CN112053198A

CN112053198A - Game data processing method, device, equipment and medium

Info

Publication number: CN112053198A
Application number: CN202010992976.7A
Authority: CN
Inventors: 马捷思; 张涵宇; 谢思发
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2020-09-21
Filing date: 2020-09-21
Publication date: 2020-12-08
Anticipated expiration: 2040-09-21
Also published as: CN112053198B

Abstract

The invention discloses a game data processing method, a game data processing device, game data processing equipment and a game data processing medium. The method comprises the following steps: obtaining target characteristic data from a game data group of a game object to be predicted, wherein the game data group comprises at least one of the following: game base data, game community data, and game derivative data; inputting the target characteristic data into a game sales prediction model; obtaining a sales forecasting result of the game object to be forecasted by utilizing the game sales forecasting model; the game sales prediction model is determined by machine learning training based on a plurality of sample characteristic data, and each sample characteristic data carries sales information labels. The invention can improve the adaptability of the sales data prediction of the game object to be predicted and can greatly improve the prediction accuracy of the sales data.

Description

Game data processing method, device, equipment and medium

Technical Field

The present application relates to the field of internet communication technologies, and in particular, to a method, an apparatus, a device, and a medium for processing game data.

Background

Electronic games refer to all interactive games that run on an electronic platform. The method is divided into five types according to different media: host games (or home games, tv games), palm games, computer games, arcade games and mobile games (mainly mobile games). The perfect electronic game appeared at the end of the 20 th century, changes the behavior mode of human playing games and the definition of a game word, and belongs to a cultural activity born along with the development of science and technology. The profit generated by the payment of the electronic game can conceptually judge the financial status, the market preference, and the like of the relevant company, and therefore, the profit determination is also an important link for the third party.

In the related art, data companies may conduct game sales statistics through physical retailers covered under lines. For the retailers who contract to share the data with the retailers, the bar codes corresponding to each selling behavior of the retailers are tracked to carry out game sales statistics. The sales data obtained in this way are limited to off-line channels, even to off-line channels in a specific area, and lack of statistics on sales of on-line channels, so that the data availability is poor and the accuracy is poor. The data company can also carry out game sales statistics based on the public data of the users in the relevant game development platform, and such sales statistics depend on the user protocol of the relevant game development platform (such as whether to disclose the user data and the degree of disclosing the user data), so that the accurate sales data can not be provided. Therefore, there is a need to provide a more accurate and efficient determination scheme for sales data.

Disclosure of Invention

In order to solve the problems of low accuracy and the like when the prior art is applied to determining sales data of games, the application provides a game data processing method, a device, equipment and a medium, wherein the game data processing method comprises the following steps:

according to a first aspect of the present application, there is provided a game data processing method, the method including:

obtaining target characteristic data from a game data group of a game object to be predicted, wherein the game data group comprises at least one of the following: game base data, game community data, and game derivative data;

inputting the target characteristic data into a game sales prediction model;

obtaining a sales forecasting result of the game object to be forecasted by utilizing the game sales forecasting model;

the game sales prediction model is determined by machine learning training based on a plurality of sample characteristic data, and each sample characteristic data carries sales information labels.

According to a second aspect of the present application, there is provided a game data processing apparatus, the apparatus comprising:

a data acquisition module: the method comprises the steps of obtaining target characteristic data from a game data group of a game object to be predicted, wherein the game data group comprises at least one of the following: game base data, game community data, and game derivative data;

a data input module: for inputting the target characteristic data into a game sales prediction model;

a result prediction module: the game sales prediction model is used for obtaining the sales prediction result of the game object to be predicted;

According to a third aspect of the present application, there is provided an electronic device comprising a processor and a memory, wherein at least one instruction or at least one program is stored in the memory, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the game data processing method according to the first aspect.

According to a fourth aspect of the present application, there is provided a computer-readable storage medium having at least one instruction or at least one program stored therein, the at least one instruction or the at least one program being loaded and executed by a processor to implement the game data processing method according to the first aspect.

According to a fifth aspect of the present application, there is provided a computer program product or computer program comprising computer instructions 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 computer device executes the game data processing method according to the first aspect.

The game data processing method, the game data processing device, the game data processing equipment and the game data processing medium have the following technical effects:

the method adopts a machine learning method to train a game sales prediction model, and utilizes the game sales prediction model to output a sales prediction result of a game object to be predicted based on input target characteristic data. The game sales prediction model obtained through training has high generalization capability, the adaptability of the sales data prediction of the game object to be predicted can be improved, and the prediction accuracy of the sales data can be greatly improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.

FIG. 1 is a schematic diagram of an application environment provided by an embodiment of the invention;

FIG. 2 is a flow chart of a game data processing method according to an embodiment of the present invention;

FIG. 3 is a schematic flow chart of a game sales prediction model obtained by training according to an embodiment of the present invention;

FIG. 4 is a schematic flow chart of a method for training a plurality of game sales prediction models according to an embodiment of the present invention;

FIG. 5 is a diagram illustrating an application scenario of a game sales prediction model according to an embodiment of the present invention;

FIG. 6 is a flow chart illustrating a method for processing game data according to an embodiment of the present invention;

FIG. 7 is an interface diagram illustrating a game sales space provided by an embodiment of the present invention;

FIG. 8 is a block diagram of a game data processing apparatus according to an embodiment of the present invention;

fig. 9 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.

