A method and system for generating digital publications with knowledge graphs integrated with XR

Abstract

The present invention discloses a method and system for generating digital publications with knowledge graphs integrated with XR, and specifically relates to the technical field of knowledge graphs. By clarifying the XR scenario types and user needs, a structured knowledge graph is constructed, the rendering content of user operations is recorded in real time, and the responsiveness of the system under different conditions is evaluated. By measuring the difference between the actual feedback of the user and the expected feedback speed, the sense of lag is judged, and the real-time response capability required for the interaction is determined based on the analysis results. The response capability is divided into high, medium, and low levels. High-frequency interactive content is preferentially rendered locally to reduce latency, while low real-time tasks are processed asynchronously. For medium real-time responses, the system dynamically adjusts the rendering details according to abnormal changes in the real-time response capability to ensure performance optimization and provide a smoother user experience. It not only effectively reduces latency, but also improves the real-time nature of the interaction and the smoothness of the user's operation.

Description

Technical Field

[0001] The present invention relates to the field of knowledge graph technology, and in particular to a method and system for generating knowledge graph digital publications integrated with XR. Background Art

[0002] The generation of digital publications with knowledge graphs that integrate XR (extended reality, including virtual reality VR, augmented reality AR, and mixed reality MR) refers to the combination of XR technology and knowledge graphs to create highly interactive and dynamically presented digital publications. Knowledge graphs are structured networks used to represent the relationship between information, which can help users understand complex knowledge points more efficiently. When XR is combined with knowledge graphs, users can intuitively explore and learn complex knowledge systems through virtual, augmented, or mixed reality environments.

[0003] When generating digital publications with knowledge graphs that integrate XR, first, by building a knowledge graph, sorting and organizing information and data in related fields, establishing nodes and relationships, designing the content framework and user interaction logic of digital publications, and clarifying the way of displaying knowledge in XR scenarios. Then, use XR development tools (such as Unity, Unreal Engine, etc.) to combine the knowledge graph with the virtual or augmented reality environment to create interactive content. Finally, test and optimize to ensure a smooth user experience on different devices, and publish digital publications for users to use.

[0004] The prior art has the following deficiencies:

[0005] In the XR environment, the interaction between users and the knowledge graph requires real-time response, especially when dynamically exploring and manipulating complex nodes. When the amount of information in the graph is large and there are many nodes, if there is a delay between operation and feedback due to system performance or network problems, the user may feel "lag" in each operation, that is, the interface is not updated in time or the response speed is too slow after the operation. And the lagging feedback not only seriously affects the user's operational fluency, but also makes the user feel frustrated and uncertain, and even completely lose trust in the system during complex interactions. In addition, too long a delay may also cause the user to mistakenly believe that the operation has failed during the operation, resulting in frequent repeated operations, further exacerbating the delay problem and a vicious cycle. Summary of the invention

[0006] The purpose of the present invention is to provide a method and system for generating digital publications with knowledge graphs integrated with XR to address the deficiencies in the background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a method for generating digital publications with knowledge graphs integrated with XR, comprising the following steps:

[0008] S1: Determine the type of XR scene according to the core goal of the XR scene and the needs of the target users. The types of XR scenes include virtual reality, augmented reality and mixed reality, and collect relevant knowledge points, data and information to form a structured knowledge graph;

[0009] S2: Record the content rendered by the system in real time when the target user operates, including the dynamic changes of nodes, the update of relationship lines, and the scene adjustment in three-dimensional space. Measure the calculation, rendering and response time of the system after each user operation. Simulate knowledge graphs of different scales through stress testing to evaluate the responsiveness of the system under different conditions.

[0010] The step S2 comprises: generating a rendering efficiency fluctuation index according to a change in rendering efficiency when the number of edges in the graph increases;

[0011] S3: Measure the real-time feedback speed when the target user performs different operations in the knowledge graph, analyze the relationship between the real-time feedback speed and the feedback speed expected by the target user, and determine the perceived lag state of the target user when performing different operations;

[0012] The step S3 includes: measuring the actual operation feedback speed of the target user in the knowledge graph, and comparing it with the user's expected feedback time, and generating a feedback delay difference deviation index according to the deviation of the feedback delay difference;

[0013] S4: Analyze the system's responsiveness under different conditions and the perceived lag state of the target user when performing different operations, and determine the real-time responsiveness required for the user's interaction with the knowledge graph based on the analysis results;

[0014] S5: The real-time response capabilities required for the interaction between users and the knowledge graph are divided into high real-time response, medium real-time response, and low real-time response. For high real-time response, the content of high-frequency interactive operations is preferentially rendered locally to reduce latency. For low real-time response, the operations of the target users are transferred to the background for asynchronous processing.

[0015] S6: For medium real-time response, dynamically adjust the rendering detail level according to the abnormal changes in the real-time response capability required for the interaction between the user and the knowledge graph within a fixed time period. If the abnormality is high, simplify the rendering.

[0016] Preferably, in S2, a rendering efficiency fluctuation index is generated according to the change of rendering efficiency when the number of edges in the graph increases, and the method for obtaining the rendering efficiency fluctuation index is:

[0017] Record the rendering efficiency data in the system and construct the rendering efficiency into time series data in chronological order. The time series of rendering efficiency is: ;in, It's time The rendering efficiency at that time, N is the number of sampling points, and the time series of rendering efficiency is converted from the time domain to the frequency domain using discrete Fourier transform. The expression is: ; In the formula, X(k) is the kth frequency component in the frequency domain, indicating the amplitude and phase corresponding to frequency k, x(n) is the nth rendering efficiency data point in the time series, and N is the total number of sampled data points, ranging from 0≤k≤ , is a complex exponential function, which represents the Fourier basis function. The amplitude corresponding to the frequency k represents the contribution of the frequency component to the fluctuation of rendering efficiency. The amplitude calculation formula is: ; Where A(k) is the amplitude of the kth frequency component, and are the real and imaginary parts of the Fourier transform results, respectively. Calculate the total energy of the high-frequency component and select the range of the high-frequency component as , Is the starting point of the high-frequency component, calculate the rendering efficiency fluctuation index, the expression is: ; Where MJ is the rendering efficiency fluctuation index.

