JP2021037060A - System, method, and program for providing predetermined game, and method for creating classification of decks - Google Patents

Abstract

To support an appropriate classification of decks in which a plurality of pieces of game content are contained.SOLUTION: A game provision system 10 provides a game using a deck in which a plurality of pieces of game content is contained to a user of a player terminal 30 communicably connected through a communication network 20. The system 10 determines the classification of the deck from a predetermined plurality of classifications. Since the predetermined plurality of classifications includes the classification corresponding to a topic extracted through machine learning with deck data of a deck group as learning data, an appropriate classification of decks in which a plurality of pieces of game content are contained is supported.SELECTED DRAWING: Figure 1

Description

The present invention relates to a system, a method, and a program for providing a predetermined game using a deck containing a plurality of game contents, and a method for creating a classification of such a deck.

Conventionally, various games using a deck composed of a plurality of game contents (for example, characters and cards, etc.) have been provided (see, for example, Patent Document 1 below). In some of these games, deck classifications (attributes) based on the combination of multiple game contents that make up the deck are predetermined, and various processes (opponent matching and various parameters) based on these attributes are defined. Correction, etc.) is performed.

Japanese Unexamined Patent Publication No. 2018-143835

However, for example, as the types of game content that can be included in a deck increase, the number of combinations of game content that make up the deck becomes enormous, and it is possible to appropriately determine the classification of decks based on such combinations in advance. It gets harder. Therefore, it is desirable to realize a mechanism to support the classification of decks.

One of the objects of the embodiment of the present invention is to support proper classification of a deck containing a plurality of game contents. Other objects of the embodiments of the present invention will become apparent by reference to the entire specification.

The system according to an embodiment of the present invention is a system for providing a predetermined game using a deck including one or a plurality of computer processors and including a plurality of game contents, and the one or a plurality of computer processors. Executes a step of determining the classification of a specific deck from a plurality of predetermined classifications according to the execution of a readable instruction, and the plurality of predetermined classifications obtain deck data of the first deck group. The topic in the first trained model of the first trained model generated through machine learning as training data, in which the game content included in the deck is treated as a word included in the document. Includes the corresponding classification.

A method according to an embodiment of the present invention is a method for providing a predetermined game using a deck that is executed by one or a plurality of computers and includes a plurality of game contents, and is included in a plurality of predetermined categories. The predetermined plurality of classifications include a step of determining the classification of a specific deck from the above, and the predetermined plurality of classifications are a first trained model generated through machine learning using the deck data of the first deck group as training data. Including the classification corresponding to the topic in the first learned model of the topic model in which the game content included in the deck is treated as a word included in the document.

The program according to one embodiment of the present invention is a program for providing a predetermined game using a deck including a plurality of game contents, and is specified for one or a plurality of computers from a plurality of predetermined classifications. The predetermined plurality of classifications are the first trained model generated through machine learning using the deck data of the first deck group as training data. , Includes the classification corresponding to the topic in the first learned model of the topic model in which the game content contained in the deck is treated as a word contained in the document.

The other method according to one embodiment of the present invention is a method for creating a classification of decks used in a predetermined game and containing a plurality of game contents, and is learning data corresponding to the deck data of the deck group. The process of preparing the training data of the topic model in which the game content contained in the deck is treated as the word contained in the document, and the process of generating the trained model of the topic model through machine learning of the prepared training data. And a step of creating a plurality of predetermined classifications including classifications corresponding to topics in the generated trained model.

Various embodiments of the present invention support proper classification of decks that include multiple game contents.

FIG. 6 is a configuration diagram schematically showing a configuration of a network including a game providing system 10 according to an embodiment of the present invention. The block diagram which shows the function which the game providing system 10 has. The figure which exemplifies the graphical model of the topic model. Diagram to illustrate machine learning of a typical topic model. Diagram to illustrate machine learning of topic models applied to deck classification. The figure for demonstrating the trained model 417. The figure which illustrates the information which the player information table 411 manages. The figure which illustrates the information managed by the deck information table 413. The figure which exemplifies the classification of a deck. The figure which illustrates the information managed by the game performance information table 415. A flow diagram exemplifying how to create a deck classification. The figure which illustrates the screen 50 for deck creation. The flow diagram which illustrates the process which system 10 executes at the time of determining the classification of the deck under creation. The figure which illustrates the screen 60 for deck automatic organization. FIG. 5 is a flow chart illustrating a process executed by the system 10 according to the selection of the automatic organization execution button 68. The figure which illustrates the event statistical information display screen 70. The figure for demonstrating the method of collectively determining the classification of a deck group.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a configuration diagram schematically showing a configuration of a network including a game providing system 10 according to an embodiment of the present invention. As shown in FIG. 1, the system 10 is communicably connected to the player terminal 30 via a communication network 20 such as the Internet. Although only one player terminal 30 is shown in FIG. 1, the system 10 is communicably connected to a plurality of player terminals 30. The system 10 provides a game to the user of the player terminal 30. The game provided by the system 10 (hereinafter, may be referred to as “provided game”) is configured as a game using a deck including a plurality of game contents.