It should be noted that the terms "comprises" and "comprising," and any variations thereof, in the description and claims of the present invention and the above-described drawings, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or server that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

Before further detailed description of the embodiments of the present invention, terms and expressions mentioned in the embodiments of the present invention are explained, and the terms and expressions mentioned in the embodiments of the present invention are applied to the following explanations.

Buy-break (B2P, Buy out game), which is distinguished between a time charge (P2P) and a free game item charge (F2P), players can experience the complete content of the game after purchase.

PS4(PlayStation 4), a home game machine by sony corporation.

NS (Nintendo Switch ), a host issued by Nintendo corporation in 2017 in 3 months, is designed integrally with a home machine and a palm machine.

Xbox One, a home gaming machine sold by Microsoft corporation.

Steam, a digital game software distribution platform.

PSN (PlayStation network), Sony corporation provides a platform for PS (PlayStation) players to provide relevant network services.

Steamspy, a data tracking website for Steam.

Gamstat, a data tracking website for PSN.

Referring to fig. 1, fig. 1 is a schematic diagram of an application environment according to an embodiment of the present invention, where the application environment may include a client 01 and a server 02, and the client 01 and the server 02 may be directly or indirectly connected through wired or wireless communication. The target object (such as a user, a user simulator; the user can comprise a common user and a worker) can send the target characteristic data of the game object to be predicted to the server through the client, and the server processes the received target characteristic data to predict sales data. It should be noted that fig. 1 is only an example.

The client may include a physical device of a type such as a smart phone, a desktop computer, a tablet computer, a notebook computer, an Augmented Reality (AR)/Virtual Reality (VR) device, a digital assistant, a smart speaker, a smart wearable device, and the like, and may also include software running in the physical device, such as a computer program. The operating system corresponding to the client may include an Android system (Android system), an IOS system (a mobile operating system developed by apple inc.), linux (an operating system), Microsoft Windows (Microsoft Windows operating system), and the like.

The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a Network service, cloud communication, a middleware service, a domain name service, a security service, a CDN (Content Delivery Network), a big data and artificial intelligence platform, and the like. Which may include a network communication unit, a processor, and memory, among others. The server may provide background services for the client.

The following describes a specific embodiment of a game data processing method according to the present invention, and fig. 2 is a schematic flow chart of a game data processing method according to an embodiment of the present invention, and the present specification provides the method operation steps as described in the embodiment or the flow chart, but may include more or less operation steps based on conventional or non-creative labor. The order of steps recited in the embodiments is merely one manner of performing the steps in a multitude of orders and does not represent the only order of execution. In practice, the system or server product may be implemented in a sequential or parallel manner (e.g., parallel processor or multi-threaded environment) according to the embodiments or methods shown in the figures. Specifically, as shown in fig. 2, the method may include:

s201: obtaining target characteristic data from a game data group of a game object to be predicted, wherein the game data group comprises at least one of the following: game base data, game community data, and game derivative data;

in the embodiment of the present invention, the target feature data may be acquired by the server itself, or acquired by the server due to the transmission of the client. The target characteristic data is from a game data group of the game object to be predicted, and the game data group comprises at least one of the following: game base data, game community data, and game derivative data:

1) the game basic data may indicate data describing game object basic information and associated information. The basic information may include basic attributes such as a sale date, a sale duration, an interval duration of a sale quantity publication date from the sale date, a player demographics, a maximum sale price, a minimum sale price, a sale platform, etc., languages such as chinese, french, spanish, japanese, korean, russian, turkish, portuguese, polish, etc., and the like. The associated information may include the nature of the distribution platform (e.g., self-selling (game developed by self), agent-selling (game developed by others), self-selling + agent-selling), the number of users of the distribution platform (e.g., the number of users on the day of game distribution, which may be obtained by least squares fitting of the financial instrument sales figures), the number of comments corresponding to the game on the distribution platform, etc.

2) The game community data may indicate feedback opinions of users (which may include professional users and general users) on the game objects. The game community data may include scores of professional users for game developers, scores of professional users for game publishers, scores of professional users for game IPs (Intellectual Property), scores of game IPs (Intellectual Property) for games, and the like. The domain professional knowledge of professional users is defined according to scoring objects, and the professional users can point to experts, professional media and the like. The game community data may include scores of ordinary users for the game (which may include scoring, adding type tags, posting comments), the amount of scored (ordinary) users corresponding to the game, the distribution of user scores corresponding to the game, and the like.

3) The game derivative data may be indicative of data relating to a game derivative. The game derivatives include real objects such as toys, foods and ornaments based on game concepts, products such as images, data, music and videos (such as commentary videos and fun videos) based on game concepts, and search statistics based on game topics (such as the number of searches of a certain game topic in a period).

In one embodiment, before the obtaining target feature data from the game data group of the game object to be predicted, the method further comprises: acquiring the sale time (sale date) of the game object to be predicted; acquiring a preset time length and determining the current time; and when the time length between the sale time and the current time meets the requirement of the preset time length, constructing the game data group.

