A method and device for processing game operation data based on player device parameters

Abstract

The present application provides a method and device for processing game operation data based on player device parameters, including: performing performance prediction on a device based on the device performance parameters of the player device; dividing the devices into different device groups, performing performance prediction on the devices within the device groups, and generating a data processing strategy; quantitatively evaluating the real-time external device parameters of a target device and adjusting the operation parameters of the game in real time; dynamically adjusting the data processing strategy parameters according to the deviation of the average external device parameters of the device group where the target device is located; evaluating the performance metrics of the data processing strategy on the target device, optimizing the data processing strategy and the device group classification model; adjusting the generation parameters of the data processing strategy according to the common performance problems of the device groups; adjusting the device performance model according to the strength of the correlation between the cooling efficiency index and the battery life data; and generating a performance optimization operation instruction corresponding to the device group.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and particularly to a method and device for processing game operation data based on player device parameters. Background Art

[0002] In the development and operation of modern games and heavy applications, how to optimize device performance to provide the best user experience is an important and challenging task. With the diversification of smart devices and the increasing performance differences, developers are faced with the complex problem of adapting and optimizing the operation performance of games and applications for various devices. Especially in resource-intensive applications such as large-scale multiplayer online games, virtual reality applications, and mobile games with high graphics requirements, the hardware performance differences among different devices are significant, and the quality of the user experience is thus greatly affected. Current device performance optimization methods usually rely on preset device classifications and static optimization strategies, and cannot respond to the dynamic changes in device performance in real time. For example, some devices may have high processing power and graphics processing capabilities, but their performance will decline due to overheating after long-term operation. Other devices may have a short battery life in high-performance mode and cannot provide a stable high-performance experience for a long time. These performance differences make it difficult for a unified optimization strategy to meet the needs of all devices, thus affecting the overall user experience. Especially during the game operation process, the real-time performance data of the device, such as CPU temperature, GPU temperature, battery power, power consumption, etc., directly affects the smoothness and stability of the game. If dynamic adjustment cannot be effectively made based on these real-time data, problems such as game lag, overheating, and fast power consumption may occur during game operation. This will not only reduce user satisfaction but may also lead to excessive wear and shortened lifespan of the device. In the prior art, although there are some optimization methods based on device performance, most of them are static or based on preset rules and lack the ability to dynamically respond to real-time data. For example, some applications may detect the hardware performance of the device during the initial installation and set an initial optimization strategy based on this. However, this static optimization strategy cannot adapt to the dynamic changes in device performance during use and is difficult to optimize the game operation performance in real time. In addition, the hardware and software environments among different devices vary greatly, such as different brands of processors, different versions of operating systems, different types of displays, etc. These differences result in different performances of devices when processing the same task. Therefore, a method that can monitor and analyze device performance data in real time and dynamically adjust the data processing strategy according to the specific performance characteristics and real-time status of the device is needed to ensure that games and applications can run in the best state on various devices. Summary of the Invention

[0003] The present invention provides a method for processing game operation data based on player device parameters, mainly including:

[0004] Obtain the device performance parameters of the player's device through the API, construct a device performance model, and perform performance prediction on the device based on the device performance parameters of the player's device to obtain the external parameter performance indicators of the device in different application scenarios;

[0005] According to the device performance parameters of the player's device, construct a device group classification model, divide the devices into different device groups, obtain the cooling efficiency index and battery life data of the devices within the device group through the device performance model, and generate corresponding data processing strategies;

[0006] Real-time monitor the real-time device external parameters when the target device runs the game, quantitatively evaluate the real-time device external parameters of the target device, control the reduction range of the rendering quality through different levels, and adjust the running parameters of the game in real time;

[0007] According to the real-time device external parameters of the target device, determine the deviation of the average device external parameters of the device group where it is located. Based on the magnitude and direction of the deviation, dynamically adjust the data processing strategy parameters, and optimize the data processing strategy according to the user satisfaction scores of different device groups;

[0008] Based on the adjusted data processing strategy, evaluate the performance indicators of the data processing strategy on the target device, obtain the parameter combination that maximizes the comprehensive performance indicators, and use it as the optimized parameter of the data processing strategy to optimize the data processing strategy and the device group classification model;

[0009] According to the real-time device external parameters of different devices in the same device group, determine the common performance problems corresponding to the device group, adjust the generation parameters of the data processing strategy based on the common performance problems, and continuously monitor the performance indicators of the device group to evaluate the effect of the data processing strategy adjustment;

[0010] According to the real-time device external parameters of the devices in different device groups, determine the correlation degree of the cooling efficiency index and battery life data, and adjust the device performance model according to the strength of the correlation degree;

[0011] According to the performance indicators and optimization requirements of the device group, generate performance optimization operation instructions corresponding to the device group and send them to each device in the device group, and adjust the performance optimization operation based on the optimization effect of the performance optimization operation.

[0012] The present invention provides a game operation data processing device based on player device parameters, mainly including:

[0013] A device performance parameter collection module, which is used to obtain the device performance parameters of the player's device, including device internal parameters and device external parameters;

[0014] A device performance modeling and optimization module, which is used to construct a device performance model and iteratively optimize the model parameters;

[0015] The device group classification and group data processing strategy generation module is used to divide device groups based on the device performance parameters of players' devices and generate data processing strategies for different device groups;

[0016] The real-time performance monitoring and strategy adjustment module is used to monitor the real-time performance of the device when running the game and adjust the data processing strategy according to the performance deviation;

[0017] The performance problem analysis and strategy optimization module is used to analyze the common problems within the device group and adjust the data processing strategy to solve the problems of high temperature and battery life;

[0018] The performance optimization operation instruction generation module is used to generate operation instructions for adjusting background programs and cooling control logic.

[0019] The technical solution provided by the embodiment of the present invention may include the following beneficial effects:

[0020] The present invention discloses a game operation data processing method and device based on players' device parameters, aiming to improve and optimize the operation performance and experience of games or heavy applications on various devices through accurate device performance models and device classification models. The present invention particularly solves the problem of how to dynamically adjust the data processing strategy of a device according to the real-time data of the device performance to adapt to the performance characteristics and limitations of different devices. The unique business scenario problem lies in how to adjust the operation strategy of an application or game in real time based on the internal processing performance and external performance indicators of the device to ensure the best user experience and device performance. The present invention can effectively improve the application performance management and optimization in a multi-device environment, especially in games and other resource-intensive applications. Through real-time monitoring and intelligent adjustment, it ensures that each device can run in the best state, thereby improving the user experience and device usage efficiency. Brief Description of the Drawings

[0021] Figure 1 It is a flowchart of a game operation data processing method based on players' device parameters of the present invention.

[0022] Figure 2 It is a schematic diagram of a game operation data processing method and device based on players' device parameters of the present invention.

[0023] Figure 3 It is another schematic diagram of a game operation data processing method and device based on players' device parameters of the present invention. Detailed Embodiment

[0024] To further understand the content of the present invention, the present invention will be described in detail in combination with the accompanying drawings and embodiments. The following further elaborates on the present application in conjunction with the drawings and embodiments. It can be understood that the specific embodiments described herein are only for explaining the relevant invention and not for limiting the invention. Additionally, it should be noted that for the sake of convenience of description, only the parts related to the invention are shown in the drawings.

[0025] As Figures 1-3 , a method and device for processing game operation data based on player device parameters in this embodiment may specifically include:

[0026] Step S101, obtain the device performance parameters of the player device through the API, construct a device performance model, and perform performance prediction on the device based on the device performance parameters of the player device to obtain the external parameter performance indicators of the device in different application scenarios.

[0027] Obtain the device performance parameters of the player device through the API, including device internal parameters and device external parameters. Among them, the device internal parameters include the processor model, number of cores, main frequency, and memory size, and the device external parameters include image quality indicators, cooling efficiency indicators, and battery life data. The image quality indicators include frame rate, rendering resolution, texture quality, image clarity, and color saturation. Standardize the obtained device performance parameters and vectorize the device performance parameters. According to the vectorized device performance parameters, use the recurrent neural network algorithm to construct a device performance model and establish a mapping relationship between the device internal parameters and the device external parameters. Based on the obtained device performance parameters of the player device, train and optimize the device performance model, and adjust the feature weights of the device performance model parameters in a cyclic iterative manner. According to the internal parameter information of the new device, use the trained device performance model to predict the performance indicators of the device to obtain the external parameter performance indicators of the device in different application scenarios. Compare the prediction results with the actual test results to obtain the prediction error and confidence level of the device performance model, and continuously monitor and improve the device performance model. Use the prediction results of the device performance model for device selection, optimization, and improvement, including optimizing the heat dissipation design of the device according to the predicted heat dissipation efficiency indicators and selecting the battery capacity according to the predicted battery life data.

