A method, system and device for constructing a player chat character

Abstract

The present application is applicable to the field of computer technology, and provides a method for constructing a player chat character, including: constructing a player emotion model to obtain the player's real-time emotion state parameters; using a clustering algorithm to perform cluster analysis on historical player game data to obtain multiple player groups, and generate emotion tendency characteristics of the corresponding player groups; generating emotion expression parameters of the player chat character according to the real-time emotion state parameters and the emotion tendency characteristics; mapping the emotion expression parameters to the character attributes of the player chat character, so that the player chat character generates a corresponding emotion expression method; establishing an emotion expression strategy model to dynamically adjust the emotion expression method for the player chat character; in the process of interaction between the player and the player chat character, capturing the player's feedback data in real time, and dynamically adjusting the interaction strategy and character attributes of the player chat character. The present application constructs an AI chat character for the player to provide more efficient and personalized services.

Description

Technical Field

[0001] The present application belongs to the field of computer technology, and in particular, relates to a method, system and device for constructing a player chat companion character. Background Art

[0002] In order to improve the gaming experience and activity of players, game companies have always adopted the method of manual chatting. However, with the continuous growth of the game user base, the traditional manual chatting method has gradually exposed its limitations. Specifically, it has the following major problems: (i) High cost: Manual chatting requires the employment of a large number of manual customer service, which leads to high operating costs. (ii) Difficult to scale: Due to the limitation of manual resources, the traditional manual chatting method is difficult to expand on a large scale to cover all game areas and meet the needs of all players. (iii) The degree of personalization of services is not high: Although manual chatting can provide personalized services to a certain extent, due to the limitation of human resources and individual differences, its degree of personalization is often limited.

[0003] In summary, the traditional artificial chat mode has obvious limitations in improving the player experience and activity of game companies. Therefore, it is necessary to build an AI chat character for players in the game to provide more efficient and personalized services. Summary of the invention

[0004] The embodiments of the present application provide a method, system, and device for constructing a player chat character, which can build an AI chat character for the player in the game to provide more efficient and personalized services.

[0005] In a first aspect, an embodiment of the present application provides a method for constructing a player chat character, comprising:

[0006] Based on historical player game data, a player emotion model is constructed, wherein the player emotion model is used to obtain the player's real-time emotion state parameters;

[0007] Using a clustering algorithm to perform cluster analysis on historical player game data to generate multiple player groups, analyzing the behavioral characteristics of each player group, and generating emotional tendency characteristics corresponding to the player group;

[0008] Based on a preset character emotion expression rule library, generating emotion expression parameters of the player's chat companion character according to the real-time emotion state parameters and the emotion tendency characteristics;

[0009] Mapping the emotion expression parameters to the character attributes of the player's companion chat character, so that the player's companion chat character generates a corresponding emotion expression method, wherein the emotion expression method includes facial expressions, tone of voice and body movements;

[0010] For different game types, an emotion expression strategy model is established based on the character emotion expression rule library. The emotion expression strategy model is used to dynamically adjust the emotion expression mode for the player's chat character according to the characteristics of the game type and the player's preferences;

[0011] During the interaction between the player and the player's companion character, the player's feedback data is captured in real time, the player's emotional state is analyzed through the feedback data, and based on the emotional state, the interaction strategy and character attributes of the player's companion character are dynamically adjusted.

[0012] Furthermore, the player emotion model is constructed based on the historical player game data, and the player emotion model is used to obtain the player's real-time emotional state parameters, including:

[0013] Based on psychology theories and game content, emotion labels and emotion characteristics are defined, wherein the emotion labels include happiness, frustration and anger, and the emotion characteristics include emotion keywords, emotion symbols and operation modes;

[0014] Extracting the emotion tag and the emotion feature from historical player game data to form an emotion feature vector;

[0015] Using a machine learning algorithm, the emotion feature vector is input into a player emotion model for training, so that the player emotion model can output real-time emotion state parameters according to real-time player behavior data, and the real-time emotion state parameters include emotion category, emotion intensity and duration;

[0016] A scoring system is constructed for the real-time emotional state parameter, and based on the scoring system, an emotional state score of the emotional state parameter is calculated.

[0017] Furthermore, the historical player game data includes historical game data generated during the game process of multiple players, and the historical game data includes operation frequency, operation type, game duration, game area, interactive object, task completion status and resource acquisition status;

[0018] The clustering algorithm is used to perform cluster analysis on historical player game data to generate multiple player groups, and the behavioral characteristics of each player group are analyzed to generate emotional tendency characteristics corresponding to the player group, including:

[0019] Preprocessing the historical player game data to obtain a player behavior data set;

[0020] Based on the operation frequency, the operation type, the game duration and the task completion status, a clustering algorithm is used to divide the multiple players into multiple player groups;

[0021] Analyze the differences in historical game data of each of the player groups to generate behavioral characteristics corresponding to the player groups, wherein the behavioral characteristics include goal-oriented characteristics, social interaction characteristics, and continuous investment characteristics;

[0022] The behavioral characteristics are subjected to correlation analysis to obtain the emotional tendency characteristics related to the flow experience characteristics in the behavioral characteristics.

[0023] Furthermore, the correlation analysis of the behavioral characteristics to obtain the emotional tendency characteristics related to the flow experience characteristics in the behavioral characteristics includes:

[0024] By calculating the correlation coefficient between the behavior characteristics and the flow experience index, obtaining the correlation degree index between the behavior characteristics and the flow experience characteristics, the flow experience index includes satisfaction, immersion and task completion efficiency;

[0025] Based on the correlation index, a machine learning algorithm is used to evaluate the importance of the behavioral characteristics in predicting the player's flow experience level, and an importance index is generated;

[0026] Based on the importance index, the plurality of behavioral features are sorted, and the first one or more behavioral features are selected as emotional tendency features.

[0027] Furthermore, the generating of the emotion expression parameters of the player's chat companion character based on the preset character emotion expression rule library according to the real-time emotion state parameters and the emotion tendency characteristics includes:

[0028] Based on the preset role emotion expression rule library, a rule matching algorithm is used to match an emotion portrait for the player's chat companion role according to the real-time emotion state parameters and the emotion tendency characteristics;

[0029] Based on the game scene, determining the degree of matching between the game scene and the emotional portrait, and dynamically updating the emotional portrait;

[0030] Obtaining a personality parameter matrix of the player's companion chat character, and determining personality parameters of the player's companion chat character based on the emotional portrait, wherein the personality parameter matrix includes extroversion, openness, and conscientiousness;

[0031] Defining the relationship parameters between the player's companion character and the player, the relationship parameters include intimacy, trust and hostility, and dynamically updating the relationship parameters according to the player's feedback behavior, the feedback behavior includes verbal communication, action interaction and decision-making selection;

[0032] Based on the personality parameter, the relationship parameter and the real-time emotional state parameter, the emotional expression parameter of the player's chatting character is obtained through weighted calculation.

