In-game advertisement putting method and device, electronic equipment and storage medium

Abstract

The invention provides an in-game advertisement putting method and device, electronic equipment and a storage medium, and relates to the technical field of machine learning. The method comprises the following steps: collecting game behavior data and advertisement interaction data of a game user and equipment feature data of user equipment; creating a heterogeneous graph based on the game behavior data, the advertisement interaction data and the equipment feature data, and training the heterogeneous graph based on a graph neural network to obtain a cross-modal association model; determining a target advertisement for the game user based on the cross-modal association model; determining an immersion coefficient of the game user based on the game behavior data, and evaluating the advertisement value of the target advertisement based on the immersion coefficient; and determining an advertisement putting strategy of the target advertisement based on the advertisement value, and pushing the target advertisement to the game user based on the advertisement putting strategy. According to the invention, cross-modal feature fusion is carried out on the multi-source data and the advertisement value is evaluated, so that the advertisement putting strategy can be dynamically adjusted, and the accuracy and effect of advertisement putting are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning, and particularly relates to a method, device, electronic device and storage medium for in-game advertisement placement. Background Art

[0002] With the rapid growth of the game and entertainment industries, the digital advertising ecosystem shows a trend of cross-platform integration. The game scenario has become a new front for advertisement placement due to its high user stickiness.

[0003] In related technologies, due to the relatively single data source relied on by the advertisement placement system, there will be a problem of low accuracy in advertisement placement. In addition, when the current advertisement placement system evaluates the value of an advertisement, it usually takes click-through rate and conversion rate as the core indicators, without considering the impact of the immersion degree of game users in the game scenario on the acceptance degree of the advertisement, which is not conducive to the success rate of advertisement conversion. Summary of the Invention

[0004] In view of this, the embodiments of the present disclosure propose a method, device, electronic device and computer-readable storage medium for in-game advertisement placement to solve the problems in related technologies, that is, due to the relatively single data source relied on by the advertisement placement system, there will be a problem of low accuracy in advertisement placement; and when the current advertisement placement system evaluates the value of an advertisement, it usually takes click-through rate and conversion rate as the core indicators, without considering the impact of the immersion degree of game users in the game scenario on the acceptance degree of the advertisement, which is not conducive to the success rate of advertisement conversion.

[0005] According to a first aspect of the present disclosure, there is provided a method for in-game advertisement placement, including: collecting game behavior data and advertisement interaction data of a game user, and device feature data of the user device of the game user; creating a heterogeneous graph based on the game behavior data, the advertisement interaction data and the device feature data, and training the heterogeneous graph based on a graph neural network to obtain a cross-modal association model; determining a target advertisement for the game user based on the cross-modal association model; determining an immersion coefficient of the game user based on the game behavior data, and evaluating the advertisement value of the target advertisement based on the immersion coefficient; determining an advertisement placement strategy for the target advertisement based on the advertisement value, and pushing the target advertisement to the game user based on the advertisement placement strategy.

[0006] According to a second aspect of the present disclosure, there is provided an in-game advertisement placement device, including: a data acquisition module, configured to acquire game behavior data and advertisement interaction data of game users, as well as device feature data of the user devices of the game users; a data processing module, configured to create a heterogeneous graph based on the game behavior data, the advertisement interaction data, and the device feature data, and train the heterogeneous graph based on a graph neural network to obtain a cross-modal association model; an advertisement screening module, configured to determine a target advertisement for the game user based on the cross-modal association model; a value evaluation module, configured to determine an immersion coefficient of the game user based on the game behavior data, and evaluate the advertisement value of the target advertisement based on the immersion coefficient; and an advertisement placement module, configured to determine an advertisement placement strategy of the target advertisement based on the advertisement value, and push the target advertisement to the game user based on the advertisement placement strategy.

[0007] According to a third aspect of the present disclosure, there is provided an electronic device, including: a processor; and a memory storing a program, wherein the program includes instructions that, when executed by the processor, cause the processor to execute the above-mentioned in-game advertisement placement method.

[0008] According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the above-mentioned in-game advertisement placement method.

[0009] The exemplary embodiments of the present disclosure may have some or all of the following beneficial effects:

[0010] In the in-game advertisement placement method provided by the exemplary embodiment of the present disclosure, game behavior data, advertisement interaction data of game users, and device feature data of the user devices of game users are collected; a heterogeneous graph is created based on the game behavior data, advertisement interaction data, and device feature data, and the heterogeneous graph is trained based on a graph neural network to obtain a cross-modal association model; a target advertisement for the game user is determined based on the cross-modal association model; an immersion coefficient of the game user is determined based on the game behavior data, and the advertisement value of the target advertisement is evaluated based on the immersion coefficient; an advertisement placement strategy for the target advertisement is determined based on the advertisement value, and the target advertisement is pushed to the game user based on the advertisement placement strategy. By collecting game behavior data, advertisement interaction data, and device feature data and constructing a heterogeneous graph based on the above multi-source data, the present disclosure can perform multi-modal feature fusion on the heterogeneous graph through a graph neural network, so as to obtain a target advertisement that better matches the game user, improving the accuracy and conversion rate of advertisement placement; in addition, the present disclosure determines the immersion coefficient of the game user through the game behavior data, and considers the immersion coefficient when evaluating the advertisement value, which helps to determine a more suitable advertisement strategy, helps to increase the willingness of users to click on advertisements, and better exerts the effect of advertisements. Brief Description of the Drawings

[0011] In the following description of exemplary embodiments with reference to the accompanying drawings, more details, features, and advantages of the present disclosure are disclosed. In the drawings:

[0012] Figure 1 A flowchart of an in-game advertisement placement method according to an exemplary embodiment of the present disclosure is shown;

[0013] Figure 2 A schematic diagram of a heterogeneous graph created in an in-game advertisement placement method according to an exemplary embodiment of the present disclosure is shown;

[0014] Figure 3 A schematic block diagram of an in-game advertisement placement device according to an exemplary embodiment of the present disclosure is shown;

[0015] Figure 4 A structural block diagram of an exemplary electronic device capable of implementing the embodiments of the present disclosure is shown. Detailed Description of the Embodiments

[0016] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Instead, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0017] It should be understood that the various steps described in the method embodiments of the present disclosure may be executed in different orders and / or executed in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0018] As used herein, the term "including" and its variations are open-ended, that is, "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description. It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0019] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0020] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0021] The embodiments of the present disclosure provide a method, device, electronic device and computer-readable storage medium for in-game advertisement placement. The following describes the solution of the present disclosure with reference to the accompanying drawings.

