MOBA game match data interaction selection method based on story line visualization

Abstract

The invention provides an MOBA game match data interaction selection method based on story line visualization. The MOBA game match data interaction selection method comprises the steps of (1) match event extraction and data preprocessing; (2) performing game similarity calculation and game hierarchy analysis; (3) performing story line layout calculation and performing round understanding by a user by using a visual analysis method of the story line; and (4) the user performs exploration and data selection of event and male and English detail information in combination with the visual elements. The method provided by the invention supports analysis of complex data in the MOBA game, the whole method can efficiently extract game key events and analyze multiple games, and rich visual elements created on the basis can help a user to visually visualize the game process; various user interactions contained in the method can promote the user to easily explore detailed information of each male, so that the research efficiency is improved, and the operation burden is reduced.

Description

Technical Field

[0001] The present invention relates to the field of visualization and human-computer interaction, and specifically proposes a MOBA game match data interactive selection method based on storyline visualization for MOBA game data. Background Art

[0002] MOBA games are an emerging type of sports competition, entertainment and social activity. Games such as League of Legends, Dota 2 and Honor of Kings have become widely popular and attracted a large number of players. Each game is known for its deep game mechanics, a wide range of heroes and a vibrant e-sports scene. For MOBA game developers, an important task is to create adaptive training modes that are suitable for players based on the characteristics of different human players. For example, players of different skill levels may require various types of agents, such as beginner, intermediate, advanced, and so on. For the training of agents, reinforcement learning algorithms can generally be used. Deep reinforcement learning combines neural networks with reinforcement learning frameworks, allowing agents to learn the best actions in complex environments by maximizing cumulative rewards through repeated trials. The reward function is crucial in reinforcement learning because it provides immediate feedback to the agent, guides its behavior in the expected direction, and measures the pros and cons of behavioral strategies by quantifying the value of each decision in the environment. However, in the context of MOBA games, long rounds, high task complexity and diverse strategies make it challenging to design a complete reward function. It is difficult to design various strategies for diverse players. Therefore, existing methods often fail to develop agents with diverse strategies required in MOBA game scenarios. Imitation learning provides an alternative approach to address the challenges in training deep reinforcement learning models. Imitation learning focuses on learning and improving from existing strategies rather than designing complex reward functions to characterize behaviors in different strategies. Since focusing on selecting representative strategy samples simplifies a lot of human design processes, it is usually easier than defining complex reward mechanisms. However, a key challenge is how to obtain suitable training data. Current methods usually select samples based on broad indicators such as reward value or win rate. However, complex sequence decision data with a large number of events cannot be measured and distinguished by only one or two final numerical indicators. In addition, relying solely on indicators may lead to biased evaluations that cannot ensure consistency with user needs. Therefore, it is necessary to introduce visual analysis methods to integrate human insights into the data selection process. However, due to contextual differences, existing sports and game visual analysis techniques are not directly applicable to MOBA games. Current sequence data visualization techniques also have difficulty in representing the evolution of individual characters and overall game dynamics in both time and space. Therefore, effectively visualizing and selecting MOBA game data remains a major challenge. Summary of the invention

[0003] In view of the defects and deficiencies in the prior art, the present invention aims to provide a method for interactive selection of MOBA game match data based on storyline visualization. This is a visualization analysis method tailored for MOBA data selection and strategy analysis, which aims to help game AI designers effectively collect target samples to train intelligent agents. Under the guidance of user design requirements, the present invention develops a data selection visualization analysis method from macro to micro. The method first presents an overview of multiple matches, with the aim of enabling data analysts to efficiently determine the matches of interest for more targeted analysis. In order to further understand the development of events in each game, the present invention introduces a new visualization technology called v-storyline. The technology encodes the position attributes of the characters on the vertical axis of the layout, allowing the abstraction and extraction of spatiotemporal sequence data from the game events, effectively illustrating the evolution of the game rounds. In addition, the method provides detailed hero statistics for comparison, thereby supporting data analysts' fine-grained hero-level data selection. Help users intuitively and conveniently understand the game strategy pattern and efficiently select data from multiple stages.