[0033] Furthermore, the emotional expression parameters are mapped to the character attributes of the player's chat character, so that the player's chat character generates a corresponding emotional expression method, and the emotional expression method includes facial expressions, tone of voice and body movements, including:

[0034] A first mapping function, a second mapping function and a third mapping function are respectively set to map the emotion expression parameters to facial expression parameters, tone of voice parameters and body movement parameters in sequence;

[0035] Based on the facial expression parameters, the tone and intonation parameters and the body movement parameters, corresponding facial expressions, tone and intonation and body movement are generated by a behavior generation module.

[0036] Furthermore, for different game types, an emotion expression strategy model is established based on the character emotion expression rule library. The emotion expression strategy model is used to dynamically adjust the emotion expression mode for the player's chat companion character according to the characteristics of the game type and the player's preferences, including:

[0037] During the process of the player chatting with the character to express emotions, the player's feedback data is continuously collected, the feedback data is analyzed using a support vector machine algorithm, and the character's emotional expression rule library is optimized, wherein the feedback data includes the player's emotional changes and behavior patterns;

[0038] A decision tree model is used to construct an emotion expression strategy model, and the game type, scene characteristics and player group are input into the emotion expression strategy model, each node of the emotion expression strategy model represents a judgment condition, and each leaf node represents an emotion expression strategy;

[0039] Based on the emotion expression strategy, selecting a corresponding emotion expression method from the optimized character emotion expression rule library;

[0040] A reinforcement learning algorithm is used to dynamically adjust the parameters of the emotion expression strategy model according to the player's feedback data.

[0041] Furthermore, during the interaction between the player and the player's companion character, the player's feedback data is captured in real time, the player's emotional state is analyzed through the feedback data, and based on the emotional state, the interaction strategy and character attributes of the player's companion character are dynamically adjusted, including:

[0042] Obtaining the conversation content between the player and the character that the player chats with, and obtaining the emotional changes of the player by analyzing the conversation content;

[0043] By analyzing the player's operation records during the game, the corresponding behavior patterns of the players are obtained;

[0044] Determining the player's emotional state and topic preference based on the emotional changes and the behavioral patterns;

[0045] Constructing an emotional knowledge graph, wherein the emotional knowledge graph includes emotional semantic information, interaction strategies, and emotional tendency characteristics under different emotional states;

[0046] According to the emotional state of the player, corresponding emotional semantic information is obtained from the emotional knowledge graph;

[0047] In the process of generating the dialogue of the player's companion chat character, the emotional semantic information is incorporated to adjust the emotional expression of the interactive language of the player's companion chat character;

[0048] Based on the topic preference of the player, the dialogue strategy of the player's chat companion role is dynamically adjusted by adjusting the role attributes of the player's chat companion role, and the dialogue strategy includes topic selection and response method.

[0049] In a second aspect, an embodiment of the present application provides a system for constructing a player chat character, including:

[0050] The first processing module is used to construct a player emotion model based on historical player game data, and the player emotion model is used to obtain the player's real-time emotion state parameters;

[0051] The second processing module is used to perform cluster analysis on historical player game data using a clustering algorithm to generate multiple player groups, analyze the behavioral characteristics of each of the player groups, and generate emotional tendency characteristics corresponding to the player groups;

[0052] The third processing module is used to generate the emotional expression parameters of the player's chat companion character based on the preset character emotional expression rule library and the real-time emotional state parameters and the emotional tendency characteristics;

[0053] The fourth processing module is used to map the emotion expression parameters to the character attributes of the player's chat companion character, so that the player's chat companion character generates a corresponding emotion expression method, and the emotion expression method includes facial expressions, intonation and body movements;

[0054] The fifth processing module is used to establish an emotion expression strategy model based on the character emotion expression rule library for different game types, and the emotion expression strategy model is used to dynamically adjust the emotion expression mode for the player's chat character according to the characteristics of the game type and the player's preference;

[0055] The sixth processing module is used to capture the player's feedback data in real time during the interaction between the player and the player's companion character, analyze the player's emotional state through the feedback data, and dynamically adjust the interaction strategy and character attributes of the player's companion character based on the emotional state.

[0056] In a third aspect, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method of constructing a player chat companion character when executing the computer program.

[0057] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0058] A method for constructing a player chat character in the present application, constructs a player emotion model based on historical player game data, and then can obtain the player's emotional state parameters in real time and accurately, greatly improving the emotional interaction between the game and the player. At the same time, combined with the real-time emotional state parameters and the emotional tendency characteristics of the player group, the preset role emotion expression rule library is used to generate personalized emotional expression parameters for the player chat character, and the emotional expression parameters are directly mapped to the character's facial expressions, tone and body movements, so that the player's companion role can show a more real and vivid emotional response when interacting with the player, thereby greatly improving the player's game experience. In addition, for different types of games, an emotional expression strategy model is established to dynamically adjust the emotional expression method for the player's chat character, ensuring that the player's chat character can maintain a highly compatible emotional interaction with the player in different game environments. In addition, in the process of interaction between the player and the chat character, by capturing and analyzing the player's feedback data in real time, the player's emotional state is further analyzed, and the interaction strategy and role attributes of the chat character are dynamically adjusted accordingly, ensuring that the character can always keep in sync with the player's emotions, thereby establishing a closer and deeper emotional connection. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0060] Figure 1 It is a flowchart of a method for constructing a player chat companion character provided by an embodiment of the present invention;

[0061] Figure 2 It is a structural diagram of a system for constructing a player chat companion character provided by an embodiment of the present invention;

[0062] Figure 3 It is a structural diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0063] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0064] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0065] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0066] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0067] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0068] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0069] See also Figure 1 As shown, the present invention is a method for constructing a player chat character, comprising the following steps:

[0070] S100, constructing a player emotion model based on historical player game data, wherein the player emotion model is used to obtain real-time emotion state parameters of the player;

[0071] In this embodiment, a player emotion model is constructed based on historical player game data, so that the player's emotional state parameters can be obtained in real time and accurately, which greatly improves the emotional interaction between the game and the player.

[0072] In some embodiments, the above step S100 includes:

[0073] Based on psychology theories and game content, emotion labels and emotion characteristics are defined, wherein the emotion labels include happiness, frustration and anger, and the emotion characteristics include emotion keywords, emotion symbols and operation modes;

[0074] Extracting the emotion tag and the emotion feature from historical player game data to form an emotion feature vector;

[0075] Using a machine learning algorithm, the emotion feature vector is input into a player emotion model for training, so that the player emotion model can output real-time emotion state parameters according to real-time player behavior data, and the real-time emotion state parameters include emotion category, emotion intensity and duration;

[0076] A scoring system is constructed for the real-time emotional state parameter, and based on the scoring system, an emotional state score of the emotional state parameter is calculated.

[0077] During the game, different game situations may trigger different emotional reactions in players. For example, in competitive games, players may feel happy because they win the game, or they may feel frustrated or angry because they lose the game. Therefore, in this embodiment, through psychological theories and game content, relevant emotion tags such as happiness, frustration and anger are defined, and emotion features related to emotion tags are defined, such as emotion keywords, emotion symbols and operation modes.

