Intelligent graph loading and rendering method and system for mobile equipment application program

Abstract

The invention provides an intelligent graph loading and rendering method and system for a mobile device application program, and the method comprises the steps: S1, initializing an SDK parameter of an intelligent graph development unit, and returning the SDK parameter to a mobile device; s2, the intelligent graph development unit collects sensor data and system performance data of the mobile device and user interaction behavior data of an application program, and uploads the data to a server; a prediction model is embedded in the server; s3, the server trains the prediction model through the sensor data and the system performance data of the mobile device and the user interaction behavior data of the application program, and predicts the user behavior trend and the device performance; s4, the intelligent graph development unit analyzes the predicted user behavior trend and equipment performance, and generates an intelligent graph loading and rendering strategy; the mobile device calls and executes the intelligent graph loading and rendering strategy, and intelligent graph loading and rendering are conducted on the application program of the mobile device. The running efficiency of the application program and the performance of the mobile equipment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of application development, and in particular to an intelligent graphics loading and rendering method and system for a mobile device application. Background Art

[0002] Mobile devices have become an indispensable tool in daily life. Their performance and functions are constantly improving, becoming an important part of life, work and entertainment. With the booming development of the mobile device application (APP) market, more and more applications need to load and render graphics resources on mobile devices to display rich graphics content and provide a better user experience. However, loading and rendering graphics resources on mobile device applications faces many challenges and problems:

[0003] Low efficiency in loading graphic resources: Traditional static loading methods cannot dynamically adjust according to user behavior and environmental changes, resulting in low loading efficiency. For example, in social media applications, a large number of images or video resources are loaded when users scroll the page, but traditional methods often load all resources in advance, resulting in resource waste and loading delays.

[0004] Rendering performance degradation: Mobile devices have low performance, and graphics resource rendering occupies a large amount of system resources, resulting in rendering performance degradation. Especially in some application scenarios with high performance requirements, such as 3D game applications, the degradation of rendering performance will directly affect the user's gaming experience.

[0005] Increased energy consumption: Graphics resource loading and rendering is one of the most energy-consuming operations on mobile devices. Traditional methods do not take energy consumption into consideration, resulting in increased energy consumption and reduced device battery life. Especially at a time when the battery life of mobile devices has become a focus of users, it is necessary to develop energy-saving loading and rendering methods.

[0006] Poor user experience: Due to the above problems, the user experience of mobile device applications is often unsatisfactory. Users may face problems such as loading delays, freezes, and rapid battery depletion, which affect the use of the application and user satisfaction.

[0007] The graphics loading and rendering of traditional applications fail to fully utilize the behavior and environmental data on mobile devices, resulting in low resource loading and rendering efficiency; the traditional graphics resource loading method requires manual configuration and adjustment, which is time-consuming and labor-intensive, and cannot respond to changes in user needs in real time. High-quality applications have high performance requirements for mobile devices. The graphics loading and rendering methods of traditional applications are prone to causing excessive load on mobile devices, affecting the smoothness of application operation. In addition, the graphics loading and rendering methods of traditional applications lack real-time monitoring and remote management capabilities, and cannot adjust resource allocation strategies in a timely manner.

[0008] Traditional applications often use static resource loading and rendering strategies for graphics loading and rendering, which cannot be adjusted in real time according to user behavior and environmental changes. This leads to inefficient resource loading and poor rendering performance, which is especially evident in scenarios where demand changes frequently.

[0009] Traditional applications usually use a fixed rendering method for graphics loading and rendering, which cannot be flexibly adjusted according to the real-time performance of mobile devices and user needs. This leads to reduced rendering performance and poor user experience, especially during peak resource demand and insufficient device performance.

[0010] The graphics loading and rendering of traditional applications do not fully consider the impact of graphics resource loading and rendering on the energy consumption of mobile devices, and lack targeted energy optimization strategies. This leads to increased energy consumption and reduced device battery life, affecting the user experience of using the device for a long time.

[0011] Graphics loading and rendering of traditional applications often lack prediction of user behavior and real-time monitoring of system performance, and are unable to load graphics resources that may be needed in advance. This leads to poor user experience and delayed resource loading, especially when quick response to user operations is required.

[0012] In summary, the loading and rendering of graphics resources for mobile device applications face many challenges and problems. We need to seek an intelligent solution that can dynamically adjust the loading and rendering strategies according to user behavior and environmental changes, improve resource loading efficiency, rendering performance, and energy utilization efficiency, thereby improving the user experience of applications on mobile devices and enhancing user satisfaction. Summary of the invention

[0013] The present invention proposes an intelligent graphics loading and rendering method and system for mobile device application programs to solve the technical problems of graphics loading and rendering of traditional application programs, low efficiency of graphics loading and rendering, and unsmooth running of application programs.

[0014] One aspect of the present invention is to provide an intelligent graphics loading and rendering method for a mobile device application, the intelligent graphics loading and rendering method comprising:

[0015] S1, initialize the SDK parameters of the intelligent graphics development unit, and return the initialized SDK parameter results to the mobile device;

[0016] S2, the intelligent graphics development unit collects sensor data of the mobile device, system performance data of the mobile device, and user interaction behavior data of the mobile device application, and uploads the collected data to the server;

[0017] Among them, the server embeds the prediction model;

[0018] S3, the server trains the prediction model through the sensor data of the mobile device, the system performance data of the mobile device and the user interaction behavior data of the mobile device application, and predicts the user behavior trend and device performance through the trained prediction model;

[0019] The server sends the predicted user behavior trends and device performance to the intelligent graphics development unit;

[0020] S4, the intelligent graphics development unit analyzes the predicted user behavior trends and device performance, and generates intelligent graphics loading and rendering strategies;

[0021] The mobile device calls and executes the intelligent graphics loading and rendering strategy to perform intelligent graphics loading and rendering on the application program of the mobile device.

[0022] In a preferred embodiment, in step S2, the intelligent graphics development unit filters and denoises the collected sensor data of the mobile device, and uploads the filtered and denoised data to the server;

[0023] The intelligent graphics development unit performs standardization processing on the collected system performance data of the mobile device and uploads the standardized data to the server;

[0024] The intelligent graphics development unit serializes the collected user interaction behavior data of the mobile device application program and uploads the serialized data to the server.

[0025] In a preferred embodiment, in step S3, the server preprocesses the sensor data of the mobile device, the system performance data of the mobile device, and the user interaction behavior data of the mobile device application, trains the prediction model using the preprocessed data, and predicts user behavior trends and device performance through the trained prediction model.

[0026] In a preferred embodiment, the prediction model includes a random forest model, a neural network model and an LSTM network model;

[0027] Among them, the random forest model uses sensor data from mobile devices to predict user behavior trends;

[0028] The neural network model uses the user interaction behavior data of mobile device applications to predict the user's interaction behavior trends;

[0029] The LSTM network model uses the system performance data of mobile devices to predict the performance and operating status of mobile devices.

[0030] In a preferred embodiment, in step S4, the intelligent graphics development unit analyzes the predicted user behavior trends and device performance, and generates intelligent graphics loading and rendering strategies, including: intelligent graphics loading strategy, intelligent graphics rendering strategy and personalized rendering strategy.

[0031] In a preferred embodiment, the intelligent graphics loading strategy performs intelligent graphics loading on the application program of the mobile device;

[0032] Intelligent graphics rendering strategy, intelligent graphics rendering for applications on mobile devices;

[0033] Personalized rendering strategy, intelligent graphics rendering for applications on mobile devices.