[0004] An interactive selection method for MOBA game match data based on storyline visualization includes the following four steps: (1) game event extraction and data preprocessing; (2) game similarity calculation and game level analysis; (3) storyline layout calculation and user use of storyline visual analysis method to understand the round level; (4) users explore event and hero details information and select multiple rounds of data in combination with visualization elements.

[0005] Specific:

[0006] (1) Game event extraction and data preprocessing:

[0007] In order to help users complete data selection tasks efficiently, this method first extracts events from a large amount of game data in units of games, and at the same time pre-processes and integrates some important information for use in subsequent visualization. The game event extraction steps can be broken down into the following process segments:

[0008] First, we read the behavior information of each hero in the data, and then calculate the indicators representing the status based on different aspects of the evaluation strategy, such as development, damage, and kills. This is a quantity that changes over time, and we use this quantity to divide the entire game into several time periods through the change point detection algorithm. Each time period corresponds to an event. Considering the background knowledge of MOBA games, we define five types of events: team battle, jungle, tower attack, movement, and death.

[0009] On the other hand, in order to analyze the spatial information of events, it is necessary to clarify the location where the events occurred. In this step, we first rely on the key locations in the MOBA game map (i.e., buildings of both sides, elite monsters in the wild area) as the basis to calculate the Voronoi diagram to segment the scene space; then, in order to avoid the area division being too trivial, we rely on the background knowledge in the MOBA game to merge the areas again, and retain a total of 7 areas corresponding to the bases of both sides, the top lane, the middle lane, the bottom lane, and the upper and lower wild areas. Finally, according to the time period information and spatial area information after segmentation, the game data is converted into storyline data. For a storyline data, it contains different numbers of events in units of heroes, and each event is a triple, which is the start time, end time and event number (where the type and location information of the event are implicitly encoded in the event number).

[0010] Game event extraction and data preprocessing mainly convert the turn data containing a lot of complex information into concise storyline data, which is used for subsequent analysis steps. This processing method has two advantages: first, it can efficiently understand the game. Users do not have to waste their energy on searching frame by frame, but can use a certain length of time period to efficiently understand the development and change process of the situation. Secondly, it can be convenient for analysis and selection. Through events, the characteristics of a game can be portrayed in a macro sense, helping users to choose from a large number of games.

[0011] (2) Game similarity calculation and game-level analysis:

[0012] In this step, the method of the present invention proposes a game similarity calculation method and designs corresponding visualization elements. First, based on the events extracted in step (1), the distribution of events in a game is counted. The horizontal axis of the distribution is the game time, and the vertical axis is the number of heroes involved in an event. Since mobile events occur in large numbers, events that are guaranteed to occur are not considered in the statistics, and death events accompanying team battle events are not divided separately. When depicting the game, considering that the strategies of the winning and losing parties in the game background are usually different, the two parties in the game are counted separately in this step. Then, by calculating the Cramer distance between the distributions and projecting the points representing the game into a two-dimensional space through the TSNE algorithm, a scatter plot representing the game is obtained in the plane.

[0013] The following visual elements and user interactions are defined in the scatter plot. First, the color represents the faction of the game; the direction of the triangle represents the win-lose relationship of the game; and the size represents the duration of the game. Second, the visualization interface of this method provides swiping interaction, zooming interaction, and clicking interaction. When using it, the user can click on the game of interest and use the methods and visualization elements in steps (3) and (4) for further analysis and selection.

[0014] (3) Storyline layout calculation and user’s understanding of the turn level using the storyline’s visual analysis method:

[0015] In this step, a turn event display based on storyline visualization is proposed, which can efficiently and intuitively display the development and evolution of the selected game situation. The specific steps of generating the storyline layout algorithm are as follows:

[0016] First, read the storyline event data generated in step (1), and then calculate the number of points in the entire layout space based on the number of key moments, the number of heroes, and the number of regions in the data. For the points in the storyline layout, their positions on the x-axis represent the time points, and their relative positional relationships on the y-axis represent the positional relationships of the regions in the method of the present invention. Then, after calculating the number of points in the layout, each point is regarded as a variable, and then the layout target based on aesthetic principles is defined. In the formula, y represents the vertical position of the layout point, i and j are loop variables, indicating that all points in the layout are counted, c is the sum of the number of heroes and the number of regions, and t is the time of the round data. The first term of the formula represents minimizing the swing of the front and rear points, and the second term represents minimizing the blanks in the layout.