Considering that sales (such as the sales in the first month) in the early period of game sale often account for a main part of the total sales of the game, especially for the game of buying and breaking, a preset time length is introduced to measure whether the time length between the sale time and the current time meets the time length condition of the new promotion game, and the preset time length corresponding to the new promotion game may be 28 days, 29 days, 30 days, 31 days, and the like. For example, the time of sale of the game object to be predicted is 8, 15 and 2020, the current time is 8, 31 and the time between the time of sale and the current time (16 days) is less than a preset time (for example, 30 days), then the game object to be predicted belongs to the new promotion game, and further, a game data group of the game object to be predicted is constructed. In addition, for business requirements, the preset duration may be used to measure whether the duration between the sale time and the current time meets a duration condition of a quarterly newly-sold game, a duration condition of a semiannual newly-sold game, a duration condition of an annual newly-sold game, and the like, and then a sales list of an online new promotion game within 30 days, a sales list of a game sold within a certain quarter, a sales list of a game sold within a semiannual, a sales list of a game sold within a year, and the like may be created according to a subsequently obtained sales prediction result. In practice, for a new promotional game, the sales forecast may be updated routinely every week for the first 30 days that the game was online. If the game is online for more than 30 days, the game is transferred to a historical game aggregate, and retrospective ranking and analysis can be performed on the games in the historical game aggregate every quarter, half year and year.

Of course, the filtering of the game object to be predicted can judge whether the game object is a game of promotion and interruption, and can also judge whether the game object is a game of purchase and interruption, and exclude a game of exemption, a game of gift and the like; whether the game is a cross-platform game or not can be judged, and games which exclusively log in small platforms such as Uply (a game platform belonging to Youbin corporation), Origin (a game platform belonging to Yi-electric corporation), Epic (a game platform belonging to EPIC corporation; EPIC corporation, a game development corporation) and the like are eliminated. The filter conditions for the above-described screening game for buy-off and cross-platform games can also be used for the determination of sample game objects.

In another embodiment, when the game data group includes the game basic data and the game community data, the method further includes constructing the game data group including: determining a target game selling site, and acquiring the game basic data from the target game selling site; determining a target game community site, and acquiring the game community data from the target game community site; and constructing the game data group based on the game basic data and the game community data.

For the game basic data, a target game distribution site may be determined, the game distribution site may be a website for providing a game distribution service, and the target game distribution site may be selected from Nintendo shop (a game distribution site belonging to Nintendo corporation), XBOX Store (a game distribution site belonging to microsoft corporation), stem and PSN. And then obtaining game basic data from the target game community site, wherein the mode of obtaining the game basic data can utilize a crawler. The game basic data here indicates data describing the basic information and the associated information of the game object to be predicted, and reference may be made to the aforementioned description on the game basic data, which is not described herein again.

For the game community data, a target game community site may be determined, where the game community site may be a website for providing game community services, and the target game community site may be a bean (a community website), rawg. The game community data is then obtained from the target game community site in a manner that may utilize a crawler. The game community data indicates feedback opinions of users (which may include professional users and ordinary users) on the game object to be predicted, and the aforementioned description about the game community data can be referred to, and is not repeated here.

In practical application, the evaluation number or the number of users who own the game object to be predicted can be respectively obtained from the bean-net, the rawg. Among other things, game type tags may include high-weight tags such as ACT, adventure, RPG (Role-playing game), narrative, strategy, FPS (First-person shooting), fighting, puzzle solving, street game, science fiction, open world, survival, and so on.

Because the labels of the game type labels on various websites are relatively disordered, each game has more than ten labels, for example, the first person shooting and the action have intersection, the shooting and the first person shooting have inclusion relationship, the first person puzzle solving and the first person shooting are easily confused, and the unclear label can increase the processing difficulty for the corresponding character type features. Chaotic tags can thus be corrected by manual intervention as required.

In another embodiment, when the game data group includes the game derivative data, the method further includes constructing the game data group including: determining a target publishing site, wherein the target publishing site is a site for publishing a derivative object corresponding to the game object to be predicted, and the derivative object comprises at least one selected from a group consisting of a media object and a statistical object; calling a corresponding application program interface to acquire the game derivative data from the target publishing site, wherein the game derivative data is associated data of the derivative object; constructing the game data group based on the game derivative data.

The game derivative data indicates the relevant data of the game derivative corresponding to the game object to be predicted, and reference may be made to the above description of the game derivative data, which is not repeated herein. In practical applications, the site for publishing the derivative object corresponding to the game object to be predicted may be a website for providing live broadcast service, a website for providing video service, or a website for providing search service. When the target publishing site is a website providing live broadcast service, the derived object may be live broadcast content corresponding to the game object to be predicted (for example, a main broadcast explains game operation skills in a live broadcast room), and the game derived data may be associated data (for example, number of viewers) of the live broadcast content. When the target publishing site is a website providing video services, the derivative object may be video content (such as a game promotion video and a funny video produced based on game materials) corresponding to the game object to be predicted, and the game derivative data may be associated data (such as a playing amount) of the video content. When the target publishing site is a website providing a search service, the derivative object may be search content corresponding to a game object to be predicted (for example, a game is used as a keyword for searching), and the game derivative data may be related data of the search content (for example, based on the number of searches corresponding to each time period in a certain period), the game is used as a search variation trend of the keyword in a certain period, and the game is used as a commonly used related word when searching is performed by using the game as the keyword).

The Youtube (a video website) can be used as a target release site, and a corresponding application program interface is called to obtain the Youtube video frequency within 48 hours of sale corresponding to the game object to be predicted as game derivative data. The Twitch (a real-time streaming media video platform facing to a video game) can be used as a target release site, and a corresponding application program interface is called to obtain the top watching number of Twitch history corresponding to a game object to be predicted as game derivative data. The Google search engine can be used as a target release site, and a corresponding application program interface is called to acquire a Google trend (Google Trends) relative value corresponding to a game object to be predicted as game derivative data.