[0028] Exemplarily, when collecting device performance parameters, a dedicated performance testing software GeekBench is used to test parameters in multiple dimensions such as the processor performance, graphics performance, and memory performance of the device. If the test results show that the processor performance score is 150,000 points and the graphics performance score is 80,000 points, these scores are used as the numerical representations of the internal and external parameters of the device. Among them, the internal parameters of the device include the processor model Snapdragon 888, the number of cores 8, the main frequency 2.84 GHz, and the memory size 8 GB. The external parameters of the device include the picture quality indicators including the frame rate 60 FPS, the rendering resolution 1080p, the high texture quality, the image clarity 80%, the color saturation 90%, the cooling efficiency indicator 70%, and the battery life data 8 hours. When cleaning and transforming the collected data, for the categorical variable of the processor model, one-hot encoding is used to convert it into a binary vector. For example, Snapdragon 888 is converted into [0, 0, 1]. For numerical variables such as the main frequency and memory size, min-max normalization can be used to scale them into the range of [0, 1]. For example, the main frequency of 2.84 GHz is normalized to 0.72, and the memory size of 8 GB is normalized to 0.4. When constructing the device performance model, a recurrent neural network is selected as the model architecture. The number of input layer nodes is 10, which is the dimension of the internal parameters of the device. The number of hidden layer nodes is set to 20, and the number of output layer nodes is 5, which is the dimension of the external parameters of the device. In the model training stage, the Adam optimizer is used, the learning rate is set to 0.001, the batch size is 32, and the number of training epochs is 100. During the training process, the model performance is evaluated on the validation set after each epoch. If the validation loss does not decrease for 5 consecutive epochs, the early stopping mechanism is triggered to prevent overfitting. For the actual performance prediction task, first, the parameters of the candidate device are input into the trained model to obtain the predicted values of various external parameter performance indicators. For example, the predicted frame rate is 55 FPS, the predicted battery life is 7.5 hours, and the predicted cooling efficiency is 65%. Then, the predicted results are compared with the actual test results. The actual test results show that the frame rate is 53 FPS, the battery life is 7 hours, and the cooling efficiency is 62%. Thus, the prediction errors are 2 FPS, 0.5 hours, and 3% respectively. The prediction results of the device performance model are used for device selection and improvement. According to the predicted cooling efficiency indicator, the heat dissipation design of the device is optimized, and the radiator area is increased to 120 cm 2 , and the fan speed is increased to 3500 RPM. According to the predicted battery life data, a battery with a capacity of 4500 mAh is selected.

[0029] Step S102: According to the device performance parameters of the player's device, construct a device group classification model to divide the devices into different device groups. Obtain the cooling efficiency indicator and battery life data of the devices within the device group through the device performance model, and generate corresponding data processing strategies.

[0030] According to the device performance parameters of the player's device, use the K-means clustering algorithm for model training to construct a device group classification model, and divide the devices into different device groups, including high similarity, medium similarity, and low similarity. For each device group, predict the cooling efficiency index and battery life data of the devices within the group through the trained device performance model to obtain the overall performance index of the group. According to the cooling efficiency and battery life of different device groups, formulate corresponding data processing strategies, including for device groups with a cooling efficiency lower than the preset cooling efficiency threshold, reducing the graphic quality by lowering the texture resolution and simplifying the lighting model, and for device groups with a battery life lower than the preset battery life threshold, limiting the frame rate by setting the maximum FPS threshold and dynamically adjusting the rendering frequency. Set the formulated data processing strategies through the API interface provided by the game engine and transfer them to the game engine to dynamically adjust the rendering parameters and physical simulation accuracy. During the game operation, continuously obtain the external parameter data of the device at a fixed time interval through the performance monitoring interface of the operating system. Compare the obtained external parameter data with the prediction results of the device performance model and adjust the data processing strategy in real time. For newly released devices or unknown devices, obtain the performance parameters of the newly released devices through quick testing, use the device group classification model to determine the device group to which they belong, and apply the existing data processing strategies. Continuously obtain the performance data of the new device through the performance monitoring interface of the operating system and update and optimize the device performance model.

[0031] Exemplarily, a batch of performance parameters of player devices are obtained through the API, including internal device parameters such as the processor model Snapdragon 888, the number of cores 8, the main frequency 2.84 GHz, the memory size 8 GB, and external device parameters such as the picture quality index, the frame rate 60 FPS, the rendering resolution 1080p, the texture quality high, the image clarity 80%, the color saturation 90%, the cooling efficiency index 70%, and the battery life data 8 hours. After standardizing and vectorizing these performance parameters, the K-means clustering algorithm is used for model training, and the devices are divided into three device groups with high similarity, medium similarity, and low similarity. For each device group, the cooling efficiency index and battery life data of the devices in the group are predicted through the trained device performance model to obtain the overall performance index of the group. For the device group with a cooling efficiency lower than 60%, the data processing strategies formulated include reducing the texture resolution to 720p and simplifying the lighting model to reduce the graphics quality. For the device group with a battery life lower than 6 hours, the maximum FPS threshold is set to 30 and the rendering frequency is dynamically adjusted to limit the frame rate. These formulated data processing strategies are set through the API interface provided by the game engine and passed to the game engine to dynamically adjust the rendering parameters and the physical simulation accuracy. During the game operation, through the performance monitoring interface of the operating system, the external parameter data of the device are continuously obtained at a frequency of once a minute, and the obtained data are compared with the prediction results of the device performance model to adjust the data processing strategy in real time. For newly released devices, such as the processor model Snapdragon 895, the number of cores 8, the main frequency 3.0 GHz, and the memory size 12 GB, their performance parameters are obtained through quick tests, and the device group classification model is used to judge the device group to which they belong. Assuming they are classified into the medium similarity group, the existing data processing strategies are applied. Through the performance monitoring interface of the operating system, the performance data of the new device are continuously obtained, and the device performance model is continuously updated and optimized to ensure that the new device can obtain the best performance index in the game.

[0032] Step S103, monitor the real-time external device parameters of the target device when running the game in real time, quantitatively evaluate the real-time external device parameters of the target device, and control the reduction range of the rendering quality through different levels to adjust the running parameters of the game in real time.

[0033] Deploy a performance monitoring module on the target device. Through the interface of the game rendering engine, obtain the external parameters data of the device in real time at a fixed frequency to form a time-series data stream. Preprocess the obtained real-time external parameters data of the device, including using the moving average method for data denoising and the Z-score normalization method for data normalization. Design a game performance evaluation index system according to the image quality index and frame rate stability during game operation, including image quality and frame rate stability indexes. Among them, the image quality index includes image sharpness index and color saturation index, the frame rate stability index is the frame rate jitter index, the image sharpness index includes sharpness and contrast, the color saturation index is the gamut coverage rate, and the frame rate jitter index is the variance of inter-frame delay. Map each index to a unified performance score through the weighted average method, and dynamically adjust the weights of each index according to the characteristics of different games and devices to obtain the real-time performance evaluation result of the target device. Judge whether the target device meets the performance requirements for game operation according to the real-time performance evaluation result of the target device and the preset performance threshold. If the evaluation result is lower than the preset performance threshold, trigger the dynamic adjustment mechanism. Control the reduction range of rendering quality at different levels by reducing texture resolution, simplifying the lighting model, and reducing the number of special effects, and adjust the running parameters of the game in real time. During the dynamic adjustment process, continuously monitor the real-time performance changes of the target device, perform weighted average on the data within the window through a sliding window of a fixed size, use the exponentially weighted moving average method to cumulatively average the historical external parameters of the device, dynamically update the performance evaluation result, and continuously monitor and predict the performance of the target device. According to the real-time external parameters data of the target device, carefully use the isolation forest algorithm for model training to detect device performance anomalies in real time. According to the real-time external parameters data of the target device, use the ARIMA algorithm for model training to predict the change trend of the external parameters data of the device and determine the change trend of game performance. Real-time display the performance status and trend of the target device through a visual dashboard and anomaly alerts to assist in game performance optimization and problem location. Continuously optimize the game rendering engine and resource loading mechanism based on the actual performance indicators and prediction results of the target device.