[0034] In a preferred embodiment, the intelligent graphics loading strategy for intelligently loading graphics for an application on a mobile device includes:

[0035] GPU resource optimization allocation, intelligent graphics preloading decisions, and mobile device performance tuning;

[0036] Intelligent graphics rendering strategy, intelligent graphics rendering of applications on mobile devices includes:

[0037] Dynamically adjust rendering resolution, adaptively adjust intelligent graphics frame rate, and dynamically allocate GPU resources;

[0038] Personalized rendering strategy, intelligent graphics rendering for mobile device applications including:

[0039] Texture management and adaptive allocation of GPU resources.

[0040] In a preferred embodiment, the intelligent graphics loading and rendering method further includes:

[0041] S5. The intelligent graphics development unit feeds back the execution data of the intelligent graphics loading and rendering to the server according to the result of the execution of the intelligent graphics loading and rendering strategy by the mobile device; the server optimizes the prediction model through the fed-back execution data of the intelligent graphics loading and rendering.

[0042] Another aspect of the present invention is to provide an intelligent graphics loading and rendering system for a mobile device application, the intelligent graphics loading and rendering system comprising a mobile device, an intelligent graphics development unit, a server and a database;

[0043] Wherein, the intelligent graphics development unit collects sensor data of the mobile device, system performance data of the mobile device and user interaction behavior data of the mobile device application, and uploads the collected data to the server;

[0044] The server has an embedded prediction model; the server trains the prediction model through the collected data, and predicts user behavior trends and device performance through the trained prediction model;

[0045] The server sends the predicted user behavior trend and device performance to the intelligent graphics development unit;

[0046] The intelligent graphics development unit analyzes predicted user behavior trends and device performance to generate intelligent graphics loading and rendering strategies;

[0047] The generated intelligent graphics loading and rendering strategies include: intelligent graphics loading strategy, intelligent graphics rendering strategy and personalized rendering strategy;

[0048] The intelligent graphics development unit includes an intelligent graphics loading strategy module, an intelligent graphics rendering strategy module and a personalized rendering strategy module;

[0049] The intelligent graphics loading strategy module uses the intelligent graphics loading strategy to perform intelligent graphics loading on the application program of the mobile device;

[0050] The intelligent graphics rendering strategy module uses the intelligent graphics rendering strategy to perform intelligent graphics rendering on the application of the mobile device;

[0051] The personalized rendering strategy module uses the personalized rendering strategy to perform intelligent graphics rendering on the application program of the mobile device.

[0052] In a preferred embodiment, the intelligent graphics loading and rendering system further comprises a completeness and compliance module;

[0053] The integrity and compliance module encrypts data transmitted between the mobile device and the intelligent graphics development unit, data transmitted between the intelligent graphics development unit and the server, data transmitted between the server and the database, and data stored in the database;

[0054] And set a retention period for all data.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The present invention proposes a method and system for intelligent graphics loading and rendering of mobile device applications, which utilizes sensor data of the mobile device, system performance data of the mobile device, and user interaction behavior data to train a prediction model in three aspects, generates an intelligent graphics loading and rendering strategy based on the user behavior trend and device performance predicted by the prediction model, and dynamically allocates resources during the intelligent graphics loading process of the mobile device application through the intelligent graphics loading and rendering strategy, thereby improving the performance, response speed and battery life of the mobile device.

[0057] The present invention proposes an intelligent graphics loading and rendering method and system for mobile device applications, which utilize sensors, system performance monitoring, and user interaction behavior data collection to comprehensively obtain mobile device status and user behavior information, filter, denoise, standardize, and serialize the collected data to ensure data quality, use autoencoders and machine learning algorithms to perform high-dimensional data processing, pattern recognition, and prediction, accurately predict user behavior and device performance changes, optimize resource loading and rendering strategies, dynamically adjust graphics resource loading and rendering strategies based on prediction results and real-time data, and improve application operation efficiency and user experience.

[0058] The intelligent graphics loading and rendering method and system for mobile device applications proposed in the present invention are different from the traditional method that only relies on a single data source. The present invention integrates sensor data, system performance data and user behavior data to provide more comprehensive information support. The autoencoder is used for high-dimensional data processing, which significantly improves the efficiency and accuracy of data processing; resource allocation is adjusted through prediction results and real-time data, which is different from the traditional static resource management method and greatly improves the efficiency of resource utilization.

[0059] The present invention proposes an intelligent graphics loading and rendering method and system for a mobile device application program, which can dynamically adjust the rendering strategy according to the performance of the mobile device and user needs, thereby achieving a highly personalized user experience.

[0060] The present invention proposes an intelligent graphics loading and rendering method and system for mobile device applications, which comprehensively adopt encryption and anonymization technologies to protect user data from data collection to processing and then to application, and manage data in compliance with relevant laws and regulations to ensure the security of data during collection, transmission and storage, comply with the requirements of relevant laws and regulations, and protect user privacy.

[0061] The present invention proposes an intelligent graphics loading and rendering method and system for mobile device applications, which can dynamically adjust the loading strategy of graphics resources according to user behavior patterns and device environment, thereby significantly improving the loading speed. Users can obtain the required graphics resources more quickly, greatly reducing waiting time and improving user experience.

[0062] The present invention proposes an intelligent graphics loading and rendering method and system for mobile device applications. The dynamic rendering strategy adopts variable rendering frame rate technology and multi-level detail level rendering strategy, and intelligently adjusts the rendering method of graphics resources according to user needs and device performance, so that users can obtain smoother and higher-quality visual effects when browsing applications, and the user experience is significantly improved.

[0063] The present invention proposes a method and system for intelligent graphics loading and rendering of mobile device applications. The graphics resource loading and rendering strategy not only improves the user experience, but also reduces the consumption of system resources and the occupation of network traffic, making users more willing to use the application for a long time, and significantly improving their satisfaction with the application. The user's investment in the application increases, and the retention rate and activity are greatly improved.

[0064] The present invention proposes an intelligent graphics loading and rendering method and system for mobile device applications. The graphics resource loading and rendering strategy can effectively save network traffic and system resource consumption without affecting user experience. This is not only conducive to improving user experience, but also reduces operating costs and improves system stability and reliability.

[0065] The intelligent graphics loading and rendering method and system of mobile device application proposed by the present invention significantly improve the performance of mobile devices, significantly improve the user experience of application, and enhance the competitiveness and brand image of application. The increase in users' favorable comments and recommendations for application helps to attract more users and increase market share. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the specific implementation of the present invention or the technical solution in the prior art, the following briefly introduces the drawings required for use in the specific implementation or the prior art description. Obviously, the drawings described below are some implementations of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0067] Figure 1 The invention discloses a flow chart of an intelligent graphics loading and rendering method for a mobile device application.

[0068] Figure 2 It is a structural block diagram of an intelligent graphics loading and rendering system for mobile device applications. DETAILED DESCRIPTION

[0069] In order to make the above and other features and advantages of the present invention more clear, the present invention is further described below in conjunction with the accompanying drawings. It should be understood that the specific embodiments given herein are for the purpose of explaining to those skilled in the art and are only exemplary and not restrictive.

[0070] Combination Figure 1 According to an embodiment of the present invention, a method for intelligent graphics loading and rendering of a mobile device application is provided, comprising:

[0071] Step S1 : Initialize SDK (Software Development Kit) parameters of the smart graphics development unit (Smart Graphics SDK) 200 , and return the initialized SDK parameter results to the mobile device 100 .

[0072] The present invention uses an intelligent graphics development unit 200 as a development toolkit for an application (e.g., an APP). The intelligent graphics development unit 200 collects data from the mobile device 100, generates an intelligent graphics loading and rendering strategy, and performs intelligent graphics loading and rendering on the application of the mobile device 100, which will be described in detail below.