In another embodiment, the target characteristic data may also come from a data analysis platform based on massive user behavior data, and the target characteristic data corresponding to the game object to be predicted may be acquired from such a data analysis platform. Such data analysis platforms often display analysis data in a graph form, and the displayed analysis data often corresponds to numerical characteristics, so that a graph recognition mode can be adopted when target characteristic data is acquired from such data analysis platforms.

S202: inputting the target characteristic data into a game sales prediction model;

in the embodiment of the invention, the game sales prediction model is determined by machine learning training based on a plurality of sample characteristic data, and each sample characteristic data carries sales information labels. The sample feature data is from the game data group of the sample game object, and the sample feature data and the game data group of the sample game object may refer to the records of the target feature data and the game data group of the game object to be predicted in step S201, and are not described again. It should be noted that, the "plurality of sample feature data" here is from a game data group of a plurality of sample game objects, and the number of sample feature data corresponding to each sample game object may be plural. A plurality of sample feature data corresponding to the same sample game object may be incorporated into one sample feature data set. In actual practice, the sample game object may indicate games that were sold in 2013 and beyond, since both PS4 and Xbox One were published in 2013. Further, the sample game may correspond to a buy-off game.

The sales volume information label carried by the sample characteristic data can be determined based on the sales volume leaked by stem in 8 months in 2018, the sales volume leaked by PSN in 12 months in 2018, the sales volume published by financial institutions of various game companies and news disclosure messages. In practical application, the sample game object corresponds to a game of 1200 surplus games on each platform up to now in 2013. In determining the sales information annotation based on the leak sales at Steam and the leak sales at PSN, the following method may be employed:

1) random sampling method: randomly acquiring public data of part of users, further presuming the possession of a certain game of the corresponding platform and characterizing the sales volume by the possession, and then optimizing parameters used for the presumption based on the leaked sales volume data, so that the optimized parameters can be utilized in subsequent sales volume determination;

2) comment number estimation method: the sales of a certain game on the corresponding platform are estimated by multiplying the number of comments by a coefficient (median 77, average 82) between 30 and 150. The earlier the game is sold, the higher the coefficient is selected, for example, the game sold in 2014 uses 110 as the coefficient, and the game sold in 2017 uses 65 as the coefficient. Of course, some game type correction parameters may be added to estimate the sales amount through the number of comments, but the calculation of introducing correction parameters is highly related to the number of comments multiplied by 65. The parameters used for the foregoing speculation are optimized based on the leaked sales data so that the optimized parameters can be utilized in subsequent sales determinations.

Of course, for the application of the reference number of players corresponding to the achievement trophy revealed by PSN, it can be determined how many proportion of players played a game based on the disclosure of the user, that is, obtained a achievement trophy of the game (e.g., MyPS4Life achievement trophy; MyPS4Life, my PS4Life, a social media planning activity of sony corporation), and then the actual number of players of the game can be known (the reference number of players is the aforementioned proportion) based on the reference number of players corresponding to the revealed achievement trophy and the sales amount can be represented by the actual number of players.

For the sales information label carried by the sample characteristic data, the sales information label indicates that at least one of the following items is included: the total sales amount of all the platforms (when only the current total sales amount of a certain platform exists, the total sales amount of all the platforms is taken as the current total sales amount of all the platforms, otherwise, the sum of at least two current total sales amounts of a certain platform is taken as the current total sales amount of all the platforms), and the total sales amount of all the platforms corresponding to a certain historical time (when only the total sales amount of a certain platform corresponding to a certain historical time exists, the total sales amount of all the platforms corresponding to a certain historical time is taken as the total sales amount of all the platforms corresponding to a certain historical time, otherwise, the sum of at least two total sales amounts of a certain platform corresponding to a certain historical time is taken as the total sales amount of all the platforms corresponding to a.

In one embodiment, as shown in fig. 3, the method further includes a process of training the game sales prediction model:

s301: obtaining the plurality of sample feature data;

before the sample feature data is input into the preset machine learning model, whether to perform further processing on the sample feature data can be determined according to the type of the sample feature data. For numerical type features, they may be used without being processed. For the character font features (such as languages, tags, etc.), One-hot encoding (One-hot encoding) can be adopted for encoding and then used.

Before the sample characteristic data is input into the preset machine learning model, the characteristic data to be input can be optimized through outlier rejection and missing value filling. Outliers may indicate sample feature data that is significantly inconsistent with the actual situation. For example, a title of a certain game has only one common word, so that return data of an abnormal number may appear when the Youtube API is called, and return data of a game of the same name may appear when data is crawled from the team, and therefore, the game needs to be screened. The common optimization method is to eliminate the abnormal sample feature data as the missing value. In consideration of the application of an eXtreme Gradient Boosting (XGBoost) model (algorithm), or a regression method (such as a squarederror method) in an XGBoost toolkit is adopted, the (XGBoost model (algorithm) or XGBoost toolkit) can treat missing values as a sparse matrix, and does not consider the missing values when nodes are split.

