontation lineup; The first strength calculation formula is: the sum value between the product of the historical recharge and payment value and the first hyperparameter and a first preset value, where the sum value is less than or equal to the second hyperparameter; or, the second hyperparameter, where the sum value is less than or equal to the second hyperparameter; The second strength calculation formula is: a second preset value, where the number of dead and wounded soldiers in the current time period is less than or equal to the third hyperparameter; or, the value corresponding to the base of the difference between the number of dead and wounded soldiers in the current time period minus the third hyperparameter and the exponent of the fourth hyperparameter, where the number of dead and wounded soldiers in the current time period is greater than the third hyperparameter; The third strength calculation formula is: the value corresponding to the base of the sum value of a third preset value and the average strength value of the confrontation lineup and the exponent of the fifth hyperparameter, minus the sixth hyperparameter.

15. The method according to claim 14, wherein The server strength value corresponding to the server is the sum of the strength values corresponding to the game objects in the server, and the strength value is the product of a first sub-strength value, a second sub-strength value, and a third sub-strength value; Among them, the first sub-strength value is calculated based on the first strength calculation formula, the second sub-strength value is calculated based on the second strength calculation formula, and the third sub-strength value is calculated based on the third strength calculation formula.

16. A data processing device, characterized in that, The device includes: An acquisition module, configured to acquire the object data of the game objects in the T-th time period; the T is greater than or equal to 1; The acquisition module is further configured to acquire the hyperparameters corresponding to the strength calculation formula; the strength calculation formula is used to calculate the strength value corresponding to the game object, and the hyperparameters are obtained by optimizing and solving an error function, and the error function is determined based on the object data of the game objects in the (T - 1)-th time period and the T-th time period, and is used to balance the error between the estimated strength and the actual strength of each server where the game objects are located; A processing module, configured to substitute the object data of the game object and the hyperparameters into the strength calculation formula to obtain the strength value corresponding to the game object; The processing module is further configured to accumulate the strength values corresponding to the game objects belonging to the same server to obtain the server strength value corresponding to each server; A merging module, configured to merge each of the server regions based on the server region strength values to obtain a server region merging result for the (T + 1)-th time period.

17. A computer device, characterized in that, The computer device includes: a processor and a memory. The memory stores a computer program, and the computer program is loaded and executed by the processor to implement the data processing method according to any one of claims 1 to 15.

18. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and the computer program is loaded and executed by a processor to implement the data processing method according to any one of claims 1 to 15.

19. A computer program product, characterized in that, The computer program product includes computer instructions. The computer instructions are stored in a computer-readable storage medium, and a processor obtains the computer instructions from the computer-readable storage medium, so that the processor loads and executes to implement the data processing method according to any one of claims 1 to 15.

Images