Cloud service predictive visualization method and platform based on AI big language model

Abstract

The invention discloses a cloud service predictive visualization method and platform based on an AI large language model. The AI big language model-based cloud service predictive visualization method comprises the following steps of S1, collecting historical data of a cloud rendering platform, and performing data cleaning, preprocessing and feature engineering; s2, establishing a model library containing a plurality of time sequence models, and providing detailed model introduction and use guide; s3, constructing an interactive interface, wherein the interactive interface supports interactive charts, real-time updating and multi-dimensional analysis functions; and S4, according to data features and analysis targets input by the user, through guidance of the AI large language model, recommending an appropriate preset time sequence model, automatically completing model training and parameter tuning, and finally, performing prediction result display. According to the method, the analysis capabilities of various time sequence models and AI big language models are integrated, so that accurate prediction and personalized analysis of cloud service key business indexes are realized.

Description

Technical Field

[0001] The present invention relates to the field of cloud computing technology, and in particular to a method and platform for predictive visualization of cloud services based on an AI large language model. Background Art

[0002] With the rapid development of cloud computing technology, cloud service platforms have accumulated a huge amount of data. How to effectively use this data for business analysis and prediction has become a common focus of cloud service providers and users. As a typical cloud service, the prediction and analysis of business indicators of cloud rendering service is crucial to optimize resource allocation, improve service quality, and enhance user experience.

[0003] Traditional cloud rendering service visualization platforms mainly rely on historical data and simple linear regression models for trend prediction, which has the following shortcomings:

[0004] Limited predictive power: Linear regression models have difficulty capturing complex data patterns and trends, resulting in low predictive accuracy.

[0005] Lack of personalized customization: Users cannot choose appropriate models and adjust parameters according to their own needs. The data dashboard display content and analysis angles are fixed, making it difficult to meet personalized needs.

[0006] Insufficient interactivity: The data dashboard displays static content and lacks real-time updates and interactive exploration functions, making it difficult for users to deeply understand the patterns behind the data.

[0007] At present, some cloud service platforms have begun to try to introduce technologies such as machine learning to predict business indicators, such as:

[0008] Amazon CloudWatch: Provides machine learning-based anomaly detection and prediction capabilities to help users identify potential problems and optimize resource utilization.

[0009] Google Cloud Monitoring: Supports user-defined metrics and alert rules, and provides predictive analysis based on historical data.

[0010] Microsoft Azure Monitor: Provides machine learning-based log analysis and anomaly detection capabilities to help users quickly identify and resolve problems.

[0011] However, most of these platforms are limited to simple machine learning models and lack a deep understanding of the characteristics of time series data, making it difficult to meet the higher requirements of cloud rendering services for prediction accuracy and personalized customization.

[0012] Disadvantages of existing technology:

[0013] Single prediction model: Most existing platforms use simple linear regression models or basic machine learning models, which are difficult to capture the complex patterns and trends of time series data, resulting in insufficient prediction accuracy.

[0014] Lack of personalized customization: Users cannot choose appropriate models and adjust parameters according to their own needs. The data dashboard display content and analysis angles are fixed, making it difficult to meet personalized needs.

[0015] Insufficient interactivity: The data dashboard displays static content and lacks real-time updates and interactive exploration functions, making it difficult for users to deeply understand the patterns behind the data.

[0016] The potential of AI big language models has not been fully tapped: Existing platforms have not effectively utilized the powerful analysis and code execution capabilities of AI big language models, and are unable to achieve intelligent recommendation and automatic training of models, limiting the platform's ease of use and prediction accuracy.

[0017] In summary, the existing cloud service predictive visualization platform has deficiencies in prediction accuracy, personalized customization, interactivity, etc., and it is difficult to meet the higher demands of cloud rendering services for data analysis and prediction.

[0018] Therefore, the prior art has defects and needs to be improved. Summary of the invention

[0019] The technical problem to be solved by the present invention is to provide a method and platform for predictive visualization of cloud services based on an AI large language model, so as to achieve a more accurate, flexible and easy-to-use data prediction and analysis experience.

[0020] The technical solution of the present invention is as follows: a method for predictive visualization of cloud services based on an AI large language model is provided, comprising the following steps.

[0021] S1: Collect historical data of the cloud rendering platform and perform data cleaning, preprocessing and feature engineering.