The method comprehensively considers the characteristic dimension of the sample characteristic data and the accuracy of the model output data to decide whether to compress the characteristic dimension of the sample characteristic data by using PCA (principal Component Analysis, a compression method). Under the condition of less characteristic dimensionality, the characteristic dimensionality can not be compressed so as to ensure the accuracy of the model output data. Accordingly, whether to perform feature dimension compression on the target feature data can be determined according to whether to perform feature dimension compression on the sample feature data.

In addition, the process of training the game sales prediction model can be regarded as a regression process, and the target of the regression is the sales. Considering that the amount of sales in a period is not normally distributed, and all the amounts of sales are larger than zero and have a stronger right deviation (the higher the amount of sales, the stronger the deviation), the amount of sales information can be labeled with the amount of sales indicated by taking the natural logarithm thereof to reduce the deviation. That is, the regression target can be set to the natural logarithm of the sales volume first, and then later reduced to a sales volume figure. And marking the indicated sales volume by combining the sales volume information, and preferentially selecting the current total sales volume of all the platforms by the regression target.

S302: based on the plurality of sample characteristic data, performing sales data prediction training by using a preset machine learning model, and adjusting model parameters of the preset machine learning model in the training until sales prediction results output by the preset machine learning model are matched with sales information labels carried by input sample characteristic data;

the preset machine learning model can adopt an XGboost model, the XGboost model is a model constructed based on a decision tree, and whether the characteristic data input into the XGboost model is subjected to characteristic scaling processing such as standardization and normalization has small influence on model training, so that the accuracy of the output data of the model cannot be effectively improved. In consideration of the efficiency of model training, feature scaling processing may not be performed on the feature data of the input model.

The preset machine learning model may be an initial model or an intermediate model. And inputting sample characteristic data carrying sales volume information labels to a preset machine learning model to perform sales volume data prediction training. In the training, the model parameters may be adjusted based on the difference between the intermediate result output by the model (the sales data of the sample game object corresponding to the sample feature data) and the sales information label carried by the sample feature data.

S303: and taking a preset machine learning model corresponding to the adjusted model parameters as the game sales prediction model.

The game sales forecasting model with high generalization ability is obtained by training the machine learning model, and the adaptability of the game sales forecasting can be improved when the sales data is forecasted by the game sales forecasting model, so that the reliability and effectiveness of the game sales forecasting can be greatly improved.

As shown in fig. 5, fig. 5 is a schematic diagram of an application scenario of a game sales prediction model according to an embodiment of the present invention. In fig. 5, the training data are sample characteristic data, and each sample characteristic data carries a sales volume information label; correspondingly, the subsequently trained game sales prediction model can predict the sales data of the target characteristic data. In fig. 5, the target feature data is input to the game sales prediction model, and the sales prediction result of the game object to be predicted corresponding to the target feature data is output via the game sales prediction model.

In another embodiment, as shown in FIG. 4, the method further comprises a process of training a plurality of game sales prediction models:

s401: obtaining the plurality of sample feature data;

reference is made to the processing content of the sample feature data in the foregoing step S301, and details are not repeated here.

S402: obtaining a plurality of random numbers, wherein the random numbers indicate the proportion of the same divided sample characteristic data, and the random numbers indicate different sample sets to form a data combination;

for example, the same division ratio indicated by the plurality of random numbers is 8:2, the plurality of sample feature data includes data 1-10, and since the plurality of random numbers indicate different sample sets to form a data combination, the sample set to form the data combination indicated by the random number a is different from the random numbers b and c.

S403: dividing the plurality of sample characteristic data into a corresponding first sample set and a second sample set respectively based on the sample set composition data combination indicated by each random number;

correspondingly, the sample set indicated by the random number forms a data combination to divide the multiple sample characteristic data into different sample sets: for the random number a, the data ratio of the first sample set and the second sample set is 8:2, the composition data of the first sample set is data 1-8, and the composition data of the second sample set is data 9-10; for the random number b, the data ratio of the first sample set and the second sample set is 8:2, the composition data of the first sample set is data 1-3, 5 and 7-10, and the composition data of the second sample set is data 4 and 6; for the random number c, the ratio of data for the first sample set and the second sample set is 8:2, the constituent data for the first sample set are data 2-4 and 6-10, and the constituent data for the second sample set are data 1 and 5.

Of course, a plurality of random numbers may indicate different division ratios. For example, the random number a indicates a division ratio of 8:2, the random number b indicates a division ratio of 7:3, and the random number c indicates a division ratio of 9: 1. Correspondingly, the plurality of sample characteristic data are divided into different sample sets according to the division ratio indicated by the random number. For example, if the random number a indicates a division ratio of 8:2, 80% of the plurality of sample feature data may be divided into the first sample set and 20% of the plurality of sample feature data may be divided into the second sample set based on the number dimension. Since "a plurality of sample feature data" are from a game data group of a plurality of sample game objects, and the number of sample feature data corresponding to each sample game object may be plural, a plurality of sample feature data corresponding to the same sample game object may be included in one sample feature data group. Then 80% of the plurality of sample feature data groups may be sorted into a first sample set and 20% of the plurality of sample feature data groups may be sorted into a second sample set based on the group dimension when the division ratio indicated by the random number a is 8: 2.