[0017]

[0018] Secondly, in order to introduce the relative position relationship in the spatial region, the variables corresponding to the storyline characters (i.e., heroes and regional boundaries) at each moment are sorted according to the layout numbers obtained in (1). Then, in order to reduce the overlap of lines in the region, the sorting relationship is adjusted again according to the principle of minimizing intersections. Then, in order to improve the alignment relationship between the corresponding points of the two moments before and after to enhance readability, for the sequences l and r of the two moments before and after, the similarity is calculated according to the formula, where LCS represents the calculation of the maximum common subsequence of the two sequences, m and n represent the number of story points contained in the two sequences, p is the indicator function with a value of 1 for the storyline points representing the boundary, and α, β, and γ are parameters.

[0019]

[0020] Next, the constraints of the y value at each moment are obtained based on the alignment relationship. Finally, a mathematical solver is used to solve the constrained quadratic optimization problem to obtain a series of layout points, where AO represents the alignment relationship order.

[0021]

[0022] Compared with the previous method, the method of the present invention introduces a relative position relationship on the y-axis, which helps users better understand the events in the MOBA scene and reduces the potential ambiguity caused by the previous method that does not perform any physical encoding on the y-axis. At the same time, compared with the complete correspondence mapping of spatial positions, it reduces the triviality and improves readability.

[0023] After completing the layout calculation step, the present invention designs the following visualization elements and provides users with multiple interactions. First, according to the layout points calculated in this section, Bezier curves are used to connect to obtain a coherent story line. Each colored line represents a hero, and the gray dotted line marks the region boundary. The gathering and deepening of the lines indicate the occurrence of events, and the movement of the lines indicates the hero changing regions and entering and leaving events. The dotted line indicates the death of a hero. On the other hand, the visualization interface allows users to interact with the story line. Users can adjust the granularity of the story line events, zoom in on the story line display by swiping the timeline, and drag the scroll bar to move the story line display range. At the same time, each story line supports floating highlighting and click interaction. Clicking can select a hero and further explore and select in step (4).

[0024] (4) Users use visualization elements to explore event and hero details and select data:

[0025] The user utilizes the visualization elements and user interaction constructed by the present invention to realize information exploration and data selection from macro to micro. The process is as follows:

[0026] First, the given data set is read and the projection data of the game obtained in step (2) is used to display it to the user. The user selects one of the games for in-depth exploration according to his or her needs.

[0027] Then, the user's interaction is read, and the evolution of this round and the development of events are displayed according to the method and data provided by the user in step (3). The user selects the appropriate story line granularity and the time period to be explored according to his or her own needs.

[0028] Next, the detailed data is read to provide the user with a statistical information view. In the statistical information view, the table headers are the values ​​of the heroes in the game and in team battles, jungle, and tower pushing. When building this visualization element, the user is provided with sorting interaction, and can click the table header to sort in descending order.

[0029] Secondly, the entire visualization method provides a hero comparison view. By default (when the user has not made a selection), the visualization elements include the positions, health points, mana points, actions performed, rewards obtained, and other data of the heroes of both sides, and the data is displayed in the form of snapshots, pie charts, bar charts, and shadow charts. When the user makes a selection, the displayed data is transformed to the object selected by the user in the visual element area of ​​the corresponding camp through a transition animation.

[0030] At the same time, when the user performs interactions such as brushing, dragging, and selecting units to change the displayed data range and objects, the corresponding visual elements in the storyline module, comparison view part, heat map, etc. in the visualization are synchronously updated and display the new selected data.

[0031] Finally, based on the information provided by the visualization elements in the entire process, the user selects events and heroes of interest to analyze their strategies, collects valuable and noteworthy data as subsequent training data, and selects new games for exploration and analysis again. The process in step 4 is repeated multiple times to complete the entire data selection task.

[0032] The method of the present invention supports the analysis of complex data in MOBA games. The entire method can efficiently extract key events of the game and analyze multiple games. The rich visualization elements created on this basis can help users visualize the game process intuitively, and include a variety of user interactions that can prompt users to easily explore the detailed information of each hero. Specifically, the effects of the present invention include:

[0033] Improve the user's ability to efficiently extract key events of the game: In step (1), the present invention supports automatic data processing, finds key time periods and summarizes them into events, avoids the user's time-consuming video playback, and reduces the user's trouble of thinking frame by frame.