[0078] In this embodiment, by introducing psychological theories and combining them with game characteristics, psychological theories are combined with game behavior data to form a theoretical basis for an emotional model suitable for the game, optimize the interactive experience between players and their companion characters, and provide players with a richer, more interesting and personalized gaming experience.

[0079] When extracting features from historical player game data, emotional keywords are usually obtained through verbal chats or text chats between players and their companion characters. Emotional keywords include happiness, satisfaction, victory, success, failure, disappointment, helplessness, sadness, anger, rage, dissatisfaction, attack, etc. Emotional symbols are obtained through emoticons or emoticons sent during text chats between players and their companion characters. The operation mode is the social operation behavior of players in the game, such as likes, shares, comments, etc.

[0080] In this embodiment, a suitable machine learning algorithm is selected as the basic model of the player emotion model, and the emotion feature vector extracted from the historical player game data is input into the above basic model for training to obtain a player emotion model that can analyze the real-time player behavior data. Specifically, an emotion recognition algorithm, such as a pattern matching method, is used to extract corresponding emotion labels and emotion features from the real-time player behavior data, which are matched with the player emotion model, and then the corresponding real-time emotion state parameters are output, wherein the real-time emotion state parameters include emotion category, emotion intensity and duration, the emotion category includes happiness, frustration and anger, etc., the emotion intensity includes strong, general and weak, and the duration indicates the duration of the corresponding emotion type, such as the player is in a strong state of happiness for 1 minute.

[0081] In one embodiment, a scoring system is defined for the above-mentioned real-time emotional state, and the above-mentioned scoring system includes emotional categories, emotional intensity and duration, and finally obtains the emotional state score. The following is a scoring system in a preferred embodiment. It can be understood that the score coefficient in the evaluation table of the evaluation system is set by the developer according to the needs of the actual game content to meet the emotional needs of the player's chat character in the diversified game. Specifically, the emotional category corresponds to the basic value of the emotional state score, the emotional intensity is the influence coefficient of the basic value, and the duration corresponds to the additional score of the basic value. In one embodiment, the score basic value corresponding to happiness is +10, the score basic value corresponding to frustration is set to -10, the score basic value corresponding to anger is set to -15, and the score basic value corresponding to calmness is set to 0. In practical applications, the corresponding basic value can be set for the emotional type according to the emotional category as positive or negative emotions, so as to improve the practicality of the scoring system; for emotional intensity, the influence coefficient of the weak corresponding basic value is 0.5, the influence coefficient of the slight basic value is 0.75, the influence coefficient of the medium basic value is 1, that is, there is no additional influence, the influence coefficient of the strong basic value is 1.5, and the influence coefficient of the extreme basic value is 2. Duration: For every 1 minute, the emotion score increases by 10% of its base value. In actual applications, the ratio can be adjusted to suit different application scenarios. In addition, if the duration is too long or too short, special processing can be performed according to the actual situation, such as setting an upper or lower limit for the score.

[0082] Specifically, in the above embodiment, the player is in a strong state of happiness for 1 minute. It can be seen that the emotion category is happiness, the emotion intensity is strong, and the duration is 1 minute. Therefore, the emotion category score is: happiness, and the basic value is +10. Emotion intensity adjustment: strong, the influence coefficient is 1.5. Duration adjustment: lasts for 1 minute, increase 10% of the basic value, that is, +1. Therefore, the emotion state score = emotion category basic value * emotion intensity influence coefficient = (10+1)*1.5=16.5.

[0083] S200, using a clustering algorithm to perform cluster analysis on historical player game data to generate multiple player groups, analyzing the behavioral characteristics of each of the player groups, and generating emotional tendency characteristics corresponding to the player groups;

[0084] In some embodiments, the historical player game data includes historical game data generated during the game process of multiple players, and the historical game data includes operation frequency, operation type, game duration, game area, interactive object, task completion status and resource acquisition status;

[0085] The above step S200 includes:

[0086] Preprocessing the historical player game data to obtain a player behavior data set;

[0087] Based on the operation frequency, the operation type, the game duration and the task completion status, a clustering algorithm is used to divide the multiple players into multiple player groups;

[0088] Analyze the differences in historical game data of each of the player groups to generate behavioral characteristics corresponding to the player groups, wherein the behavioral characteristics include goal-oriented characteristics, social interaction characteristics, and continuous investment characteristics;

[0089] The behavioral characteristics are subjected to correlation analysis to obtain the emotional tendency characteristics related to the flow experience characteristics in the behavioral characteristics.

[0090] In this embodiment, the historical player game data is the historical game data of multiple players. By performing cluster analysis on the historical game data of the above multiple players, multiple player groups can be obtained. By performing feature analysis on each player group, the corresponding behavioral characteristics of the player group are obtained. At the same time, the behavioral characteristics are correlated with the emotion-related flow experience characteristics to obtain the corresponding emotional tendency characteristics. Based on the emotional fragrance characteristics, the player's chat character generates corresponding emotional expressions to meet the player's chat needs.

[0091] Flow experience usually occurs when an individual is fully absorbed in an activity, at which time a balance is reached between their skills and challenges. Therefore, in this embodiment, the emotional tendency characteristics related to the flow experience in the behavioral characteristics are identified, and the player's behavioral characteristics are associated with the emotional changes during the game, so that the player's chatting character can analyze and obtain the player's corresponding emotional tendency characteristics based on the player's behavioral characteristics, so that the player's chatting character can adopt corresponding interactive strategies based on the corresponding emotional tendency characteristics, guide the player into the flow state, and improve the player's gaming experience.

[0092] In this embodiment, historical game data of multiple players are obtained, wherein the historical game data includes dimensions such as operation frequency, operation type, game duration, game area, interactive object, task completion status, and resource acquisition status. Based on the above historical game data, a structured player behavior data set is constructed for multiple players. The above player behavior data set is cleaned to remove invalid data with obvious errors, incompleteness, or abnormalities to ensure the accuracy and integrity of the data. At the same time, for different data types and value ranges, methods such as maximum and minimum value normalization or Z-score standardization are used to normalize or standardize the data to eliminate differences in dimensions and value ranges.

[0093] In this embodiment, based on the standardized player behavior data set, the K-means clustering algorithm is used to divide players into different player groups, such as casual players, advanced players, veteran players, and hardcore players, according to dimensions such as operation type, game time, and task completion.