S404: respectively constructing a reference model corresponding to each random number based on a first sample set corresponding to each random number and a preset machine learning algorithm;

the first sample set may be regarded as a training set, and since each random number corresponds to one training set, a reference model is constructed based on each training set and a preset machine learning algorithm. For example, the random number a corresponds to a first sample set a and a second sample set a, and a reference model a is constructed based on the first sample set a and a preset machine learning algorithm; the random number b corresponds to a first sample set b and a second sample set b, and a reference model b is constructed based on the first sample set b and a preset machine learning algorithm; and the random number c corresponds to the first sample set c and the second sample set c, and a reference model c is constructed based on the first sample set c and a preset machine learning algorithm.

S405: and optimizing the corresponding reference model by using the corresponding second sample set based on each random number to obtain a game sales prediction model corresponding to each random number.

And combining the reference model a, the reference model b and the reference model c obtained in the step S404. The reference model a can be optimized by using a second sample set a, and the optimized reference model a is used as a game sales prediction model a; a second sample set b can be used for optimizing a reference model b, and the optimized reference model b is used as a game sales prediction model b; the reference model c may be optimized using the second sample set c, and the optimized reference model c may be used as the game sales prediction model c.

Correspondingly, the inputting the target characteristic data into a game sales prediction model comprises: the target feature data is input into a plurality of the game sales predicting models respectively, and the difference between the plurality of the game sales predicting models is in a combination of constituent data divided into a first sample set and a second sample set in the same proportion, or in a proportion of the plurality of sample feature data divided into the first sample set and the second sample set.

The process of optimizing the corresponding reference model with the corresponding second sample set may be:

the first method comprises the following steps: firstly, acquiring a preset goodness of fit; and then performing sales data prediction training by using the corresponding reference model based on the corresponding second sample set, and adjusting the model parameters of the corresponding reference model in the training to meet the requirement of the preset goodness of fit on the sales prediction result output by the corresponding reference model and the sales information label carried by the input sample characteristic data.

The preset goodness of fit serves as a model evaluation index of the reference model, the preset goodness of fit corresponds to an R2 score (namely, a determination coefficient in statistics and translated into a judgment coefficient), and the R2 score can reflect the proportion that all variation of a dependent variable can be explained by an independent variable through a regression relationship. For example, if the R2 score is 0.8, this indicates that 80% of the variability of the dependent variable can be explained based on the regression relationship. The larger the value of the R2 score, the higher the interpretation of the dependent variable by the independent variable, the higher the percentage of total variation by the dependent variable, and the denser the observation points are near the regression line. The value of the preset goodness of fit can be 0.8, 0.88, 0.9 and the like, and can be flexibly adjusted according to needs. This optimization may be viewed as the second sample set both being used as a test set to test the reference model and as another training set to guide parameter adjustment of the reference model.

And the second method comprises the following steps: the second sample set may be regarded as a test set to test the reference model, the preset goodness of fit (the aforementioned related description of the preset goodness of fit may be referred to) is regarded as a model evaluation index of the reference model, and whether the reference model meets the requirement of the model evaluation index is determined according to the sales prediction result output by the model. When not satisfied, the parameters of the reference model are continuously adjusted using the first sample set.

In practical applications, as a third-party module commonly used in machine learning, sklern (an open-source python language-based machine learning toolkit) may partition a plurality of sample feature data into training sets and test sets (see the following codes) based on train _ test _ split of the sklern toolkit, and calculate an R2 score based on R2_ score of the sklern toolkit:

X_train，X_test，y_train，y_test＝train_test_split(X，y，test_size＝0.2，random_state＝1)

regress_model＝xgb.XGBRegressor(objective＝”reg：squarederror”，nthread＝-1，seed＝1234)

the random _ state and seed parameters may jointly determine the generation of the random number, or may be regarded as a tag of the random number, and then different random numbers may be located through different assignment combinations of the random _ state and the seed parameters. The parameter test _ size may indicate a division ratio. The number of random numbers may be 1000, and accordingly, 1000 game sales prediction models may be obtained.

The machine learning training is carried out based on the sample characteristic data corresponding to the games of the 1200 remaining platforms in 2013 till now, and the r2 score of the model is about 0.88 after multiple tests, which can show that the calculation of the model is relatively accurate. Of course, the model may be updated and retrained based on new financial instrument published sales, new sales revealed by news, and sales revealed by subsequent platforms.

S203: obtaining a sales forecasting result of the game object to be forecasted by utilizing the game sales forecasting model;

in the embodiment of the invention, the sales prediction result of the game object to be predicted can be output by using the game sales prediction model. In conjunction with the description of steps S401 to S405, the target feature data may be input into each of the plurality of game sales prediction models, and then the candidate results corresponding to each of the plurality of game sales prediction models may be obtained, and the sales prediction result may be obtained based on the candidate results corresponding to the plurality of game sales prediction models. Statistical analysis may be performed on the plurality of candidate results to obtain a sales volume interval for the game object to be predicted by calculating the corresponding average value μ and standard deviation σ. Using a sales interval instead of a single sales figure gives a more representative data. For each game, 1000 randomized models are substituted for calculation, and the sales average μ and standard deviation σ of the corresponding game can be obtained. According to the Layida criterion, the actual sales corresponding to the game object to be predicted has 65% of probability of falling within (mu-sigma, mu + sigma), 95% of probability of falling within (mu-2 sigma, mu +2 sigma) and 99% of probability of falling within (mu-3 sigma, mu +3 sigma), which is also called 3 sigma principle. In practical applications, when the front-end presentation is performed by using the sales prediction result of the game object to be predicted, the 6 σ interval is considered to be too large, and sometimes even exceeds an order of magnitude, because the sales interval for the game object to be predicted is determined by using (μ - σ, μ + σ) selectively at the time of the front-end presentation.