[0034] Exemplarily, a performance monitoring module is deployed on the target device. Through the interface of the game rendering engine, the external parameter data of the device is obtained in real time at a frequency of once per second, including frame rate of 60 FPS, image clarity of 80%, color saturation of 90%, and other image quality and frame rate stability metrics, forming a time-series data stream. The obtained real-time external parameter data of the device is preprocessed. The moving average method is used to denoise the frame rate data, and the Z-score normalization method is used to normalize the image clarity and color saturation data. According to the image quality metrics and frame rate stability during game operation, a game performance evaluation index system is designed, including an image clarity index, including sharpness and contrast, a color saturation index of gamut coverage, and a frame rate stability index of inter-frame delay variance. Among them, the image clarity is 80%, the contrast is 70%, the color saturation is 90%, and the frame rate jitter, that is, the inter-frame delay variance is 5 ms. Through the weighted average method, each index is mapped to a unified performance score, and the weights of each index are dynamically adjusted according to the characteristics of different games and devices to obtain the real-time performance evaluation result of the target device. For example, the comprehensive score is 85 points. According to the real-time performance evaluation result of the target device and the preset performance threshold of 80 points, it is judged whether the target device meets the performance requirements for game operation. If the evaluation result is lower than the preset performance threshold, a dynamic adjustment mechanism is triggered. By reducing the texture resolution from 1080p to 720p, simplifying the lighting model, reducing the number of special effects, etc., the reduction amplitude of the rendering quality is controlled at different levels, and the operation parameters of the game are adjusted in real time. During the dynamic adjustment process, the real-time performance changes of the target device are continuously monitored. By setting a sliding window every 5 minutes, the data within the window is weighted averaged, and the exponential weighted moving average method is used to cumulatively average the historical external parameter data of the device, and the performance evaluation result is dynamically updated. To detect device performance anomalies in real time, based on the real-time external parameter data of the target device, the isolation forest algorithm is used for model training to identify the abnormal points in the performance data. At the same time, the ARIMA algorithm is used for model training on the external parameter data of the device to predict the change trend of the future external parameter data of the device to determine the change trend of the game performance. The performance status and trend of the target device are displayed in real time through a visualization dashboard, and abnormal alarms are set to assist in game performance optimization and problem location. Based on the actual performance metrics of the target device with the real-time frame rate dropping to 50 FPS and the prediction result that the frame rate may continue to decline in the future, continuous optimization is carried out on the game rendering engine and resource loading mechanism, such as further reducing the complexity of special effects or adjusting the resource loading order, to ensure that the target device can maintain the best game performance metrics under different application scenarios.

[0035] Step S104: Determine the deviation of the average external parameters of the device group based on the real-time external parameters of the target device. Dynamically adjust the data processing strategy parameters based on the magnitude and direction of the deviation, and optimize the data processing strategy according to the user satisfaction scores of different device groups.

[0036] Monitor the real-time external parameters of the target device, including frame rate, rendering resolution, texture quality, image clarity, and color saturation, and perform normalization processing on each external parameter index. Calculate the Euclidean distance between the normalized index vectors to obtain the deviation between the external parameters of the target device and the average external parameters of the group. If the deviation exceeds the preset deviation threshold, dynamically adjust the data processing strategy parameters according to the magnitude and direction of the deviation. Continuously monitor the performance indicators of the target device after the data processing strategy parameters are adjusted, and dynamically update the calculation results of the external parameter deviation by means of a sliding window. According to the performance indicators and user feedback of the target device, use the Q-learning algorithm, with the device performance indicators and user satisfaction as the reward signals, to automatically learn and optimize the data processing strategy parameters. If the adjusted deviation still exceeds the preset deviation threshold, continue to adjust the data processing strategy. If the number of times of adjusting the data processing strategy exceeds the preset number threshold and the external parameter deviation of the target device still exceeds the preset deviation threshold, re-divide the target device into device groups and match the device group with the highest similarity. Obtain user experience feedback data through the in-game feedback interface, including user ratings and comments, and preprocess the user experience feedback data, including removing noise, word segmentation, and labeling sentiment tags. Fine-tune the BERT model on the labeled sentiment data, and use the fine-tuned BERT model to perform sentiment classification on the new user comment data to determine the sentiment tendency of each comment. Calculate the user satisfaction scores of different device groups according to the sentiment classification results. Optimize the data processing strategy based on the device performance indicators and user experience feedback, and feedback the optimization results of the data processing strategy to the strategy adjustment module. Dynamically adjust the preset deviation threshold using the EWMA algorithm according to the historical external parameter data of the device and user satisfaction.

[0037] Exemplarily, when monitoring the real-time external parameters of the target device, including a frame rate of 60 FPS, a rendering resolution of 1080p, high texture quality, an image clarity of 80%, and a color saturation of 90%, after performing Z-score normalization on each external parameter index, the Euclidean distance between the normalized index vectors is calculated, and it is found that the deviation between the external parameters of the target device and the group average external parameters is 0.7. If the deviation exceeds the preset deviation threshold of 0.5, then according to the magnitude and direction of the deviation, the data processing strategy parameters are dynamically adjusted, including reducing the texture quality to medium or adjusting the rendering resolution to 720p. Continuously monitor the performance metrics of the target device after the data processing strategy parameters are adjusted. Through the method of a sliding window, the calculation result of the external parameter deviation is dynamically updated every minute. According to the performance metrics of the target device and user feedback, using the Q-learning algorithm, with the frame rate, image clarity, and user satisfaction as reward signals, automatically learn and optimize the data processing strategy parameters, thereby reducing the maximum FPS limit from 60 to 50. If the adjusted deviation still exceeds the preset deviation threshold of 0.5, then continue to adjust the data processing strategy. If the number of times of adjusting the strategy exceeds 5 times and the external parameter deviation still exceeds the preset deviation threshold, then re-divide the target device into device groups and match the device group with the highest similarity to it. Obtain user experience feedback data through the in-game feedback interface, including user ratings and comments. Use the BERT model to classify the sentiment data and determine the sentiment tendency of each comment. For example, 80% of the comments are positive. According to the sentiment classification results, calculate the user satisfaction scores of different device groups. For example, the satisfaction score of the high-similarity group is 0.9. According to the device performance metrics and user experience feedback, optimize the data processing strategy, such as further reducing unnecessary special effects, improving the frame rate stability, and feedback the optimization results to the strategy adjustment module. According to the historical device external parameter data and user satisfaction, use the EWMA algorithm to dynamically adjust the preset deviation threshold to make it more flexible to adapt to the needs of different devices and scenarios.

[0038] Step S105, based on the adjusted data processing strategy, evaluate the performance metrics of the data processing strategy on the target device, obtain the parameter combination that maximizes the comprehensive performance metrics, and use it as the optimization parameter of the data processing strategy and the device group classification model to optimize the data processing strategy and the device group classification model.

[0039] According to the adjusted data processing strategy, extract the key parameters in the data processing strategy and determine the data processing strategy parameter vector. The key parameters include the rendering quality level and the frame rate limit threshold. According to the strategy parameter vector, use the K-means clustering algorithm for model training to group the data processing strategy parameter vector. For each data processing strategy parameter group, including using the Gaussian process as the prior distribution, using the expected improvement as the acquisition function, and updating the hyperparameters of the Gaussian process by the stochastic gradient descent method, evaluate the performance metrics of the data processing strategy on the target device, and obtain the parameter combination that maximizes the comprehensive performance metrics as the optimized parameters of the data processing strategy. Among them, the comprehensive performance metrics include device performance parameters, game revenue, and user retention rate. According to the key parameters of the data processing strategy, use the recurrent neural algorithm to construct a strategy generation model and determine the optimized data processing strategy. Use the optimized data processing strategy parameters as training data to retrain the strategy generation model, and by modifying the loss function of the strategy generation model and adding a parameter offset penalty term, obtain the optimized strategy generation model. According to the performance metrics of the target device after adjusting the data processing strategy, evaluate the similarity between the target device and the original device group. By calculating the Mahalanobis distance between the device extrinsic parameter vector and the group center vector, determine whether the target device matches the current group. If the similarity between the target device and the original group is lower than the preset similarity threshold, update the parameters and decision boundary of the device group classification model by incremental learning, fine-tune some parameters of the device group classification model by the gradient descent method, and re-classify and predict the target device to obtain its new device group. Use the device performance parameters of the new device as the new goal of data processing strategy optimization to form a two-way feedback mechanism between data processing strategy optimization and device classification. By establishing a feedback channel between data processing strategy optimization and game operation, feedback the impact of the data processing strategy on business metrics to the strategy optimization module to achieve a closed-loop of business-driven data processing strategy optimization. By continuously monitoring the execution effect of the optimized data processing strategy on the target device, set the evaluation period and evaluation index system, regularly evaluate the data for improving the device performance and user experience by the data processing strategy, and apply the evaluation results to the iteration of data processing strategy optimization and device group classification.