[0018] Preferably, in S3, the actual operation feedback speed of the target user in the knowledge graph is measured and compared with the user's expected feedback time, and a feedback delay difference deviation index is generated according to the deviation of the feedback delay difference. The method for obtaining the feedback delay difference deviation index is:

[0019] The actual feedback time and the expected feedback time are considered as two time series respectively. and , where each element in the sequence corresponds to the feedback time of each operation; create an N×M two-dimensional matrix D, where D(i,j) represents the distance between the i-th element in the actual feedback time sequence and the j-th element in the expected feedback time sequence, initialize the distance matrix, and calculate the similarity between the actual feedback time and the expected feedback time. The specific calculation expression is: D(i,j)= ; Where D(i,j) is the similarity between the actual feedback time and the expected feedback time, is the actual feedback time of the system in the i-th operation, is the expected feedback time of the jth operation; construct a p×q cumulative matrix M, where M(i,j) represents the minimum cumulative distance from (1,1) to (i,j). The specific calculation expression is: ; Start from M(p,q) and backtrack to find the optimal alignment path, which represents the alignment method with the minimum cumulative distance. The steps of the backtracking path are: start from (n,m) and look forward for the minimum value: from M(i,j) to ; Repeat the steps until reaching (1,1); Calculate the total cumulative distance of the optimal alignment path, that is, calculate the power efficiency fluctuation index, the expression is: ; Where L is the length of the optimal alignment path, which represents the number of matching path points between the expected feedback sequence and the actual feedback sequence, is the feedback delay difference deviation index.

[0020] Preferably, the rendering efficiency fluctuation index and the feedback delay difference deviation index are converted into a first eigenvector, and the first eigenvector is used as the input of the machine learning model. The machine learning model uses each group of first eigenvectors to predict the real-time response capability value label required for the interaction between the user and the knowledge graph as the prediction target, and takes minimizing the sum of prediction errors of the real-time response capability value labels required for the interaction between all users and the knowledge graph as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The real-time response capability value required for the interaction between the user and the knowledge graph is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0021] Preferably, in S5, the real-time response capability required for the interaction between the user and the knowledge graph is divided into high real-time response, medium real-time response and low real-time response, specifically:

[0022] Compare the acquired real-time response capability value required for the interaction between the user and the knowledge graph with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and compare the real-time response capability value required for the interaction between the user and the knowledge graph with the first standard threshold and the second standard threshold respectively;

[0023] If the real-time response capability value required for the interaction between the user and the knowledge graph is greater than the second standard threshold, it means that the real-time response capability required for the interaction between the user and the knowledge graph is high, and a high real-time response signal is generated at this time, and it is classified as a high real-time response;

[0024] If the real-time response capability value required for the interaction between the user and the knowledge graph is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the real-time response capability required for the interaction between the user and the knowledge graph is average, and a medium real-time response signal is generated at this time, and it is classified as a medium real-time response;

[0025] If the real-time response capability value required for the interaction between the user and the knowledge graph is less than the first standard threshold, it means that the real-time response capability required for the interaction between the user and the knowledge graph is low. At this time, a low real-time response signal is generated and it will be classified as a low real-time response.

[0026] Preferably, for medium real-time response, the rendering detail level is dynamically adjusted according to the abnormal changes in the real-time response capability required for the interaction between the user and the knowledge graph within a fixed time period, specifically:

[0027] For medium real-time response, that is, the real-time response capability value required for the interaction between the user and the knowledge graph generated in a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the real-time response capability value required for the interaction between the user and the knowledge graph generated in a subsequent fixed time period that is greater than or equal to the first standard threshold and less than or equal to the second standard threshold is collected, and a corresponding data set is established, the mean and standard deviation of the data set are calculated, and after analysis, the abnormal changes in the real-time response capability required for the interaction between the user and the knowledge graph in the fixed time period are determined, and the rendering detail level is dynamically adjusted according to the judgment result.

[0028] Preferably, if the mean of the real-time response capability values ​​in the data set is greater than or equal to the reference threshold of the mean of the real-time response capability values, and the standard deviation of the real-time response capability values ​​is less than the reference threshold of the standard deviation of the real-time response capability values, then the system response is stable, the user-perceived delay is small, and high-quality rendering details continue to be maintained;

[0029] If the mean of the real-time response capability value is greater than or equal to the reference threshold of the mean of the real-time response capability value, and the standard deviation of the real-time response capability value is greater than or equal to the reference threshold of the standard deviation of the real-time response capability value, the overall response capability of the system is high, but the response fluctuates greatly and the user experience is unstable. Reduce the rendering detail level to reduce the system load;

[0030] If the mean of the real-time responsiveness value is less than the reference threshold of the mean of the real-time responsiveness value, and the standard deviation of the real-time responsiveness value is greater than or equal to the reference threshold of the standard deviation of the real-time responsiveness value, the average responsiveness of the system is poor and fluctuates greatly, indicating that the system performs poorly in responding to user operations and needs to reduce the rendering detail level;

[0031] If the mean of the real-time response capability value is less than the reference threshold of the mean of the real-time response capability value, and the standard deviation of the real-time response capability value is less than the reference threshold of the standard deviation of the real-time response capability value, the system's response capability is poor but the fluctuation is small, which means that although the overall performance of the system is lagging but stable, the lag perceived by the user exists but is consistent, simplifying the rendering detail level.

[0032] The present invention also provides a knowledge graph digital publication generation system integrated with XR, including a knowledge graph construction module, a real-time rendering module, a lag analysis module, a responsiveness evaluation module, a responsiveness classification module, and a dynamic rendering adjustment module;

[0033] Knowledge graph construction module: Determine the type of XR scenario according to the core objectives of the XR scenario and the needs of the target users. The types of XR scenarios include virtual reality, augmented reality and mixed reality, and collect relevant knowledge points, data and information to form a structured knowledge graph;

[0034] Real-time rendering module: records the content rendered by the system in real time when the target user operates, including the dynamic changes of nodes, the update of relationship lines, and the scene adjustment in three-dimensional space. It measures the calculation, rendering and response time of the system after each user operation, simulates knowledge graphs of different scales through stress testing, and evaluates the responsiveness of the system under different conditions.

[0035] Lag analysis module: measures the real-time feedback speed when the target user performs different operations in the knowledge graph, analyzes the relationship between the real-time feedback speed and the feedback speed expected by the target user, and determines the perceived lag state of the target user when performing different operations;

[0036] Responsiveness evaluation module: Analyzes the responsiveness of the system under different conditions and the perceived lag state of the target user when performing different operations, and determines the real-time responsiveness required for the interaction between the user and the knowledge graph based on the analysis results;

[0037] Response capability classification module: The real-time response capability required for the interaction between users and knowledge graphs is divided into high real-time response, medium real-time response and low real-time response. For high real-time response, the content of high-frequency interactive operations is preferentially rendered locally to reduce latency. For low real-time response, the target user's operations are transferred to the background for asynchronous processing.