[0022] The embodiments of the present disclosure first provide a method for in-game advertisement placement. Refer to Figure 1 As shown, the method for in-game advertisement placement specifically includes the following steps:

[0023] S110: Collect the game behavior data and advertisement interaction data of game users, as well as the device feature data of the user devices of game users;

[0024] S120: Create a heterogeneous graph based on the game behavior data, advertisement interaction data and device feature data, and train the heterogeneous graph based on a graph neural network to obtain a cross-modal association model;

[0025] S130: Determine the target advertisement for game users based on the cross-modal association model;

[0026] S140: Determine the immersion coefficient of game users based on the game behavior data, and evaluate the advertisement value of the target advertisement based on the immersion coefficient;

[0027] S150: Determine the advertising placement strategy for the target advertisement based on the advertising value, and push the target advertisement to the game users based on the advertising placement strategy.

[0028] The in-game advertising placement method provided by the exemplary embodiments of the present disclosure collects game behavior data, advertisement interaction data, and device feature data of game users, and constructs a heterogeneous graph based on the above multi-source data. Thus, multi-modal feature fusion can be performed on the heterogeneous graph through a graph neural network, so that a target advertisement more suitable for the game users can be obtained, improving the accuracy and conversion rate of advertising placement. In addition, the present disclosure determines the immersion coefficient of game users through game behavior data, and considers this immersion coefficient when evaluating the advertising value, which helps to determine a more suitable advertising strategy, helps to increase the willingness of users to click on advertisements, and better exerts the effect of advertisements.

[0029] Next, in another embodiment, the above steps will be described in more detail.

[0030] In step S110, collect the game behavior data and advertisement interaction data of game users, as well as the device feature data of the user devices of game users.

[0031] In the embodiments of the present application, the above game users are users participating in video games, and the above video games refer to game forms carried out through electronic devices such as televisions, computers, game consoles, mobile phones, etc. The above user devices are the televisions, computers, game consoles, mobile phones, etc. used by the above users when participating in video games.

[0032] In the embodiments of the present application, the above game behavior data refers to various data generated by game users during the game process. Exemplarily, the game behavior data may include basic information data, game operation data, social interaction data, and economic data. Among them, the above basic information data may include information such as the age, gender, region, and registration time of game users; the above game operation data may include information such as the frequency and position of clicks, swipes, and key operations of game users, the movement speed and direction of characters, and the timing and sequence of skill releases; the above social interaction data may include information such as the number of friends added by game users, the frequency of teaming up with friends to play games, chat content and times, and the number of posts and replies in the game community; the above economic data may include information such as the quantity and acquisition channels of virtual currencies such as gold coins and diamonds obtained by game users, the types and frequencies of purchased items, skins, and gift packages, and the transaction data in the in-game trading market.

[0033] In the embodiments of the present application, the above-mentioned advertisement interaction data is data used to measure the advertisement effect and the engagement of game users. Exemplarily, the advertisement interaction data may include click data, viewing duration data, interaction behavior data, and conversion data. Among them, the above-mentioned click data may include information such as the number of times, time, and location of a user clicking on an advertisement. By analyzing the click data, the impact of different advertisement creatives and placement positions on the user's click behavior can be understood; the above-mentioned viewing duration data is used to reflect the degree of attention of the user to the advertisement content; the above-mentioned interaction behavior data may include behavior data such as the user's like, comment, share, and favorite on the advertisement; the above-mentioned conversion data is used to record the process of the user from seeing the advertisement to completing the target behavior, and the above-mentioned target behavior refers to specific behaviors such as purchasing a product, registering an account, and downloading an application that the advertisement guides the user to perform.

[0034] In the embodiments of the present application, the above-mentioned device feature data refers to data information that can reflect various characteristics and attributes of the above-mentioned user device. Exemplarily, the device feature data may include hardware parameters, network connection data, and sensor data. Among them, the above-mentioned hardware parameters may include information such as the brand, model, processor model, memory capacity, storage capacity, screen size, screen resolution, and camera parameters of the user device; the above-mentioned network connection data may include information such as network type, network speed, and signal strength; the user device may be equipped with sensors such as an accelerometer, gyroscope, magnetometer, and light sensor, and the above-mentioned sensor data is the data generated by the sensors in the user device and can be used for the interactive design of the game, such as a motion-sensing game using the accelerometer and gyroscope to achieve action control.

[0035] In the embodiments of the present application, collecting the game behavior data, advertisement interaction data of game users, and device feature data of the user devices of game users can be implemented as follows by way of example: collecting the above-mentioned game behavior data through a game behavior data logger; collecting the above-mentioned advertisement interaction data through an advertisement interaction probe; collecting the above-mentioned device feature data through the above-mentioned device feature collector. Among them, the above-mentioned game behavior data logger collects the above-mentioned game behavior data by implanting specific code snippets at key positions in the game code. Specifically, the scene stay duration, virtual item interaction trajectory, and achievement completion path can be collected through the game behavior data logger. The above-mentioned achievement completion path is a tree-like data structure used to describe the task logic relationship in the game, representing the dependency relationship between tasks; the above-mentioned advertisement interaction probe is a tool for monitoring and analyzing the interaction between game users and advertisements. Specifically, the cross-platform Software Development Kit (SDK) integration data, advertisement contact coordinates, and interaction depth classification during the interaction between game users and advertisements can be obtained through the advertisement interaction probe. The above-mentioned SDK integration tool is used to collect data such as the system, game coordinate points, memory, frame rate, model, click position, and viewport size in real time; the above-mentioned device feature collector is a tool for collecting various feature information of user devices. Specifically, the hardware fingerprint and dynamic performance metrics can be obtained by calling the interfaces provided by the user device operating system through the device feature collector. Among them, the hardware fingerprint is an identifier based on the hardware features of the user device. By way of example, it can include GPU model, memory capacity, and network type. The dynamic performance metrics are used to measure the performance of the user device during operation. By way of example, it can include data such as frame rate volatility, battery temperature, and storage space margin.