[0073] In one embodiment, the SDK parameters for initializing the intelligent graphics development unit 200 include but are not limited to the operating system API interface parameters and key parameters of the mobile device 100 .

[0074] The SmartGraphics SDK 200 of the present invention is integrated into the operating system of the mobile device 100 .

[0075] In one embodiment, the intelligent graphics development unit 200 is integrated into the IOS system, and the SDK parameters of the intelligent graphics development unit 200 are initialized:

[0076] Import the SmartGraphics SDK 200 header file and add the static library or dynamic framework to the Xcode project.

[0077] Add the SDK's .framework file to your Xcode project.

[0078] Set Framework Search Paths in Build Settings.

[0079] Add the initialization code in the AppDelegate.swift file.

[0080] Call the SDK initialization method in the didFinishLaunchingWithOptions method of AppDelegate.swift (or AppDelegate.m in Objective-C) and set the necessary configuration parameters.

[0081] In one embodiment, the intelligent graphics development unit 200 is integrated into the Android system, and the SDK parameters of the intelligent graphics development unit 200 are initialized:

[0082] Add the AAR or Maven dependency of the intelligent graphics development unit 200 to the project build file (build.gradle).

[0083] Add SDK dependencies to the project's build.gradle file.

[0084] Add initialization code in the Application class.

[0085] Initialize the SDK in the lifecycle method of the Application class and configure related parameters.

[0086] Step S2: The intelligent graphics development unit 200 collects sensor data of the mobile device 100, system performance data of the mobile device 100, and user interaction behavior data of the application program of the mobile device 100, and uploads the collected data to the server 300. The server 300 has a built-in prediction model.

[0087] The present invention collects sensor data of the mobile device 100 to capture the three-dimensional motion trajectory, posture change, acceleration change and other information of the mobile device 100 in real time, so as to identify the user's behavior pattern and environmental status, and provide data support for the subsequent prediction model to accurately predict the user's behavior trend (the prediction model predicts the user's behavior trend is explained in detail below).

[0088] In a specific embodiment, when the intelligent graphics development unit 200 collects sensor data of the mobile device 100, it uses the official sensor interface and SDK provided by the operating system of the mobile device 100 to connect to the sensor of the mobile device 100 for data access.

[0089] Specifically, the accelerometer, gyroscope, and magnetometer, as motion sensors, continuously output raw data on acceleration, angular velocity, and magnetic field strength of the mobile device 100 in three-axis directions.

[0090] The accelerometer can measure the acceleration of the mobile device 100 in three axes (X, Y, Z), usually in meters per square second (m / s 2 When the mobile device 100 is stationary, the accelerometer will be affected by the earth's gravity, and the output acceleration data will be close to 9.8m / s on a certain axis. 2 (Gravity acceleration) When the mobile device 100 moves in any direction, the accelerometer will sense the corresponding acceleration change and output corresponding data.

[0091] The gyroscope measures the rotation rate or angular velocity of the mobile device 100 around three axes (X, Y, Z), usually in radians per second (rad / s). The gyroscope can sense the rotational motion of the mobile device 100, including rotation around the X axis (pitch), Y axis (roll), and Z axis (yaw). When the mobile device 100 remains stationary, the angular velocity data output by the gyroscope should be close to zero; when the mobile device 100 rotates, the gyroscope will output non-zero angular velocity data in the corresponding direction.

[0092] The magnetometer measures the magnetic field strength to which the mobile device 100 is subjected, and outputs data on the magnetic field strength in three axes (X, Y, Z), usually in microtesla (μT). The magnetometer can sense the earth's magnetic field and the surrounding magnetic field interference, so it can be used in applications such as navigation and direction perception. When the mobile device 100 rotates or moves, the magnetometer detects the change in the magnetic field in the corresponding direction and outputs the corresponding data.

[0093] The intelligent graphics development unit 200 utilizes the official sensor interface and SDK provided by the operating system of the mobile device 100 to connect to the sensors (accelerometer, gyroscope and magnetometer) of the mobile device 100 for data access.

[0094] In one embodiment, the smart graphics development unit (SmartGraphics SDK) 200 provides a unified access interface to the Android SensorManager class and the iOS CoreMotion framework, which facilitates developers to quickly integrate and collect real-time data from the accelerometer, gyroscope, magnetometer and GPS of the mobile device 100.

[0095] The present invention collects system performance data of the mobile device 100, obtains and accurately monitors the core performance indicators and network status parameters of the mobile device 100 in real time, and provides data support for subsequent prediction models to predict the performance and operating status of the mobile device 100 (the prediction model for measuring the performance and operating status of the mobile device 100 is described in detail below).

[0096] In a specific embodiment, the SmartGraphics SDK 200 provides startPerformanceMonitoring() and stopPerformanceMonitoring() methods to implement real-time monitoring of the CPU usage, memory usage, and battery power of the mobile device.

[0097] The intelligent graphics development unit 200 provides startNetworkMonitoring() and stopNetworkMonitoring() methods for obtaining information such as network connection status, type, bandwidth rate, packet loss rate, and delay to ensure consistency of network status across platforms.

[0098] In a specific embodiment, when the intelligent graphics development unit 200 collects the system performance data of the mobile device 100, it uses deep system integration technology to directly call the operating system API interface to collect data.

[0099] Specifically, by calling the CPU time slice API interface provided by the operating system, the CPU occupancy ratio of each process and system service is calculated, thereby obtaining the overall CPU usage rate.

[0100] Use the memory management API to obtain detailed data such as the total memory, used memory, free memory, and cache memory of the current operating system. In addition, the memory allocation and release of specific applications or processes should be monitored to prevent memory leaks.

[0101] With the help of the battery management API interface, the remaining power, charging status, discharge rate and other information of the mobile device 100 battery can be monitored in real time, and the remaining usage time can be estimated based on factors such as battery health.

[0102] The network API interface is used to obtain the network connection type (Wi-Fi, 4G / 5G, 3G, 2G, etc.), signal strength, real-time upload and download speed, and data traffic consumption of the mobile device 100. At the same time, key performance indicators such as network delay and packet loss rate are monitored to ensure that corresponding adjustments can be made in time when network conditions change.

[0103] The intelligent graphics development unit 200 uses deep system integration technology to directly call the operating system API interface to collect system performance data of the mobile device 100.

[0104] During the intelligent graphics loading and rendering process of the mobile device 100 application, accurate user interaction behavior is crucial to optimizing the graphics processing flow. The user interaction behavior directly affects the intelligent graphics loading and rendering strategy, thereby affecting the rendering efficiency and graphics display effect of the application.

[0105] The present invention collects user interaction behavior data of the mobile device 100 application to track the user's operation path, page access and jump and other interaction behaviors in the application, and provides data support for subsequent prediction models to predict the user's interaction behavior trends (the prediction model predicts the user's interaction behavior trends is described in detail below).

[0106] In a specific embodiment, when the intelligent graphics development unit 200 collects the user interaction behavior data of the application of the mobile device 100, the data is collected through the application embedding technology and event monitoring mechanism.

[0107] Various event listeners, such as OnClickListener and OnPageChangeListener, are registered in the application of the mobile device 100. When a user triggers an event, the application automatically records the event details and configures filtering rules through the setMinimumStayTime() and setMinimumClicks() methods.

[0108] Event-driven tracking technology is the core of user interaction behavior tracking. It monitors and records user activities by inserting tiny code snippets at key interaction points in the application.

[0109] Intelligent code generation and deployment.

[0110] Use automated tools to dynamically generate and deploy tracking code based on predefined user interaction models, significantly improving development efficiency and reducing errors.