[0022] S2: Establish a model library containing various time series models, and provide detailed model introduction and usage guide.

[0023] S3: Build an interactive interface to provide a friendly user experience. The interactive interface supports interactive charts, real-time updates, and multi-dimensional analysis functions.

[0024] S4: Use the tool calling and code execution capabilities of the AI ​​big language model to perform data prediction and analysis; specifically, based on the data features and analysis objectives entered by the user, the AI ​​big language model is guided to recommend a suitable preset time series model, automatically complete model training and parameter tuning, and finally display the prediction results.

[0025] Furthermore, in step S1, the historical data includes: submitted job data, user data, cost data, and rendering time data.

[0026] Furthermore, in step S2, the time series model includes ARIMA, Prophet, and LSTM.

[0027] Further, in step S3, the interactive interface is constructed using the React framework or the Vue framework.

[0028] Furthermore, the prediction result display includes: interactive charts, multi-dimensional analysis, and prediction result interpretation.

[0029] Furthermore, the present invention also provides a cloud service predictive visualization platform based on an AI large language model, comprising: a data layer, a model layer, a service layer and a user interface layer.

[0030] The data layer is used to collect, clean, and pre-process the time series data of the cloud rendering platform, and perform feature engineering to provide high-quality data input for model training.

[0031] The model layer uses the tool calling and code execution capabilities of the AI ​​big language model. According to the data characteristics and analysis objectives input by the user, the AI ​​big language model analyzes the data, recommends a suitable preset time series model, and automatically completes model training and parameter tuning.

[0032] The service layer obtains data through the API interface, then calls the model, processes the data through the model, predicts the results, and visualizes the predicted results.

[0033] The user interface layer is used to provide functions such as interactive charts, real-time updates, and multi-dimensional analysis, so that users can intuitively observe data changes and conduct interactive exploration.

[0034] Furthermore, the data collected by the data layer include: submitted job data, submitted job user data, job rendering cost data, total job rendering time data, applied machine data, total machine billing time data, failed job data, frame task data, project data, etc.; the data cleaning and preprocessing performed by the data layer include: missing value processing, outlier detection, data standardization, etc.; the feature engineering performed by the data layer is to extract valuable features according to business needs, such as user type, job type, machine configuration, etc.

[0035] Furthermore, the model layer presets time series models, including ARIMA, Prophet, and LSTM; the model training process and results performed by the model layer support visualization. ARIMA: Autoregressive Integrated Moving Average Model, LSTM: Long Short-Term Memory.

[0036] Furthermore, interactive charts support multiple chart types such as line charts, bar charts, scatter charts, etc. Users can customize chart styles and data dimensions; real-time updates are used to update forecast result charts based on user input and operations to show data change trends; multi-dimensional analysis supports users to analyze data from multiple dimensions, such as time, users, projects, machines, etc., and provides data filtering, sorting, grouping and other functions.

[0037] Furthermore, the service layer provides explanations of the model prediction results, such as which features have a greater impact on the prediction results, to help users understand the prediction logic of the time series model.

[0038] Compared with the prior art, the present invention has the following characteristics:

[0039] Use AI big language models to guide model selection and training: Through the analysis capabilities of AI big language models, suitable models are recommended according to user needs, and model training and parameter tuning are automatically completed, which lowers the usage threshold and improves prediction accuracy.

[0040] Provide personalized customization function: users can choose appropriate models, adjust parameters, and customize the display content and analysis angle of the data dashboard according to their own needs to obtain analysis results that better meet their needs.

[0041] Realize interactive display of prediction results: The prediction result chart is updated in real time based on user input and operation, and users can conduct interactive exploration and deeply understand the rules behind the data.

[0042] The present invention aims to solve the problems of insufficient prediction accuracy, weak personalized customization capability, and poor interactivity of existing cloud service predictive visualization platforms, and provide a cloud service predictive visualization platform based on an AI large language model to achieve a more accurate, flexible, and easy-to-use data prediction and analysis experience, which can truly reflect the future trends of key business indicators in cloud rendering services and meet users' personalized data analysis needs.

[0043] The present invention improves prediction accuracy: it overcomes the limitation of the single prediction model of the existing platform, introduces multiple time series models and combines the analysis capabilities of the AI ​​large language model to more accurately predict future data trends.