[0036] In the embodiments of the present application, by comparing the viewport size and click position data of different platforms, the advertisement click hotspots of game users can be determined and the visual focus position of the advertisement material can be optimized accordingly. By way of example, if it is found through comparing the viewport size and click position data of different platforms that the advertisement click hotspots of users of a certain type of device are concentrated in the upper half of the screen, then the visual focus position of the advertisement material can be focused on the upper half of the screen area.

[0037] In the embodiments of the present application, the above-mentioned interaction depth classification is used to clarify different interaction depth levels and their corresponding user behaviors. For example, the first-level interaction can be defined as an advertisement click, the second-level interaction can be defined as staying on a new page for more than a certain period of time after clicking an advertisement, and the third-level interaction can be defined as completing specific operations such as downloading a game and registering an account under the guidance of an advertisement. Further, after obtaining the interaction depth classification, the advertisement placement strategy can also be optimized based on the interaction depth classification. For example, when the exposure click conversion rate of an advertisement material is lower than the industry average, the following strategies are automatically triggered: inserting an advertisement during the 3-second prime time after the user completes a BOSS battle; replacing the advertisement material with a version that matches the current game scene style.

[0038] In step S120, a heterogeneous graph is created based on game behavior data, advertisement interaction data, and device feature data, and the heterogeneous graph is trained based on a graph neural network to obtain a cross-modal association model.

[0039] In the embodiments of the present application, the above-mentioned heterogeneous graph is a graph structure including various types of nodes and edges. Exemplarily, the heterogeneous graph constructed in the embodiments of the present application can be as Figure 2 shown, including game nodes, advertisement nodes, and device nodes created based on the above-mentioned game behavior data, advertisement interaction data, and device feature data. Among them, the game nodes can contain information such as game IDs and scene IDs, representing different games or game scenes; the device nodes are identified by device fingerprint hashes and contain hardware configuration information such as GPU models and memory capacities, representing different user devices; the advertisement nodes carry advertisement IDs, placement IDs, etc., representing different advertisement contents and placement positions; the connection relationship of the edges is determined according to the interaction behaviors between the game nodes, advertisement nodes, and device nodes. For example, based on the interaction behaviors generated by game users with advertisements during the game process, such as clicking an advertisement, a "click" edge is established between the game node and the advertisement node; when an advertisement is exposed on a user device, an "exposure" edge is formed between the advertisement node and the device node.

[0040] Exemplarily, creating a heterogeneous graph based on game behavior data, advertisement interaction data, and device feature data can be achieved as follows: extracting the user behavior sequence data of game users from the game behavior data, and encoding the user behavior sequence data through a long short-term memory network to obtain a first multi-dimensional vector; extracting the advertisement material features from the advertisement interaction data, and encoding the advertisement material features through word embedding encoding to obtain a second multi-dimensional vector; performing dimensionality reduction encoding on the device feature data through principal component analysis to obtain a third multi-dimensional vector; creating game nodes, advertisement nodes, and device nodes, taking the first multi-dimensional vector, the second multi-dimensional vector, and the third multi-dimensional vector as the node attributes of the game nodes, advertisement nodes, and device nodes respectively, and creating edges between the game nodes, advertisement nodes, and device nodes based on the game behavior data, advertisement interaction data, and device feature data to obtain a heterogeneous graph.

[0041] Specifically, extracting the user behavior sequence data of game users from the game behavior data and encoding the user behavior sequence data through a long short-term memory network to obtain a first multi-dimensional vector can be achieved as follows: constructing a user behavior sequence from the game behavior data collected by the game behavior tracer. Exemplarily, the above user behavior sequence data may include a scene switching time series and a task completion path. For example, the above scene switching time series may include a timestamp sequence of main city → dungeon → store, and the above task completion path may be the node order of task A → task B → hidden task C; capturing the long-term dependencies (such as continuous 3-hour high-activity behavior) in the constructed sequence through long short-term memory network (LSTM) modeling, and encoding the variable-length user behavior sequence data into a fixed 128-dimensional vector (i.e., the above first multi-dimensional vector), which is used to represent the user's game habits. Identifying core game users through this first multi-dimensional vector and pushing high-value advertisements to core game users at the end of the dungeon can help improve the advertisement conversion rate.

[0042] Extracting the advertisement material features in the above-mentioned advertisement interaction data and encoding the advertisement material features by means of word embedding encoding to obtain a second multi-dimensional vector can achieve the following: for the data collected through the advertisement interaction probe and the advertisement material features, in terms of structured features, information such as advertisement type (video / banner), ad position size (full screen / sidebar), etc. is extracted; in terms of unstructured features, the term frequency-inverse document frequency (TF-IDF) value of the advertisement copy in the advertisement material features is calculated, and the convolutional neural network (CNN) features of the material pictures in the advertisement material features are extracted. The above features cover the basic attributes and content characteristics of the advertisement; the above discrete advertisement attributes are mapped into 64-dimensional vectors in the continuous vector space by using the word vector model (Word2Vec) (i.e., the above-mentioned second multi-dimensional vector). In practical applications, scene-based advertisement matching can be achieved according to the cosine similarity between the first multi-dimensional vector and the second multi-dimensional vector, improving the fit between the advertisement and the game scene as well as the interests of game users. Also, the similarity between corresponding advertisements can be calculated according to the cosine similarity between the second multi-dimensional vectors, such as the vector distances of similar game advertisements being close.