[0111] The following details the key steps for implementing intelligent code generation and deployment:

[0112] Metadata and template definition: Use Swift / Objective-C on iOS and Java / Kotlin on Android to define the event tracking metadata model. The metadata file (stored in JSON format) describes the event type, interface element ID, and triggering conditions in detail. These metadata are used to design templates for code generation, which are responsible for event monitoring and data reporting.

[0113] Code Generation Tools:

[0114] For Android, write scripts using the Gradle plugin to automatically read data from metadata and generate Java or Kotlin code snippets through the template engine.

[0115] For iOS, use Fastlane or a custom script with the Xcode command line tools to read data from the same metadata file and generate Swift or Objective-C code snippets.

[0116] CI / CD Integration:

[0117] Automate the code generation and update process in GitLab CI / CD.

[0118] Leverage platform-specific build systems (gradle.properties for Android and xcconfig for iOS) to automatically integrate generated code into your project.

[0119] Use the blue-green deployment strategy to ensure a smooth transition of the new tracking logic:

[0120] Set up two identical production environments, called the blue environment and the green environment. The two environments include the same hardware configuration, operating system, network settings, database replication, and other dependent services to ensure that applications running in the two environments behave consistently.

[0121] Keep the data and configuration of the two environments synchronized, and use automated tools to synchronize database and application configurations regularly or in real time to ensure data consistency and service continuity at any time when switching.

[0122] Dynamic binding and advanced event capture.

[0123] Through framework-level integration, dynamically bind event listeners to capture and process user interactions in real time. These listeners adjust their capture strategies based on the application state and user history to ensure that the most critical data is recorded at the most appropriate moment.

[0124] The following details the key steps to achieve dynamic binding and advanced event capture:

[0125] Native event monitoring:

[0126] iOS: Use UIGestureRecognizer to implement dynamic binding of touch events.

[0127] Android: Use View.setOnTouchListener directly in Activity or Fragment to implement dynamic management of events.

[0128] Responsive Programming:

[0129] iOS: Use RxSwift to process event streams through chain calls and optimize event monitoring logic.

[0130] Android: Use RxJava to manage user interaction events and apply Observable and Observer patterns.

[0131] The event monitoring mechanism can effectively track the data of user click events and page browsing events.

[0132] Click event.

[0133] Deploy event listeners on all interactive elements of the application to capture and record every click event in real time. Automatically record key information such as the element location, ID, and class name where the click occurred.

[0134] Page view event.

[0135] Embed event recording logic in the life cycle of application pages to ensure that every action from user access to leaving is recorded. Record the page URL, title, and user entry and exit timestamps in detail to provide data support for analyzing user traffic and page popularity.

[0136] The intelligent graphics development unit 200 collects user interaction behavior data of the mobile device 100 application through application embedding technology and event monitoring mechanism.

[0137] According to a specific embodiment of the present invention, the intelligent graphics development unit 200 filters and denoises the collected sensor data of the mobile device 100, and uploads the filtered and denoised data to the server 300. Filtering and denoising the sensor data of the mobile device 100 is the key to accurately predicting the user's behavior trend.

[0138] Specifically, the intelligent graphics development unit 200 uses a Butterworth filter to perform low-pass filtering on the collected sensor data of the mobile device 100, retaining useful low-frequency signals, achieving data filtering and denoising, and reducing noise and random fluctuations in the data.

[0139] The Butterworth filter was chosen for its flat frequency response in the passband, optimizing the data accuracy of the accelerometer and gyroscope.

[0140] In a specific embodiment, a Butterworth filter is used to filter and remove noise from the sensor data collected by the mobile device 100 by the following method:

[0141]

[0142] Where, f represents the signal frequency, f c represents the cut-off frequency, and n represents the order of the Butterworth filter. Preferably, the cut-off frequency is set to 5 Hz, which can be suitable for sensor data on most mobile devices 100.

[0143] In one embodiment, in addition to low-pass filtering the sensor data of the mobile device 100, moving average filtering may be used to further smooth the data, reduce the impact of random fluctuations on performance, and provide a more stable data basis for subsequent prediction of user behavior trends.

[0144] Moving average filtering is used to reduce random fluctuations and noise in data. It can smooth the data curve by calculating the average value of the data window and continuously sliding this window through the entire data set.

[0145] Specifically, within a fixed time window, key statistical features such as mean, variance, maximum and minimum values ​​are extracted from the filtered or smoothed data for subsequent data analysis and pattern recognition.

[0146] The steps are as follows:

[0147] Define the window size, for example each window contains all data points within 1 second.

[0148] For each window, the following statistical features are calculated and recorded:

[0149] Mean: the average value of all data points in the window;

[0150] Variance: The variance of the data points within the window, indicating the degree of fluctuation;

[0151] Maximum and Minimum: The highest and lowest data points within the window.

[0152] The above calculations are repeated for each data window until the entire data set is processed.

[0153] The intelligent graphics development unit 200 filters and denoises the collected sensor data of the mobile device 100 , and uploads the filtered and denoised data to the server 300 .

[0154] Furthermore, the intelligent graphics development unit 200 serializes the filtered and denoised data into a JSON format and uploads it to the server 300 via an HTTP POST request.

[0155] According to an embodiment of the present invention, the intelligent graphics development unit 200 performs standardization processing on the collected system performance data of the mobile device 100, and uploads the standardized data to the server 300. Standardization processing on the collected system performance data of the mobile device 100 can effectively simplify the data processing of the subsequent prediction model predicting the performance and operating status of the mobile device 100.

[0156] Specifically, the intelligent graphics development unit 200 normalizes the collected system performance data of the mobile device 100 (eg, CPU usage, memory occupancy, etc.) to a unified metric (ranging from 0 to 1), thereby eliminating the impact of the dimension.

[0157] In a specific embodiment, the collected system performance data of the mobile device 100 is normalized to a unified metric (range from 0 to 1) by the following method:

[0158] The collected system performance data of the mobile device 100 (each performance indicator / data point, such as CPU usage, memory usage, etc.) is used to determine the maximum and minimum values ​​of all data point sets; the collected system performance data of the mobile device 100 (each performance indicator / data point) is converted to between 0 and 1 using the following method:

[0159]

[0160] Where x represents the original data point, X represents the set of all data points, min(X) and max(X) represent the minimum and maximum values ​​of the data point set X, and x norm Represents the normalized value, ranging from 0 to 1.

[0161] In some embodiments, a smoothing method is used to identify long-term trends and periodic fluctuations in the collected system performance data of the mobile device 100, so as to identify and predict performance fluctuations of the mobile device 100 and provide a basis for resource management.

[0162] In a specific embodiment, the long-term trend and periodic fluctuation of the collected system performance data of the mobile device 100 are identified by the following smoothing method:

[0163] Select a smoothing coefficient α (for example, 0.2), start from the smoothing value S0 of the original data point x at time t=0, and use the following recursive method to calculate the smoothing value of the original data point x at each time:

[0164] S t =αx t +(1-α)S t-1 ,

[0165] Among them, S t represents the smoothed value at time t, x t represents the original data point at time t, and α represents the smoothing coefficient, which is between 0 and 1.

[0166] The smoothed value S0 of the original data point x at time t=0 selects the first observed value of the data point sequence. By calculating the smoothed value of the original data point x at each time, the overall trend of the collected system performance data of the mobile device 100 is depicted.

[0167] In other embodiments, outliers are identified on the collected system performance data of the mobile device 100 , and these outliers may be early indicators of the performance of the mobile device 100 .

[0168] In a specific embodiment, the Z-Score method is used to identify abnormal points in the collected system performance data of the mobile device 100:

[0169] The mean and standard deviation of all data points of the collected system performance data of the mobile device 100 are calculated.