[0038] Dynamic rendering adjustment module: For medium real-time response, the rendering detail level is dynamically adjusted according to the abnormal changes in the real-time response capabilities required for the interaction between the user and the knowledge graph within a fixed time period. If the abnormality is high, the rendering is simplified.

[0039] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0040] 1. The present invention determines the type of XR scenario, constructs a structured knowledge graph, and records the system response of user interaction operations in real time, combines stress testing and lag analysis, and accurately evaluates the system's responsiveness under different conditions. By measuring the difference between the user's actual feedback and expected feedback, a feedback delay difference deviation index is generated, and combined with the rendering efficiency fluctuation index, a polynomial regression model is used to predict the real-time response capability required for the user to interact with the knowledge graph, thereby achieving accurate responsiveness classification. This method dynamically optimizes the rendering strategies for high, medium, and low real-time operations based on the system's real-time response capability, ensuring that the system can prioritize local rendering in high real-time scenarios to reduce latency, and asynchronously process in low real-time operations to save resources.

[0041] 2. The present invention can dynamically adjust the rendering detail level according to abnormal changes in real-time response capabilities, ensuring a stable and smooth user experience under different load conditions. This method effectively solves the interaction delay problem caused by system performance or network problems, improves the operation smoothness in the XR environment, avoids users' frustration or misoperation due to delayed feedback, and significantly enhances users' trust and satisfaction with the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0043] Figure 1 is a flow chart of the method of the present invention;

[0044] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] Example 1, please refer to Figure 1 and Figure 2 As shown, the method for generating a knowledge graph digital publication integrated with XR described in this embodiment includes the following steps:

[0047] S1: Determine the type of XR scene according to the core goal of the XR scene and the needs of the target users. The types of XR scenes include virtual reality, augmented reality and mixed reality, and collect relevant knowledge points, data and information to form a structured knowledge graph;

[0048] S2: Record the content rendered by the system in real time when the target user operates, including the dynamic changes of nodes, the update of relationship lines, and the scene adjustment in three-dimensional space. Measure the calculation, rendering and response time of the system after each user operation. Simulate knowledge graphs of different scales through stress testing to evaluate the responsiveness of the system under different conditions.

[0049] S3: Measure the real-time feedback speed when the target user performs different operations in the knowledge graph, analyze the relationship between the real-time feedback speed and the feedback speed expected by the target user, and determine the perceived lag state of the target user when performing different operations;

[0050] S4: Analyze the system's responsiveness under different conditions and the perceived lag state of the target user when performing different operations, and determine the real-time responsiveness required for the user's interaction with the knowledge graph based on the analysis results;

[0051] S5: The real-time response capabilities required for the interaction between users and the knowledge graph are divided into high real-time response, medium real-time response, and low real-time response. For high real-time response, the content of high-frequency interactive operations is preferentially rendered locally to reduce latency. For low real-time response, the operations of the target users are transferred to the background for asynchronous processing.

[0052] S6: For medium real-time response, dynamically adjust the rendering detail level according to the abnormal changes in the real-time response capability required for the interaction between the user and the knowledge graph within a fixed time period. If the abnormality is high, simplify the rendering.

[0053] Among them, in S1, the type of XR scene is determined according to the core goal of the XR scene and the needs of the target users. The types of XR scenes include virtual reality, augmented reality and mixed reality, and relevant knowledge points, data and information are collected and organized into a structured knowledge graph, specifically:

[0054] XR scene types are divided into virtual reality (VR), augmented reality (AR) and mixed reality (MR). Determining the appropriate XR type based on project goals and user needs is the primary task, and first clarify the goals of the XR project. For example, whether it is for education, training, data visualization, commercial display, etc. The core functional requirements of the scene will determine the choice of XR type. VR scene: suitable for a fully immersive experience, such as a virtual education environment, panoramic display, etc., where users need to completely leave the real world. AR scene: suitable for superimposing digital content with the real environment, such as interactive teaching and maintenance guidance, where users need to interact with the real environment in real time. MR scene: used in situations where virtual objects and real environments need to be operated simultaneously, such as design, simulation, etc.

[0055] Analyze the target users' familiarity with technology, their habits of using the device, and their specific expectations. For example: VR: Suitable for user groups that need deep immersion, such as scientific research, gamers, or professional learners in specific industries. AR: Suitable for users who need to stay connected to the real environment, such as workers, engineers, teachers, etc., who hope to be able to overlay virtual information in the actual working environment. MR: Suitable for users who need to operate virtual and real objects at the same time, such as designers, engineers, or simulation operations in the medical field.

[0056] Determine the types of devices used by the target users of the XR project, such as whether a head-mounted display (VR headset), smartphone or AR glasses are needed. Different technical equipment determines the choice of scene type: VR devices: such as Oculus Rift, HTC Vive, suitable for completely virtual immersive experience. AR devices: such as smartphones, tablets, HoloLens, etc., suitable for information display and operation of augmented reality. MR devices: MR headsets, such as HoloLens 2, combine virtual and real information interaction, suitable for complex simulation and design scenarios.

[0057] Evaluate whether the operating environment of the XR application requires the user to be completely in the virtual world or to interact with the real environment. For example: VR: If the user needs a completely virtual world to focus on specific content, such as virtual training, simulation, etc., VR is the best choice. AR: If the user needs to receive virtual information directly in the real environment during work and study, such as surgical guidance or AR navigation, AR is the first choice. MR: If the application scenario involves operating real and virtual objects at the same time, such as industrial design or architectural display, MR is more suitable.

[0058] Analyze the user's interaction needs in the XR scene, whether a large number of gestures, voice, and body movements are required to interact with virtual content: VR: Suitable for complex interactive content, such as panoramic roaming, virtual operations, etc. AR: Suitable for simplified interactions, such as clicking and scanning real objects to activate virtual information display. MR: Users can directly interact with virtual and real objects, suitable for multi-dimensional interactions that require high complexity.