[0094] For different groups of players, the differences in various dimensions of each group of players in the historical game data are analyzed to characterize the behavioral characteristics of different groups of players. In a preferred embodiment, for the goal-oriented feature, the task completion status of each group of players in the game is obtained, and the strength and direction of their goal orientation are evaluated. Specifically, novice players: weak goal orientation, mainly focusing on the basic operation and gameplay of the game, casual players: moderate goal orientation, pursuing the entertainment and relaxation of the game. Advanced players: strong goal orientation, pursuing higher game achievements and rankings. Senior players: very strong goal orientation, pursuing various mechanisms and gameplay in the game, and willing to invest a lot of time and energy. Hardcore players: extremely strong goal orientation, pursuing the ultimate game experience and competitive level, and may conduct professional training. For social interaction features, the social behavior of the player group with the interactive object in the game, such as chat frequency, team situation, friend relationship, etc., is obtained. Novice players: less social interaction, mainly focusing on the game itself, casual players: moderate social interaction, willing to have simple communication and cooperation with other players, advanced players: frequent social interaction, like to team up with other players to challenge difficult tasks. Senior players: They have very frequent social interactions and have established a stable social network. Hardcore players: Their social interactions are mainly competitive and confrontational, and they like to engage in PvP battles with other players or participate in various leaderboards and competitions. For the continuous investment feature, the data features of the player's operation frequency, operation type, game time, game area and other dimensions are obtained to evaluate their continuous investment in the game. Novice players: They have less continuous investment and may give up due to the difficulty of the game or lack of interest. Casual players: They have moderate continuous investment and will play the game when they have free time. Advanced players: They have high continuous investment and are willing to invest more time and energy to improve their game level. Senior players: They have very high continuous investment and have already regarded games as a part of their lives. Hardcore players: They have extremely high continuous investment and may pursue games as a profession or career. Therefore, different player groups have significant differences in goal-oriented characteristics, social interaction characteristics and continuous investment characteristics.

[0095] In some embodiments, the correlation analysis of the behavioral characteristics to obtain the emotional tendency characteristics related to the flow experience characteristics in the behavioral characteristics includes:

[0096] By calculating the correlation coefficient between the behavior characteristics and the flow experience index, obtaining the correlation degree index between the behavior characteristics and the flow experience characteristics, the flow experience index includes satisfaction, immersion and task completion efficiency;

[0097] Based on the correlation index, a machine learning algorithm is used to evaluate the importance of each of the behavioral features in predicting the player's flow experience level, and an importance index is generated;

[0098] Based on the importance index, the plurality of behavioral features are sorted, and the first one or more behavioral features are selected as emotional tendency features.

[0099] In this embodiment, the historical game data corresponding to the behavior feature is assigned a value to construct a behavior feature vector corresponding to each behavior feature. Each element in the behavior feature vector corresponds to the numerical value of a behavior feature. Based on the above behavior feature vector, the behavior feature data of each behavior feature is obtained. Specifically, the goal-oriented feature in the behavior feature is used to reflect the player's task completion and goal pursuit in the game; the social interaction feature is used to reflect the player's social activities and interaction with other players in the game; the continuous investment feature is used to measure the player's continuous participation and investment time in the game.

[0100] When assigning values ​​to historical game data, it is necessary to obtain the data of various data in the historical game data. Specifically, the operation frequency is the number of operations performed by the player in the game, such as clicking, moving, etc. The operation type is the specific operation performed by the player, such as attacking, building, collecting, etc. The game duration is the duration of each time the player logs into the game. The game area is the game map or area where the player is active. The interactive object is the communication object of the player in the game, including other players, NPCs (non-player characters), etc. The task completion status is the number, quality and progress of the tasks completed by the player. The resource acquisition status is the type and quantity of resources obtained by the player in the game.

[0101] In this embodiment, the goal-oriented feature is related to the numerical values ​​of task completion rate, task completion speed and task completion quality. Specifically, according to the task completion status, the ratio of the number of tasks completed by the player to the total number of tasks is calculated to obtain the task completion rate, the average time for the player to complete a single task is calculated to obtain the task completion speed, and the quality of task completion is evaluated according to the evaluation or reward after the task is completed to obtain the task completion quality. The above three numerical values ​​are combined into a goal-oriented feature vector, for example: [task completion rate, task completion speed, task completion quality]. For the social interaction feature, it is related to the interaction frequency, interaction diversity and interaction influence. Specifically, the interaction frequency is obtained by calculating the number of interactions between the player and other players or NPCs, and the interaction diversity index is calculated by counting the number of interactions between the player and different interaction objects to assign a value to the interaction diversity. The social influence of the player is quantified according to the importance of the interaction object (such as senior players, key NPCs) and the interaction results (such as successful transactions, getting help), and the above three numerical values ​​are combined into a social interaction feature vector, for example: [interaction frequency, interaction diversity, interaction influence]. The continuous investment feature is related to the login frequency, game duration stability and resource acquisition efficiency. Specifically, the login frequency is obtained by calculating the number of times a player logs in within a certain period of time, the stability of the player's game duration each time is evaluated by the standard deviation, and the resource acquisition efficiency is measured by calculating the player's resource acquisition speed in the game. The above three values ​​are combined into a continuous investment feature vector, such as [login frequency, game duration stability, resource acquisition efficiency]. The above goal-oriented feature vector, social interaction feature vector and continuous investment feature vector are combined to form a player's behavior feature vector. For example: [task completion rate, task completion speed, task completion quality, interaction frequency, interaction diversity, interaction influence, login frequency, game duration stability, resource acquisition efficiency].

[0102] In this embodiment, for the flow experience index, a questionnaire containing multiple questions is designed to evaluate the player's satisfaction and immersion. The questionnaire questions involve aspects such as the game's graphics, sound effects, plot, difficulty, and the player's emotional experience and involvement in the game. By obtaining the scores of the corresponding options, the values ​​related to satisfaction and immersion are calculated. For task completion efficiency, the task completion time and task completion rate in the historical game data are used to measure it, and the above satisfaction value, immersion value and task completion efficiency are combined to form a flow experience feature vector.

[0103] In this embodiment, the Pearson correlation coefficient formula is used to calculate the correlation coefficient between the behavioral characteristics and the flow experience index. The two variables X and Y in the Pearson correlation coefficient formula represent the behavioral characteristic vector and the flow experience characteristic vector, respectively. The correlation coefficient is used to represent the correlation degree index. That is, when the correlation coefficient is 1, it indicates that the above two variables are completely positively correlated, and the corresponding correlation degree index is 1; when the correlation coefficient is -1, it indicates that the two variables are completely negatively correlated, and the corresponding correlation degree index is -1; when the correlation coefficient is 0, it indicates that there is no linear correlation between the two variables, and the corresponding correlation degree index is 0.

[0104] In this embodiment, the behavioral feature data corresponding to the above-mentioned correlation index of 1 is input into the random forest model for training. During the training process, the performance of the model is optimized by adjusting the parameters of the random forest, such as the number of decision trees, the maximum depth, etc. At the same time, the feature importance evaluation function provided by the random forest algorithm is used to calculate the contribution of each behavioral feature data in the model construction process. Specifically, the information entropy or Gini index of the feature in the decision tree splitting process is calculated to measure it, and then the importance index is generated. It is worth noting that information gain is an indicator to measure the predictive ability of a feature for a target variable, which can be obtained by calculating the change in information entropy or Gini index before and after the feature splitting.