The system 10 is configured as a general computer, and as shown in FIG. 1, a computer processor 11 configured as a CPU, a GPU, and the like, a main memory 12, an input / output I / F13, and a communication I / F14. And a storage (storage device) 15, and each of these components is electrically connected via a bus or the like (not shown).

The computer processor 11 reads various programs stored in the storage 15 or the like into the main memory 12 and executes various instructions included in the programs. The main memory 12 is composed of, for example, a DRAM or the like.

The input / output I / F 13 includes various input / output devices for exchanging information with a user or the like. The input / output I / F 13 includes, for example, an information input device such as a keyboard and a pointing device (for example, a mouse, a touch panel, etc.), a voice input device such as a microphone, and an image input device such as a camera. Further, the input / output I / F 13 includes an image output device such as a display and an audio output device such as a speaker.

The communication I / F 14 is implemented as hardware such as a network adapter, various communication software, or a combination thereof, and is configured to realize wired or wireless communication via the communication network 20 or the like.

The storage 15 is composed of, for example, a magnetic disk, a flash memory, or the like. The storage 15 stores various programs including an operating system, various data, and the like.

In this embodiment, the system 10 can be configured using a plurality of computers, each having the hardware configuration described above. For example, the system 10 may consist of one or more server devices.

The game providing system 10 configured in this way has functions as a web server and an application server, and is installed from a web browser or other application (for example, an application for the provided game) installed on the player terminal 30. Various processes are executed in response to the request, and screen data (for example, HTML data), control data, and the like corresponding to the results of the processes are transmitted to the player terminal 30. The player terminal 30 displays a web page or other screen based on the received data.

The player terminal 30 is configured as a general computer, and as shown in FIG. 1, a computer processor 31 configured as a CPU, a GPU, and the like, a main memory 32, input / output I / F 33, and communication I / The F34 and the storage (storage device) 35 are provided, and each of these components is electrically connected via a bus or the like (not shown).

The computer processor 31 reads various programs stored in the storage 35 or the like into the main memory 32 and executes various instructions included in the programs. The main memory 32 is composed of, for example, a DRAM or the like.

The input / output I / F 33 includes various input / output devices for exchanging information with a user or the like. The input / output I / F 33 includes, for example, an information input device such as a keyboard and a pointing device (for example, a mouse, a touch panel, etc.), a voice input device such as a microphone, and an image input device such as a camera. Further, the input / output I / F 33 includes an image output device such as a display and an audio output device such as a speaker.

The communication I / F 34 is implemented as hardware such as a network adapter, various communication software, and a combination thereof, and is configured to realize wired or wireless communication via the communication network 20 or the like.

The storage 35 is composed of, for example, a magnetic disk, a flash memory, or the like. The storage 35 stores various programs including an operating system, various data, and the like. The program stored in the storage 35 can be downloaded and installed from the application market or the like.

In the present embodiment, the player terminal 30 can be configured as a smartphone, a tablet terminal, a wearable device, a personal computer, a game-dedicated terminal, or the like.

The user (game player) of the player terminal 30 configured in this way is provided by the system 10 by executing communication with the system 10 via a web browser or other application installed in the storage 35 or the like. You can play the game to play.