[0059] Knowledge graph is a structured way to represent the complex relationships between data and information. In XR scenarios, knowledge graph can help users explore complex information more intuitively and effectively. According to the content theme of the XR project, collect knowledge points, information and data sources in related fields. These sources may include: Text materials: such as academic papers, manuals, books, etc. Multimedia content: including pictures, videos, 3D models. External databases: such as open semantic networks, industry databases, etc. Ensure data quality: Ensure that the collected knowledge points and data are accurate and credible, and the content is hierarchical and relevant.

[0060] Define each independent knowledge point or information fragment as a "node" in the knowledge graph. These nodes can be specific concepts, events, objects, etc. Edges (representing relationships): Establish the association between nodes and define the relationship between different nodes. Relationship types include: Hierarchical relationship: Represents the relationship between superiors and subordinates, parent-child structures, such as the relationship between concepts and sub-concepts. Parallel relationship: Indicates that two or more nodes belong to the same category or level. Association relationship: Indicates the intersection or connection between different concepts, such as causal relationship or influence relationship.

[0061] Connect different nodes with edges to clarify the type of relationship between them. Relationships can represent causal relationships, correlations, time sequences, etc. of knowledge points. Example: In a historical knowledge graph, a historical event is a first-level node, related people and places are second-level nodes, and the timeline between events is a relationship line. Data visualization tool selection: Use graph building tools (such as Neo4j, Gephi, Cytoscape, etc.) to visualize the organized knowledge points and relationship data for easy subsequent operations in XR scenarios.

[0062] Ensure that the hierarchy of the knowledge graph is clear and easy to understand, that key information (such as first-level nodes) is highlighted, and that secondary information can be gradually revealed through interactive exploration. Choose an appropriate visualization layout based on the complexity of knowledge points and relationships. For example, use a grid, tree, star, or hierarchical layout so that users can intuitively understand the relationship between each knowledge point. Design the interactive mode of the graph for XR scenarios. For example, when a user clicks on a node, the relevant information of the node is displayed, and the associated content of the node is dynamically loaded.

[0063] In this application, when determining the type of XR scene, by analyzing the core goals, user needs, device capabilities, environmental requirements and interaction complexity, it is possible to effectively decide whether to use virtual reality, augmented reality or mixed reality. When collecting and organizing relevant knowledge points and data, the information is converted into a structured knowledge graph through steps such as defining nodes and edges, classification hierarchy, and relationship visualization, so that users can efficiently explore and understand complex knowledge systems in the XR environment.

[0064] S2: Record the content rendered by the system in real time when the target user operates, including the dynamic changes of nodes, the update of relationship lines, and the scene adjustment in three-dimensional space. Measure the calculation, rendering, and response time of the system after each user operation. Simulate knowledge graphs of different sizes through stress testing to evaluate the system's responsiveness under different conditions.

[0065] Record user operations and system real-time rendering content, and classify them according to the typical interactive operations of the target users. Common operations include: Clicking a node: When a user selects a specific knowledge point, the system needs to display relevant information. Drag the graph: The user drags the graph to view different areas or adjust the viewing angle. Zooming the scene: The user zooms in or out of the knowledge graph to view overall or local details. Relationship exploration: The user expands or collapses nodes to show or hide the relationship lines between nodes.

[0066] Whenever the user performs these operations, the system performs a series of rendering tasks, including: Dynamic changes of nodes: When the user clicks or drags a node, the system updates its position, size or display status in real time. Update of relationship lines: The relationship between different nodes in the knowledge graph needs to be displayed dynamically. As the node moves or is selected, the relationship lines will be adjusted synchronously. Scene adjustment in three-dimensional space: If the knowledge graph is a 3D structure, the system will adjust the perspective, distance and rotation angle of the entire graph according to the user's interactive actions so that the user can get a suitable view in the virtual space.

[0067] After each user operation, the system needs to process data and calculate. For example, when a user clicks on a node, the system calculates the node's associations, possible data loading, etc. The duration of this process can be recorded by monitoring the system's computing performance. Once the calculation is completed, the system needs to convert the calculation results into visual elements and present them to the user, which involves graphics rendering. In knowledge graphs, especially when complex 3D images are involved, rendering time is critical. Rendering time can be measured by frame rate (FPS), and higher FPS indicates higher rendering efficiency. After each user interaction, record whether the update speed of the graph nodes and relationship lines is within the acceptable range for the user. Ideally, the response time should be short enough to maintain a smooth operation experience. Response time is the delay time from the user performing an operation (such as clicking, dragging) to seeing the result on the screen, usually measured in milliseconds. For a good user experience, the response time should usually be controlled between 100 milliseconds and 500 milliseconds.

[0068] Stress testing is used to simulate knowledge graphs of different sizes, including: dividing the scale of the knowledge graph into: Small graph: a small number of nodes and relationship lines (such as dozens of nodes). Medium-sized graph: contains multiple sub-nodes and a large number of nodes (hundreds of nodes). Large graph: contains a large number of nodes and complex relationships (thousands of nodes or even more). By gradually increasing the complexity of the graph, the system's processing capabilities when dealing with different scales are evaluated. You can use simulated data or actual large-scale knowledge graphs for testing. Specifically: gradually increase the number of nodes and monitor the system's computing, rendering, and response performance when processing multiple nodes. Evaluate the system's performance in processing complex relationships between a large number of nodes, especially the rendering efficiency when the number of edges in the graph increases. Simulate user operations in complex knowledge graphs, such as frequent clicking, dragging, zooming, etc., and observe the system's performance under high load.

[0069] The rendering efficiency fluctuation index is generated according to the change of rendering efficiency when the number of edges in the graph increases, and the responsiveness of the system under different conditions is evaluated. The method for obtaining the rendering efficiency fluctuation index is as follows:

[0070] Record rendering efficiency data in the system, such as the frame rate per second (FPS) or the time of each rendering (rendering delay, in milliseconds). Construct the rendering efficiency into time series data in chronological order, assuming that the sampling frequency is to record the rendering efficiency data once per second (FPS). The time series of rendering efficiency is: ;in, It's time The rendering efficiency at the time, N is the number of sampling points, and the discrete Fourier transform (DFT) is used to transform the time series of the rendering efficiency from the time domain to the frequency domain. The expression is: ; In the formula, X(k) is the kth frequency component (complex form) in the frequency domain, which represents the amplitude and phase corresponding to frequency k, x(n) is the nth rendering efficiency data point in the time series, N is the total number of sampled data points, and k is the frequency index, with a value range of 0≤k≤ , is a complex exponential function, which represents the Fourier basis function. The amplitude corresponding to the frequency k represents the contribution of the frequency component to the fluctuation of rendering efficiency. The amplitude calculation formula is: ; Where A(k) is the amplitude of the kth frequency component, and are the real and imaginary parts of the Fourier transform result respectively. The fluctuation of rendering efficiency is mainly reflected in the high-frequency component in the frequency domain, because the high-frequency component represents the part where the rendering efficiency fluctuates dramatically over time. In order to quantify the fluctuation index of rendering efficiency, the total energy or contribution of the high-frequency component can be calculated. The range of the high-frequency component is selected as , that is, ignoring the low-frequency part (low-frequency mainly reflects the long-term trend), Is the starting point of the high frequency component, usually selected = N / 4 or N / 3, adjust according to the specific situation to ensure that only the high-frequency components with violent fluctuations are captured. Calculate the rendering efficiency fluctuation index, the expression is: ; Where MJ is the rendering efficiency fluctuation index.

[0071] The larger the rendering efficiency fluctuation index, the more dramatic the fluctuation of the rendering efficiency of the system under different conditions, reflecting the poor responsiveness of the system, especially when dealing with complex or high-load scenarios, the frame rate or rendering time of the system is unstable. A larger fluctuation index means that users may frequently encounter delays, freezes, or frame rate drops during the interaction process, which indicates that the system cannot maintain consistent rendering performance in different scenarios, affecting the user experience. Usually, this situation is more common when the number of graph nodes increases, the relationships are complex, or the system resources are limited.

[0072] The smaller the rendering efficiency fluctuation index, the more stable the rendering efficiency of the system under different conditions, which means that the system can adapt well to various load conditions and maintain consistent responsiveness. A smaller fluctuation index reflects that the frame rate or rendering time fluctuates less. Even when processing complex graphs or high loads, the system can complete the rendering task in a relatively stable manner and provide a smooth user experience. This shows that the system has strong processing capabilities and can effectively cope with the rendering needs of complex operations or large amounts of data.

[0073] S3: Measure the real-time feedback speed when the target user performs different operations in the knowledge graph, analyze the relationship between the real-time feedback speed and the feedback speed expected by the target user, and determine the perceived lag state of the target user when performing different operations.

[0074] Different types of operations are divided according to the common interactive operations between users and knowledge graphs. Common operations include: Clicking a node: Users click on a knowledge point and expect the system to quickly feedback relevant information. Drag the graph: Users drag the graph and expect the system to update the display in real time. Zoom view: Users adjust the display ratio of the graph by zooming in or out. Relationship display: Users expand the associated information of a node to display adjacent nodes and connecting lines.

[0075] After the user performs these operations, the system records the time consumed from the operation triggering to the system giving feedback, which is called the actual feedback time. The real-time feedback speed is usually measured in milliseconds, indicating the time interval from the user performing the operation to the visual feedback appearing on the interface. Use internal logs to record the timestamp of each operation. Record the time point when the system completes rendering and feedback. Real-time feedback speed = feedback completion time - operation execution time.

[0076] Common feedback times include: Click feedback: The expected feedback time is usually between 100-300 milliseconds. Drag the graph: The expected drag feedback should be less than 50 milliseconds to ensure smoothness. Zoom operation: The feedback time should ideally be controlled at 100-200 milliseconds. Relationship expansion: Depending on the amount of data, the expected feedback time varies from 200-500 milliseconds.

[0077] Depending on the type of operation, set the target user's expected feedback time, which can be determined through user research, industry standards, or empirical values. Visual feedback delays after an operation that are less than 100 milliseconds are usually not noticeable by users (the minimum delay perception ability of human vision). Continuous operations such as dragging and zooming should provide feedback within 50 milliseconds to maintain a smooth experience. More complex operations (such as expanding a large number of related nodes) can have longer expected feedback times, usually no more than 500 milliseconds.

[0078] Analyze the difference between the actual feedback time and the expected feedback time, and calculate the lag degree of each operation. The expression is: ; In the formula, is the actual feedback time of the system, is the user's expected feedback time, is the feedback delay difference. ΔT≤0: The system feedback time is consistent with the user's expectations or faster, and there is no lag. 0<ΔT≤100ms: Slight lag, the user may occasionally perceive the delay of the operation, but it has little impact on the experience. ΔT>100ms: Obvious lag, the user perceives an obvious delay in the operation feedback, which may cause frustration or cognitive discomfort.

[0079] Measure the actual operation feedback speed of the target user in the knowledge graph and compare it with the user's expected feedback time. Generate a feedback delay difference deviation index based on the deviation of the feedback delay difference. The method for obtaining the feedback delay difference deviation index is:

[0080] The actual feedback time and the expected feedback time are considered as two time series respectively. and , where each element in the sequence corresponds to the feedback time of each operation; create an N×M two-dimensional matrix D, where D(i,j) represents the distance between the i-th element in the actual feedback time sequence and the j-th element in the expected feedback time sequence, initialize the distance matrix, and calculate the similarity between the actual feedback time and the expected feedback time. The specific calculation expression is: D(i,j)= ; Where D(i,j) is the similarity between the actual feedback time and the expected feedback time, is the actual feedback time of the system in the i-th operation, is the expected feedback time of the jth operation; construct a p×q cumulative matrix M, where M(i,j) represents the minimum cumulative distance from (1,1) to (i,j). The specific calculation expression is: ; Start from M(p,q) and backtrack to find the optimal alignment path, which represents the alignment method with the minimum cumulative distance. The steps of the backtracking path are: start from (n,m) and look forward for the minimum value: from M(i,j) to ; Repeat the steps until reaching (1,1); Calculate the total cumulative distance of the optimal alignment path, that is, calculate the power efficiency fluctuation index, the expression is: ; Where L is the length of the optimal alignment path, which represents the number of matching path points between the expected feedback sequence and the actual feedback sequence, is the feedback delay difference deviation index.

[0081] The larger the deviation index of the feedback delay difference, the greater the deviation between the actual feedback time of the system and the feedback time expected by the user, and the target user perceives obvious lag when performing operations. An increase in the index means that the fluctuation of the feedback delay is larger, and the user's operation experience becomes less smooth, which may cause users to frequently perceive system freezes or delayed operation responses. This sense of lag will significantly reduce user satisfaction, especially in frequent operations (such as dragging and zooming), where users may mistakenly believe that the operation failed or the system performance is poor.