[0040] Exemplarily, key parameters are extracted according to the adjusted data processing strategy. The rendering quality level is 3 and the frame rate limit threshold is 45 FPS. These key parameters are determined as the data processing strategy parameter vector (3, 45). The K-means clustering algorithm is used to group these data processing strategy parameter vectors, and the strategy parameter vectors are divided into 3 groups. For each data processing strategy parameter group, the Gaussian process is used as the prior distribution, and the expected improvement is used as the acquisition function. The hyperparameters of the Gaussian process are updated by the stochastic gradient descent method, and the performance metrics of the data processing strategy on the target device are evaluated to obtain the parameter combination that maximizes the comprehensive performance metric. If the optimized parameter combination of a certain group is a rendering quality level of 2 and a frame rate limit threshold of 50 FPS, the comprehensive performance metrics include a 10% increase in the frame rate stability of the device performance parameter, a 5% increase in game revenue, and an 8% increase in user retention rate. According to these key parameters, a strategy generation model is constructed using the recurrent neural network algorithm, and the optimized data processing strategy is determined. The optimized data processing strategy parameters are used as training data to retrain the strategy generation model. By modifying the loss function of the strategy generation model and adding a parameter offset penalty term, an optimized strategy generation model is obtained. If the performance metrics of the target device after adjusting the data processing strategy are that the frame rate is stable at 48 FPS and the rendering quality remains at a medium level. Next, the similarity between the target device and the original device group is evaluated. By calculating the Mahalanobis distance between the external parameter index vector of the device and the group center vector, it is determined whether the target device matches the current group. If the similarity between the target device and the original group is lower than the preset similarity threshold of 0.6, then through incremental learning, the parameters and decision boundaries of the device group classification model are updated, some parameters of the device group classification model are fine-tuned by the gradient descent method, and the target device is re-classified and predicted to obtain its new device group to which it belongs. Therefore, the target device is re-classified as a high-similarity group. The device performance parameters of the new device, a rendering quality level of 2 and a frame rate limit of 50 FPS, are used as the new target for data processing strategy optimization, forming a two-way feedback mechanism between data processing strategy optimization and device classification. By establishing a feedback channel between data processing strategy optimization and game operation, the impact of the data processing strategy on business metrics, including user retention rate and game revenue, is fed back to the strategy optimization module to achieve a closed-loop of business-driven data processing strategy optimization. By continuously monitoring the execution effect of the optimized data processing strategy on the target device, setting an evaluation cycle of once a week and an evaluation index system including frame rate stability, user satisfaction, etc., regularly evaluating the data on the improvement of the device performance and user experience by the data processing strategy, and applying the evaluation results to the iteration of data processing strategy optimization and device group classification to ensure that the device maintains the best performance and user experience in different application scenarios.

[0041] Step S106: Determine the common performance issues corresponding to the device group based on the real-time external parameters of different devices in the same device group, adjust the generation parameters of the data processing strategy based on the common performance issues, and continuously monitor the performance metrics of the device group to evaluate the effect of the data processing strategy adjustment.

[0042] Sampling is performed at a fixed frequency through device sensors to obtain the device operation data during the operation of each device in the device group in real time, forming device operation time series data. The device operation data includes CPU temperature, GPU temperature, battery power, and power consumption. Preprocess the device operation time series data, including performing outlier detection using the interquartile range method, setting the outlier threshold to Q1 - 1.5IQR and Q3 + 1.5IQR, identifying and removing outliers. Use exponential weighted moving average for data smoothing to eliminate short-term fluctuations. Set a time window of a fixed size and slide it every certain time. For each parameter of the device operation data, calculate the statistics of different devices within the same device group on each time window, including mean, variance, and peak value. By comparing the differences in statistics between different devices, identify the common performance problems within the group. Calculate the speed of the device battery power decline. If the power decline speed is greater than the preset speed threshold within the preset time, then determine the device as a device with battery life problem. According to the device operation time series data, construct feature vectors, including mean, standard deviation, kurtosis, and wavelet transform coefficients, train an isolation forest model, set an outlier threshold, and mark and alarm the devices that exceed the outlier threshold. According to the identified performance problems, conduct problem classification and attribution identification to determine whether the problem is a hardware problem or a software problem. If it is determined to be a hardware problem, solve it by optimizing the hardware design or selecting high-end components. If it is determined to be a software problem, relieve the performance bottleneck of the hardware by reducing the rendering load and lowering the model complexity. Map the common performance problems identified within the group to the generation parameters of the data processing strategy, and determine the parameter items to be adjusted and the adjustment range according to the problem type and severity. According to different performance problems, adjust the parameters of the data processing strategy. For high-temperature problems, give priority to adjusting the rendering resolution and frame rate limit. For battery life problems, give priority to adjusting the screen brightness and data transmission frequency, and set several levels according to the problem severity. Input the adjusted data processing generation parameters into the optimized strategy generation model to obtain the optimized data processing strategy. After the optimized data processing strategy is deployed, continuously monitor the performance indicators of the device group, and evaluate the effect of the data processing strategy adjustment through the comparison of the performance indicators before and after optimization. Set quantitative evaluation indicators, and set differentiated evaluation goals for different game types and device groups. In the A / B test, select a preset number of devices as the experimental group, adopt the optimized data processing strategy, and the remaining devices as the control group, adopt the old data processing strategy, and obtain the changes in the key indicators of the two groups of devices within the preset time period. The key indicators include device performance parameters, game revenue, and user retention rate, and verify the effectiveness and security of the optimized data processing strategy. Promote the data processing strategy with good adjustment effect to other device groups, conduct attribution identification on the data processing strategy with poor adjustment effect, and iteratively optimize the data processing strategy generation parameters until the optimal data processing strategy configuration is determined.

[0043] Exemplarily, samples are taken at a frequency of once per second through device sensors to obtain the runtime data of each device in the device group in real time, including CPU temperature of 75°C, GPU temperature of 70°C, battery power of 85%, and power consumption of 10W, forming device runtime time series data. When preprocessing this data, the interquartile range method is used for outlier detection, and the outlier thresholds are set to Q1 - 1.5IQR and Q3 + 1.5IQR. After identifying and removing outliers, exponential weighted moving average is used for data smoothing to eliminate short-term fluctuations. A fixed-size time window of 5 minutes is set and slides once per minute. Statistics of different devices within the same device group are calculated for each time window, including mean of 70°C, variance of 5°C, and peak value of 80°C. By comparing the statistical differences between different devices, common high-temperature performance problems within the group are identified. Calculate the rate of battery power decline of the device and find that the power of a certain device drops at a rate of 10% per hour within the preset time, exceeding the preset speed threshold of 5% per hour, and determine that this device is a device with a battery life problem. According to the device runtime time series data, a feature vector is constructed, including mean, standard deviation, kurtosis, and wavelet transform coefficients. An isolation forest model is trained, and the outlier threshold is set to 0.7. Devices exceeding the outlier threshold are marked and alarmed for problems. Through problem classification and attribution recognition, it is determined that the high-temperature problem of a certain device is a hardware problem caused by poor heat dissipation design, while the battery life problem is caused by too many background programs. For hardware problems, they are solved by optimizing the heat dissipation design or selecting high-end heat dissipation devices. For software problems, the hardware performance bottleneck is alleviated by reducing the rendering load and lowering the model complexity. Map the identified common high-temperature and battery life performance problems within the group to the generation parameters of the data processing strategy, and determine the parameter items to be adjusted and the adjustment range according to the problem type and severity. For high-temperature problems, the rendering resolution is preferentially adjusted from 1080p to 720p and the frame rate limit is adjusted from 60FPS to 30FPS. For battery life problems, the screen brightness is preferentially adjusted from 100% to 70% and the data transmission frequency is changed from once per second to once every 5 seconds, and several gears are set according to the problem severity. Input the adjusted data processing generation parameters into the optimized strategy generation model to obtain the optimized data processing strategy. After the optimized data processing strategy is deployed, continuously monitor the performance indicators of the device group, and evaluate the effect of the data processing strategy adjustment through comparison of the performance indicators before and after optimization. If the average CPU temperature of the optimized device drops to 65°C and the battery life is extended to 10 hours. Set quantitative evaluation indicators, and set differentiated evaluation targets for different game types and device groups.In the A / B test, 100 devices are selected as the experimental group, and the optimized data processing strategy is adopted. The remaining 100 devices are used as the control group, and the old data processing strategy is adopted. The changes in key indicators of the two groups of devices within one week are obtained. The key indicators include device performance parameters, a 10% increase in game revenue, and an 8% increase in user retention rate. The device performance parameters include CPU temperature and battery life. The effectiveness and security of the optimized data processing strategy are verified. The data processing strategy with good adjustment effect is promoted to other device groups, and the data processing strategy with poor adjustment effect is attributed and identified. It is found that a certain device cannot achieve the expected effect due to hardware limitations. The data processing strategy parameters are further iteratively optimized until the optimal data processing strategy configuration is determined.