[0105] In this embodiment, multiple behavioral feature data with a correlation index of 1 are sorted according to the numerical value of the importance index, and the first one or more behavioral features are selected as emotional tendency features for obtaining emotional expression parameters of the subsequent player chat character.

[0106] S300, based on a preset role emotion expression rule library, generating emotion expression parameters of the player's chat companion role according to the real-time emotion state parameters and the emotion tendency characteristics;

[0107] In this embodiment, by combining the real-time emotional state parameters and the emotional tendency characteristics of the player group, a preset character emotional expression rule library is used to generate personalized emotional expression parameters for the player's companion character, and the emotional expression parameters are directly mapped to the character's facial expressions, tone of voice and body movements, so that the player's companion character can show more real and vivid emotional reactions when interacting with other players, thereby greatly enhancing the player's gaming experience.

[0108] In one embodiment, the step S300 includes:

[0109] Based on the preset role emotion expression rule library, a rule matching algorithm is used to match an emotion portrait for the player's chat companion role according to the real-time emotion state parameters and the emotion tendency characteristics;

[0110] Based on the game scene, determining the degree of matching between the game scene and the emotional portrait, and dynamically updating the emotional portrait;

[0111] Obtaining a personality parameter matrix of the player's companion chat character, and determining personality parameters of the player's companion chat character based on the emotional portrait, wherein the personality parameter matrix includes extroversion, openness, and conscientiousness;

[0112] Defining the relationship parameters between the player's companion character and the player, the relationship parameters include intimacy, trust and hostility, and dynamically updating the relationship parameters according to the player's feedback behavior, the feedback behavior includes verbal communication, action interaction and decision-making selection;

[0113] Based on the personality parameter, the relationship parameter and the real-time emotional state parameter, the emotional expression parameter of the player's chatting character is obtained through weighted calculation.

[0114] In this embodiment, a preset character emotion expression rule library is provided with a plurality of expression method libraries corresponding to emotions, and one emotion expression method library contains a plurality of expression methods corresponding to emotions. For example, there is a happy expression method library in the character emotion expression rule library, wherein the happy expression method library includes a plurality of types of expression methods, such as animation, sound, text, expression, etc. In actual operation, the player's chatting character expresses happy emotions by playing happy animations, making laughter, showing happy expressions, etc.

[0115] It can be understood that the situation status score corresponding to the real-time emotional state parameter can be obtained by step S100, and the emotional portrait role emotional expression rule library provides a variety of emotional portraits for the chat character according to the player's emotional state score and behavioral characteristics. Each emotional portrait includes a set of specific emotional expressions and language styles to meet the needs of different players. For example, in one embodiment, the emotional category of the player's real-time emotional state parameters is obtained as happy, the emotional intensity is strong, and the duration is 1 minute. In the emotional tendency characteristics of the player group corresponding to the player, the importance index of the goal-oriented feature is 0.35, the social interaction feature is 0.5, and the continuous investment feature is 0.15. Then, through step S100, the emotional state score can be obtained as 16.5. In addition, the corresponding behavioral feature is the social interaction feature. Therefore, from the role emotional expression rule library, it can be obtained that the emotional portrait of the player's chat character is cheerful and optimistic, and interacts with the player in a more positive and optimistic attitude.

[0116] In this embodiment, for environmental parameters, the atmosphere of the game scene is matched with the emotional portrait of the player. If the environmental influencing factor is positive, an enhancement coefficient is set for the above emotional state score. If the environmental influencing factor is negative, a reduction coefficient is set for the above emotional state score. The corresponding enhancement coefficient and reduction coefficient are pre-set by the developer according to the needs of the actual game content. If the emotional category in the real-time emotional state parameters of the player is happy, the emotional intensity is strong, and the duration is 1 minute in the above embodiment, the current game scene is a relaxed and happy festival celebration, then the scene atmosphere matches the emotional portrait of the player, so the environmental influencing factor is positive, and the pre-set enhancement coefficient is 1.2, then the emotional state score at this time is 16.5×1.2=19.8. It can be understood that according to the above updated emotional state score, a new emotional portrait is matched for the player's chat character from the character emotional expression rule library.

[0117] In this embodiment, the personality parameter matrix is ​​pre-set by the developer according to the needs of the actual game content. Specifically, for the above emotional portrait, the personality parameter of the cheerful and optimistic type is 9, the personality parameter of openness is 8, and the personality parameter of conscientiousness is 7. Therefore, the personality parameter of the above cheerful and optimistic type is 9+8+7=24. In some other emotional portraits, the personality parameter matrix also includes agreeableness and neuroticism.

[0118] In this embodiment, the relationship parameters between the player's chat character and the player are dynamically updated through the player's feedback behavior during the game. For example, the relationship parameters between a player and the player's chat character are as follows: intimacy = 50, trust = 60, hostility = 20. During the player's game, the relationship between the verbal communication, action interaction and decision-making selection between the player and the player's chat character is calculated within the preset interaction time to dynamically update the relationship parameters between the player and the player's chat character. For example, if the player and the player's chat character have an active verbal communication, the corresponding weight values ​​of intimacy and trust are +5, and if an attack behavior is performed, the corresponding weight value of hostility is -10. Therefore, the relationship parameter = intimacy parameter + trust parameter + hostility parameter = 55 + 65 + 10 = 130. It can be understood that the weight values ​​of the corresponding relationship parameters are pre-set by the developer according to the needs of the actual game content to improve the practicality of the chat character finally generated.

[0119] In this embodiment, according to the game design requirements and the characteristics of the player's chat character, a weight is assigned to the personality parameter and the relationship parameter respectively. For example, in a role-playing game with the theme of adventure and exploration, the player's chat character is a loyal partner of the player, has an independent emotional system, and can make emotional responses according to the player's behavior. Then the weight of the relationship parameter related to the player's behavior is correspondingly higher than the personality parameter, and the weight value of the player's real-time emotional state parameter should be higher than the personality parameter and the relationship parameter, so that the player's chat character can make emotional responses in time. In one embodiment, the weight of the real-time emotional state parameter is 45%, the weight of the relationship parameter is 40%, and the weight of the personality parameter is 15%. Thus, the emotional expression parameter of the player's chat character is obtained by weighted calculation. Specifically, taking the above-mentioned obtained player's real-time emotional state parameter as the emotional category of happy, the emotional intensity of strong, and the duration of 1 minute as an example, the numerical calculation of its emotional expression parameter is, emotional expression parameter = 19.8 × 45% + 130 × 40% + 24 × 15% = 64.51.

[0120] S400, mapping the emotion expression parameters to the role attributes of the player's companion chat character, so that the player's companion chat character generates a corresponding emotion expression method, wherein the emotion expression method includes facial expressions, intonation and body movements;

[0121] In this embodiment, the emotion expression parameters are passed to the behavior generation module, and the behavior generation module is provided with parameter values ​​of the emotion expression methods corresponding to the emotion expression parameters. Specifically, the parameter ranges corresponding to facial expressions, intonations and body movements are all 0-1.