[0082] The smaller the feedback delay difference deviation index is, the smaller the deviation between the system feedback and the user's expected feedback time is. The user hardly feels any obvious lag when performing an operation, and the operation experience is smoother and more natural. A smaller index means that the system can respond within the time expected by the user, and the user's operation feedback perception is more timely and in line with expectations. In this case, the user experience will be smoother, the continuity of the operation can be maintained, the system performance perceived by the user is relatively stable, and the lag is weak or non-existent.

[0083] S4: Analyze the system’s responsiveness under different conditions and the perceived lag of target users when performing different operations, and determine the real-time responsiveness required for the user’s interaction with the knowledge graph based on the analysis results.

[0084] The rendering efficiency fluctuation index and the feedback delay difference deviation index are converted into the first eigenvector, and the first eigenvector is used as the input of the machine learning model. The machine learning model uses each group of first eigenvectors to predict the real-time response capability value label required for the interaction between the user and the knowledge graph as the prediction target, and takes minimizing the sum of prediction errors of the real-time response capability value labels required for the interaction between all users and the knowledge graph as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The real-time response capability value required for the interaction between the user and the knowledge graph is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0085] The method for obtaining the real-time response capability value required for the interaction between the user and the knowledge graph is: from the first feature vector training data of the trained machine learning model, obtain the corresponding function expression: ; In the formula, is the output function of the model, MJ is the rendering efficiency fluctuation index, is the feedback delay difference deviation index, The real-time responsiveness value required for the interaction between users and the knowledge graph.

[0086] S5: The real-time response capabilities required for the interaction between users and knowledge graphs are divided into high real-time response, medium real-time response and low real-time response. For high real-time response, the content of high-frequency interactive operations is preferentially rendered locally to reduce latency. For low real-time response, the target user's operations are transferred to the background for asynchronous processing.

[0087] Compare the acquired real-time response capability value required for the interaction between the user and the knowledge graph with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and compare the real-time response capability value required for the interaction between the user and the knowledge graph with the first standard threshold and the second standard threshold respectively;

[0088] If the real-time response capability value required for the interaction between the user and the knowledge graph is greater than the second standard threshold, it means that the real-time response capability required for the interaction between the user and the knowledge graph is high, and a high real-time response signal is generated at this time, and it is classified as a high real-time response;

[0089] If the real-time response capability value required for the interaction between the user and the knowledge graph is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the real-time response capability required for the interaction between the user and the knowledge graph is average, and a medium real-time response signal is generated at this time, and it is classified as a medium real-time response;

[0090] If the real-time response capability value required for the interaction between the user and the knowledge graph is less than the first standard threshold, it means that the real-time response capability required for the interaction between the user and the knowledge graph is low. At this time, a low real-time response signal is generated and it will be classified as a low real-time response.

[0091] Local rendering with high real-time response: When the system detects that the user's operation is subject to high real-time response requirements, such as frequent node clicks, dragging, zooming and other high-interaction density operations, local rendering is preferred to handle these tasks. By rendering on the user's device instead of relying on a remote server, the system can significantly reduce latency and reduce the risk of delays caused by network transmission, thereby improving response speed. Local rendering can ensure that users get instant feedback during operation, ensure smooth operation, avoid users from perceiving any lag, and improve the overall interactive experience.

[0092] Asynchronous processing of low real-time response: For low real-time response operations that do not require immediate feedback, such as global search, complex relationship display, or background data loading, the system transfers these tasks to the background for asynchronous processing. Asynchronous processing allows the system to complete data calculation or content loading in the background, and users can continue to perform other operations and wait for the task to be completed before receiving the result or notification. In this way, the system avoids blocking the user interface, improves parallel processing capabilities, and ensures the continuity of the user experience. Even if there is a certain delay, the user will not feel a significant operation jam.

[0093] S6: For medium real-time response, dynamically adjust the rendering detail level according to the abnormal changes in the real-time response capability required for the interaction between the user and the knowledge graph within a fixed time period. If the abnormality is high, simplify the rendering.

[0094] For medium real-time response, that is, the real-time response capability value required for the interaction between the user and the knowledge graph generated in a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the real-time response capability value required for the interaction between the user and the knowledge graph generated in a subsequent fixed time period that is greater than or equal to the first standard threshold and less than or equal to the second standard threshold is collected, and a corresponding data set is established, the mean and standard deviation of the data set are calculated, and after analysis, the abnormal changes in the real-time response capability required for the interaction between the user and the knowledge graph in the fixed time period are determined, and the rendering detail level is dynamically adjusted according to the judgment result.

[0095] If the mean of the real-time response capability values ​​in the data set is greater than or equal to the reference threshold of the mean of the real-time response capability values, and the standard deviation of the real-time response capability values ​​is less than the reference threshold of the standard deviation of the real-time response capability values, then the system's response capability is high, and the response is very stable with small fluctuations. This means that the system performs well within the fixed time period, and the user-perceived delay is small. The rendering detail level can be maintained at a high level. Due to the light system load and stable response, high-quality rendering details can continue to be maintained to provide rich visual effects and user experience.

[0096] If the mean of the real-time response capability value is greater than or equal to the reference threshold of the mean of the real-time response capability value, and the standard deviation of the real-time response capability value is greater than or equal to the reference threshold of the standard deviation of the real-time response capability value, the overall response capability of the system is high, but the response fluctuation is large, indicating that the system may have a large delay in certain operations. The user experience is unstable, sometimes good and sometimes bad. You can moderately reduce the rendering detail level to reduce the system load, especially simplifying the rendering during the period of large fluctuations. This can reduce the fluctuations caused by rendering and improve the stability of the system under high load.

[0097] If the mean of the real-time response capability value is less than the reference threshold of the mean of the real-time response capability value, and the standard deviation of the real-time response capability value is greater than or equal to the reference threshold of the standard deviation of the real-time response capability value, the average response capability of the system is poor and fluctuates greatly, indicating that the system performs poorly in responding to user operations. Users may frequently experience freezes or response delays. The rendering detail level needs to be significantly reduced. At this time, in order to ensure the stability of the system and improve the response speed, the complexity of the image can be greatly reduced, and the display and animation effects of the nodes can be simplified to reduce the processing burden of the system.