In one embodiment, based on the idea that the regression target can be set to be the natural logarithm of the sales volume first and then be restored to the sales volume figure at the later stage, the sales volume prediction result of the game object to be predicted, which is output by the game sales volume prediction model, is the natural logarithm of the sales volume, and needs to be restored to the sales volume figure.

Referring to fig. 6 and 7, the game sales interval calculated by the game sales prediction model is shown on a specific page. The embodiment of the invention can develop a set of models for carrying out sales volume calculation aiming at any purchase and break PC or host computer games by utilizing machine learning and data mining technologies, and integrates the models into the background of the applet in the instant messaging application so as to carry out game sales volume display through the applet. Therefore, when a new online game is found, the automatic processes of data mining, sales calculation and display can be realized. The small program product has the characteristics of high visualization degree, strong usability, high data precision and the like, and has great irreplaceable advantages on calculation aiming at cross-platform game sales.

Considering that sales (such as first month sales) before the game is sold often account for a major part of the total sales of the game, especially for buy-and-break games, the sales interval of the game sold for nearly 30 days can be shown on the main page of the applet product and ranked from high to low. After the time of selling the game exceeds 30 days, the sales volume interval of the game can be put into a historical archive, and then a summary list is reviewed quarterly, semiannually and annually, so that the sales volume interval can be conveniently checked and corrected with the numbers of the financial reports published by various manufacturers, and certainly, the sales volume published by the financial reports, the sales volume disclosed by news and the sales volume leaked by a subsequent platform can be conveniently compared to update the model. In addition to the list updated in real time, the game icon on the right side of the graph 7 is triggered to jump into the game detail page of the corresponding game, and the sales volume interval obtained through model calculation is also integrated in the game detail page, so that the intuitive comprehensive information display is convenient for the user.

In practical application, the embodiment of the invention adopts a machine learning method, and can train each feature of the game of the purchase and breakdown of the PC + host platform in the global market and model and calculate the sales of other games by using known real sales figures (sources such as financial reports, news, data leakage and the like). The model obtained by the embodiment of the invention can provide sales figures summarized by each platform in the global market, has no limit of regions and platforms, can calculate sales of platforms (such as NS) without API support, and fills the blank of the current game field calculation tool. Since most games are registered on several platforms, it is difficult to give the overall game performance by considering the sales volume of the PC alone. The sales volume understanding of host platforms such as PS4 and NS is also lacked for a long time, so that the model obtained by the embodiment of the invention can realize the sales volume calculation tasks of a plurality of platforms such as PC, PS4, XBOX ONE and NS. Meanwhile, as the financial reports of each game company count the total sales volume of all platforms, a sales volume analysis tool for verifying the data is lacked at present. If the sales of any single game on a plurality of platforms can be calculated, the income of the game and the financial condition of the corresponding company can be calculated to a certain extent.

The game sales prediction model provided by the embodiment of the invention not only can provide the cross-platform sales corresponding to the game to be predicted, but also can provide a sales interval with a smaller range and higher accuracy. The game sales predicting model provided by the embodiment of the invention can ensure that the real sales of most games fall within the range of the standard deviation of plus or minus three times of the average sales calculated by using the model, even the real sales of a plurality of games fall within a narrow range. Compared with wide and fixed intervals such as 50-100 ten thousand and 200-.

According to the technical scheme provided by the embodiment of the present specification, the game sales prediction model is trained by adopting a machine learning method in the embodiment of the present specification, and the sales prediction result of the game object to be predicted is output based on the input target characteristic data by using the game sales prediction model. The game sales prediction model obtained through training has high generalization capability, the adaptability of the sales data prediction of the game object to be predicted can be improved, and the prediction accuracy of the sales data can be greatly improved.

An embodiment of the present invention further provides a game data processing apparatus, as shown in fig. 8, the apparatus includes:

the data acquisition module 810: the method comprises the steps of obtaining target characteristic data from a game data group of a game object to be predicted, wherein the game data group comprises at least one of the following: game base data, game community data, and game derivative data;

data input module 820: for inputting the target characteristic data into a game sales prediction model;

the result prediction module 830: the game sales prediction model is used for obtaining the sales prediction result of the game object to be predicted;

It should be noted that the device and method embodiments in the device embodiment are based on the same inventive concept.

The embodiment of the invention provides electronic equipment, which comprises a processor and a memory, wherein at least one instruction or at least one program is stored in the memory, and the at least one instruction or the at least one program is loaded and executed by the processor to realize the game data processing method provided by the method embodiment.

Further, fig. 9 is a schematic diagram of a hardware structure of an electronic device for implementing the game data processing method provided by the embodiment of the present invention, and the electronic device may participate in forming or including the game data processing apparatus provided by the embodiment of the present invention. As shown in fig. 9, the electronic device 90 may include one or more (shown here as 902a, 902b, … …, 902 n) processors 902 (the processors 902 may include, but are not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 904 for storing data, and a transmission device 906 for communication functions. Besides, the method can also comprise the following steps: a display, an input/output interface (I/O interface), a Universal Serial Bus (USB) port (which may be included as one of the ports of the I/O interface), a network interface, a power source, and/or a camera. It will be understood by those skilled in the art that the structure shown in fig. 9 is only an illustration and is not intended to limit the structure of the electronic device. For example, the electronic device 90 may also include more or fewer components than shown in FIG. 9, or have a different configuration than shown in FIG. 9.