[0122] In some embodiments, the step S400 includes:

[0123] A first mapping function, a second mapping function and a third mapping function are respectively set to map the emotion expression parameters to facial expression parameters, tone of voice parameters and body movement parameters in sequence;

[0124] Based on the facial expression parameters, the tone and intonation parameters and the body movement parameters, corresponding facial expressions, tone and intonation and body movement are generated by a behavior generation module.

[0125] Specifically, the first mapping function corresponding to the facial expression parameters is as follows:

[0126] FEP=a(e bx -1)

[0127] The second mapping function corresponding to the tone and intonation parameters is as follows:

[0128] TVP=c(1-e -dx )

[0129] The third mapping function corresponding to the limb movement parameters is as follows:

[0130] BMP=e fx -g

[0131] Among them, FEP is the facial expression parameter, TVP represents the tone and intonation parameter, BMP represents the body movement parameter, x represents the emotion expression parameter, a, b, c, d, f and g are all constants, which are used to adjust the shape and range of each mapping function. The second mapping function increases with the increase of the emotion expression parameter, but the growth rate gradually slows down to simulate the gradual change of tone and intonation from calm to cheerful. Similarly, the third mapping function increases rapidly with the increase of EEP to simulate the drastic change of body movement from static to active.

[0132] In one embodiment, in order to ensure that the values ​​of FEP, TVP and BMP are between 0 and 1, the corresponding parameter values ​​greater than 1 need to be normalized.

[0133] In this embodiment, the specific values ​​of a, b, c, d, f and g are determined by experiments or optimization algorithms. In some preferred embodiments, the developer makes detailed adjustments according to the overall style of the game, the character of the player's chat companion, and the need for emotional expression.

[0134] In a preferred embodiment, a=0.8, b=0.01, c=0.7, d=0.02, f=0.03, g=5, then when the above-mentioned emotion expression parameter is 64.51, the facial expression parameter FEP=0.72, the tone and intonation parameter TVP=0.5, and the body movement parameter BMP=1.93 are calculated, and the normalization algorithm is used to normalize the body movement language so that its parameter range value is within the range of 0-1. Specifically, after normalization calculation, the body movement parameter BMP=0.81, and then, the behavior generation module generates corresponding facial expressions, tone and intonation, and body movements according to the calculated values ​​of FEP, TVP and BMP.

[0135] S500: for different game types, based on the character emotion expression rule library, establish an emotion expression strategy model, wherein the emotion expression strategy model is used to dynamically adjust the emotion expression mode for the player's chat companion character according to the characteristics of the game type and the player's preference;

[0136] In this embodiment, an emotion expression strategy model is established for different types of games to dynamically adjust the emotion expression method of the player's chat character, ensuring that the player's chat character can maintain highly consistent emotional interaction with the player in different game environments.

[0137] In some embodiments, the step S500 includes:

[0138] During the process of the player chatting with the character to express emotions, the player's feedback data is continuously collected, the feedback data is analyzed using a support vector machine algorithm, and the character's emotional expression rule library is optimized, wherein the feedback data includes the player's emotional changes and behavior patterns;

[0139] A decision tree model is used to construct an emotion expression strategy model, and the game type, scene characteristics and player group are input into the emotion expression strategy model, each node of the emotion expression strategy model represents a judgment condition, and each leaf node represents an emotion expression strategy;

[0140] Based on the emotion expression strategy, selecting a corresponding emotion expression method from the optimized character emotion expression rule library;

[0141] A reinforcement learning algorithm is used to dynamically adjust the parameters of the emotion expression strategy model according to the player's feedback data.

[0142] In this embodiment, the player's feedback data, including the player's emotional changes, behavior patterns, etc., is continuously collected, and the feedback data is analyzed through the support vector machine algorithm to continuously optimize and improve the emotion expression rule library.

[0143] In this embodiment, the game type, scene features and player group type are obtained. Specifically, the current game type is read from the game database, the current scene feature data is obtained from the game engine, and the player group type is obtained from the user database, and then the emotion expression strategy is selected based on the decision tree model. A pre-trained decision tree model is used, and the input of the decision tree model is the game type, scene features and player group, and the output is the corresponding emotion expression strategy. Each node of the decision tree represents a judgment condition, such as whether the game type is RPG, whether the scene feature is a combat state, whether the player's emotion is an excited state, etc. Each leaf node represents an emotion expression strategy, such as expressing happiness, expressing sadness, expressing anger, etc., which corresponds to the emotional portrait of the player's chat character. Then, according to the selected emotion expression strategy, the specific emotion expression method is determined. If the emotion expression strategy is to express happiness, an expression method is randomly selected from the preset happy expression method library, such as playing a happy animation, making a laugh, showing a happy expression, etc. It can be understood that the happy expression method library contains various types of expressions, such as animation, sound, text, expression, etc. Then, the game character is controlled by the game engine to execute the selected emotion expression method. For example, if the emotional expression is to play a happy animation, the game engine will play the preset happy animation. If the emotional expression is to laugh, the game engine will play the preset laughter sound effect.

[0144] In this embodiment, the emotional expression strategy model is dynamically adjusted according to player feedback. Specifically, the player's feedback data on the emotional expression of the player's chat character during the game, such as emotional changes and behavior patterns, is collected. A reinforcement learning algorithm is used to dynamically adjust the parameters of the decision tree model according to the player's feedback data to improve the accuracy and effectiveness of the emotional expression strategy selection.

[0145] In some embodiments, the expression library corresponding to multiple emotions is regularly updated according to game updates and player needs, such as adding new expressions and deleting outdated expressions, to keep the emotional expressions of game characters fresh and attractive.

[0146] S600: During the interaction between the player and the player's companion character, the player's feedback data is captured in real time, the player's emotional state is analyzed through the feedback data, and based on the emotional state, the interaction strategy and character attributes of the player's companion character are dynamically adjusted.

[0147] In this embodiment, during the interaction between the player and the companion character, the player's feedback data is captured and analyzed in real time, the player's emotional state is further analyzed, and the interaction strategy and role attributes of the companion character are dynamically adjusted accordingly, ensuring that the character can always keep in sync with the player's emotions, thereby establishing a closer and deeper emotional connection.

[0148] In some embodiments, the step S600 includes:

[0149] Obtaining the conversation content between the player and the character that the player chats with, and obtaining the emotional changes of the player by analyzing the conversation content;

[0150] By analyzing the player's operation records during the game, the corresponding behavior patterns of the players are obtained;

[0151] Determining the player's emotional state and topic preference based on the emotional changes and the behavioral patterns;

[0152] Constructing an emotional knowledge graph, wherein the emotional knowledge graph includes emotional semantic information, interaction strategies, and emotional tendency characteristics under different emotional states;

[0153] According to the emotional state of the player, corresponding emotional semantic information is obtained from the emotional knowledge graph;

[0154] In the process of generating the dialogue of the player's companion chat character, the emotional semantic information is incorporated to adjust the emotional expression of the interactive language of the player's companion chat character;

[0155] Based on the topic preference of the player, the dialogue strategy of the player's chat companion role is dynamically adjusted by adjusting the role attributes of the player's chat companion role, and the dialogue strategy includes topic selection and response method.