Next, the functions of the game providing system 10 of the present embodiment will be described. FIG. 2 is a block diagram schematically showing the functions of the system 10. As shown in the figure, the system 10 has an information storage management unit 41 that stores and manages various information, a basic function control unit 43 that controls the basic functions of the provided game, and a game progress control unit 45 that controls the progress of the game. And a deck management unit 47 that manages decks used in the game. These functions are realized by the hardware such as the computer processor 11 and the main memory 12, and various programs and data stored in the storage 15 and the like operating in cooperation with each other. For example, the main memory 12 This is realized by the computer processor 11 executing the instructions included in the program read in. Further, a part or all of the functions shown in FIG. 2 may be realized by the cooperation of the system 10 and the player terminal 30, or may be realized by the player terminal 30.

The information storage management unit 41 of the system 10 stores and manages various information in the storage 15 and the like. For example, as shown in FIG. 2, the information storage management unit 41 has a player information table 411 that manages information about the player, a deck information table 413 that manages information about the deck, and a game performance that manages information about the game performance. It has an information table 415. Further, in the present embodiment, the information storage management unit 41 has a trained model 417 used for classifying decks.

The basic function control unit 43 of the system 10 executes various processes related to the control of the basic function. For example, the basic function control unit 43 transmits screen data or control data of various screens related to the basic function to the player terminal 30, and responds to various operation inputs by the player via the screen displayed on the player terminal 30. Processing is executed, and screen data or control data according to the result of the processing is transmitted to the player terminal 30. Basic function The basic function controlled by the control unit 43 includes, for example, login processing (user authentication), billing control, and player management (for example, updating the player information table 411).

The game progress control unit 45 of the system 10 executes various processes related to the control of the progress of the game. For example, the game progress control unit 45 transmits screen data or control data of various screens for advancing the game to the player terminal 30, and responds to an operation input by the player via the screen displayed on the player terminal 30. Then, various processes are executed, and screen data or control data according to the result of the process is transmitted to the player terminal 30.

The deck management unit 47 of the system 10 executes various processes related to deck management. For example, the deck management unit 47 transmits screen data or control data of various screens related to deck management (for example, updating the deck information table 413) to the player terminal 30, and displays the screen displayed on the player terminal 30. Various processes are executed in response to the operation input by the player via the player, and screen data or control data according to the result of the processes is transmitted to the player terminal 30.

In the present embodiment, the deck management unit 47 is configured to determine the classification of the deck from a plurality of predetermined classifications. The determined deck classification is managed, for example, in the deck information table 413.

In the present embodiment, the predetermined plurality of classifications include classifications corresponding to topics in the trained model 417 of the topic model in which the game content included in the deck is treated as a word included in the document. The trained model 417 is generated through machine learning using deck data of a deck group (plurality of decks) as training data. It can also be said that the topic in the trained model 417 is a topic extracted through machine learning of the training data. A deck group is composed of a large number of decks for creating a deck classification, for example, decks actually used by players (may be limited to players of a specific level or higher) in a specific period of time. Will be done. Machine learning can be performed on a computer such as system 10 or other device.

Here, the topic model will be described. The topic model is a model that assumes that the document is probabilistically generated from a plurality of potential topics, and it is assumed that each word in the document appears according to the probability distribution of the words that the topic has. FIG. 3 illustrates a graphical model of LDA (Latent Dirichlet Allocation) as an example of a topic model. In the graphical model, D is the number of documents, T is the number of topics, and N _d is the number of words in the document d. Further, W is the word itself, Z is the topic of the word, θ is the probability distribution of the topic in each document, and α is a parameter of the Dirichlet distribution for determining θ. In addition, φ is the probability distribution of words in each topic, and β is a parameter of the Dirichlet distribution for determining φ.

When performing machine learning using such LDA, the number of topics T, which is a hyperparameter, and the parameters α and β of the Dirichlet distribution are manually adjusted. Then, the probability distribution θ of the topic in each document and the probability distribution φ of the word in each topic estimated through machine learning are acquired.

FIG. 4 is a diagram for explaining machine learning using a typical topic model. As shown in the figure, in machine learning using a topic model, a large amount of document data is used as learning data, and the learning data is expressed in the form of "Bag of Words" (the number of occurrences of each of a plurality of words is expressed for each document). It is prepared in the "BoW format"). Then, when machine learning is performed on the training data, the probability distribution (φ) of the word in each topic and the probability distribution (θ) of the topic in each document are acquired (estimated) as shown in the figure.