[0098] If the mean of the real-time response capability value is less than the reference threshold of the mean of the real-time response capability value, and the standard deviation of the real-time response capability value is less than the reference threshold of the standard deviation of the real-time response capability value, the system's response capability is poor, but the fluctuation is small, which means that although the overall performance of the system is lagging, it is relatively stable, and the user-perceived lag exists but is relatively consistent. The rendering detail level can be moderately simplified. Because the system's response capability is weak, the rendering complexity must be reduced, but because the fluctuation is small, some details can be slightly maintained without over-simplification.

[0099] High abnormality, that is, low mean and large fluctuations. When the system's responsiveness is seriously insufficient and the fluctuations are large, the rendering details should be greatly reduced. Specifically, it includes: reducing the complexity of graphics rendering: reducing the display details of nodes in the graph, simplifying or canceling animation effects, and reducing unnecessary rendering of graphic elements. Reduce rendering frame rate: moderately reduce the rendering frequency so that the system can focus on processing key interactive tasks.

[0100] Low abnormality, that is, high mean and small fluctuations, when the system responsiveness is relatively good and stable, you can maintain a high rendering quality, and only make slight adjustments when necessary to ensure continued stability: Maintain high-detail rendering: Continue to show high-quality details in the graph, including complex nodes and relationship lines. Enable high-quality textures and animations: When the user operation is less, high-resolution textures and smooth animation effects can be maintained to improve the visual experience.

[0101] In this application, by analyzing the mean and standard deviation of the interaction data set between users and the knowledge graph within a fixed time period, the system can intelligently determine the abnormal changes in real-time response capabilities and dynamically adjust the rendering details according to different abnormality levels. When the abnormality level is high, the rendering is simplified to ensure a smooth user experience; when the abnormality level is low, high-quality rendering is maintained to ensure visual effects and interaction quality.

[0102] In this embodiment, in the XR scenario, the XR type is first determined according to the core goals and target user needs, and relevant knowledge points are collected and organized to form a structured knowledge graph. Next, the real-time rendering content of the system during user operation is recorded, and the system responsiveness under different conditions is evaluated through calculation, rendering, response time and stress testing. Then, the real-time feedback speed of the user operation is measured, and its relationship with the expected feedback speed is analyzed to determine the user's perceived lag state. Based on this analysis, the real-time response capability required for the user to interact with the knowledge graph is determined, and it is divided into three response levels: high, medium, and low. When the real-time response is high, high-frequency interactive operations are rendered locally first; when the real-time response is low, the task is transferred to the background asynchronous processing. For medium real-time response, the rendering details are dynamically adjusted by analyzing the abnormal changes in the response capability within a fixed time period. If the degree of abnormality is high, the rendering is simplified to optimize performance.

[0103] Embodiment 2, a knowledge graph digital publication generation system integrated with XR described in this embodiment includes a knowledge graph construction module, a real-time rendering module, a lag analysis module, a responsiveness evaluation module, a responsiveness classification module and a dynamic rendering adjustment module;

[0104] Knowledge graph construction module: Determine the type of XR scenario according to the core objectives of the XR scenario and the needs of the target users. The types of XR scenarios include virtual reality, augmented reality and mixed reality, and collect relevant knowledge points, data and information to form a structured knowledge graph;

[0105] Real-time rendering module: records the content rendered by the system in real time when the target user operates, including the dynamic changes of nodes, the update of relationship lines, and the scene adjustment in three-dimensional space. It measures the calculation, rendering and response time of the system after each user operation, simulates knowledge graphs of different scales through stress testing, and evaluates the responsiveness of the system under different conditions.

[0106] Lag analysis module: measures the real-time feedback speed when the target user performs different operations in the knowledge graph, analyzes the relationship between the real-time feedback speed and the feedback speed expected by the target user, and determines the perceived lag state of the target user when performing different operations;

[0107] Responsiveness evaluation module: Analyzes the responsiveness of the system under different conditions and the perceived lag state of the target user when performing different operations, and determines the real-time responsiveness required for the interaction between the user and the knowledge graph based on the analysis results;

[0108] Response capability classification module: The real-time response capability required for the interaction between users and knowledge graphs is divided into high real-time response, medium real-time response and low real-time response. For high real-time response, the content of high-frequency interactive operations is preferentially rendered locally to reduce latency. For low real-time response, the target user's operations are transferred to the background for asynchronous processing.

[0109] Dynamic rendering adjustment module: For medium real-time response, the rendering detail level is dynamically adjusted according to the abnormal changes in the real-time response capabilities required for the interaction between the user and the knowledge graph within a fixed time period. If the abnormality is high, the rendering is simplified.