[0044] Step S107: Determine the correlation degree between the cooling efficiency index and the battery life data according to the real-time external device parameters of the devices in different device groups, and adjust the device performance model according to the strength of the correlation degree.

[0045] The cooling efficiency index and the battery life data of the devices in different device groups are obtained in real time at a fixed frequency through the temperature sensor and the fuel gauge of the device, forming a time series data set reflecting the device performance. The obtained cooling efficiency index and battery life data are preprocessed, including outlier detection, data normalization, and time alignment. The correlation degree between the cooling efficiency index and the battery life data is calculated through the Pearson correlation coefficient, a correlation coefficient matrix is generated, a significance threshold for the correlation coefficient is set, and a significance level for the p-value is set. When the absolute value of the correlation coefficient is greater than the preset significance threshold for the correlation coefficient and the p-value is less than the significance level, the correlation relationship between the two indicators is considered strong; otherwise, the correlation relationship is considered weak. According to the strength of the correlation relationship, the structure and parameters of the model are adjusted. For the indicators with a strong correlation relationship, the corresponding influence path is added to the device performance model, and the path weight is scaled according to the size of the correlation degree. For the indicators with a weak correlation relationship, the weight of the corresponding influence path can be reduced. The correlation degree is introduced into the regularization term of the model. When the correlation degree is greater than the preset correlation degree threshold, the regularization strength is reduced; when the correlation degree is less than the preset threshold, the regularization strength is increased. Based on the adjusted device performance model, the gradient descent optimization algorithm is used to update the weight parameters of the model through backpropagation, and the adaptive moment estimation of the Adam optimizer is used to balance the convergence speed and generalization performance of the model. The adjusted device performance model is applied to different device groups, and the performance improvement effect of the device performance model is evaluated by comparing the prediction errors of the device performance model before and after adjustment on the test set or the validation set. The training error and validation error of the device performance model in different iteration rounds are recorded, a learning curve is drawn, and the convergence speed and overfitting degree of the model are determined. The visualization tool TensorBoard is used to monitor the structure and performance changes of the device performance model in real time to assist in the correlation degree analysis and the tuning of the device performance model.

[0046] Exemplarily, through the temperature sensor and the battery meter of the device, the cooling efficiency index and the battery life data of the devices in different device groups are obtained in real time at a frequency of once per second, forming a time series data set reflecting the device performance. For the obtained cooling efficiency index and battery life data, the cooling efficiency index includes a CPU temperature of 65 °C and a GPU temperature of 60 °C, and the battery life data includes a battery power drop rate of 10% per hour. For preprocessing, the interquartile range method is used for outlier detection, Z-score normalization, and linear interpolation method for time alignment. By calculating the Pearson correlation coefficient, the correlation coefficient between the cooling efficiency and the battery life data is -0.6, and the p-value is 0.01, generating a correlation coefficient matrix. The significance threshold of the correlation coefficient is set to 0.5, and the significance level of the p-value is 0.05. When the absolute value of the correlation coefficient is greater than the preset threshold and the p-value is less than the significance level, the correlation relationship between the two indicators is considered strong. According to the strong correlation relationship, the correlation degree between the cooling efficiency and the battery life is introduced into the regularization term of the device performance model, reducing the regularization strength, and adding corresponding influence paths in the model and scaling the path weights. For the indicators with weak correlation relationships, the weights of the corresponding influence paths can be reduced. The gradient descent optimization algorithm is used to update the weight parameters of the model through backpropagation, and the adaptive moment estimation of the Adam optimizer is used to balance the convergence speed and generalization performance of the model. If the prediction error of the adjusted device performance model on the training set drops from 0.05 to 0.03, and the prediction error on the validation set drops from 0.06 to 0.04, it shows a significant performance improvement. Record the training error and validation error of the device performance model at different iteration rounds, draw a learning curve, and determine the convergence speed and overfitting degree of the model. Use the visualization tool TensorBoard to monitor the structure and performance changes of the device performance model in real time, assist in correlation analysis and device performance model tuning. At the 50th iteration, the training error and validation error of the model are 0.03 and 0.04 respectively, indicating that the model converges well and there is no overfitting phenomenon. Apply the adjusted device performance model in different device groups, and evaluate the performance improvement effect of the device performance model by comparing the prediction errors of the device performance model before and after adjustment on the test set or the validation set, ensuring that the device maintains the best performance and user experience in different application scenarios.

[0047] Step S108, according to the performance indicators and optimization requirements of the device group, generate performance optimization operation instructions corresponding to the device group and send them to each device in the device group, and adjust the performance optimization operation based on the optimization effect of the performance optimization operation.

[0048] According to the performance indicators and optimization requirements of the device group, determine the performance optimization operation instructions for different groups, including adjusting the background program running strategy to enhance battery life and adjusting the control logic of the cooling components to improve cooling efficiency, to form a set of optimization operation instructions. Encode and parameterize the optimization operation instructions, and convert the instructions into a format that can be directly executed by the device, including Protocol Buffers and MessagePack. According to the type and priority attributes of the instructions, assign a unique identifier and metadata label to each instruction, and define the instruction semantics and syntax specifications, including the instruction type, parameter list, and execution conditions. According to the network topology structure and communication protocol of the device group, design an instruction distribution strategy, and use multicast, broadcast, and point-to-point methods to send the optimization operation instructions to each device node in the group, and ensure the arrival rate of the instructions through retransmission and confirmation mechanisms. For services that require real-time response, use an event-driven instruction distribution method, and for optimization operations that require synchronous execution, use an instruction synchronization mechanism based on timestamps or version numbers. The device parses the content and parameters of the received optimization operation instructions, and triggers corresponding performance optimization actions according to the type and priority of the instructions. According to the system type and version information of the device, dynamically load the corresponding API module and call parameters, and decouple the optimization operation instructions from the system environment. If the instruction is for the background program running strategy, call the adapted system API to dynamically adjust the scheduling strategy and resource allocation of the background process, including reducing the priority of non-critical tasks and clearing the memory occupancy of non-critical tasks. If the optimization operation instruction is for the cooling component control logic, obtain the temperature data of the device through sensors, including CPU temperature and battery temperature, and judge the current heat dissipation state of the device according to the trend and amplitude of the temperature change. The heat dissipation state includes normal, warning, and dangerous. According to the heat dissipation state and the preset temperature threshold, determine the target value of the PID controller, and the target value is the desired temperature range. According to the difference between the current temperature and the target temperature, set the proportional, integral, and differential parameters of the PID controller, calculate the output value of the controller, including the adjustment amount of the fan speed and the liquid cooling pump power, and dynamically adjust the heat dissipation components. Continuously obtain the performance and energy consumption data of the device, and based on the performance evaluation standard AppStartupTime, obtain the operation behavior and experience feedback of users in different business scenarios through data embedding and log collection. Conduct a correlation analysis between the device performance and experience feedback, calculate the battery life, heat generation, average operation completion time, page loading delay, crash rate, and stutter rate before and after performance optimization, evaluate the effect of the performance optimization operation through before-and-after comparison and incremental analysis, generate an optimization effect evaluation report, and send the evaluation results back to the optimization decision-making module. According to the battery life, heat generation, average operation completion time, page loading delay, crash rate, and stutter rate in the optimization effect evaluation report, judge whether the current optimization operation has achieved the expected goal.If the evaluation report shows that the battery life, heat generation, average operation completion time, page loading latency, crash rate, and jitter rate after the device performance optimization operation do not reach the preset optimization threshold, it is considered that the performance optimization operation fails to achieve the expected goal, and the performance optimization operation is adjusted again, including reducing the priority of background tasks and redesigning the PID controller parameters. By adjusting the instruction parameters and modifying the algorithm logic, the performance indicators of the device are iteratively optimized until the optimization goal is achieved. If the optimization operation has achieved the expected goal, the configuration information of the optimization operation is stored in the non-volatile memory of the device, and a unified configuration file format and read / write interface are defined. When the device is started or restored to the factory settings, the configuration of the optimization operation is automatically loaded and applied. Design a version management and upgrade mechanism for the configuration file to match the iterative update of the optimization strategy, and take data verification and exception handling measures to ensure the integrity and security of the configuration information. After the performance optimization operation is solidified as the default configuration of the device, the performance optimization strategy is dynamically adjusted according to the real-time evaluation feedback, including adjusting the threshold and weight of the optimization operation, or generating a new optimization operation, forming a long-term performance optimization strategy, and adding the performance optimization case to the knowledge base to guide subsequent optimization decisions.