FIG. 5 is a diagram for explaining machine learning of a topic model applied to deck classification in the present embodiment. In machine learning using a topic model applied to deck classification, a large amount of deck data is used as learning data, and game content in the deck is treated as a word in a document in a typical topic model. The learning data is prepared in the BoW format (the number of appearances of each of the plurality of game contents is expressed for each deck). Then, when machine learning is performed on the training data, the probability distribution (φ) of the game content in each topic and the probability distribution (θ) of the topic in each deck are acquired as shown in the figure. These acquired values are stored in the trained model 417.

As shown in FIG. 6, the trained model 417 of the topic model generated in this way is input with deck data (prepared in BoW format) of a new deck (deck X in the example of FIG. 6). Then, it is configured to output the probability distribution of the topic in the new deck.

A given classification used to classify a deck consists of a plurality of classifications corresponding to a plurality of topics in the trained model 417, or a plurality of classifications corresponding to each of the multiple topics and other viewpoints. It consists of one or more classifications set.

As described above, in the game providing system 10 of the present embodiment, the classification of the deck is determined from the predetermined plurality of classifications, and the predetermined plurality of classifications perform machine learning using the deck data of the deck group as learning data. It includes classifications that correspond to topics extracted through it, helping to properly classify decks that contain multiple game content.

The game provided by the system 10 of the present embodiment may include various types of games using a deck. In addition, the game content included in the deck may include various game contents such as characters, cards, and items associated with them (for example, equipment items equipped by the character).

In the present embodiment, the deck management unit 47 inputs the deck data of a specific deck into the trained model 417 and outputs the deck data from the trained model 417 based on at least the probability distribution of the topic in the specific deck. It can be configured to determine the classification of a particular deck, for example, the classification corresponding to the topic identified (eg, the most probable topic) based on at least the probability distribution of the topic in that particular deck. It may be configured to determine as a classification for that particular deck. Such a configuration makes it possible to determine the classification of each deck based on the probability distribution of topics in each deck.

Further, the deck management unit 47 may be configured to determine the classification of the deck being created by the player and present it to the player. For example, the deck management unit 47 accepts the player's creation of a deck (including selection of game content to be included in the deck) via the screen displayed on the player terminal 30, and deck data of the deck being created. Is input to the trained model 417, and the classification of the deck is determined based on at least the probability distribution of the topic in the deck output from the trained model 417, and the determined classification is displayed on the screen. To. Such a configuration allows the player to know the classification of the deck being created and assists the player in creating an appropriate deck.

Further, the deck management unit 47 accepts designation by a player of a classification included in a plurality of predetermined classifications, and can be used by a plurality of game contents (for example, the player) based on at least the designated classification (for example, the player). It may be configured to select the game content to be included in the deck from among (multiple game contents) owned by. For example, the deck management unit 47 is configured to accept the classification designation by the player via the screen displayed on the player terminal 30 and to display the game content selected based on the designated classification on the screen. obtain. In this case, the deck management unit 47 selects the game content to be included in the deck based on at least the probability distribution of the game content in the topic of the trained model 417 corresponding to the classification specified by the player (for example, a game with a high probability). It can be configured to prioritize and select content, etc.). Such a configuration enables automatic selection of game content based on a designated deck classification and assists the player in creating a deck.

Further, the deck management unit 47 may be configured to present to the player statistical information of the provided game based on at least a plurality of predetermined classifications. The statistical information is generated based on the information managed in the game performance information table 415, for example. For example, the deck management unit 47 may be configured to transmit screen data of the screen including statistical information to the player terminal 30 in response to a request from the player via the player terminal 30. Such a configuration allows the user to know the stats of the game based on the deck classification.

Further, the deck management unit 47 determines the classification of each deck of the deck group based on at least the probability distribution of the topic in each deck of the deck group corresponding to the training data when the trained model 417 is generated. Can be configured. Such a configuration makes it possible to collectively determine the classification of the deck group corresponding to the training data when the trained model 417 is generated.

In the present embodiment, the classification of each deck of the second deck group different from the first deck group corresponding to the training data when the trained model 417 is generated may be collectively determined. For example, the first deck group consists of a plurality of decks used in the provided game in the first period, and the second deck group is in the provided game in the second period after the first period. Consists of multiple decks used.