[0170] Use the Z-Score algorithm to calculate the Z-Score value of each data point:

[0171] Z = (x-μ) / σ, where Z represents the Z-Score value, x represents the original data point, μ represents the mean of the data point set, and σ represents the standard deviation of the data point set.

[0172] The intelligent graphics development unit 200 performs standardization processing on the collected system performance data of the mobile device 100 , and then uploads the standardized data to the server 300 .

[0173] Furthermore, the intelligent graphics development unit 200 serializes the standardized data into JSON format and uploads it to the server 300 via an HTTP POST request.

[0174] According to an embodiment of the present invention, the intelligent graphics development unit 200 performs serialization processing on the collected user interaction behavior data of the mobile device 100 application, and uploads the serialized data to the server 300. The collected user interaction behavior data of the mobile device 100 application is serialized to provide a suitable input format for the subsequent prediction model to predict the user's interaction behavior trend.

[0175] In a specific embodiment, first, the intelligent graphics development unit 200 performs data encoding on the collected user interaction behavior data of the application program of the mobile device 100:

[0176] One-hot encoding technology is used to assign a unique binary code to each user interaction behavior data point of the mobile device 100 application, and each user interaction behavior data is converted into binary form, so that the representation of each interaction behavior event in the vector is mutually exclusive.

[0177] Then, a time series data model is constructed so that each time point corresponds to an encoding vector (binary code vector) of one or more user interaction behavior data points;

[0178] The Python Pandas library is used for time series processing to ensure that the data is accurately organized in chronological order and to complete the serialization processing of the user interaction behavior data of the collected mobile device 100 applications.

[0179] In one embodiment, the serialized user interaction behavior data is grouped to identify typical interaction behavior events.

[0180] Specifically, based on the serialized user interaction behavior data (such as the frequency and duration of the interaction behavior events, and the above-mentioned factors (such as network status));

[0181] Statistical methods are used to evaluate the correlation between each data and user satisfaction, and the most representative data are selected for clustering.

[0182] Preferably, the clustering method uses K-means algorithm for clustering, which is highly efficient and suitable for processing large data sets.

[0183] In a specific embodiment, the optimal number of clusters K is determined by the elbow method, and the cluster centers are iteratively updated until the change in the cluster centers is less than a predetermined threshold.

[0184] Identify the user interaction behavior events represented by each cluster, and use these clusters in subsequent prediction models to predict user interaction behavior trends to optimize the user experience of the application.

[0185] The intelligent graphics development unit 200 serializes the collected user interaction behavior data of the application program of the mobile device 100 and uploads the serialized data to the server 300 .

[0186] Furthermore, the intelligent graphics development unit 200 serializes the serialized data into a JSON format, and uploads it to the server 300 via an HTTP POST request.

[0187] Step S3: The server 300 trains the prediction model through the sensor data of the mobile device 100, the system performance data of the mobile device 100 and the user interaction behavior data of the application of the mobile device 100, and predicts the user behavior trend and device performance through the trained prediction model.

[0188] The server 300 sends the predicted user behavior trend and device performance to the intelligent graphics development unit 200 .

[0189] In step S2, the intelligent graphics development unit 200 filters and denoises the collected sensor data of the mobile device 100 and uploads it to the server 300, standardizes the collected system performance data of the mobile device 100 and uploads it to the server 300, and serializes the collected user interaction behavior data of the application of the mobile device 100 and uploads it to the server 300.

[0190] The data used by the server 300 to train the prediction model are: the sensor data of the mobile device 100 after filtering and denoising, the system performance data of the mobile device 100 after standardization, and the user interaction behavior data of the application of the mobile device 100 after serialization.

[0191] In some embodiments, the server 300 stores the filtered and denoised sensor data of the mobile device 100 , the standardized system performance data of the mobile device 100 , and the serialized user interaction behavior data of the application of the mobile device 100 in the database 400 .

[0192] When the server 300 trains the prediction model, it extracts the filtered and denoised sensor data of the mobile device 100, the standardized system performance data of the mobile device 100, and the serialized user interaction behavior data of the application of the mobile device 100 from the database 400.

[0193] In some embodiments, the server 300 cleans and preprocesses the sensor data of the mobile device 100, the system performance data of the mobile device 100, and the user interaction behavior data of the application of the mobile device 100, trains the prediction model with the preprocessed data, and predicts the user behavior trend and device performance through the trained prediction model.

[0194] Specifically, the present invention uses an autoencoder to preprocess sensor data of the mobile device 100, system performance data of the mobile device 100, and user interaction behavior data of the application of the mobile device 100. The autoencoder is implemented using the TensorFlow framework.

[0195] An autoencoder is a neural network for unsupervised learning that compresses input data (such as sensor data of the mobile device 100, performance data of the mobile device 100, and user interaction behavior data of the mobile device 100 application) into a low-dimensional vector through an encoding process, and then attempts to reconstruct the input data through a decoding process, with the goal of capturing the most important features in the data. This low-dimensional vector not only improves the efficiency of data processing, but also enhances the representativeness and comprehensiveness of the data, providing support for the prediction model.

[0196] In a specific embodiment, the encoding process of the autoencoder is: r=ρ(W·y+b),

[0197] Among them, r represents the encoded low-dimensional vector, W represents the encoder weight matrix, y represents the input data (a combination of sensor data of the mobile device 100, performance data of the mobile device 100 and user interaction behavior data of the application of the mobile device 100), b represents the encoder bias vector, and ρ represents the activation function of the encoder, which is used to determine whether a neuron should be activated.

[0198] The decoding process of the autoencoder is: y′=ρ′(W′·r+b′),

[0199] Among them, r represents the encoded low-dimensional vector, W' represents the decoder weight matrix, y' represents the reconstructed input data (as close to the original input data y as possible), b' represents the decoder bias vector, and ρ' represents the activation function of the decoder, which can be the same as or different from the activation function of the encoder.

[0200] The present invention preprocesses the sensor data of the mobile device 100, the system performance data of the mobile device 100, and the user interaction behavior data of the application of the mobile device 100 through the autoencoder, and reconstructs the original input data in as few dimensions as possible, so as to learn the most important features in the data. The reconstructed input data y' is used as the input of the prediction model, which is very valuable for training the prediction model and can significantly improve the processing speed and efficiency.

[0201] The present invention pre-processes the sensor data of the mobile device 100, the system performance data of the mobile device 100, and the user interaction behavior data of the application of the mobile device 100 through an autoencoder, reconstructs the original input data in as few dimensions as possible, and can effectively reduce the space and resources occupied by the data in storage and processing.

[0202] According to an embodiment of the present invention, the server 300 has a built-in prediction model, which includes a random forest model, a neural network model (eg, a deep neural network model, DNN), and an LSTM network model.

[0203] As a powerful ensemble learning model, random forest handles complex classification and regression problems by building multiple decision trees, making it particularly suitable for pattern recognition and feature importance assessment in feature-rich scenarios.

[0204] Neural network models are particularly suitable for processing nonlinear and high-dimensional data, such as user interaction behavior data, and can reveal deep patterns of user behavior.

[0205] The LSTM network model processes time series data through its unique gating mechanism, which can capture long-term temporal dependencies excellently, making it ideal for predicting mobile device performance trends.

[0206] The server 300 of the present invention trains a random forest model through the reconstructed input data y', including:

[0207] Data processing: Perform missing value processing on the reconstructed input data y' to ensure that the data format is suitable for decision tree processing.

[0208] Model training: Use the training dataset to build multiple decision trees, each tree using randomly selected features and samples.