[0034] Reducing the difficulty for users to select data from multiple levels: In step (2), the present invention combines the process results in step (1), integrates the distance calculation method, and designs visualization elements and user interactions, providing users with a starting point for exploration. Users do not need to perform aimless sequential analysis and random sampling, thereby improving exploration efficiency and reducing operational burden.

[0035] Enhance the user's ability to understand the development of the game: In step (3), the present invention uses a new storyline visualization form, which can intuitively and efficiently display the key events obtained in step (1). It can more intuitively display the development of multiple heroes and multiple events at the time and space level, facilitate the analysis of changes in the situation, and provide multiple interactions for user exploration, which is an important part of realizing hierarchical exploration.

[0036] Reducing the difficulty of users comparing hero strategy information: In step (4), the visualization of the present invention provides comparison of hero levels, comparison within a camp, and comparison between different camps. At the same time, the user interaction implied in steps (3) and (4) allows users to freely choose the event period and hero content to explore, reducing the user's consumption of searching and locating from a large amount of statistical data, and reducing the difficulty of users comparing different hero strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A flowchart of interactive selection of MOBA game match data from macro to micro in a specific implementation manner of the present invention;

[0038] Figure 2 It is a schematic diagram of the steps of game event extraction and data preprocessing in a specific implementation manner of the present invention;

[0039] Figure 3 It is the meaning and main visual elements of the story line method in the specific implementation mode of the present invention;

[0040] Figure 4 A specific page of the human-computer interaction visual analysis involved in the present invention in the specific implementation mode of the present invention;

[0041] Figure 5 It is a logic block diagram in a specific implementation manner of the present invention. DETAILED DESCRIPTION

[0042] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0043] Figure 1 The flowchart of interactive selection of MOBA game match data from macro to micro in a specific embodiment of the present invention is shown: First, data preprocessing and event extraction need to be performed in advance, such as Figure 2 ; Then the user will conduct three progressive exploration stages in sequence: analysis at the game level, understanding at the event level, and comparison at the hero level. Figure 1 The analysis methods and corresponding visualization elements provided in the method flow are shown in Figure 2. Users complete the interaction progressively according to their own needs, select a game of interest, understand the game strategy pattern through the events in the game, and compare the similarities and differences of different hero strategies. Users continue to cycle through this process, explore new data, and complete the entire MOBA game data interactive selection process.

[0044] Figure 2A schematic diagram of the steps of game event extraction and data preprocessing in a specific embodiment of the present invention is shown. The entire event extraction and preprocessing is divided into three main steps, corresponding to the three parts from left to right in the figure. First, data is extracted in units of heroes based on the information in the original data, and the original behavior is reflected by calculating state quantities based on development, KDA, damage, etc. Then, a model-free change point detection algorithm is used to divide the entire game round and obtain several segments in time. Finally, a category label is added to the event and a regional representation is added according to the location where it occurred.

[0045] Figure 3 The representation meaning and main visual elements of the storyline method in the specific implementation mode of the present invention are shown. The horizontal gray dotted lines in the figure divide the space of the entire storyline layout into several areas, which corresponds to the map space divided by the Voronoi diagram. Each colored solid line in the figure represents a hero, and the gathering of lines corresponds to the corresponding time and the corresponding area. The relevant heroes jointly participate in an event. Depending on the different events, there can be mainly team battles, wild fighting and tower pushing, etc.

[0046] Figure 4 The specific page of the human-computer interaction visual analysis involved in the present invention is shown.

[0047] (a) View is the game mode selection view, which is at the very beginning of each user's exploration process. The visualization elements show multiple games and project them into two-dimensional space based on their features. Users can freely select one or a group of games to explore through interaction.

[0048] (b) is a hero heat map view, where the dark lines show the result of dividing the space based on the Voronoi diagram, and the darker black is the merged area boundary. The heat map in the view shows the position distribution of the selected hero in the selected time period. By default, it shows the position statistics of the entire camp during the entire round.