[0156] In this embodiment, the player's conversation content and operation records are obtained. The conversation content data is obtained from the chat records and voice records between the player and the player's companion character in the game. The operation record is the operation frequency and operation mode in the historical game data, such as movement, attack, transaction, click, share, like, etc. The natural language processing technology is used to perform semantic understanding and sentiment analysis on the player's conversation content to capture the player's emotional changes. By analyzing the player's operation records during the game, the corresponding behavior pattern of the player is obtained, such as frequent attacking behavior, long-term stillness, sharing or clicking, etc.

[0157] In this embodiment, the conversation content in which players participate is analyzed, and topic models such as LDA are used to identify hot topics and points of interest. The behavior patterns of players after discussing specific topics, such as clicks, sharing, etc., are examined to determine the topic preferences of players.

[0158] In this embodiment, for the emotional knowledge graph, common emotional words, phrases and expressions in different emotional states are sorted out to form an emotional semantic library that includes a variety of emotional semantic information. Different interaction strategies are formulated according to the emotional state, such as using light-hearted and humorous language under positive emotions and providing comfort and support under negative emotions.

[0159] In this embodiment, emotional computing technology is used to analyze the content of player conversations, obtain emotional words related to emotional changes in the conversation content, match the emotional words in the emotional knowledge graph, obtain emotional semantic information, and incorporate emotional semantic information when generating conversations for the player's chatting character to make the conversation more closely aligned with the player's emotional state.

[0160] According to the player's topic preferences, the character attributes of the chat character, such as personality, interests, occupation and corresponding emotional expression parameters, are adjusted to better communicate with the player on the topic. At the same time, the dialogue strategy is dynamically adjusted according to the topic preferences, including topic selection and response methods. For example, if the player prefers game topics, the chat character can have more conversations around the game, provide game strategies, share game experiences, etc. In addition, the knowledge base of the player's chat character can be dynamically expanded according to the number of conversation rounds on a certain topic by the player to meet the player's chat needs.

[0161] See also Figure 2 As shown, the present invention also provides a system for constructing a player chat companion character, the system comprising:

[0162] The first processing module 201 is used to construct a player emotion model based on historical player game data, and the player emotion model is used to obtain the player's real-time emotion state parameters;

[0163] The second processing module 202 is used to perform cluster analysis on historical player game data using a clustering algorithm to generate multiple player groups, analyze the behavioral characteristics of each of the player groups, and generate emotional tendency characteristics corresponding to the player groups;

[0164] The third processing module 203 is used to generate the emotion expression parameters of the player's chat companion character based on the preset character emotion expression rule library and the real-time emotion state parameters and the emotion tendency characteristics;

[0165] The fourth processing module 204 is used to map the emotion expression parameters to the character attributes of the player's chat companion character, so that the player's chat companion character generates a corresponding emotion expression method, and the emotion expression method includes facial expressions, intonation and body movements;

[0166] The fifth processing module 205 is used to establish an emotion expression strategy model based on the character emotion expression rule library for different game types, and the emotion expression strategy model is used to dynamically adjust the emotion expression mode for the player's chat character according to the characteristics of the game type and the player's preference;

[0167] The sixth processing module 206 is used to capture the player's feedback data in real time during the interaction between the player and the player's companion character, analyze the player's emotional state through the feedback data, and dynamically adjust the interaction strategy and character attributes of the player's companion character based on the emotional state.

[0168] It is understandable that if Figure 1 The contents of the method embodiment for constructing a player chat companion character shown in the figure are applicable to the system embodiment for constructing a player chat companion character. The functions specifically implemented by the system embodiment for constructing a player chat companion character are similar to those in the example above. Figure 1 The method for constructing a player chat character is the same as the embodiment shown in the figure, and the beneficial effects achieved are the same as those of the embodiment shown in the figure. Figure 1 The beneficial effects achieved by the method embodiment of constructing a player chat companion character shown are also the same.

[0169] It should be noted that the information interaction, execution process and other contents between the above-mentioned systems are based on the same concept as the embodiment of the method of the present invention. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0170] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0171] See also Figure 3 As shown, an embodiment of the present invention further provides a computer device 3, comprising: a memory 302 and a processor 301 and a computer program 303 stored on the memory 302. When the computer program 303 is executed on the processor 301, a method for constructing a player chat character as described in any one of the above methods is implemented.

[0172] The computer device 3 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The computer device 3 may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art will appreciate that Figure 3 It is only an example of computer device 3 and does not constitute a limitation on computer device 3. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components, for example, it may also include input and output devices, network access devices, etc.

[0173] The processor 301 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0174] In some embodiments, the memory 302 may be an internal storage unit of the computer device 3, such as a hard disk or memory of the computer device 3. In other embodiments, the memory 302 may also be an external storage device of the computer device 3, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device 3. Further, the memory 302 may also include both an internal storage unit of the computer device 3 and an external storage device. The memory 302 is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program, etc. The memory 302 may also be used to temporarily store data that has been output or is to be output.

[0175] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for constructing a player chat companion character as described in any one of the above methods is implemented.

[0176] In this embodiment, if the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the camera / computer device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. For example, USB flash drive, mobile hard disk, disk or optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0177] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for constructing a player chat character, characterized in that: include: Based on historical player game data, a player emotion model is constructed, wherein the player emotion model is used to obtain the player's real-time emotion state parameters; Using a clustering algorithm to perform cluster analysis on historical player game data to generate multiple player groups, analyzing the behavioral characteristics of each player group, and generating emotional tendency characteristics corresponding to the player group; Based on a preset character emotion expression rule library, generating emotion expression parameters of the player's chat companion character according to the real-time emotion state parameters and the emotion tendency characteristics; Mapping the emotion expression parameters to the character attributes of the player's companion chat character, so that the player's companion chat character generates a corresponding emotion expression method, wherein the emotion expression method includes facial expressions, tone of voice and body movements; For different game types, an emotion expression strategy model is established based on the character emotion expression rule library. The emotion expression strategy model is used to dynamically adjust the emotion expression mode for the player's chat character according to the characteristics of the game type and the player's preferences; During the interaction between the player and the player's companion character, the player's feedback data is captured in real time, the player's emotional state is analyzed through the feedback data, and based on the emotional state, the interaction strategy and character attributes of the player's companion character are dynamically adjusted.

2. The method according to claim 1, characterized in that The player emotion model is constructed based on the historical player game data, and the player emotion model is used to obtain the player's real-time emotion state parameters, including: Based on psychology theories and game content, emotion labels and emotion characteristics are defined, wherein the emotion labels include happiness, frustration and anger, and the emotion characteristics include emotion keywords, emotion symbols and operation modes; Extracting the emotion tag and the emotion feature from historical player game data to form an emotion feature vector; Using a machine learning algorithm, the emotion feature vector is input into a player emotion model for training, so that the player emotion model can output real-time emotion state parameters according to real-time player behavior data, and the real-time emotion state parameters include emotion category, emotion intensity and duration; A scoring system is constructed for the real-time emotional state parameter, and based on the scoring system, an emotional state score of the emotional state parameter is calculated.