[0209] Feature Importance Evaluation: Use the feature importance scores output by the random forest model to select the most influential features, which have the greatest impact on the output of the random forest model.

[0210] In some embodiments, the number and depth of trees are adjusted to optimize the accuracy of the random forest model and prevent overfitting.

[0211] The server 300 of the present invention trains the neural network model through the reconstructed input data y', including:

[0212] Build a neural network model: Design a neural network structure with multiple hidden layers, and use a Sigmoid activation function in each layer to enhance the nonlinear learning ability of the neural network model.

[0213] Back-propagation learning: Use the back-propagation algorithm for training to minimize the prediction error by adjusting weights layer by layer.

[0214] Hyperparameter Tuning: Tune hyperparameters like learning rate, batch size, and number of iterations to find the best neural network performance.

[0215] Feature processing: The input is processed through an embedding layer or normalized to suit the needs of a neural network.

[0216] The server 300 of the present invention trains the LSTM network model through the reconstructed input data y', including:

[0217] Sequence data processing: Ensure that all time series data are arranged in chronological order and perform appropriate window cutting to form training samples.

[0218] LSTM cell configuration: Configure the number of LSTM layers and the number of cells in each layer to capture long-term dependencies in the data.

[0219] Gating mechanism: Optimize the parameters of the input gate, forget gate, and output gate to control the inflow, retention, and outflow of information.

[0220] Prediction and Regression: Use LSTM networks to perform regression analysis to predict future mobile device usage trends and performance requirements.

[0221] According to an embodiment of the present invention, the server 300 predicts user behavior trends and device performance through trained prediction models (random forest model, neural network model and LSTM network model).

[0222] The user behavior trend and device performance predicted by the prediction model of the present invention include: the user's behavior trend, the user's interactive behavior trend, and the performance and operating status of the mobile device 100.

[0223] According to an embodiment of the present invention, the random forest model of the prediction model uses the sensor data of the mobile device 100 to predict the user's behavior trend (user behavior pattern recognition).

[0224] The neural network model of the prediction model uses the user interaction behavior data of the application program of the mobile device 100 to predict the user's interaction behavior trend.

[0225] The LSTM network model of the prediction model uses the system performance data of the mobile device 100 to predict the performance and operating status of the mobile device 100 .

[0226] The server 300 sends the predicted user behavior trends and device performance (user behavior trends, user interaction behavior trends, performance and operating status of the mobile device 100 ) to the intelligent graphics development unit 200 .

[0227] Furthermore, the server 300 sends the predicted user behavior trends and device performance (user behavior trends, user interaction behavior trends, performance and operating status of the mobile device 100) to the intelligent graphics development unit 200 through the API interface.

[0228] In one embodiment, the server 300 stores the predicted user behavior trends and device performance (user behavior trends, user interaction behavior trends, performance and operating status of the mobile device 100 ) in the database 400 .

[0229] Step S4: The intelligent graphics development unit 200 analyzes the predicted user behavior trend and device performance, and generates an intelligent graphics loading and rendering strategy.

[0230] The mobile device 100 calls and executes the intelligent graphics loading and rendering strategy to perform intelligent graphics loading and rendering on the application of the mobile device 100 .

[0231] In step S3, the server 300 predicts user behavior trends and device performance (user behavior trends, user interaction behavior trends, performance and operating status of the mobile device 100) through prediction models (random forest model, neural network model and LSTM network model).

[0232] In step S4, the intelligent graphics development unit 200 generates an intelligent graphics loading and rendering strategy according to the predicted user behavior trend and device performance (user behavior trend, user interaction behavior trend, performance and operation status of the mobile device 100).

[0233] According to an embodiment of the present invention, the intelligent graphics development unit 200 analyzes predicted user behavior trends and device performance (user behavior trends, user interaction behavior trends, performance and operating status of the mobile device 100) and generates intelligent graphics loading and rendering strategies.

[0234] The generated intelligent graphics loading and rendering strategies include: intelligent graphics loading strategy, intelligent graphics rendering strategy and personalized rendering strategy.

[0235] According to an embodiment of the present invention, the intelligent graphics loading strategy performs intelligent graphics loading on the application program of the mobile device 100 .

[0236] According to an embodiment of the present invention, the intelligent graphics loading strategy for intelligently loading graphics for an application of the mobile device 100 includes:

[0237] Optimized GPU resource allocation, intelligent graphics preloading decisions, and performance tuning for mobile devices.

[0238] Optimized allocation of GPU resources.

[0239] Dynamically adjust GPU resource optimization allocation based on the user behavior trends predicted by the random forest model and the user interaction behavior trends predicted by the neural network model.

[0240] For example, when the random forest model predicts that the user's behavior trend leads to GPU resource-intensive activities (high resource consumption), and / or the neural network model predicts that the user's interactive behavior trend leads to GPU resource-intensive activities (high resource consumption), the allocation of GPU resources is increased to ensure that the application of the mobile device 100 performs intelligent graphics rendering smoothly.

[0241] When the random forest model predicts that the user's behavior trend leads to GPU resource-sparse activities (low resource consumption), and / or the neural network model predicts that the user's interactive behavior trend leads to GPU resource-sparse activities (low resource consumption), the allocation of GPU resources is reduced to save energy and improve the performance of other applications.

[0242] Intelligent graphics preloading decisions.

[0243] Intelligent graphics preloading is implemented based on the performance and operating status of the mobile device 100 predicted by the LSTM network model.

[0244] For example, when users launch games or large applications, necessary intelligent graphics are preloaded to reduce waiting time and improve response speed.

[0245] Mobile device performance tuning.

[0246] The load of the mobile device is adjusted according to the performance and operating status of the mobile device 100 predicted by the LSTM network model.

[0247] For example, when the LSTM network model predicts that the performance and operating status of the mobile device 100 will cause the CPU and memory of the mobile device 100 to run at a high load, the power supply is increased; when the LSTM network model predicts that the performance and operating status of the mobile device 100 will cause the CPU and memory of the mobile device 100 to run at a low load, the power supply is reduced.

[0248] The performance and operating status of the mobile device 100 predicted by the LSTM network model are used to adjust the load of the mobile device 100, thereby balancing the performance of the mobile device 100 and the battery usage efficiency.

[0249] According to an embodiment of the present invention, an intelligent graphics rendering strategy performs intelligent graphics rendering on an application of the mobile device 100 .

[0250] According to an embodiment of the present invention, the intelligent graphics rendering strategy for intelligently rendering the application of the mobile device 100 includes:

[0251] Dynamically adjust rendering resolution, adaptively adjust intelligent graphics frame rate, and dynamically allocate GPU resources.

[0252] Dynamically adjust rendering resolution.

[0253] According to the performance and operating status of the mobile device 100 predicted by the LSTM network model, the resolution of the intelligent graphics rendering is dynamically adjusted to ensure the best balance between fluency and visual effects.

[0254] Specifically, the resolution of intelligent graphics rendering is dynamically adjusted by the following method:

[0255]

[0256] Among them, R represents the resolution of intelligent graphics rendering after adjustment, P max represents the maximum performance load of the mobile device 100, P current represents the predicted current performance load of the mobile device 100, R base Indicates the base resolution for smart graphics rendering.

[0257] Adaptively adjust intelligent graphics frame rate.

[0258] According to the performance and operating status of the mobile device 100 predicted by the LSTM network model, the graphics frame rate of the intelligent graphics rendering is adaptively adjusted to ensure that the frame rate is reduced to reduce lag when the load of the mobile device 100 is high, and the frame rate is increased to improve visual fluency when the load is low.