[0049] Exemplarily, according to the performance metrics and optimization requirements of different device groups, performance optimization operation instructions for the high-temperature group and the low-battery group are determined. Among them, the cooling efficiency of the high-temperature group is lower than 70%, and the battery life of the low-battery group is less than 5 hours. An instruction set for optimization operations is formed, including reducing the priority of background programs, adjusting the fan speed and the power of the liquid cooling pump. These optimization operation instructions are encoded and parameterized into the format Protocol Buffers that can be directly executed by the device, and a unique identifier and metadata label are assigned to each instruction. The semantics and syntax specifications of the instructions are defined, including instruction types, parameter lists, and execution conditions. According to the network topology and communication protocol of the device group, an instruction distribution strategy is designed, and the optimization operation instructions are sent to each device node in the group by multicast, and the arrival rate of the instructions is ensured through a retransmission and confirmation mechanism. For services that require real-time response, an event-driven instruction distribution method is adopted, and for optimization operations that need to be executed synchronously, a timestamp-based instruction synchronization mechanism is used. When the device receives the optimization operation instruction, it parses the content and parameters of the instruction, and triggers corresponding performance optimization actions according to the type and priority of the instruction. If the instruction is for the background program running policy, the adapted system API is called to dynamically adjust the scheduling policy and resource allocation of the background process, including reducing the priority of non-critical tasks and clearing the memory occupancy of non-critical tasks. If the optimization operation instruction is for the cooling component control logic, the temperature data of the device is obtained through sensors, the CPU temperature is 75°C, and the battery temperature is 50°C. According to the trend and amplitude of the temperature change, it is judged that the current heat dissipation state of the device is a warning state. According to the heat dissipation state and the preset temperature threshold, which is normally below 70°C, in the warning range of 70°C to 85°C, and dangerous above 85°C, the target value of the PID controller is determined to be 70°C. According to the difference between the current temperature and the target temperature, the three parameters of the PID controller, proportional P = 0.1, integral I = 0.05, and derivative D = 0.01, are set, and the output value of the controller is calculated, such as adjusting the fan speed to 30,000 RPM and the power of the liquid cooling pump to 50% to dynamically adjust the heat dissipation components. Continuously obtain the performance and energy consumption data of the device, and through the performance evaluation standard AppStartupTime, obtain the operation behaviors and experience feedback of users in different business scenarios through data embedding and log collection. The device performance is correlated with the user experience feedback, and the battery life before and after performance optimization is calculated. The battery life before optimization is 4 hours, and after optimization is 6 hours; the heat generation before optimization is 80°C, and after optimization is 70°C; the average operation completion time before optimization is 5 seconds, and after optimization is 3 seconds; the page loading delay before optimization is 2 seconds, and after optimization is 1 second; the crash rate before optimization is 2%, and after optimization is 1%; the stuttering rate before optimization is 5%, and after optimization is 2%. The effect of the performance optimization operation is evaluated through before-and-after comparison and incremental analysis, an optimization effect evaluation report is generated, and the evaluation results are sent back to the optimization decision-making module.Based on the indicators such as battery life, heat generation, average operation completion time, page loading latency, crash rate, and stutter rate in the optimization effect evaluation report, determine whether the current optimization operation has achieved the expected goal. If the evaluation report shows that the indicators after the device performance optimization operation do not reach the preset optimization threshold, such as the battery life not reaching 5 hours, it is considered that the performance optimization operation has not achieved the expected goal, and the performance optimization operation is adjusted again, including further reducing the background task priority, redesigning the PID controller parameters, etc. Iteratively optimize the device's performance indicators by adjusting instruction parameters and modifying algorithm logic until the optimization goal is achieved. If the optimization operation has achieved the expected goal, store the configuration information of the optimization operation in the non-volatile memory of the device, and define a unified configuration file format and read-write interface. When the device starts up or restores to the factory settings, automatically load and apply the configuration of the optimization operation, design a version management and upgrade mechanism for the configuration file to match the iterative update of the optimization strategy, and take data verification and exception handling measures to ensure the integrity and security of the configuration information. After the performance optimization operation is solidified as the default configuration of the device, dynamically adjust the performance optimization strategy according to the real-time evaluation feedback, including adjusting the threshold and weight of the optimization operation, or generating new optimization operations, forming a long-term performance optimization strategy, and adding the performance optimization cases to the knowledge base to guide subsequent optimization decisions. On the optimized device, the user satisfaction score has increased by 20%, and the game revenue has increased by 15%.