[0044] The present invention enhances personalized customization: it enables users to select appropriate models and adjust parameters according to their own needs, and customize the display content and analysis angle of the data dashboard to obtain analysis results that better meet their needs.

[0045] The present invention improves interactivity: by updating the prediction result chart in real time and providing multi-dimensional analysis tools, users can intuitively observe data changes, conduct interactive exploration, and deeply understand the laws behind the data.

[0046] The present invention fully utilizes the potential of the AI ​​big language model: utilizing the tool calling and code execution capabilities of the AI ​​big language model to achieve intelligent recommendation and automatic training of the model, thereby improving the platform's ease of use and prediction accuracy.

[0047] The present invention realizes a powerful, easy-to-use cloud service predictive visualization method and platform that can meet personalized needs, helping users to better understand and predict business trends, optimize resource allocation, and improve service quality.

[0048] By adopting the above scheme, the present invention provides a method and platform for predictive visualization of cloud services based on AI large language model, which has the following advantages:

[0049] 1. Higher prediction accuracy: By introducing the analytical capabilities of multiple time series models and AI large language models, data patterns and trends can be captured more accurately, significantly improving prediction accuracy. For example, in the prediction of job data submitted to the cloud rendering service, the prediction accuracy of the present invention is improved by more than 20% compared with the traditional linear regression model.

[0050] 2. Stronger customization capabilities: Users can choose appropriate models, adjust parameters, and customize the display content and analysis angle of the data dashboard according to their own needs to obtain analysis results that better meet their needs, meeting the personalized needs of different users for data analysis.

[0051] 3. Better interactivity: The prediction result chart is updated in real time based on user input and operation. Users can conduct interactive exploration, observe data changes more intuitively, and understand the rules behind the data more deeply.

[0052] 4. Greater ease of use: The guidance function of the AI ​​large language model simplifies the model selection and training process, lowers the user threshold, and even users without professional knowledge can easily get started.

[0053] Specifically, the technical effects brought about by the present invention include:

[0054] 1. More accurately predict future trends of key business indicators of cloud rendering services, such as the number of submitted jobs, rendering costs, rendering time, etc., to help users better plan resources, optimize costs, and improve service quality.

[0055] 2. Provide more flexible and personalized data analysis tools to meet the personalized needs of different users for data analysis. For example, operations personnel can focus on user behavior analysis, financial personnel can focus on cost-benefit analysis, and managers can focus on overall business trend analysis.

[0056] 3. Improve user participation and interest in data analysis. Through interactive charts and visual analysis tools, users can understand data more intuitively and participate more actively in data analysis, thereby tapping into data value and improving decision-making efficiency.

[0057] In summary, the present invention realizes accurate prediction and personalized analysis of cloud service data by combining AI large language models and multiple time series models, providing users with more accurate, flexible and easy-to-use data analysis tools, which has significant technical advantages and application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic diagram of the process of data prediction and analysis of the present invention. DETAILED DESCRIPTION

[0059] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0060] The present invention provides a method for predictive visualization of cloud services based on an AI large language model, comprising the following steps.

[0061] S1: Collect historical data of the cloud rendering platform, perform data cleaning, preprocessing and feature engineering. The historical data includes: submitted job data, user data, cost data, and rendering time data.

[0062] S2: Establish a model library containing multiple time series models and provide detailed model introduction and usage guide. The time series models include ARIMA, Prophet, and LSTM.

[0063] S3: Build an interactive interface to provide a friendly user experience. The interactive interface supports interactive charts, real-time updates, and multi-dimensional analysis functions. Use the React framework or Vue framework to build the interactive interface.

[0064] S4: Use the tool calling and code execution capabilities of the AI ​​big language model to perform data prediction and analysis; specifically, based on the data features and analysis objectives entered by the user, the AI ​​big language model is guided to recommend a suitable preset time series model, automatically complete model training and parameter tuning, and finally display the prediction results.

[0065] The forecast result display includes: interactive charts, multi-dimensional analysis, and forecast result interpretation.

[0066] The present invention also provides a cloud service predictive visualization platform based on an AI large language model, comprising: a data layer, a model layer, a service layer and a user interface layer.

[0067] The data layer is used to collect, clean, and pre-process the time series data of the cloud rendering platform, and perform feature engineering to provide high-quality data input for model training.