[0259] Specifically, the graphics frame rate of intelligent graphics rendering is adaptively adjusted by the following method:

[0260]

[0261] Among them, F represents the graphics frame rate of the adjusted intelligent graphics rendering, C opt represents the computing power (CPU ratio) of the mobile device 100, C current represents the predicted computing capability of the mobile device 100, F base Indicates the base graphics frame rate for smart graphics rendering.

[0262] GPU resources are dynamically allocated.

[0263] The intelligent graphics development unit 200 collects detailed information of the current intelligent graphics rendering task (eg, texture resolution, shader complexity, and estimated rendering time required for the intelligent graphics rendering).

[0264] According to the user's behavior trend predicted by the random forest model and the user's interactive behavior trend predicted by the neural network model, the current GPU usage of the mobile device 100 is evaluated to determine the available GPU resources.

[0265] For example, when the random forest model predicts that the user's behavior trend leads to GPU resource-intensive activities (high resource consumption), and / or the neural network model predicts that the user's interaction behavior trend leads to GPU resource-intensive activities (high resource consumption), the allocation of GPU resources is increased, and fewer GPU resources are available.

[0266] When the random forest model predicts that the user's behavior trend leads to GPU resource sparse activities (low resource consumption), and / or the neural network model predicts that the user's interactive behavior trend leads to GPU resource sparse activities (low resource consumption), the allocation of GPU resources is reduced, and more GPU resources are available.

[0267] When more GPU resources are available (GPU resource load is low), improve the current intelligent graphics rendering task. For example, increase texture resolution, shader complexity and rendering time.

[0268] When the available GPU resources are low (GPU resource load is high), reduce the current intelligent graphics rendering task. For example, reduce texture resolution, shader complexity and rendering time to ensure smooth rendering.

[0269] In some embodiments, the priorities of different intelligent graphics rendering tasks are determined based on the visual importance of the intelligent graphics rendering tasks and user feedback, and intelligent graphics rendering tasks with high visual importance and high user feedback satisfaction are rendered first.

[0270] In some embodiments, dependencies are determined for different intelligent graphics rendering tasks to improve rendering efficiency. For example, if the rendering of a certain intelligent graphics rendering task requires rendering the texture or data of another intelligent graphics rendering task first, then the other intelligent graphics rendering task is rendered first.

[0271] According to an embodiment of the present invention, a personalized rendering strategy is used to perform intelligent graphics rendering on the application program of the mobile device 100 .

[0272] According to an embodiment of the present invention, the personalized rendering strategy for performing intelligent graphics rendering on an application of the mobile device 100 includes:

[0273] Texture management and adaptive allocation of GPU resources.

[0274] Texture management.

[0275] According to the user's behavior trend predicted by the random forest model and the user's interactive behavior trend predicted by the neural network model, the current GPU usage of the mobile device 100 is evaluated to determine the available GPU resources.

[0276] According to the performance and operating status of the mobile device 100 predicted by the LSTM network model, the current CPU and memory usage of the mobile device 100 is evaluated to determine the available CPU and memory resources.

[0277] Dynamically adjusts the compression level and loading level of textures based on available CPU, memory resources, and available GPU resources.

[0278] For example, when available CPU and memory resources are less (CPU and memory of mobile device 100 are running at high load) and available GPU resources are less (GPU resource load is high), the compression level and loading level of the texture are reduced.

[0279] When more CPU and memory resources are available (CPU and memory of mobile device 100 are running at low load) and more GPU resources are available (GPU resource load is low), the compression level and loading level of the texture are increased.

[0280] In some preferred embodiments, textures that have the greatest impact on user experience are loaded first.

[0281] Adaptive allocation of GPU resources.

[0282] According to the user's behavior trend predicted by the random forest model and the user's interactive behavior trend predicted by the neural network model, the current GPU usage of the mobile device 100 is evaluated to determine the available GPU resources.

[0283] When more GPU resources are available (GPU resource load is low), the rendering details of the current intelligent graphics rendering task are improved. For example, the shading accuracy and light and shadow processing are increased.

[0284] When the available GPU resources are less (the GPU resource load is high), the rendering details of the current intelligent graphics rendering task are reduced, for example, shading accuracy and light and shadow processing, to ensure smooth rendering.

[0285] According to an embodiment of the present invention, the mobile device 100 calls and executes the intelligent graphics loading and rendering strategy to perform intelligent graphics loading and rendering on the application program of the mobile device 100 .

[0286] Specifically, the mobile device 100 calls the intelligent graphics development unit 200 to parse the predicted user behavior trends and device performance (user behavior trends, user interaction behavior trends, performance and operating status of the mobile device 100) to generate intelligent graphics loading and rendering strategies (intelligent graphics loading strategy, intelligent graphics rendering strategy and personalized rendering strategy), and performs intelligent graphics loading and rendering on the application of the mobile device 100.

[0287] Step S5: The intelligent graphics development unit 200 feeds back the execution data of the intelligent graphics loading and rendering to the server 300 according to the result of the execution of the intelligent graphics loading and rendering strategy by the mobile device.

[0288] Furthermore, the intelligent graphics development unit 200 serializes the execution data of the intelligent graphics loading and rendering into a JSON format, and uploads it to the server 300 via an HTTP POST request.

[0289] The server 300 stores the fed-back intelligent graphics loading and rendering execution data in the database 400 , and optimizes and adjusts the prediction models (random forest model, neural network model, and LSTM network model) through the fed-back intelligent graphics loading and rendering execution data.

[0290] In order to ensure the security of data during collection, transmission and storage, the present invention encrypts all transmitted and stored data. For example, a strong encryption algorithm (SM4) that complies with national standards is used to ensure the security of data during transmission and storage, and before processing user interaction behavior data, data anonymization technology is used to remove all information that may identify personal identity to ensure that the user's privacy is protected during the analysis process. Regularly check and evaluate the effectiveness of data protection measures, monitor in real time, detect and respond to any possible security threats or abnormal activities in a timely manner, enhance data security and system stability, and ensure that all systems and operations meet national security standards.

[0291] The data transmission encryption of the present invention uses the HTTPS protocol to ensure that all data transmission is carried out through the HTTPS protocol and uses the TLS (Transport Layer Security) protocol to encrypt the transmitted data.

[0292] Data transmission encryption configures and manages SSL certificates to ensure the validity and security of the certificates.

[0293] The data storage encryption of the present invention uses a strong encryption algorithm (SM4) that complies with national standards to ensure the security of data during transmission and storage.

[0294] Before storing and processing user data, the present invention removes all information that may identify an individual (such as name, address, telephone number, etc.), performs hash processing on sensitive fields to ensure that the original data cannot be restored, and processes the data using an appropriate anonymization algorithm (such as k-anonymity, l-diversity, t-proximity), thereby anonymizing the data before processing the user interaction behavior data.

[0295] In order to ensure that data complies with the requirements of relevant laws and regulations during the collection, transmission and storage process, protect user privacy, formulate a strict data retention policy, comply with the provisions of domestic laws on data retention period, determine which data needs to be kept for a long time and which should be safely deleted when no longer needed, and implement a data backup strategy that complies with the requirements of the "Cybersecurity Law" to ensure that data can be quickly restored in the event of data loss or system failure.

[0296] The data retention policy of the present invention uses a configuration file or a database table to record the data retention policy, clearly defines the retention period and deletion time of the data, and ensures that the system follows these rules.

[0297] The present invention formulates a data backup strategy to regularly back up the data in the database 400 to ensure the integrity and security of the data, and uses an automated backup tool or script to regularly back up the data to local and remote storage.

[0298] The present invention writes a data recovery script to quickly recover data from backups and regularly performs data recovery tests to ensure that data can be quickly recovered in the event of data loss or system failure.