[0050] As described above, the above embodiments are only used to illustrate the technical solutions of the present application and are not intended to limit them. Although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for processing game operation data based on player device parameters, characterized in that The method includes: Obtaining the device performance parameters of the player's device through the API, constructing a device performance model, predicting the performance of the device based on the device performance parameters of the player's device, and obtaining the external device parameters of the device under different application scenarios; Constructing a device group classification model according to the device performance parameters of the player's device, dividing the devices into different device groups, obtaining the cooling efficiency index and battery life data of the devices within the device group through the device performance model, and generating corresponding data processing strategies; Real-time monitoring the real-time external device parameters of the target device when running the game, quantitatively evaluating the real-time external device parameters of the target device, controlling the reduction range of the rendering quality through different levels, and adjusting the running parameters of the game in real time; Determining the deviation of the average external device parameters of the device group where the target device is located according to the real-time external device parameters of the target device, dynamically adjusting the data processing strategy parameters based on the magnitude and direction of the deviation, and optimizing the data processing strategy according to the user satisfaction scores of different device groups; Based on the adjusted data processing strategy, evaluating the performance index of the data processing strategy on the target device, obtaining the parameter combination that maximizes the comprehensive performance index, and using it as the optimization parameter of the data processing strategy to optimize the data processing strategy and the device group classification model; Determining the common performance problems corresponding to the device group according to the real-time external device parameters of different devices in the same device group, adjusting the generation parameters of the data processing strategy based on the common performance problems, and continuously monitoring the performance index of the device group to evaluate the effect of the data processing strategy adjustment; Determining the correlation degree of the cooling efficiency index and the battery life data according to the real-time external device parameters of the devices in different device groups, and adjusting the device performance model according to the strength of the correlation degree; Generating a performance optimization operation instruction corresponding to the device group according to the performance index and optimization requirements of the device group, sending it to each device in the device group, and adjusting the performance optimization operation based on the optimization effect of the performance optimization operation; Among them, the obtaining the device performance parameters of the player's device through the API, constructing a device performance model, predicting the performance of the device based on the device performance parameters of the player's device, and obtaining the external device parameters of the device under different application scenarios includes: Obtain the device performance parameters of the player's device through the API, including internal device parameters and external device parameters. Among them, the internal device parameters include the processor model, number of cores, main frequency, and memory size, and the external device parameters include the image quality index, cooling efficiency index, and battery life data. The image quality index includes frame rate, rendering resolution, texture quality, image clarity, and color saturation; standardize the obtained device performance parameters and vectorize the device performance parameters; construct a device performance model using the recurrent neural network algorithm based on the vectorized device performance parameters, and establish a mapping relationship between the internal device parameters and the external device parameters; based on the obtained device performance parameters of the player's device, train and optimize the device performance model, and adjust the feature weights of the device performance model parameters in a cyclic iterative manner; according to the internal parameter information of the new device, use the trained device performance model to predict the performance parameters of the device, and obtain the external device parameters of the device in different application scenarios; compare the prediction results with the actual test results to obtain the prediction error and confidence level of the device performance model, and continuously monitor and improve the device performance model; use the prediction results of the device performance model for device selection, optimization, and improvement, including optimizing the heat dissipation design of the device according to the predicted heat dissipation efficiency index and selecting the battery capacity according to the predicted battery life data; Among them, determining the deviation of the average device external parameters of the device group where the target device is located based on the real-time device external parameters of the target device, dynamically adjusting the data processing strategy parameters based on the magnitude and direction of the deviation, and optimizing the data processing strategy according to the user satisfaction scores of different device groups, including: Monitor the real-time external parameters of the target device, including frame rate, rendering resolution, texture quality, image clarity, and color saturation, normalize each external parameter index, calculate the Euclidean distance between the normalized index vectors, and obtain the deviation between the external parameters of the target device and the group average external parameters; if the deviation exceeds the preset deviation threshold, dynamically adjust the data processing strategy parameters according to the magnitude and direction of the deviation, and the data processing strategy parameters include texture quality and rendering resolution; continuously monitor the performance indicators of the target device after the data processing strategy parameters are adjusted, and dynamically update the calculation result of the external parameter deviation through the sliding window method; according to the performance indicators of the target device and user feedback, use the Q-learning algorithm, with the device performance indicators and user satisfaction as the reward signals, to automatically learn and optimize the data processing strategy parameters; if the adjusted deviation still exceeds the preset deviation threshold, continue to adjust the data processing strategy; if the number of times of adjusting the data processing strategy exceeds the preset number threshold and the external parameter deviation of the target device still exceeds the preset deviation threshold, re-group the target device and match the device group with the highest similarity; obtain user experience feedback data through the built-in feedback interface of the game, including user ratings and comments, preprocess the user experience feedback data, including removing noise, word segmentation, and labeling sentiment tags; fine-tune the BERT model on the labeled sentiment data, use the fine-tuned BERT model to perform sentiment classification on the new user comment data, and determine the sentiment tendency of each comment; calculate the user satisfaction scores of different device groups according to the sentiment classification results; optimize the data processing strategy according to the device performance indicators and user experience feedback, and feedback the optimization result of the data processing strategy to the strategy adjustment module; dynamically adjust the preset deviation threshold using the EWMA algorithm according to the historical device external parameter data and user satisfaction; Among them, based on the adjusted data processing strategy, evaluate the performance indicators of the data processing strategy on the target device, obtain the parameter combination that maximizes the comprehensive performance indicators, and use it as the optimization parameter of the data processing strategy. Optimize the data processing strategy and the device group classification model, including: According to the adjusted data processing strategy, extract the key parameters in the data processing strategy to determine the data processing strategy parameter vector; the key parameters include the rendering quality level and the frame rate limit threshold; according to the strategy parameter vector, use the K-means clustering algorithm for model training to group the data processing strategy parameter vector; for each data processing strategy parameter group, including using the Gaussian process as the prior distribution, using expected improvement as the acquisition function, and updating the hyperparameters of the Gaussian process by the stochastic gradient descent method, evaluate the performance metrics of the data processing strategy on the target device to obtain the parameter combination that maximizes the comprehensive performance metrics as the optimized parameters of the data processing strategy, where the comprehensive performance metrics include device performance parameters, game revenue, and user retention rate; according to the key parameters of the data processing strategy, use the recurrent neural algorithm to construct a strategy generation model to determine the optimized data processing strategy; use the optimized data processing strategy parameters as training data to retrain the strategy generation model, and by modifying the loss function of the strategy generation model and adding a parameter offset penalty term, obtain the optimized strategy generation model; according to the performance metrics of the target device after adjusting the data processing strategy, evaluate the similarity between the target device and the original device group, and by calculating the Mahalanobis distance between the device extrinsic parameter vector and the group center vector, determine whether the target device matches the current group; if the similarity between the target device and the original group is lower than the preset similarity threshold, then update the parameters and decision boundaries of the device group classification model through incremental learning, fine-tune some parameters of the device group classification model by the gradient descent method, and re-classify and predict the target device to obtain its new device group; use the device performance parameters of the new device as the new goal of data processing strategy optimization to form a two-way feedback mechanism between data processing strategy optimization and device classification; by establishing a feedback channel between data processing strategy optimization and game operation, feedback the impact of the data processing strategy on business metrics to the strategy optimization module to achieve a closed-loop of business-driven data processing strategy optimization; by continuously monitoring the execution effect of the optimized data processing strategy on the target device, set the evaluation period and evaluation index system, regularly evaluate the data for improving the device performance and user experience by the data processing strategy, and use the evaluation results for the iteration of data processing strategy optimization and device group classification; Among them, determining the common performance problems corresponding to the device group according to the real-time device extrinsic parameters of different devices in the same device group, adjusting the generation parameters of the data processing strategy based on the common performance problems, and continuously monitoring the performance metrics of the device group to evaluate the effect of the data processing strategy adjustment, including: Sample at a fixed frequency through device sensors to obtain the device operation data of each device in the device group in real time during operation, and form device operation time series data. The device operation data includes CPU temperature, GPU temperature, battery power, and power consumption. Preprocess the device operation time series data, including performing outlier detection using the interquartile range method, setting the outlier threshold to Q1 - 1.5IQR and Q3 + 1.5IQR, identifying and removing outliers; performing data smoothing using exponential weighted moving average to eliminate short-term fluctuations; setting a time window of a fixed size and sliding it every certain period of time; for each parameter of the device operation data, calculate the statistics of different devices within the same device group on each time window, including mean, variance, and peak value, and identify the common performance problems within the group by comparing the statistical differences between different devices; calculate the rate of battery power decline of the device. If the power decline rate is greater than the preset speed threshold within the preset time, determine the device as a device with battery life problems; construct a feature vector based on the device operation time series data, including mean, standard deviation, kurtosis, and wavelet transform coefficients, train an isolation forest model, set an outlier threshold, and mark and alarm the devices that exceed the outlier threshold; classify and attribute the identified performance problems to determine whether the problem is a hardware problem or a software problem; if it is determined to be a hardware problem, solve it by optimizing the hardware design or selecting high-end components; if it is determined to be a software problem, relieve the performance bottleneck of the hardware by reducing the rendering load and lowering the model complexity; map the identified common performance problems within the group to the generation parameters of the data processing strategy, and determine the parameter items to be adjusted and the adjustment range according to the problem type and severity; perform parameter adjustment of the data processing strategy according to different performance problems, including for high-temperature problems, giving priority to adjusting the rendering resolution and frame rate limit, for battery life problems, giving priority to adjusting the screen brightness and data transmission frequency, and setting several levels according to the problem severity; input the adjusted data processing generation parameters into the optimized strategy generation model to obtain the optimized data processing strategy; after the optimized data processing strategy is deployed, continuously monitor the performance indicators of the device group, and evaluate the effect of the data processing strategy adjustment by comparing the performance indicators before and after optimization; set quantitative evaluation indicators, and set different evaluation targets for different game types and device groups; in the AB test, select a preset number of devices as the experimental group and adopt the optimized data processing strategy, and the remaining devices as the control group and adopt the old data processing strategy, and obtain the changes in the key indicators of the two groups of devices within the preset time period. The key indicators include device performance parameters, game revenue, and user retention rate, and verify the effectiveness and security of the optimized data processing strategy; promote the data processing strategy with good adjustment effect to other device groups, attribute the identified data processing strategy with poor adjustment effect, and iteratively optimize the data processing strategy generation parameters until the optimal data processing strategy configuration is determined.

2. The method according to claim 1, wherein Construct an equipment group classification model based on the equipment performance parameters of the player's device, divide the devices into different equipment groups, obtain the cooling efficiency index and battery life data of the devices within the equipment group through the equipment performance model, and generate corresponding data processing strategies, including: Based on the equipment performance parameters of the player's device, use the K-means clustering algorithm for model training to construct an equipment group classification model, and divide the devices into different equipment groups, including high similarity, medium similarity, and low similarity; for each equipment group, predict the cooling efficiency index and battery life data of the devices within the group through the trained equipment performance model to obtain the overall performance index of the group; according to the cooling efficiency and battery life of different equipment groups, formulate corresponding data processing strategies, including for equipment groups with a cooling efficiency lower than the preset cooling efficiency threshold, reducing the graphic quality by reducing the texture resolution and simplifying the lighting model, and for equipment groups with a battery life lower than the preset battery life threshold, limiting the frame rate by setting the maximum FPS threshold and dynamically adjusting the rendering frequency; set the formulated data processing strategies through the API interface provided by the game engine and transfer them to the game engine to dynamically adjust the rendering parameters and physical simulation accuracy; During the game operation, continuously obtain the external parameter data of the device at a fixed time interval through the performance monitoring interface of the operating system; compare the obtained external parameter data with the prediction results of the equipment performance model and adjust the data processing strategy in real time; For newly released devices or unknown devices, obtain the performance parameters of the newly released devices through quick testing, use the equipment group classification model to determine the equipment group to which they belong, and apply the existing data processing strategies; Continuously obtain the performance data of the new device through the performance monitoring interface of the operating system and update and optimize the equipment performance model.