[0068] The model layer uses the tool calling and code execution capabilities of the AI ​​big language model. According to the data characteristics and analysis objectives input by the user, the AI ​​big language model analyzes the data, recommends a suitable preset time series model, and automatically completes model training and parameter tuning.

[0069] The service layer obtains data through the API interface, then calls the model, processes the data through the model, predicts the results, and visualizes the predicted results.

[0070] The user interface layer is used to provide functions such as interactive charts, real-time updates, and multi-dimensional analysis, so that users can intuitively observe data changes and conduct interactive exploration.

[0071] The data collected by the data layer include: submitted job data, submitted job user data, job rendering cost data, total job rendering time data, applied machine data, total machine billing time data, failed job data, frame task data, project data, etc.; the data cleaning and preprocessing performed by the data layer include: missing value processing, outlier detection, data standardization, etc.; the feature engineering performed by the data layer is to extract valuable features according to business needs, such as user type, job type, machine configuration, etc.

[0072] The model layer presets time series models, including ARIMA, Prophet, and LSTM; the model training process and results performed by the model layer support visualization.

[0073] Interactive charts support multiple chart types such as line charts, bar charts, scatter charts, etc. Users can customize chart styles and data dimensions; real-time updates are used to update forecast result charts in real time based on user input and operations to show data change trends; multi-dimensional analysis supports users to analyze data from multiple dimensions, such as time, users, projects, machines, etc., and provides data filtering, sorting, grouping and other functions.

[0074] The service layer provides explanations of the model prediction results, such as which features have a greater impact on the prediction results, to help users understand the prediction logic of the time series model.

[0075] The following is a specific embodiment:

[0076] See also Figure 1 , assuming that the user wants to predict the number of jobs submitted to the cloud rendering platform in the next week, the specific steps are as follows:

[0077] User input: Users enter data features on the platform interface, such as historical job submission data, user data, holiday information, etc., and select the analysis target as "predict the number of jobs submitted in the next week."

[0078] AI model guidance: The AI ​​large language model analyzes the data features of user input and recommends using the Prophet model for prediction. This is because the Prophet model is good at processing time series data with seasonality and emergencies and is suitable for predicting the number of submitted jobs affected by holidays.

[0079] Model training and parameter tuning: The platform automatically uses historical data to train the Prophet model and perform parameter tuning to improve forecasting accuracy.

[0080] Forecast result display: The platform displays the forecast results for the next week in the form of an interactive line chart, allowing users to intuitively observe the forecast trend.

[0081] Multi-dimensional analysis: Users can select different time granularities (such as days or hours) to view forecast results, or perform data analysis by dimensions such as user type and job type.

[0082] Interpretation of prediction results: The platform provides explanations of the model’s prediction results, such as which features have a greater impact on the prediction results, to help users understand the model’s prediction logic.

[0083] The above embodiment can accurately predict the number of jobs submitted in the next week, helping users plan resources in advance to avoid insufficient or wasted resources. The above embodiment can provide a multi-dimensional analysis tool to help users gain an in-depth understanding of the changing trend of the number of submitted jobs, such as which user groups and which types of jobs have a greater impact on the number of submitted jobs. The above embodiment provides an interpretation of the prediction results to help users understand the prediction logic of the model and enhance the user's trust in the prediction results.

[0084] In summary, the present invention provides a method and platform for predictive visualization of cloud services based on an AI large language model, which has the following advantages:

[0085] 1. Higher prediction accuracy: By introducing the analytical capabilities of multiple time series models and AI large language models, data patterns and trends can be captured more accurately, significantly improving prediction accuracy. For example, in the prediction of job data submitted to the cloud rendering service, the prediction accuracy of the present invention is improved by more than 20% compared with the traditional linear regression model.

[0086] 2. Stronger customization capabilities: Users can choose appropriate models, adjust parameters, and customize the display content and analysis angle of the data dashboard according to their own needs to obtain analysis results that better meet their needs, meeting the personalized needs of different users for data analysis.

[0087] 3. Better interactivity: The prediction result chart is updated in real time based on user input and operation. Users can conduct interactive exploration, observe data changes more intuitively, and understand the rules behind the data more deeply.

[0088] 4. Greater ease of use: The guidance function of the AI ​​large language model simplifies the model selection and training process, lowers the user threshold, and even users without professional knowledge can easily get started.