[0049] (c) The view is a storyline view. After the layout points are calculated by the v-storyline visualization method introduced in step (3), these layout points are smoothly connected using Bezier curves. The present invention provides multiple interactions such as granularity selection, storyline magnification, time period selection, and storyline click suspension.

[0050] (d) The view is the economic experience curve, through which the difference between the economic experience curves of the two sides in a game is plotted in a discounted form.

[0051] (e) View is a statistical information comparison view. This view uses a bar chart to represent the data of the corresponding items in the table, and the colors are bound to the camps. To facilitate comparison, the view provides sorting functions by row and column, supporting sorting data from high to low.

[0052] The (f) view is a hero comparison view that divides the time period into ten sections. This view provides detailed match data, including hero actions, spatial distribution, health and mana status, and the difference between the two teams in resource acquisition and status indicators. This view consists of four sections.

[0053] (f1) shows snapshots at each time interval. Each snapshot shows the spatial distribution of heroes at a given time as gray circles. The centers and radii of these circles are calculated using a clustering algorithm based on the average hero positions.

[0054] (f2) and (f4) show pie charts representing the action choices of the two teams in the same period: white represents idle, green represents movement, pink represents attack, and blue-purple represents skill use. The shaded area below each pie chart shows the fluctuation of health (pink) and mana (purple) over time, allowing users to track team actions and the status of each hero.

[0055] View (f3) illustrates the changes in resource acquisition and status metric differences between the two teams using a bar graph and shaded area. The height of the bar graph represents the total resources each team has acquired, while the shaded area below shows the real-time difference in status metrics. Positive (upper) shading indicates that the red team is dominant, while negative (lower) shading indicates that the blue team is dominant.

[0056] Figure 5 The logical block diagram of the present invention is shown. It corresponds to the proposed analysis steps and the above-mentioned flow chart, and integrates the visual analysis method to form a page for user interaction. The working logic of the method in the process is shown in the figure, which is mainly divided into three steps, corresponding to the game selection from macro to micro, event understanding and hero level comparison.

Claims

1. A MOBA game match data interactive selection method based on storyline visualization, characterized in that: The following steps are involved: (1) Game event extraction and data preprocessing; (2) Game similarity calculation and game-level analysis; (3) storyline layout calculation and user-level understanding of storyline visual analysis methods; (4) Users combine visualization elements to explore event and hero details and conduct multiple rounds of data selection.

2. The MOBA game match data interactive selection method based on storyline visualization according to claim 1 is characterized in that: The specific method of step (1) is: First, we extract events from a large amount of game data in units of games, and pre-process and integrate important information to facilitate use in the subsequent visualization steps. For the game event extraction step, the process is subdivided as follows: First, we read the behavior information of each hero in the data, and then calculate the indicators representing the status based on different aspects of the evaluation strategy. This is a quantity that changes over time. We use this quantity to divide the entire game into several time periods through the change point detection algorithm. Each time period corresponds to an event. Considering the background knowledge of MOBA games, we define five types of events: team battle, jungle, tower attack, movement, and death. On the other hand, in order to analyze the spatial information of the event and clarify the location where the event occurred, in this step, we firstly divide the scene space based on the key positions in the MOBA game map, and calculate the Voronoi diagram; then, in order to avoid the area division being too trivial, we merge the areas again based on the background knowledge in the MOBA game, and retain a total of 7 areas corresponding to the bases of both sides, the top lane, the middle lane, the bottom lane, and the upper and lower jungle areas; Finally, according to the segmented time period information and spatial area information, the entire game data is divided into storyline data; for a storyline data, it contains a different number of events in units of heroes, and each event is a triplet, which includes the start time, end time and event number.

3. The MOBA game match data interactive selection method based on storyline visualization according to claim 1, characterized in that: The specific method of step (2) is: First, based on the events extracted in step (1), the distribution of events in a game is counted; the horizontal axis of the distribution is the game time, and the vertical axis is the number of heroes involved in an event; when describing the game, the two sides of the game are counted separately; Then, by calculating the Cramer distance between the distributions and projecting the points representing the games into two-dimensional space using the TSNE algorithm, a scatter plot representing the games is obtained in the plane; The following visual elements and user interactions are defined in the scatter plot visualization. First, the color represents the camp of the game. The direction of the triangle represents the win-loss relationship of the game. The size represents the duration of the game. Secondly, the visualization interface provides swiping interaction, zooming interaction, and clicking interaction; when the user uses it, he clicks on the game of interest and uses the methods and visualization elements in steps (3) and (4) for further analysis and selection.