3. The method according to claim 1, characterized in that The historical player game data includes historical game data generated during the game process of multiple players, and the historical game data includes operation frequency, operation type, game duration, game area, interactive object, task completion status and resource acquisition status; The clustering algorithm is used to perform cluster analysis on historical player game data to generate multiple player groups, and the behavioral characteristics of each player group are analyzed to generate emotional tendency characteristics corresponding to the player group, including: Preprocessing the historical player game data to obtain a player behavior data set; Based on the operation frequency, the operation type, the game duration and the task completion status, a clustering algorithm is used to divide the multiple players into multiple player groups; Analyze the differences in historical game data of each of the player groups to generate behavioral characteristics corresponding to the player groups, wherein the behavioral characteristics include goal-oriented characteristics, social interaction characteristics, and continuous investment characteristics; The behavioral characteristics are subjected to correlation analysis to obtain the emotional tendency characteristics related to the flow experience characteristics in the behavioral characteristics.

4. The method according to claim 3, characterized in that The performing of correlation analysis on the behavioral characteristics to obtain the emotional tendency characteristics in the behavioral characteristics that are related to the flow experience characteristics includes: By calculating the correlation coefficient between the behavior characteristics and the flow experience index, obtaining the correlation degree index between the behavior characteristics and the flow experience characteristics, the flow experience index includes satisfaction, immersion and task completion efficiency; Based on the correlation index, a machine learning algorithm is used to evaluate the importance of the behavioral characteristics in predicting the player's flow experience level, and an importance index is generated; Based on the importance index, the plurality of behavioral features are sorted, and the first one or more behavioral features are selected as emotional tendency features.

5. The method according to claim 1, characterized in that The method of generating the emotion expression parameters of the player's chat companion character based on the preset character emotion expression rule library according to the real-time emotion state parameters and the emotion tendency characteristics includes: Based on the preset role emotion expression rule library, a rule matching algorithm is used to match an emotion portrait for the player's chat companion role according to the real-time emotion state parameters and the emotion tendency characteristics; Based on the game scene, determining the degree of matching between the game scene and the emotional portrait, and dynamically updating the emotional portrait; Obtaining a personality parameter matrix of the player's companion chat character, and determining personality parameters of the player's companion chat character based on the emotional portrait, wherein the personality parameter matrix includes extroversion, openness, and conscientiousness; Defining the relationship parameters between the player's companion character and the player, the relationship parameters include intimacy, trust and hostility, and dynamically updating the relationship parameters according to the player's feedback behavior, the feedback behavior includes verbal communication, action interaction and decision-making selection; Based on the personality parameter, the relationship parameter and the real-time emotional state parameter, the emotional expression parameter of the player's chatting character is obtained through weighted calculation.

6. The method according to claim 1, characterized in that The mapping of the emotion expression parameters to the character attributes of the player's companion chat character so that the player's companion chat character generates a corresponding emotion expression method, wherein the emotion expression method includes facial expressions, intonation and body movements, including: A first mapping function, a second mapping function and a third mapping function are respectively set to map the emotion expression parameters to facial expression parameters, tone of voice parameters and body movement parameters in sequence; Based on the facial expression parameters, the tone and intonation parameters and the body movement parameters, corresponding facial expressions, tone and intonation and body movement are generated by a behavior generation module.

7. The method according to claim 1, characterized in that For different game types, based on the character emotion expression rule library, an emotion expression strategy model is established, and the emotion expression strategy model is used to dynamically adjust the emotion expression mode for the player's chat companion character according to the characteristics of the game type and the player's preference, including: During the process of the player chatting with the character to express emotions, the player's feedback data is continuously collected, the feedback data is analyzed using a support vector machine algorithm, and the character's emotional expression rule library is optimized, wherein the feedback data includes the player's emotional changes and behavior patterns; A decision tree model is used to construct an emotion expression strategy model, and the game type, scene characteristics and player group are input into the emotion expression strategy model, each node of the emotion expression strategy model represents a judgment condition, and each leaf node represents an emotion expression strategy; Based on the emotion expression strategy, selecting a corresponding emotion expression method from the optimized character emotion expression rule library; A reinforcement learning algorithm is used to dynamically adjust the parameters of the emotion expression strategy model according to the player's feedback data.

8. The method according to claim 1, characterized in that During the interaction between the player and the player's companion character, the player's feedback data is captured in real time, the player's emotional state is analyzed through the feedback data, and the interaction strategy and character attributes of the player's companion character are dynamically adjusted based on the emotional state, including: Obtaining the conversation content between the player and the character that the player chats with, and obtaining the emotional changes of the player by analyzing the conversation content; By analyzing the player's operation records during the game, the corresponding behavior patterns of the players are obtained; Determining the player's emotional state and topic preference based on the emotional changes and the behavioral patterns; Constructing an emotional knowledge graph, wherein the emotional knowledge graph includes emotional semantic information, interaction strategies, and emotional tendency characteristics under different emotional states; According to the emotional state of the player, corresponding emotional semantic information is obtained from the emotional knowledge graph; In the process of generating the dialogue of the player's companion chat character, the emotional semantic information is incorporated to adjust the emotional expression of the interactive language of the player's companion chat character; Based on the topic preference of the player, the dialogue strategy of the player's chat companion role is dynamically adjusted by adjusting the role attributes of the player's chat companion role, and the dialogue strategy includes topic selection and response method.

9. A system for constructing a player chat character, characterized in that: include: The first processing module is used to construct a player emotion model based on historical player game data, and the player emotion model is used to obtain the player's real-time emotion state parameters; The second processing module is used to perform cluster analysis on historical player game data using a clustering algorithm to generate multiple player groups, analyze the behavioral characteristics of each of the player groups, and generate emotional tendency characteristics corresponding to the player groups; The third processing module is used to generate the emotional expression parameters of the player's chat companion character based on the preset character emotional expression rule library and the real-time emotional state parameters and the emotional tendency characteristics; The fourth processing module is used to map the emotion expression parameters to the character attributes of the player's chat companion character, so that the player's chat companion character generates a corresponding emotion expression method, and the emotion expression method includes facial expressions, intonation and body movements; The fifth processing module is used to establish an emotion expression strategy model based on the character emotion expression rule library for different game types, and the emotion expression strategy model is used to dynamically adjust the emotion expression mode for the player's chat character according to the characteristics of the game type and the player's preference; The sixth processing module is used to capture the player's feedback data in real time during the interaction between the player and the player's companion character, analyze the player's emotional state through the feedback data, and dynamically adjust the interaction strategy and character attributes of the player's companion character based on the emotional state.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

Images