[0043] The above-mentioned dimensionality reduction encoding of the device feature data by means of principal component analysis to obtain a third multi-dimensional vector can achieve the following: collecting hardware metric data such as GPU rendering frame rate, memory occupancy fluctuation value, network latency, device model encoding, etc.; using the principal component analysis (PCA) method to perform dimensionality reduction processing on the collected hardware metrics to obtain a 32-dimensional vector (i.e., the above-mentioned third multi-dimensional vector). PCA can eliminate the multicollinearity among the above-mentioned collected hardware metric data, such as the strong correlation relationship that may exist between the GPU model and the memory capacity. In addition, the dimensionality-reduced vector can be used to discover users of low-end devices, automatically reduce the advertisement resolution to improve fluency, and at the same time, it also helps to improve the efficiency and stability of model training.

[0044] After obtaining the above-mentioned first multi-dimensional vector, second multi-dimensional vector, and third multi-dimensional vector, the obtained vectors can be used as the node attributes of the game node, advertisement node, and device node respectively. Among them, the above-mentioned first multi-dimensional vector is an important attribute of the game node, describing the features and roles of the game node in the heterogeneous graph; the above-mentioned second multi-dimensional vector is an important attribute of the advertisement node, used to measure the association and similarity between the advertisement and other nodes; the above-mentioned third multi-dimensional vector is an important attribute of the device node, used to judge the device performance level (determining the advertisement material loading speed) and user usage habits (predicting the best advertisement display period).

[0045] In the embodiments of the present application, in the process of creating the heterogeneous graph based on the game behavior data, advertisement interaction data, and device feature data, the node weights of the nodes in the heterogeneous graph can also be determined. Exemplarily, the process of determining the nodes may include determining the weights of the game nodes and device nodes, and can be specifically implemented as follows: Obtain the scene residence duration of the game user in the current game scene from the game behavior data, and determine the game node weight of the game node based on the immersion coefficient and the scene residence duration; obtain the first performance coefficient of the image processing unit of the user device, the second performance coefficient of the device memory, and the third performance coefficient of the device network, and determine the device performance coefficient of the user device based on the first performance coefficient, the second performance coefficient, and the third performance coefficient; obtain the average daily usage duration of the user device, and determine the usage frequency coefficient of the user device based on the average daily usage duration; determine the device node weight of the device node based on the device performance coefficient and the usage frequency coefficient.

[0046] Among them, the above immersion coefficient is used to describe the immersion degree of the game user in the game. Exemplarily, the immersion coefficient can be determined in the following manner: Obtain the average single-game duration of the game user, the number of completed tasks in the game, and the total number of game tasks; determine the task completion density coefficient based on the number of completed tasks and the total number of game tasks of the game user; determine the immersion coefficient based on the average single-game duration and the task completion density coefficient.

[0047] Specifically, the above immersion coefficient (Immersion Index, Imm) can be determined by the following formula:

[0048] Imm = (average single-game duration / industry benchmark value) * task completion density coefficient

[0049] Among them, the above average single-game duration is used to reflect the average time consumed by the game user in one game process, and is a key indicator for measuring the game behavior and game immersion degree of the game user. It can be determined by the following formula: average single-game duration = total game duration in a specific period / total number of games. The specific period can be set to different durations such as one day, one week, one month, or one year according to the actual situation; the above industry benchmark value can be the value of the immersion coefficient obtained with reference to the industry report of a third-party data platform in the game industry; the above task completion density coefficient refers to the ratio of the number of tasks completed by the game user in a specific period to the total game duration in that period. It is an indicator for measuring the task completion density of the game user in the game.

[0050] After calculating the immersion coefficient through the above process, the node weight of the above game node can be determined by the following formula: Game node weight = Immersion coefficient * log(Scene stay duration); where the scene stay duration refers to the length of time that a game user stays in a specific scene in the game, which is generally recorded by the game background data statistics system. The timing starts when the game user enters a scene and stops when the game user leaves the scene.

[0051] The node weight of the above device node can be determined by the following formula: Device node weight = Performance coefficient * Usage frequency coefficient; where the above performance coefficient is used to measure the hardware performance level of the device and can be determined by the following formula: Performance coefficient = sigmoid(GPU score * 0.6 + Memory score * 0.3 + Network score * 0.1); the above usage frequency coefficient reflects the frequency of device usage and can be determined by the following formula: Usage frequency coefficient = 1 / (1 + e^(-0.1 * (Average daily usage duration - 4))).

[0052] In the embodiment of the present application, after the above heterogeneous graph is created, the created heterogeneous graph can be input into a graph neural network model (Graph Neural Network, GNN), and the DeepWalk algorithm is used to train the heterogeneous graph to obtain the above cross-modal association model.

[0053] In the embodiment of the present application, the input of the above cross-modal association model is the heterogeneous graph structure, node information, and node feature vectors. The node feature vectors include the above first multi-dimensional vector, second multi-dimensional vector, and third multi-dimensional vector. The output is a 128-dimensional node embedding vector generated by processing the heterogeneous graph using the DeepWalk algorithm. This node embedding vector integrates the node's own features, the relationship with neighbor nodes, and the graph structure information, and can be used to calculate node similarity to achieve advertisement matching.

[0054] In the embodiment of the present application, the above cross-modal association model can also be updated regularly. Specifically, the model update can be achieved by combining weekly incremental training and quarterly full-scale training. During the update process, the input is incremental data and historical data, and the output is the updated model. Processing new data through incremental training can maintain the timeliness of the model, and full-scale training can enable the model to relearn all data to adapt to long-term pattern changes.

[0055] In step S130, a target advertisement for the game user is determined based on the cross-modal association model.

[0056] In an embodiment of the present application, after training the cross-modal association model, the target advertisement for the game user can be determined through the node embedding vector output by the cross-modal association model, and the above target advertisement is the advertisement to be pushed to the game user in the game.

[0057] In an embodiment of the present application, exemplarily, determining the target advertisement for the game user based on the cross-modal association model can be implemented as follows: generating a user profile of the game user based on the cross-modal association model; determining the matching degree between each advertisement in the advertisement library and the user profile, and when the matching degree is greater than the first preset threshold, determining the corresponding advertisement as the target advertisement for the game user.

[0058] Specifically, generating the user profile of the game user based on the cross-modal association model may include the following steps:

[0059] S1: Feature extraction and fusion.