4. The MOBA game match data interactive selection method based on storyline visualization according to claim 1, characterized in that: The specific method of step (3) is: The specific steps of the storyline layout algorithm are as follows: First, read the storyline event data generated in step (1), and then calculate the number of points in the entire layout space based on the number of key moments, the number of heroes, and the number of regions in the data; for the points in the storyline layout, their positions on the x-axis represent the time points, and their relative positions on the y-axis represent the positional relationships of the regions; then, after calculating the number of points in the layout, treat each point as a variable, and then define the layout target based on aesthetic principles, where y represents the vertical position of the layout point, i and j are loop variables, indicating that all points in the layout are counted, c is the sum of the number of heroes and the number of regions, and t is the time of the round data; the first term of the formula represents minimizing the swing of the front and back points, and the second term represents minimizing the blank space in the layout; Secondly, according to the layout number obtained in (1), the variables corresponding to the storyline characters at each moment are sorted; then, the sorting relationship is adjusted again according to the principle of minimizing crossover; then, for the sequences l and r at the previous and next moments, the similarity is calculated according to the formula, where LCS means calculating the largest common subsequence of the two sequences, m and n represent the number of story points contained in the two sequences, p is the indicator function with a value of 1 for the storyline points representing the boundary, and α, β, and γ are parameters; Next, the constraints of the y value at each moment are obtained based on the alignment relationship; finally, the mathematical solver is used to solve the constrained quadratic optimization problem to obtain a series of layout points, where AO represents the alignment relationship order; After completing the layout calculation step, the following visualization elements are designed and multiple user interactions are provided. First, the layout points calculated in this section are connected using Bezier curves to obtain a coherent storyline. Each colored line represents a hero, and the gray dashed line marks the region boundary. The gathering and deepening of the lines indicate the occurrence of events, and the movement of the lines indicates the hero changing regions and entering and leaving events. The dashed line indicates the death of a hero. On the other hand, the user is provided with the function of interacting with the storyline during the visualization process. The user can adjust the granularity of the storyline event, zoom in on the storyline display by swiping the timeline, and drag the scroll bar to move the storyline display range. At the same time, each storyline supports floating highlighting and click interaction. Clicking can select a hero, and further exploration and selection can be performed in step (4).

5. The MOBA game match data interactive selection method based on storyline visualization according to claim 1, characterized in that: The specific method of step (4) is: First, continue to read the given data set and use the projection data of the game obtained in step (2) to display it to the user. The user selects one of the games for in-depth exploration according to his or her needs. Then, continue to read the user's interaction, and display the evolution of this round and the development of events according to the method and data provided by the user in step (3). The user selects the appropriate story line granularity and the time period to be explored according to his own needs; Next, continue to read the detailed data and provide the user with a statistical information view. In the statistical information view, the table headers are the values ​​of the heroes in the game and the values ​​in team battles, jungle, and tower attacking. When constructing this visualization element, provide users with sorting interactions, and click on the table header to sort in descending order. Secondly, the whole visualization method provides a hero comparison view. By default, the visualization elements include the positions, health points, mana points, actions performed, and rewards of the heroes of both teams. The data is displayed in the form of snapshots, pie charts, bar charts, and shadow charts. When the user makes a selection, the displayed data is transformed to the object selected by the user in the visual element area of ​​the corresponding camp through a transition animation. At the same time, when the user performs the interaction of brushing, dragging, and selecting units to change the displayed data range and objects, the corresponding storyline module, comparison view part, and visual elements in the heat map in the visualization method are synchronously updated and display the new selected data; Finally, based on the information provided by the visualization elements in the entire process, the user selects events and heroes of interest to analyze their strategies, collects valuable and noteworthy data as subsequent training data, and selects new games for exploration and analysis again. The process in step 4 is repeated multiple times to complete the entire data selection task.

Images