For example, the deck management unit 47 is a second learned model generated through machine learning using the deck data of the second deck group as learning data, and the game content included in the deck is included in the document. The topic of the second trained model identified and identified based on at least the probability distribution of the topic in each deck of the second deck group in the second trained model of the topic model treated as a word. The classification corresponding to the topic of the first trained model (trained model 417) corresponding to the topic may be configured to be determined as the classification of each deck of the second deck group. The correspondence between the topics of the first and second trained models is set based on, for example, a comparison of the probability distributions of the game contents in each topic of both models (for example, the similarity of the probability distributions of the game contents is high). The topics are set as corresponding topics.) Such a configuration makes it possible to collectively determine the classification of each of a plurality of deck groups.

Next, a specific example of the game providing system 10 of the present embodiment having such a function will be described. The system 10 in this specific example provides a social game in which a quest is executed using a deck composed of a plurality of characters. In addition, as an event in the social game, a battle between players using the same deck is also performed.

FIG. 7 illustrates the information managed by the player information table 411 in this example. The player information table 411 in this example manages information for each player, and as shown in the figure, is associated with a "player account" that identifies an individual player, and increases according to the execution of the quest (clearing the stage). Information such as "experience value", "rank" that gradually increases each time the experience value reaches a predetermined value, and "possessed character information" that is information about a character possessed by the player is managed.

The possessed character information includes the character information of each of the plurality of characters possessed by the player. The character information includes, for example, a character ID that identifies an individual character, an experience value, a level, a cost, an HP, an attack power, and the like of the character. The experience value of the character is increased by strengthening the character, for example, and the level is gradually increased each time the experience value reaches a predetermined value.

FIG. 8 illustrates the information managed by the deck information table 413 in this example. The deck information table 413 in this example manages information for each player's deck, and as shown in the figure, is a combination of a "player account" that identifies an individual player and a "deck ID" that identifies an individual deck for each player. In association with, information such as "composition character information" which is information about the characters constituting this deck and "deck classification" indicating the classification of this deck is managed. The constituent character information includes the character information as described above for each of the plurality of characters constituting the deck.

In this example, the deck is composed of two or more characters, and the player can select (create) a predetermined number of characters from the characters owned by the player. In this example, the player can create a plurality of decks (plural variations of decks) and can select the deck to be used from among the plurality of decks when executing a quest or a match.

FIG. 9 illustrates the deck classification (several predetermined classifications) in this example. In this example, as shown in the figure, eight deck classifications from "A" to "H" are set. Further, these classifications "A" to "H" correspond to eight topics (topics 0 to 7) in the trained model 417, respectively. A value corresponding to any of these classifications is set in the "deck classification" of the deck information table 413.

FIG. 10 illustrates the information managed by the game performance information table 415 in this example. The game performance information table 415 in this example manages information on the performance of battles between players, and as shown in the figure, is associated with a "match ID" that identifies individual battles, and has a "match date and time" and a turn system. "First-attack player account" that identifies the first-attack player in the battle, "First-attack deck composition character information" that is information about the characters that make up the deck of the first-attack player, "First-attack deck classification" that is the classification of the first-attack player's deck , "Second-attack player account" that identifies the second-attack player in turn-based battles, "Second-attack deck composition character information" that is information about the characters that make up the deck of the second-attack player, It manages information such as "second-attack deck classification" which is a classification, and "competition content information" which is information on battle content (win / loss, number of turns, etc.).

Here, a method of creating a deck classification illustrated in FIG. 9 will be described. FIG. 11 is a flow diagram illustrating a method for creating a deck classification. First, as shown in the figure, training data of a topic model corresponding to deck data of a target deck group for creating a deck classification is prepared (step S100). The target deck group is a plurality of decks (for example, hundreds of thousands of decks) actually used in a battle between players in a certain period in the past (before the deck classification was created and operated). Consists of. As described above, the training data of the topic model is prepared in the BoW format.

Next, the trained model 417 of the topic model is generated through machine learning of the prepared training data (step S110). Machine learning is performed in system 10 or other equipment. Machine learning is repeated with adjustment of the number of topics T, which is a hyperparameter, and the parameters α and β of the Dirichlet distribution, and finally, the probability distribution (φ) of the character in each topic and each deck. The probability distribution (θ) of the topic in is acquired and saved as a trained model 417. The final number of topics T in this example is 8.

Subsequently, a classification corresponding to the topic (topic extracted through machine learning) in the trained model 417 is created (step S120). When creating a classification, the meaning / definition of each topic may be examined with reference to the probability distribution of the character of each topic, and the name of the corresponding classification may be set.