[0110] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0111] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0112] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0113] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for generating digital publications with knowledge graphs integrated with XR, characterized in that: The steps include: S1: Determine the type of XR scene according to the core goal of the XR scene and the needs of the target users. The types of XR scenes include virtual reality, augmented reality and mixed reality, and collect relevant knowledge points, data and information to form a structured knowledge graph; S2: Record the content rendered by the system in real time when the target user operates, including the dynamic changes of nodes, the update of relationship lines, and the scene adjustment in three-dimensional space. Measure the calculation, rendering and response time of the system after each user operation. Simulate knowledge graphs of different scales through stress testing to evaluate the responsiveness of the system under different conditions. The step S2 comprises: generating a rendering efficiency fluctuation index according to a change in rendering efficiency when the number of edges in the graph increases; S3: Measure the real-time feedback speed when the target user performs different operations in the knowledge graph, analyze the relationship between the real-time feedback speed and the feedback speed expected by the target user, and determine the perceived lag state of the target user when performing different operations; The step S3 includes: measuring the actual operation feedback speed of the target user in the knowledge graph, and comparing it with the feedback time expected by the user, and generating a feedback delay difference deviation index according to the deviation of the feedback delay difference; S4: Analyze the system's responsiveness under different conditions and the perceived lag state of the target user when performing different operations, and determine the real-time responsiveness required for the interaction between the user and the knowledge graph based on the analysis results; specifically, it includes: The rendering efficiency fluctuation index and the feedback delay difference deviation index are converted into the first eigenvector, and the first eigenvector is used as the input of the machine learning model. The machine learning model predicts the real-time response capability value label required for the interaction between the user and the knowledge graph by each group of first eigenvectors as the prediction target, and minimizes the sum of the prediction errors of the real-time response capability value labels required for the interaction between all users and the knowledge graph as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence, and the model training is stopped. The real-time response capability value required for the interaction between the user and the knowledge graph is determined according to the model output result, wherein the machine learning model is a polynomial regression model; S5: The real-time response capabilities required for the interaction between users and the knowledge graph are divided into high real-time response, medium real-time response, and low real-time response. For high real-time response, the content of high-frequency interactive operations is preferentially rendered locally to reduce latency. For low real-time response, the operations of the target users are transferred to the background for asynchronous processing. S6: For medium real-time response, dynamically adjust the rendering detail level according to the abnormal changes in the real-time response capability required for the interaction between the user and the knowledge graph within a fixed time period. If the abnormality is high, simplify the rendering; Specifically: For medium real-time response, that is, the real-time response capability value required for the interaction between the user and the knowledge graph generated in a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the real-time response capability value required for the interaction between the user and the knowledge graph generated in a subsequent fixed time period that is greater than or equal to the first standard threshold and less than or equal to the second standard threshold is collected, and a corresponding data set is established, the mean and standard deviation of the data set are calculated, and after analysis, the abnormal changes in the real-time response capability required for the interaction between the user and the knowledge graph in the fixed time period are determined, and the rendering detail level is dynamically adjusted according to the judgment result; If the mean of the real-time response capability values ​​in the data set is greater than or equal to the reference threshold of the mean of the real-time response capability values, and the standard deviation of the real-time response capability values ​​is less than the reference threshold of the standard deviation of the real-time response capability values, high-quality rendering details are continued; If the mean of the real-time response capability values ​​is greater than or equal to the reference threshold of the mean of the real-time response capability values, and the standard deviation of the real-time response capability values ​​is greater than or equal to the reference threshold of the standard deviation of the real-time response capability values, simplify the rendering detail level to reduce the system load; If the mean of the real-time response capability values ​​is less than the reference threshold of the mean of the real-time response capability values, and the standard deviation of the real-time response capability values ​​is greater than or equal to the reference threshold of the standard deviation of the real-time response capability values, the rendering detail level is reduced and the reduction is greater than that of simplifying the rendering detail level; If the mean of the real-time responsiveness values ​​is less than a reference threshold of the mean of the real-time responsiveness values, and the standard deviation of the real-time responsiveness values ​​is less than a reference threshold of the standard deviation of the real-time responsiveness values, the rendering detail level is simplified.

2. According to the method for generating digital publications with knowledge graph integrated with XR in claim 1, it is characterized by: In S2, a rendering efficiency fluctuation index is generated according to the rendering efficiency change when the number of edges in the graph increases. The rendering efficiency fluctuation index is obtained as follows: Record the rendering efficiency data in the system and construct the rendering efficiency into time series data in chronological order. The time series of rendering efficiency is: ;in, It's time The rendering efficiency at that time, N is the number of sampling points, and the time series of rendering efficiency is converted from the time domain to the frequency domain using discrete Fourier transform. The expression is: ; In the formula, X(k) is the kth frequency component in the frequency domain, indicating the amplitude and phase corresponding to frequency k, x(n) is the nth rendering efficiency data point in the time series, N is the total number of sampled data points, and k ranges from 0≤k≤ , is a complex exponential function, which represents the Fourier basis function. The amplitude corresponding to the frequency k represents the contribution of the frequency component to the fluctuation of rendering efficiency. The amplitude calculation formula is: ; Where A(k) is the amplitude of the kth frequency component, and are the real and imaginary parts of the Fourier transform results, respectively. Calculate the total energy of the high-frequency component and select the range of the high-frequency component as , Is the starting point of the high-frequency component, calculate the rendering efficiency fluctuation index, the expression is: ; Where MJ is the rendering efficiency fluctuation index.

3. The method for generating digital publications with knowledge graph integrated with XR according to claim 2, characterized in that: In S3, the actual operation feedback speed of the target user in the knowledge graph is measured and compared with the user's expected feedback time. The feedback delay difference deviation index is generated according to the deviation of the feedback delay difference. The method for obtaining the feedback delay difference deviation index is: The actual feedback time and the expected feedback time are considered as two time series respectively. and , where each element in the sequence corresponds to the feedback time of each operation; Create an N×M two-dimensional matrix D, where D(i,j) represents the distance between the i-th element in the actual feedback time series and the j-th element in the expected feedback time series. Initialize the distance matrix and calculate the similarity between the actual feedback time and the expected feedback time. The specific calculation expression is: D(i,j)= ; Where D(i,j) is the similarity between the actual feedback time and the expected feedback time, is the actual feedback time of the system in the i-th operation, is the expected feedback time of the jth operation; Construct a p×q cumulative matrix M, where M(i,j) represents the minimum cumulative distance from (1,1) to (i,j). The specific calculation expression is: ; Starting from M(p,q), we can backtrack to find the optimal alignment path, which represents the alignment with the minimum cumulative distance. The steps of the backtracking path are: starting from (n,m), looking forward for the minimum value: from M(i,j) to ; Repeat the steps until reaching (1,1); Calculate the total cumulative distance of the optimal alignment path, that is, calculate the feedback delay difference deviation index, the expression is: ; Where L is the length of the optimal alignment path, which represents the number of matching path points between the expected feedback sequence and the actual feedback sequence, is the feedback delay difference deviation index.

4. The method for generating digital publications with knowledge graph integrated with XR according to claim 3, characterized in that: In S5, the real-time response capability required for the interaction between users and knowledge graphs is divided into high real-time response, medium real-time response and low real-time response, specifically: Compare the acquired real-time response capability value required for the interaction between the user and the knowledge graph with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and compare the real-time response capability value required for the interaction between the user and the knowledge graph with the first standard threshold and the second standard threshold respectively; If the real-time response capability value required for the interaction between the user and the knowledge graph is greater than the second standard threshold, it means that the real-time response capability required for the interaction between the user and the knowledge graph is high, and a high real-time response signal is generated at this time, and the real-time response capability required for the interaction between the user and the knowledge graph is classified as a high real-time response; If the real-time response capability value required for the interaction between the user and the knowledge graph is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the real-time response capability required for the interaction between the user and the knowledge graph is average, and a medium real-time response signal is generated at this time, and the real-time response capability required for the interaction between the user and the knowledge graph is classified as a medium real-time response; If the real-time response capability value required for the interaction between the user and the knowledge graph is less than the first standard threshold, it means that the real-time response capability required for the interaction between the user and the knowledge graph is low. At this time, a low real-time response signal is generated, and the real-time response capability required for the interaction between the user and the knowledge graph is classified as a low real-time response.

Images