[0060] In this step, the trained cross-modal association model can learn the internal connection between different modal data, extract the fused feature vector from the middle layer or output layer of the model, and this feature vector contains multi-dimensional information of the game user.

[0061] S2: Portrait dimension division.

[0062] In this step, the user profile is divided into different dimensions according to actual needs and user characteristics. Exemplarily, the above portrait dimensions may include basic attribute dimensions, game behavior dimensions, and social attribute dimensions. Among them, the above basic attribute dimensions include information such as age, gender, and region, the above game behavior dimensions include information such as game preferences, payment habits, and activity levels, and the above social attribute dimensions may include information such as the number of friends, social circles, and teaming frequencies.

[0063] S3: Label assignment and portrait presentation.

[0064] In this step, for the user profile of each dimension above, analyze the game user according to the fused feature vector and assign corresponding labels to the user. For example, if the game behavior characteristics of the game user show that they often play role-playing games and pay a lot, the label of "heavy-paying player of role-playing games" can be assigned to them; integrate the labels of each dimension and present the portrait of the game user in a visual way.

[0065] In an embodiment of the present application, the above advertisement library is used to store the advertisements to be pushed to the game user in the game.

[0066] In the embodiments of the present application, after generating the user profile of the game user, the target advertisement suitable for the game user can be selected from the candidate advertisements in the advertisement library according to the user profile of the game user and pushed to the game user. Exemplarily, the matching degree between each advertisement in the advertisement library and the user profile is determined, and when the matching degree is greater than the first preset threshold, the corresponding advertisement is determined as the target advertisement for the game user, which can be realized as follows: the matching degree between each advertisement in the advertisement library and the user profile tags is determined, and when the matching degree of a certain advertisement and the user profile is greater than 80%, the similar advertisements of this advertisement are preferentially pushed to the game user.

[0067] In the embodiments of the present application, exemplarily, the determination of the target advertisement for the game user based on the cross-modal association model can also be realized as follows: the relevance degree between each advertisement in the advertisement library and the current game scene is determined, and when the relevance degree is greater than the second preset threshold, the corresponding advertisement is determined as the target advertisement for the game user. Specifically, the relevance degree between each advertisement in the advertisement library and the current game scene can be calculated, and when the relevance degree of a certain advertisement and the current game scene is greater than 70%, this advertisement is pushed to the game user.

[0068] In step S140, the immersion coefficient of the game user is determined based on the game behavior data, and the advertising value of the target advertisement is evaluated based on the immersion coefficient.

[0069] In the embodiments of the present application, after determining the target advertisement, the advertising value of the target advertisement also needs to be determined.

[0070] Exemplarily, the advertising value Ad_Value of the target advertisement evaluated based on the immersion coefficient can be determined by the following formula:

[0071] Ad_Value = (CTR × 0.4 + CVR × 0.6) × immersion coefficient + device adaptation correction value

[0072] Among them, the above CTR (Click-Through Rate) is the click-through rate of the target advertisement, which is determined by the ratio of the click volume and the display volume of the target advertisement; the above CVR (Conversion Rate) is the conversion rate of the target advertisement, which is determined by the ratio of the conversion volume and the click volume of the target advertisement; the determination process of the above immersion coefficient has been described in detail at the corresponding position above, so it will not be repeated here; the above device adaptation correction value refers to some parameter adjustment values for enabling the application program or interface to be correctly displayed and run on different user devices, and can be determined by the following formula: device correction value = (GPU_score × 0.4) + (RAM_score × 0.3) + (Screen_ratio × 0.3).

[0073] In step S150, an advertising strategy for the target advertisement is determined based on the advertising value, and the target advertisement is pushed to the game users based on the advertising strategy.

[0074] In the embodiment of the present application, the above advertising strategy includes the placement information and the placement time of the target advertisement. Exemplarily, determining the advertising strategy of the target advertisement based on the advertising value and pushing the target advertisement to the game users based on the advertising strategy can be achieved as follows: determining the placement information and the placement time of the target advertisement based on the advertising value, and placing the target advertisement at the interface position corresponding to the placement information in the user device at the placement time.

[0075] Specifically, the above placement information may include a placement identifier and a placement type. Among them, the above placement identifier can be determined as follows: Placement ID = Scene ID + "_" + UI element position code. For example, in the scene of the lower right corner button area (POS_RB) of the "battle scene" (SCN_002), the corresponding placement ID is S02_RB; the above placement type may include different types such as full-screen illustration, floating banner, and bottom information flow.

[0076] In a specific embodiment, after determining the advertising value of the target advertisement through the above advertising value calculation formula, the advertising strategy determined based on the calculated advertising value can be as follows: when the score of the advertising value is greater than 90 points, push the target advertisement to the game users in the form of a full-screen illustration during the scene switching interval of more than 3 seconds; when the score of the advertising value is in the range of (70, 89), push the target advertisement to the game users in the form of a floating banner in the non-combat state; when the score of the advertising value is in the range of (50, 69), push the target advertisement to the game users in the form of a bottom information flow at any time.

[0077] In another embodiment of the present application, in each step of the above in-game advertising placement method, differential privacy technology can also be applied to determine the data security in the data collection, transmission, and processing stages. Specifically, in the data collection stage, the local differential privacy method is used to add Laplace noise to the buried point SDK; in the data transmission stage, the AES-256-GCM encryption algorithm is used to establish a TLS1.3 channel and add an HMAC signature to the data packet; in the data processing stage, the K-anonymization processing algorithm is used to protect the device fingerprint data.

[0078] In another embodiment of the present application, the in-game advertising can also be realized by establishing a game achievement - advertising incentive dynamic redemption system. Exemplarily, it can be achieved as follows: quantitatively evaluating the virtual achievements of the game users in the game; redeeming the scores obtained from the quantitative evaluation into the right to watch advertisements based on the preset redemption rules, so that the game users can watch the corresponding advertisement rules to obtain rewards.