So far, we have explained how to create a deck classification. Next, the operation related to the creation of the deck by the player will be described. FIG. 12 illustrates a deck creation screen 50 displayed on the player terminal 30. The screen 50 is a screen for the player to create an individual deck, and as shown in the figure, the HP display area 52 for displaying the HP of the deck, the attack power display area 54 for displaying the attack power of the deck, and the like. It has a classification display area 55 for displaying the classification (arch type) of the deck, a character setting area 56 for setting a plurality of characters constituting the deck, and an automatic organization button 58.

In the character setting area 56, a plurality of individual areas 561 (16 in the example of FIG. 12), each of which corresponds to an individual character, are arranged in a grid pattern. When the player selects any of the individual areas 561, a screen for setting (including changing) the corresponding character is displayed, and the player sets the deck from the characters owned by the player through the screen. You can select the character you want to play. Further, in this example, the player can create a plurality of variations of the deck, and the character used in one variation can be duplicated in other variations.

The HP of the deck displayed in the HP display area 52 is calculated based on the HP of each of the plurality of characters included in the deck. When the character included in the deck changes due to the character setting via the character setting area 56, the HP display of the deck in the HP display area 52 also changes in real time.

Similarly, the attack power of the deck displayed in the attack power display area 54 is calculated based on the attack power of each of the plurality of characters included in the deck, and is included in the deck by the character setting via the character setting area 56. When the character is changed, the display of the attack power of the deck in the attack power display area 54 also changes in real time.

Further, the classification (arch type) of the deck displayed in the classification display area 55 is automatically determined and displayed when all of the plurality of characters constituting the deck are set. Further, the classification is automatically determined and displayed (updated) every time the character included in the deck changes due to the setting of the character via the character setting area 56.

FIG. 13 is a flow chart illustrating a process executed by the system 10 when determining the classification of the deck being created. First, the system 10 inputs the deck data of the deck being created into the trained model 417 as shown in the figure (step S200). Specifically, BoW format deck data is created based on a plurality of characters set via the character setting area 56 of the deck creation screen 50 and input to the trained model 417.

Next, the system 10 acquires the probability distribution of the topic output from the trained model 417 (step S210), and determines the classification of the deck being created based on the probability distribution of the acquired topic (step S220). .. Specifically, in the probability distribution of acquired topics, the classification corresponding to the topic with the highest probability is determined as the classification of the deck being created. For example, if the topic with the highest probability is "Topic 3", the classification of the deck being created is determined to be "D" (see FIG. 9).

In this way, when the classification of the deck being created is determined, the determined classification is displayed in the classification display area 55 of the deck creation screen 50.

The automatic organization button 58 on the deck creation screen 50 is an object for automatically organizing a deck (automatic selection of characters). When the button 58 is selected by the player, the deck automatic organization screen 60 illustrated in FIG. 14 is displayed superimposed on the deck creation screen 50. As shown in the figure, the screen 60 is displayed as a classification designation area 62 for designating the deck classification (archetype), a total cost designation area 64 for designating the total cost, and "deck creation". It has an automatic knitting execution button 68.

The player can automatically organize the deck by selecting the automatic organization execution button 68 after designating the classification and the total cost of the deck through the classification designated area 62 and the total cost designated area 64. FIG. 15 is a flow chart illustrating a process executed by the system 10 in response to the selection of the automatic organization execution button 68. The system 10 first acquires the probability distribution of the characters in the topic corresponding to the designated deck classification, as illustrated (step S300). The probability distribution of the characters in each topic is stored in the trained model 417.

Then, the system 10 selects a plurality of characters to be included in the deck based on the acquired character probability distribution and the total cost designated via the total cost designation area 64 (step S310). Specifically, it is confirmed whether or not the player possesses the character in order from the character having the highest probability in the probability distribution of the character, and if the player possesses the character, the character is selected as a character to be included in the deck. In addition, even if the character is owned by the player, if the total cost of the characters included in the deck exceeds the specified total cost, the character is not selected and then next. Proceed to a character with a high probability.

When all of the plurality of characters constituting the deck are selected in this way, a screen (not shown) displaying a list of the selected characters is displayed on the player terminal 30, and the selected characters are displayed on the deck through a confirmation operation by the player. It is set in the character setting area 56 of the creation screen 50. When the player confirms the creation of the deck, the information (including the deck classification) regarding the created deck is registered in the deck information table 413. Thus, in this example, the deck classification is determined and registered each time a deck is created.