[0089] Specifically, the technical effects brought about by the present invention include:

[0090] 1. More accurately predict future trends of key business indicators of cloud rendering services, such as the number of submitted jobs, rendering costs, rendering time, etc., to help users better plan resources, optimize costs, and improve service quality.

[0091] 2. Provide more flexible and personalized data analysis tools to meet the personalized needs of different users for data analysis. For example, operations personnel can focus on user behavior analysis, financial personnel can focus on cost-benefit analysis, and managers can focus on overall business trend analysis.

[0092] 3. Improve user participation and interest in data analysis. Through interactive charts and visual analysis tools, users can understand data more intuitively and participate more actively in data analysis, thereby tapping into data value and improving decision-making efficiency.

[0093] In summary, the present invention realizes accurate prediction and personalized analysis of cloud service data by combining AI large language models and multiple time series models, providing users with more accurate, flexible and easy-to-use data analysis tools, which has significant technical advantages and application value.

[0094] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predictive visualization of cloud services based on AI large language model, characterized in that: The following steps are involved: S1: Collect historical data of the cloud rendering platform and perform data cleaning, preprocessing and feature engineering; S2: Establish a model library containing various time series models and provide detailed model introduction and usage guide; S3: Build an interactive interface that supports interactive charts, real-time updates, and multi-dimensional analysis functions; S4: Use the tool calling and code execution capabilities of the AI ​​big language model to perform data prediction and analysis; specifically, based on the data features and analysis objectives entered by the user, the AI ​​big language model is guided to recommend a suitable preset time series model, automatically complete model training and parameter tuning, and finally display the prediction results.

2. The method for predictive visualization of cloud services based on an AI large language model according to claim 1, characterized in that: In step S1, the historical data includes: submitted job data, user data, cost data, and rendering time data.

3. The method for predictive visualization of cloud services based on an AI large language model according to claim 1, characterized in that: In step S2, the time series model includes ARIMA, Prophet, and LSTM.

4. The method for predictive visualization of cloud services based on an AI large language model according to claim 1, characterized in that: In step S3, the interactive interface is constructed using the React framework or the Vue framework.

5. The method for predictive visualization of cloud services based on an AI large language model according to claim 1, characterized in that: The forecast result display includes: interactive charts, multi-dimensional analysis, and forecast result interpretation.

6. A cloud service predictive visualization platform based on AI large language model, characterized in that: include: Data layer, model layer, service layer, and user interface layer; The data layer is used to collect, clean, and pre-process the time series data of the cloud rendering platform, and perform feature engineering to provide high-quality data input for model training; The model layer uses the tool calling and code execution capabilities of the AI ​​big language model. According to the data features and analysis objectives input by the user, the AI ​​big language model analyzes the data, recommends a suitable preset time series model, and automatically completes model training and parameter tuning. The service layer obtains data through the API interface, then calls the model, processes the data through the model, predicts the results, and visualizes the predicted results; The user interface layer is used to provide interactive charts, real-time updates, and multi-dimensional analysis functions, allowing users to intuitively observe data changes and conduct interactive exploration.

7. The cloud service predictive visualization platform based on AI large language model according to claim 6, characterized in that: The data collected by the data layer includes: submitted job data, submitted job user data, job rendering fee data, total job rendering time data, application machine data, total machine billing time data, failed job data, frame task data, and project data; Data cleaning and preprocessing performed at the data layer include: missing value processing, outlier detection, and data standardization; The feature engineering performed at the data layer is to extract valuable features based on business needs.

8. The cloud service predictive visualization platform based on AI large language model according to claim 6, characterized in that: The model layer presets time series models, including ARIMA, Prophet, and LSTM; the model training process and results performed by the model layer support visualization.

9. The cloud service predictive visualization platform based on AI large language model according to claim 6, characterized in that: Interactive charts support line charts, bar charts, and scatter chart types. Users can customize chart styles and data dimensions. Real-time updates are used to update forecast result charts based on user input and operations, showing data change trends. Multi-dimensional analysis supports users to analyze data from multiple dimensions, and provides data filtering, sorting, and grouping functions.

10. The cloud service predictive visualization platform based on AI large language model according to claim 6, characterized in that: The service layer provides explanations of the model prediction results to help users understand the prediction logic of the time series model.

Images