[0299] like Figure 2 As shown, according to an embodiment of the present invention, a smart graphics loading and rendering system for a mobile device application is provided, including a mobile device 100, a smart graphics development unit 200, a server 300 and a database 400.

[0300] The intelligent graphics development unit 200 collects sensor data of the mobile device 100 , system performance data of the mobile device 100 , and user interaction behavior data of the application program of the mobile device 100 , and uploads the collected data to the server 300 .

[0301] The server 300 has a built-in prediction model. The server 300 trains the prediction model through the collected data, and predicts user behavior trends and device performance through the trained prediction model.

[0302] The server 300 sends the predicted user behavior trend and device performance to the intelligent graphics development unit 200 .

[0303] The intelligent graphics development unit 200 analyzes predicted user behavior trends and device performance, and generates intelligent graphics loading and rendering strategies.

[0304] The generated intelligent graphics loading and rendering strategies include: intelligent graphics loading strategy, intelligent graphics rendering strategy and personalized rendering strategy.

[0305] The intelligent graphics development unit 200 includes an intelligent graphics loading strategy module 205 , an intelligent graphics rendering strategy module 206 and a personalized rendering strategy module 207 .

[0306] The intelligent graphics loading strategy module 205 performs intelligent graphics loading on the application of the mobile device 100 using the intelligent graphics loading strategy.

[0307] The intelligent graphics rendering strategy module 206 performs intelligent graphics rendering on the application of the mobile device 100 using the intelligent graphics rendering strategy.

[0308] The personalized rendering strategy module 207 uses the personalized rendering strategy to perform intelligent graphics rendering on the application of the mobile device 100 .

[0309] According to an embodiment of the present invention, a smart graphics loading and rendering system for a mobile device application further includes a completeness and compliance module 204 .

[0310] The integrity and compliance module 204 encrypts data transmitted between the mobile device 100 and the intelligent graphics development unit 200, data transmitted between the intelligent graphics development unit 200 and the server 300, data transmitted between the server 300 and the database 400, and data stored in the database 400, and establishes a retention period and data backup for all data.

[0311] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and deform the above embodiments within the scope of the present invention.

Claims

1. A method for intelligent graphics loading and rendering of a mobile device application, characterized in that: The intelligent graphics loading and rendering method comprises: S1, initialize the SDK parameters of the intelligent graphics development unit, and return the initialized SDK parameter results to the mobile device; S2, the intelligent graphics development unit collects sensor data of the mobile device, system performance data of the mobile device, and user interaction behavior data of the mobile device application, and uploads the collected data to the server; Among them, the server embeds the prediction model; S3, the server trains the prediction model through the sensor data of the mobile device, the system performance data of the mobile device and the user interaction behavior data of the mobile device application, and predicts the user behavior trend and device performance through the trained prediction model; The server sends the predicted user behavior trends and device performance to the intelligent graphics development unit; S4, the intelligent graphics development unit analyzes the predicted user behavior trends and device performance, and generates intelligent graphics loading and rendering strategies; The mobile device calls and executes the intelligent graphics loading and rendering strategy to perform intelligent graphics loading and rendering on the application program of the mobile device.

2. The intelligent graphics loading and rendering method according to claim 1, characterized in that: In step S2, the intelligent graphics development unit filters and denoises the collected sensor data of the mobile device, and uploads the filtered and denoised data to the server; The intelligent graphics development unit performs standardization processing on the collected system performance data of the mobile device and uploads the standardized data to the server; The intelligent graphics development unit serializes the collected user interaction behavior data of the mobile device application program and uploads the serialized data to the server.

3. The intelligent graphics loading and rendering method according to claim 1, characterized in that: In step S3, the server preprocesses the sensor data of the mobile device, the system performance data of the mobile device, and the user interaction behavior data of the mobile device application, trains the prediction model using the preprocessed data, and predicts user behavior trends and device performance through the trained prediction model.

4. The intelligent graphics loading and rendering method according to claim 3, characterized in that: Prediction models include random forest model, neural network model and LSTM network model; Among them, the random forest model uses sensor data from mobile devices to predict user behavior trends; The neural network model uses the user interaction behavior data of mobile device applications to predict the user's interaction behavior trends; The LSTM network model uses the system performance data of mobile devices to predict the performance and operating status of mobile devices.

5. The intelligent graphics loading and rendering method according to claim 1, characterized in that: In step S4, the intelligent graphics development unit analyzes the predicted user behavior trend and device performance, and generates intelligent graphics loading and rendering strategies, including: intelligent graphics loading strategy, intelligent graphics rendering strategy and personalized rendering strategy.

6. The intelligent graphics loading and rendering method according to claim 5, characterized in that: Intelligent graphics loading strategy, intelligent graphics loading for applications on mobile devices; Intelligent graphics rendering strategy, intelligent graphics rendering for applications on mobile devices; Personalized rendering strategy, intelligent graphics rendering for applications on mobile devices.

7. The intelligent graphics loading and rendering method according to claim 6, characterized in that: Intelligent graphics loading strategy, intelligent graphics loading for mobile device applications includes: GPU resource optimization allocation, intelligent graphics preloading decisions, and mobile device performance tuning; Intelligent graphics rendering strategy, intelligent graphics rendering of applications on mobile devices includes: Dynamically adjust rendering resolution, adaptively adjust intelligent graphics frame rate, and dynamically allocate GPU resources; Personalized rendering strategy, intelligent graphics rendering for mobile device applications including: Texture management and adaptive allocation of GPU resources.

8. The intelligent graphics loading and rendering method according to claim 1, characterized in that: The intelligent graphics loading and rendering method further includes: S5. The intelligent graphics development unit feeds back the execution data of the intelligent graphics loading and rendering to the server according to the result of the execution of the intelligent graphics loading and rendering strategy by the mobile device; the server optimizes the prediction model through the fed-back execution data of the intelligent graphics loading and rendering.

9. An intelligent graphics loading and rendering system for mobile device applications, characterized in that: The intelligent graphics loading and rendering system includes a mobile device, an intelligent graphics development unit, a server and a database; Wherein, the intelligent graphics development unit collects sensor data of the mobile device, system performance data of the mobile device and user interaction behavior data of the mobile device application, and uploads the collected data to the server; The server has an embedded prediction model; the server trains the prediction model through the collected data, and predicts user behavior trends and device performance through the trained prediction model; The server sends the predicted user behavior trend and device performance to the intelligent graphics development unit; The intelligent graphics development unit analyzes predicted user behavior trends and device performance to generate intelligent graphics loading and rendering strategies; The generated intelligent graphics loading and rendering strategies include: intelligent graphics loading strategy, intelligent graphics rendering strategy and personalized rendering strategy; The intelligent graphics development unit includes an intelligent graphics loading strategy module, an intelligent graphics rendering strategy module and a personalized rendering strategy module; The intelligent graphics loading strategy module uses the intelligent graphics loading strategy to perform intelligent graphics loading on the application program of the mobile device; The intelligent graphics rendering strategy module uses the intelligent graphics rendering strategy to perform intelligent graphics rendering on the application of the mobile device; The personalized rendering strategy module uses the personalized rendering strategy to perform intelligent graphics rendering on the application program of the mobile device.

10. The intelligent graphics loading and rendering system according to claim 9, characterized in that: The intelligent graphics loading and rendering system also includes a completeness and compliance module; The integrity and compliance module encrypts data transmitted between the mobile device and the intelligent graphics development unit, data transmitted between the intelligent graphics development unit and the server, data transmitted between the server and the database, and data stored in the database; And set a retention period for all data.

Images