The operation related to the creation of the deck by the player has been described above. Next, the operation related to the display of the statistical information of the provided game will be described. In this example, statistical information regarding the performance of the battle between the players managed in the game performance information table 415 is provided to the player. FIG. 16 illustrates an event statistical information display screen 70 displayed on the player terminal 30. The screen 70 displays statistical information on the competition results in the target event held in a predetermined period, and has a classification analysis button 72 displayed as "by archetype". The statistical information displayed on the event statistical information display screen 70 includes four rankings related to deck classification (architecture type) as shown in the figure. Specifically, the four rankings are "(1) Popular Archetype", which is a ranking of categories with a large number of decks used, and "(2) High win rate (overall)", which is a ranking of categories with a high win rate. "Architype", "(3) Architype with a high win rate (first attack)", which is a ranking of the classification with a high win rate when the turn is the first attack, and "(3) Ranking of the classification with a high win rate when the turn is the second attack" (4) Architype with a high win rate (second attack) ".

When the analysis button 72 by classification is selected, a screen (not shown) for performing various analyzes based on the classification of the deck is displayed on the player terminal 30, and the player can play the battle record in the target event through the screen. You can enjoy the analysis. The screen shows, for example, the probability distribution of characters (character usage rate), the number of decks used, the number of users used, the number of battles used, the winning rate (total), the winning rate (first attack), the winning rate (second attack), according to the deck classification. Information such as the winning percentage by classification of the opponent's deck and the number of turns until the settlement is displayed.

In the above-mentioned example, the classification of the deck is determined and registered each time the deck is created, but in the other example of the present embodiment, the classification of the deck group is collectively determined.

For example, as illustrated in FIG. 17 (A), first, the deck data of the first deck group of the first period (for example, the deck group actually used in the first period (for example, N month)). Is generated as the training data of the topic model, the first trained model 417A is generated (1), and a plurality of topics (first topic group) corresponding to the plurality of topics (first topic group) in the first trained model 417A are generated. (2), and then determine the classification of the first deck group (3). The classification of the first deck group is determined based on the probability distribution of the topic in each deck of the first deck group, and specifically, the classification corresponding to the topic with the highest probability is determined as the classification of each deck. Will be done.

Then, as illustrated in FIG. 17B, it was actually used in a second deck group (eg, N + 1 January) in a second period after the first period. A second trained model 417B is generated through machine learning using the deck data of the deck group) as the training data of the topic model (4), and a plurality of topics (second topic) in the second trained model 417B are generated. The correspondence between the group) and the first topic group in the first trained model 417A is analyzed (5). The correspondence between the first and second topic groups is analyzed based on the comparison between the probability distribution of the character in each topic of the first topic group and the probability distribution of the character in each topic of the second topic group. Specifically, topics having similar probability distributions of such characters are determined as corresponding topics. After that, the classification of the second deck group is determined (6). The classification of the second deck group is determined based on the probability distribution of the topics in each deck of the second deck group and the correspondence between the first and second topic groups, and specifically, the most probable. The classification corresponding to the topic of the first topic group corresponding to the topic of the higher second topic group is determined as the classification of each deck.

In the above-mentioned example, only the classification corresponding to the topic (topic extracted through machine learning) in the trained model 417 was created, but in the other examples of the present embodiment, in addition to this, the determination is made based on the rule. The classification to be done is created. For example, a classification corresponding to a deck containing a predetermined character can be created as a rule-based classification, in which case the deck containing the predetermined character is trained when generating the trained model 417. Excluded from the decks corresponding to the data.

In the game providing system 10 according to the present embodiment described above, the classification of decks is determined from a plurality of predetermined classifications, and the predetermined plurality of classifications perform machine learning using the deck data of the deck group as learning data. It includes classifications that correspond to topics extracted through it, helping to properly classify decks that contain multiple game content.

The processes and procedures described herein are implemented by software, hardware, or any combination thereof, other than those expressly described. For example, the processes and procedures described herein are realized by implementing logic corresponding to the processes and procedures on a medium such as an integrated circuit, a volatile memory, a non-volatile memory, or a magnetic disk. Further, the processes and procedures described in the present specification can be implemented as a computer program corresponding to the processes and procedures, and can be executed by various computers.