3. The method according to claim 1, wherein The real-time device external parameters during the game operation of the target device are monitored in real time, the real-time device external parameters of the target device are quantitatively evaluated, and the reduction amplitude of the rendering quality is controlled through different levels to adjust the game operation parameters in real time, including: Deploy a performance monitoring module on the target device, and obtain the device external parameter data in real time at a fixed frequency through the interface of the game rendering engine to form a time-series data stream; Preprocess the obtained real-time device external parameter data, including using the moving average method for data denoising and using the Z-score normalization method for data normalization; Design a game performance evaluation index system according to the picture quality index and frame rate stability during game operation, including picture quality and frame rate stability indexes. Among them, the picture quality index includes image clarity index and color saturation index, the frame rate stability index is the frame rate jitter index, the image clarity index includes sharpness and contrast, the color saturation index is the gamut coverage rate, and the frame rate jitter index is the inter-frame delay variance; through the method of weighted average, map each index to a unified performance score, and dynamically adjust the weights of each index according to the characteristics of different games and devices to obtain the real-time performance evaluation result of the target device; according to the real-time performance evaluation result of the target device and the preset performance threshold, judge whether the target device meets the performance requirements for game operation. If the evaluation result is lower than the preset performance threshold, trigger the dynamic adjustment mechanism; by reducing the texture resolution, simplifying the lighting model, and reducing the number of special effects, control the reduction range of rendering quality in different levels and adjust the game operation parameters in real time; during the dynamic adjustment process, continuously monitor the real-time performance changes of the target device, perform weighted average on the data within the window through a sliding window of a fixed size, use the exponentially weighted moving average method to cumulatively average the historical device external parameters, dynamically update the performance evaluation result, and continuously monitor and predict the performance of the target device; according to the real-time device external parameter data of the target device, carefully use the isolation forest algorithm for model training to detect device performance anomalies in real time; according to the real-time device external parameter data of the target device, use the ARIMA algorithm for model training to predict the change trend of device external parameter data and determine the game performance change trend; through the visualization dashboard and anomaly warning, display the performance status and trend of the target device in real time to assist in game performance optimization and problem location; based on the actual performance indicators and prediction results of the target device, continuously optimize the game rendering engine and resource loading mechanism.

4. The method according to claim 1, wherein Determine the correlation degree between the cooling efficiency index and the battery life data according to the real-time device external parameters of the devices in different device groups, and adjust the device performance model according to the strength of the correlation degree, including: Through the temperature sensor and power meter of the device, obtain the cooling efficiency index and battery life data of the devices in different device groups in real time at a fixed frequency to form a time series data set reflecting the device performance; preprocess the obtained cooling efficiency index and battery life data, including outlier detection, data normalization, and time alignment; Calculate the correlation degree between the cooling efficiency index and the endurance data through the Pearson correlation coefficient, generate a correlation coefficient matrix, set a significance threshold for the correlation coefficient, and set the significance level of the p-value. When the absolute value of the correlation coefficient is greater than the preset correlation coefficient significance threshold and the p-value is less than the significance level, it is considered that the correlation between the two indicators is strong; otherwise, the correlation is considered weak. Adjust the structure and parameters of the model according to the strength of the correlation. For indicators with a strong correlation, add corresponding influence paths to the device performance model and scale the path weights according to the magnitude of the correlation degree. For indicators with a weak correlation, the weights of the corresponding influence paths can be reduced. Introduce the correlation degree into the regularization term of the model. When the correlation degree is greater than the preset correlation degree threshold, reduce the regularization strength; when the correlation degree is less than the preset threshold, increase the regularization strength. Based on the adjusted device performance model, use the gradient descent optimization algorithm to update the weight parameters of the model through backpropagation, and adopt the adaptive moment estimation of the Adam optimizer to balance the convergence speed and generalization performance of the model. Apply the adjusted device performance model in different device groups, and evaluate the performance improvement effect of the device performance model by comparing the prediction errors of the device performance model before and after adjustment on the test set or validation set. Record the training error and validation error of the device performance model at different iteration rounds, draw a learning curve, and determine the convergence speed and overfitting degree of the model. Use the visualization tool TensorBoard to monitor the structure and performance changes of the device performance model in real time, and assist in correlation analysis and device performance model tuning.

5. The method according to claim 1, wherein, Generate performance optimization operation instructions corresponding to the device group according to the performance indicators and optimization requirements of the device group, send them to each device in the device group, and adjust the performance optimization operation based on the optimization effect of the performance optimization operation, including: Based on the performance metrics and optimization requirements of the device group, determine the performance optimization operation instructions for different groups, including adjusting the background program running strategy to enhance battery life and adjusting the control logic of the cooling components to improve cooling efficiency, to form a set of optimization operation instructions; encode and parameterize the optimization operation instructions, and convert the instructions into a format directly executable by the device, including ProtocolBuffers and MessagePack; according to the type and priority attributes of the instructions, assign a unique identifier and metadata tag to each instruction, and define the instruction semantics and syntax specifications, including instruction type, parameter list, and execution conditions; according to the network topology structure and communication protocol of the device group, design an instruction distribution strategy, and use multicast, broadcast, and point-to-point methods to send the optimization operation instructions to each device node in the group, and ensure the arrival rate of the instructions through retransmission and confirmation mechanisms; for services that require real-time response, adopt an event-driven instruction distribution method, and for optimization operations that require synchronous execution, adopt an instruction synchronization mechanism based on timestamps or version numbers; the device parses the content and parameters of the received optimization operation instructions, and triggers corresponding performance optimization actions according to the type and priority of the instructions; according to the system type and version information of the device, dynamically load the corresponding API modules and call parameters to decouple the optimization operation instructions from the system environment; If the instruction is for the background program running strategy, call the adapted system API to dynamically adjust the scheduling strategy and resource allocation of the background process, including reducing the priority of non-critical tasks and clearing the memory occupancy of non-critical tasks; if the optimization operation instruction is for the cooling component control logic, obtain the temperature data of the device through sensors, including CPU temperature and battery temperature, and judge the current heat dissipation state of the device according to the trend and amplitude of the temperature change, and the heat dissipation state includes normal, warning, and dangerous; according to the heat dissipation state and the preset temperature threshold, determine the target value of the PID controller, and the target value is the desired temperature range; according to the difference between the current temperature and the target temperature, set the proportional, integral, and differential parameters of the PID controller, calculate the output value of the controller, including the adjustment amount of the fan speed and the liquid cooling pump power, and dynamically adjust the heat dissipation components; Continuously obtain the performance and energy consumption data of the device, and based on the performance evaluation criterion AppStartupTime, obtain the operation behaviors and experience feedback of users in different business scenarios through data embedding and log collection; perform correlation analysis on the device performance and experience feedback, calculate the battery life, heat generation, average operation completion time, page loading delay, crash rate, and stutter rate before and after performance optimization. Through before-and-after comparison and incremental analysis, evaluate the effect of the performance optimization operation, generate an optimization effect evaluation report, and transmit the evaluation result back to the optimization decision-making module; according to the battery life, heat generation, average operation completion time, page loading delay, crash rate, and stutter rate in the optimization effect evaluation report, determine whether the current optimization operation has achieved the expected goal; if the evaluation report shows that the battery life, heat generation, average operation completion time, page loading delay, crash rate, and stutter rate after the device performance optimization operation do not reach the preset optimization threshold, it is considered that the performance optimization operation has not achieved the expected goal, and the performance optimization operation is adjusted again, including reducing the background task priority and redesigning the PID controller parameters; through adjusting instruction parameters and modifying algorithm logic, iteratively optimize the performance indicators of the device until the optimization goal is achieved; If the optimization operation has achieved the expected goal, store the configuration information of the optimization operation in the non-volatile memory of the device, and define a unified configuration file format and read-write interface; when the device is started or restored to factory settings, automatically load and apply the configuration of the optimization operation; design a version management and upgrade mechanism for the configuration file to match the iterative update of the optimization strategy, and take data verification and exception handling measures to ensure the integrity and security of the configuration information; after the performance optimization operation is solidified as the default configuration of the device, dynamically adjust the performance optimization strategy according to the real-time evaluation feedback, including adjusting the thresholds and weights of the optimization operation, or generating new optimization operations, forming a long-term performance optimization strategy, and adding the performance optimization cases to the knowledge base to guide subsequent optimization decisions.

6. A game operation data processing device based on player device parameters, which executes the steps of a game operation data processing method based on player device parameters according to any one of claims 1-5, characterized in that The device includes: A device performance parameter collection module, which is used to obtain the device performance parameters of the player's device, including device internal parameters and device external parameters; A device performance modeling and optimization module, which is used to construct a device performance model and iteratively optimize the model parameters; A device group classification and group data processing strategy generation module, which is used to divide device groups based on the device performance parameters of the player's device and generate data processing strategies for different device groups; A real-time performance monitoring and strategy adjustment module, which is used to monitor the real-time performance of the device when running the game and adjust the data processing strategy according to the performance deviation; A performance problem analysis and strategy optimization module, which is used to analyze the common problems within the device group and adjust the data processing strategy to solve the problems of high temperature and battery life; A performance optimization operation instruction generation module, which is used to generate operation instructions for adjusting background programs and cooling control logic.

Images