[0079] Specifically, the above-mentioned virtual achievement quantification evaluation algorithm can be as follows:

[0080] Virtual achievement quantification evaluation score = Σ(task difficulty coefficient * completion time efficiency coefficient) * social dissemination weight

[0081] The above-mentioned task difficulty coefficient can be determined by the following formula: D = log2(average number of failures + 1) × type correction factor + equipment requirement index. The benchmark value range of the above-mentioned type correction factor is 0.5 (new player tutorial) - 3.0 (legendary). Taking the main storyline task (corresponding type correction factor 0.8), which requires defeating 3 BOSSes, with an historical average of 2.3 failures and an equipment level requirement of 15 as an example, D = 1.2 × 0.8 + 0.3 = 1.26; among them, the above-mentioned equipment requirement index can be determined according to the corresponding relationship between the equipment level requirement and the equipment requirement index in Table 1 below.

[0082] Table 1:

[0083] Equipment level requirement Equipment demand index 5 0.1 10 0.2 15 0.3 20 0.4 25 0.5 30 0.6 35 0.7 40 0.8 45 0.9 50 1

[0084] The above-mentioned completion time efficiency coefficient can be determined by the following formula: T = e^(-0.1×Δt) × time compression ratio, where Δt = actual time consumed / recommended time consumed. In the process of calculating the completion time efficiency coefficient, the following special rules are involved: In the case of the first kill achievement, the time compression ratio is fixed at 2.0; in the case of seasonal tasks, when Δt exceeds 1.5, exponential decay is triggered (T = 0.8^Δt). Specifically, in the scenario where a task recommended for 2 hours is completed in 1.5 hours, the corresponding completion time efficiency coefficient T = e^(-0.1×0.75) × 1 = 0.86; in the scenario of the first kill in the arena (recommended 8 hours, actual 5 hours), the corresponding completion time efficiency coefficient T = e^(-0.1×0.625) × 2 = 1.84.

[0085] The above-mentioned social dissemination weight can be determined by the following formula: S = 0.3 × number of shares + 0.5 × invitation conversion rate + 0.2 × cross-platform exposure volume. Specifically, if a game user shares a dungeon achievement to 3 social platforms, bringing 5 effective clicks and a cross-platform exposure volume of 1000, then S = 0.3 × 3 + 0.5 × (5 / 1000) + 0.2 × 1000 ≈ 0.9.

[0086] In the embodiment of the present application, after calculating the virtual achievement quantification score of the game user through the virtual achievement quantification evaluation algorithm, an advertising voucher redemption interface can be opened based on the virtual achievement quantification score. Exemplarily, it can be implemented as follows: Every 100 achievements can be exchanged for 30 seconds of customized advertisement viewing rights to obtain rare items (including but not limited to skins, defeat effects, medals, etc.).

[0087] Taking a role-playing mobile game as an example, the above process can be implemented as follows: The game user completes achievements such as "continuous boss battles" or "first kill" (e.g., the system automatically calculates that the achievement points increase by 200); triggers an advertisement redemption pop-up window: The game user can watch a 15-second car advertisement to redeem rare items such as limited costumes and backpacks; the system monitors the player's immersion in real time; dynamically increases the advertising value weight during the user's game time and preferentially pushes high-unit-price advertisements.

[0088] The in-game advertisement placement method provided by the embodiments of this application has the following beneficial effects: (1) Precise value evaluation through multi-modal data fusion: By constructing a three-dimensional data cube of game behavior (character operation trajectory + scene stay duration), advertisement interaction (exposure frequency + click depth), and device characteristics (hardware configuration + network environment), and using a graph neural network for cross-domain feature association, compared with traditional single-dimensional evaluation, the prediction accuracy of advertisement value is improved; (2) Real-time feedback mechanism for dynamic scene perception: By calculating the scene immersion, real-time monitoring of more than 20-dimensional indicators such as the game scene switching frequency (e.g., the conversion interval from the battle scene to the settlement scene) and operation response delay, dynamically adjusting the advertisement placement strategy. Compared with the fixed-time period placement mode, the reach rate of advertisements during non-interference periods for game users is improved; (3) Privacy-enhanced data processing architecture: Adopting a differential privacy protection mechanism, effectively solving the privacy leakage risk existing in traditional advertisement systems while ensuring the accuracy of user portraits; (4) Construction of a game-based incentive ecological closed-loop: Based on an original virtual achievement point redemption system, by converting advertisement interaction behaviors into virtual assets such as game item redemption vouchers and limited skin fragments, the active advertisement participation rate of game users is increased, forming a positive cycle of "advertisement experience - game rewards - continuous interaction".

[0089] The above mainly introduced the solution provided by the embodiments of the present invention from the perspective of the method. To implement the above functions, it includes the corresponding hardware structure and / or software module for executing each function. Those skilled in the art should easily realize that, combining the units and algorithm steps of each example described in the embodiments disclosed in this article, the present invention can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described function for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0090] In the embodiments of the present invention, the in-game advertisement placement device can be divided into functional modules according to the above method examples. For example, each functional module can be corresponding to each function, or two or more functions can be integrated into one processing module. The above integrated module can be implemented in the form of hardware or in the form of a software functional module. It should be noted that the division of modules in the embodiments of the present invention is illustrative, only a logical function division, and there can be other division methods in actual implementation.

[0091] Correspondingly, the embodiments of the present disclosure also provide an in-game advertisement placement device. Referring to Figure 3 As shown, the in-game advertisement placement device 300 may include a data collection module 310, a data processing module 320, an advertisement screening module 330, a value evaluation module 340, and an advertisement placement module 350, where:

[0092] The data collection module 310 is configured to collect the game behavior data and advertisement interaction data of game users, as well as the device feature data of the user devices of game users;

[0093] The data processing module 320 is configured to create a heterogeneous graph based on the game behavior data, advertisement interaction data, and device feature data, and train the heterogeneous graph based on a graph neural network to obtain a cross-modal association model;

[0094] The advertisement screening module 330 is configured to determine target advertisements for game users based on the cross-modal association model;

[0095] The value evaluation module 340 is configured to determine the immersion coefficient of game users based on the game behavior data, and evaluate the advertisement value of the target advertisements based on the immersion coefficient;

[0096] The advertisement placement module 350 is configured to determine an advertisement placement strategy for the target advertisements based on the advertisement value, and push the target advertisements to game users based on the advertisement placement strategy.