It should be noted that the one or more processors 902 and/or other data processing circuitry described above may be referred to generally herein as "data processing circuitry". The data processing circuitry may be embodied in whole or in part in software, hardware, firmware, or any combination thereof. Further, the data processing circuit may be a single stand-alone processing module, or incorporated in whole or in part into any of the other elements in the electronic device 90 (or mobile device). As referred to in the embodiments of the application, the data processing circuit acts as a processor control (e.g. selection of a variable resistance termination path connected to the interface).

The memory 904 may be used for storing software programs and modules of application software, such as program instructions/data storage devices corresponding to the game data processing method described in the embodiment of the present invention, and the processor 902 executes various functional applications and data processing by running the software programs and modules stored in the memory 94, so as to implement the above-mentioned game data processing method. The memory 904 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 904 may further include memory located remotely from the processor 902, which may be connected to the electronic device 90 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The transmitting means 906 is used for receiving or sending data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider of the electronic device 90. In one example, the transmission device 906 includes a network adapter (NIC) that can be connected to other network devices through a base station so as to communicate with the internet. In one embodiment, the transmitting device 906 can be a Radio Frequency (RF) module, which is used for communicating with the internet in a wireless manner.

The display may be, for example, a touch screen type Liquid Crystal Display (LCD) that may enable a user to interact with a user interface of the electronic device 90 (or mobile device).

Embodiments of the present invention also provide a computer-readable storage medium, which may be disposed in an electronic device to store at least one instruction or at least one program for implementing a game data processing method in the method embodiments, where the at least one instruction or the at least one program is loaded and executed by a processor to implement the game data processing method provided in the method embodiments.

Alternatively, in this embodiment, the storage medium may be located in at least one network server of a plurality of network servers of a computer network. Optionally, in this embodiment, the storage medium may include, but is not limited to: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.

It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the device and electronic apparatus embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference may be made to some descriptions of the method embodiments for relevant points.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims

1. A game data processing method, characterized in that the method comprises:

inputting the target characteristic data into a game sales prediction model;

2. The method of claim 1, wherein said inputting the target feature data into a game sales prediction model comprises:

inputting the target characteristic data into a plurality of game sales predicting models respectively, wherein the game sales predicting models are different in that the sample characteristic data are divided into composition data combinations of a first sample set and a second sample set according to the same proportion;

correspondingly, the obtaining of the sales prediction result of the game object to be predicted by using the game sales prediction model includes:

respectively utilizing each game sales prediction model to obtain corresponding candidate results;

and obtaining the sales prediction result based on the candidate results corresponding to the plurality of game sales prediction models.

3. The method of claim 2, further comprising training a process of deriving the plurality of game sales prediction models by:

obtaining the plurality of sample feature data;

obtaining a plurality of random numbers, wherein the random numbers indicate the proportion of the same divided sample characteristic data, and the random numbers indicate different sample sets to form a data combination;

dividing the plurality of sample characteristic data into a corresponding first sample set and a second sample set respectively based on the sample set composition data combination indicated by each random number;

respectively constructing a reference model corresponding to each random number based on a first sample set corresponding to each random number and a preset machine learning algorithm;

and optimizing the corresponding reference model by using the corresponding second sample set based on each random number to obtain a game sales prediction model corresponding to each random number.

4. The method of claim 3, wherein optimizing the corresponding reference model using the corresponding second set of samples comprises:

acquiring a preset goodness of fit;

and performing sales data prediction training by using the corresponding reference model based on the corresponding second sample set, and adjusting the model parameters of the corresponding reference model in the training until the sales prediction result output by the corresponding reference model and the sales information label carried by the input sample characteristic data meet the requirement of the preset goodness of fit.

5. The method of claim 1, wherein prior to obtaining target feature data from the game data group of the game object to be predicted, the method further comprises:

acquiring the selling time of the game object to be predicted;

acquiring a preset time length and determining the current time;

and when the time length between the sale time and the current time meets the requirement of the preset time length, constructing the game data group.

6. The method of claim 1, wherein when the game data group includes the game base data and the game community data, the method further comprises constructing the game data group including:

determining a target game selling site, and acquiring the game basic data from the target game selling site;

determining a target game community site, and acquiring the game community data from the target game community site;

and constructing the game data group based on the game basic data and the game community data.

7. The method of claim 1, wherein when the game data group includes the game derivative data, the method further comprises constructing the game data group comprising:

determining a target publishing site, wherein the target publishing site is a site for publishing a derivative object corresponding to the game object to be predicted, and the derivative object comprises at least one selected from a group consisting of a media object and a statistical object;

calling a corresponding application program interface to acquire the game derivative data from the target publishing site, wherein the game derivative data is associated data of the derivative object;

constructing the game data group based on the game derivative data.

8. A game data processing apparatus, characterized in that the apparatus comprises:

9. An electronic device, comprising a processor and a memory, wherein at least one instruction or at least one program is stored in the memory, and the at least one instruction or the at least one program is loaded by the processor and executed to implement the game data processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, in which at least one instruction or at least one program is stored, the at least one instruction or the at least one program being loaded and executed by a processor to implement the game data processing method according to any one of claims 1 to 7.