Even if it is explained that the processes and procedures described herein are performed by a single device, software, component, module, such processes or procedures are performed by multiple devices, multiple software, multiple devices. Can be executed by a component of, and / or multiple modules. The software and hardware components described herein can also be realized by integrating them into fewer components or by breaking them down into more components.

In the present specification, even if the components of the invention are described as either singular or plural, or even if they are described without limitation to either singular or plural, they should be understood separately in the context. Except for, the component may be singular or plural.

10 Game providing system 20 Communication network 30 Player terminal 41 Information storage management unit 411 Player information table 413 Deck information table 415 Game performance information table 417, 417A, 417B Learned model 43 Basic function control unit 45 Game progress control unit 47 Deck management unit 50 Deck creation screen 60 Deck automatic organization screen 70 Event statistics information display screen

Claims (12)

A system for providing a predetermined game using a deck including one or more computer processors and containing a plurality of game contents.
The one or more computer processors depend on the execution of readable instructions.
Perform the step of determining the classification of a specific deck from a plurality of predetermined classifications,
The predetermined plurality of classifications are first trained models generated through machine learning using the deck data of the first deck group as training data, and the game content included in the deck is included in the document. Including the classification corresponding to the topic in the first trained model of the topic model treated as a word,
system. The determination step inputs deck data of the particular deck into the first trained model and at least based on the probability distribution of topics in the particular deck output from the first trained model. Including determining the classification of the particular deck,
The system of claim 1. The determination step comprises determining a classification corresponding to a topic identified based on at least the probability distribution of the topic in the particular deck as the classification of the particular deck.
The system of claim 2. The determination step includes determining the classification of the deck being created by the player and presenting it to the player.
The system according to any one of claims 1 to 3. The one or more computer processors further include a game to be included in the deck from among a plurality of game contents based on at least a step of accepting designation by a player of a classification included in the predetermined plurality of classifications and at least the designated classification. Steps to select content, and perform
The selection step comprises selecting the game content to include in the deck, at least based on the probability distribution of the game content in the topic of the first trained model corresponding to the specified classification.
The system according to any one of claims 1 to 4. The determination step includes determining the classification of each deck in the first deck group based on at least the probability distribution of topics in each deck of the first deck group in the first trained model. ,
The system according to any one of claims 1 to 5. The determination step is a second learned model generated through machine learning using the deck data of the second deck group as training data, and the game content included in the deck is used as a word included in the document. The first trained model corresponding to the topic of the second trained model identified at least based on the probability distribution of the topic in each deck of the second deck group in the second trained model of the topic model being dealt with. Including determining the classification corresponding to the topic of the trained model as the classification of each deck of the second deck group.
The system according to any one of claims 1 to 6. The first deck group comprises a plurality of decks used in the predetermined game in the first period.
The second deck group comprises a plurality of decks used in the predetermined game in a second period after the first period.
The system of claim 7. The one or more computer processors further perform a step of presenting to the player the statistical information of the predetermined game based on at least the predetermined plurality of classifications.
The system according to any one of claims 1 to 8. A method for providing a predetermined game using a deck that is executed by one or more computers and contains a plurality of game contents.
It includes a step to determine the classification of a specific deck from a plurality of predetermined classifications.
The predetermined plurality of classifications are first trained models generated through machine learning using the deck data of the first deck group as training data, and the game content included in the deck is included in the document. Including the classification corresponding to the topic in the first trained model of the topic model treated as a word,
Method. A program for providing a predetermined game using a deck containing a plurality of game contents, and is used on one or more computers.
Executes the process of determining the classification of a specific deck from a plurality of predetermined classifications,
The predetermined plurality of classifications are first trained models generated through machine learning using the deck data of the first deck group as training data, and the game content included in the deck is included in the document. Including the classification corresponding to the topic in the first trained model of the topic model treated as a word,
program. A method for creating a deck classification that is used in a given game and contains multiple game contents.
The process of preparing the learning data of the topic model, which is the learning data corresponding to the deck data of the deck group and in which the game content contained in the deck is treated as a word contained in the document.
The process of generating a trained model of the topic model through machine learning of the prepared training data,
It comprises a step of creating a plurality of predetermined classifications including classifications corresponding to topics in the generated trained model.
Method.

Images