[0097] In the embodiments of the present application, the above data processing module is specifically configured to: extract the user behavior sequence data of game users in the game behavior data, and encode the user behavior sequence data through a long short-term memory network to obtain a first multi-dimensional vector; extract the advertisement material features in the advertisement interaction data, and encode the advertisement material features through a word embedding encoding method to obtain a second multi-dimensional vector; perform dimensionality reduction encoding on the device feature data through principal component analysis dimensionality reduction to obtain a third multi-dimensional vector; create game nodes, advertisement nodes, and device nodes, use the first multi-dimensional vector, the second multi-dimensional vector, and the third multi-dimensional vector as the node attributes of the game nodes, advertisement nodes, and device nodes respectively, and create edges between the game nodes, advertisement nodes, and device nodes based on the game behavior data, advertisement interaction data, and device feature data to obtain a heterogeneous graph.

[0098] In the embodiment of the present application, the above-mentioned value evaluation module is specifically configured to: obtain the average single-game duration of the game user, the number of completed tasks in the game, and the total number of game tasks; determine the task completion density coefficient based on the number of completed tasks and the total number of game tasks of the game user; and determine the immersion coefficient based on the average single-game duration and the task completion density coefficient.

[0099] In the embodiment of the present application, the above-mentioned data processing module is specifically configured to: obtain the scene stay duration of the game user in the current game scene from the game behavior data, and determine the game node weight of the game node based on the immersion coefficient and the scene stay duration; obtain the first performance coefficient of the image processing unit of the user device, the second performance coefficient of the device memory, and the third performance coefficient of the device network, and determine the device performance coefficient of the user device based on the first performance coefficient, the second performance coefficient, and the third performance coefficient; obtain the average daily usage duration of the user device, and determine the usage frequency coefficient of the user device based on the average daily usage duration; and determine the device node weight of the device node based on the device performance coefficient and the usage frequency coefficient.

[0100] In the embodiment of the present application, the above-mentioned advertisement screening module is specifically configured to: generate a user portrait of the game user based on the cross-modal association model; determine the matching degree between each advertisement in the advertisement library and the user portrait, and when the matching degree is greater than the first preset threshold, determine the corresponding advertisement as the target advertisement for the game user.

[0101] In the embodiment of the present application, the above-mentioned advertisement screening module is specifically configured to: determine the relevance between each advertisement in the advertisement library and the current game scene, and when the relevance is greater than the second preset threshold, determine the corresponding advertisement as the target advertisement for the game user. In the embodiment of the present application, the above-mentioned advertisement placement module is specifically configured to: determine the ad placement information and the placement time of the target advertisement based on the advertisement value, and place the target advertisement at the interface position corresponding to the ad placement information in the user device at the placement time.

[0102] The specific implementation details of the above-mentioned in-game advertisement placement device have been described in detail at the corresponding position of the in-game advertisement placement method, so they will not be elaborated here.

[0103] Figure 4 It is a schematic structural diagram of an electronic device in an embodiment of the present disclosure. Specifically refer to the following Figure 4 , which shows a schematic structural diagram of the electronic device 400 suitable for implementing the embodiment of the present disclosure. Figure 4 The shown electronic device is only an example and should not bring any limitation to the functions and usage scope of the embodiment of the present disclosure.

[0104] As Figure 4As shown, the electronic device 400 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 401, which may perform various appropriate actions and processes according to a program stored in the read-only memory (ROM) 402 or a program loaded from the storage device 408 into the random access memory (RAM) 403 to implement the in-game advertisement placement method of the embodiments as described in the present disclosure. In the RAM 403, various programs and data required for the operation of the electronic device 400 are also stored. The processing device 401, the ROM 402, and the RAM 403 are connected to each other through a bus 404. The input / output (I / O) interface 405 is also connected to the bus 404.

[0105] Generally, the following devices may be connected to the I / O interface 405: an input device 406 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 407 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 408 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 409. The communication device 409 may allow the electronic device 400 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 4 an electronic device 400 with various devices is shown, it should be understood that it is not required to implement or have all the shown devices. Instead, more or fewer devices may be implemented or had.

[0106] Specifically, according to an embodiment of the present disclosure, the process described above with reference to the flowchart may be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes program codes for executing the method shown in the flowchart, so as to implement the in-game advertisement placement method as described above. In such an embodiment, the computer program may be downloaded and installed from a network through the communication device 409, or installed from the storage device 408, or installed from the ROM 402. When the computer program is executed by the processing device 401, the above functions defined in the method of the embodiment of the present disclosure are executed.

[0107] It should be noted that the above-mentioned computer-readable medium in the present disclosure can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device. In the present disclosure, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, in which computer-readable program code is carried. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, and this computer-readable signal medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0108] In some embodiments, the client and the server can communicate using any currently known or future-developed network protocol such as HTTP (Hyper Text Transfer Protocol), and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed networks.

[0109] The above-mentioned computer-readable medium can be included in the above-mentioned electronic device; or it can exist separately and not be assembled into the electronic device.

[0110] The above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by the electronic device, the electronic device is caused to:

[0111] Collect the game behavior data and advertisement interaction data of game users, as well as the device feature data of the user devices of game users;

[0112] Create a heterogeneous graph based on the game behavior data, advertisement interaction data and device feature data, and train the heterogeneous graph based on a graph neural network to obtain a cross-modal association model;

[0113] Determine a target advertisement for a game user based on the cross-modal association model;

[0114] Determine the immersion coefficient of a game user based on the game behavior data, and evaluate the advertisement value of the target advertisement based on the immersion coefficient;

[0115] Determine an advertisement placement strategy for the target advertisement based on the advertisement value, and push the target advertisement to the game user based on the advertisement placement strategy.

[0116] Optionally, when one or more of the above programs are executed by the electronic device, the electronic device may further execute the other steps described in the above embodiments.

[0117] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages or combinations thereof. The above programming languages include, but are not limited to, object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may execute entirely on the user's computer, partially on the user's computer, execute as a stand-alone software package, execute partially on the user's computer and partially on a remote computer, or execute entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0118] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code that contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0119] The units involved in the embodiments described in the present disclosure can be implemented in software or in hardware. Among them, the name of the unit does not constitute a limitation on the unit itself in some cases.

[0120] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that can be used include: Field Programmable Gate Arrays (FPGA), Application Specific Integrated Circuits (ASIC), Application Specific Standard Products (ASSP), System on a Chip (SOC), Complex Programmable Logic Devices (CPLD), and so on.

[0121] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a Random Access Memory (RAM), a Read-Only Memory (ROM), an Erasable Programmable Read-Only Memory (EPROM or Flash Memory), an optical fiber, a portable Compact Disc Read-Only Memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0122] The above description is only a preferred embodiment of the present disclosure and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the disclosure involved in the present disclosure is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosure concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions disclosed in the present disclosure.

[0123] In addition, although the operations are depicted in a particular order, this should not be construed as requiring that the operations be performed in the particular order shown or in sequential order. In certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although a number of specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the present disclosure. Certain features described in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, the various features described in the context of a single embodiment may also be implemented separately or in any suitable sub-combination in multiple embodiments.

[0124] Although the subject matter has been described in language specific to structural features and / or methodological logical acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are merely example forms of implementing the claims.

Claims

1. A method for in-game advertisement placement, characterized in that, The method includes: Collecting the game behavior data and advertisement interaction data of the game user, as well as the device feature data of the user device of the game user; Creating a heterogeneous graph based on the game behavior data, the advertisement interaction data, and the device feature data, and training the heterogeneous graph based on a graph neural network to obtain a cross-modal association model; Determining a target advertisement for the game user based on the cross-modal association model; Determining the immersion coefficient of the game user based on the game behavior data, and evaluating the advertisement value of the target advertisement based on the immersion coefficient; Determining an advertisement placement strategy for the target advertisement based on the advertisement value, and pushing the target advertisement to the game user based on the advertisement placement strategy.

2. The in-game advertisement placement method according to claim 1, wherein The creating of the heterogeneous graph based on the game behavior data, the advertisement interaction data, and the device feature data includes: Extracting the user behavior sequence data of the game user from the game behavior data, and encoding the user behavior sequence data through a long short-term memory network to obtain a first multi-dimensional vector; Extracting the advertisement material features from the advertisement interaction data, and encoding the advertisement material features through a word embedding encoding method to obtain a second multi-dimensional vector; Performing dimensionality reduction encoding on the device feature data through principal component analysis dimensionality reduction to obtain a third multi-dimensional vector; Creating game nodes, advertisement nodes, and device nodes, taking the first multi-dimensional vector, the second multi-dimensional vector, and the third multi-dimensional vector as the node attributes of the game nodes, the advertisement nodes, and the device nodes respectively, and creating edges between the game nodes, the advertisement nodes, and the device nodes based on the game behavior data, the advertisement interaction data, and the device feature data to obtain the heterogeneous graph.

3. The in-game advertisement placement method according to claim 2, wherein The determining of the immersion coefficient of the game user based on the game behavior data includes: Obtaining the average single-game duration of the game user, the number of completed tasks in the game, and the total number of game tasks; Determining a task completion density coefficient based on the number of completed tasks and the total number of game tasks of the game user; Determining the immersion coefficient based on the average single-game duration and the task completion density coefficient.

4. The in-game advertisement placement method according to claim 3, wherein The creating of the heterogeneous graph based on the game behavior data, the advertisement interaction data, and the device feature data further includes: Obtaining the scene stay duration of the game user in the current game scene from the game behavior data, and determining the game node weight of the game node based on the immersion coefficient and the scene stay duration; Obtaining a first performance coefficient of the image processing unit of the user device, a second performance coefficient of the device memory, and a third performance coefficient of the device network, and determining a device performance coefficient of the user device based on the first performance coefficient, the second performance coefficient, and the third performance coefficient; Obtaining the average daily usage duration of the user device, and determining a usage frequency coefficient of the user device based on the average daily usage duration; Determining the device node weight of the device node based on the device performance coefficient and the usage frequency coefficient.

5. The in-game advertisement placement method according to claim 4, characterized in that, Determining a target advertisement for the game user based on the cross-modal association model includes: Generating a user profile of the game user based on the cross-modal association model; Determining the matching degree between each advertisement in the advertisement library and the user profile, and when the matching degree is greater than a first preset threshold, determining the corresponding advertisement as the target advertisement for the game user.

6. The in-game advertisement placement method according to claim 4, wherein Determining a target advertisement for the game user based on the cross-modal association model further includes: Determining the association degree between each advertisement in the advertisement library and the current game scene, and when the association degree is greater than a second preset threshold, determining the corresponding advertisement as the target advertisement for the game user.

7. The in-game advertisement placement method according to claim 1, wherein Determining an advertisement placement strategy for the target advertisement based on the advertisement value, and pushing the target advertisement to the game user based on the advertisement placement strategy includes: Determining the placement position information and the placement time of the target advertisement based on the advertisement value, and placing the target advertisement at the interface position corresponding to the placement position information in the user device at the placement time.

8. A device for in-game advertisement placement, characterized in that, The device includes: A data collection module, configured to collect game behavior data and advertisement interaction data of a game user, and device feature data of the user device of the game user; A data processing module, configured to create a heterogeneous graph based on the game behavior data, the advertisement interaction data, and the device feature data, and train the heterogeneous graph based on a graph neural network to obtain a cross-modal association model; An advertisement screening module, configured to determine a target advertisement for the game user based on the cross-modal association model; A value evaluation module, configured to determine an immersion coefficient of the game user based on the game behavior data, and evaluate the advertisement value of the target advertisement based on the immersion coefficient; An advertisement placement module, configured to determine an advertisement placement strategy for the target advertisement based on the advertisement value, and push the target advertisement to the game user based on the advertisement placement strategy.

9. An electronic device, comprising: A processor; And a memory storing a program, wherein the program includes instructions for causing the processor to execute the method according to any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the method according to any one of claims 1-7.

Images