Method, system and equipment for optimizing page layout based on large model and medium

Abstract

The invention provides a method, a system, equipment and a medium for optimizing page layout based on a large model, and belongs to the technical field of webpage design, the method comprises the following steps: creating an intelligent agent for page optimization, and installing server deployment, Web application construction and Web test tools; responding to a page generation demand, calling a Web application construction tool through a large model to create an initial page code with a burying point, and performing modification confirmation according to user opinions; responding to the implementation request, calling a server deployment tool through a large model to deploy simulation and formal environments, and running page codes in the simulation and formal environments in sequence; and in response to an optimization demand, regularly collecting user operation data through burying points, inputting the user operation data into the large model to generate an optimization suggestion, calling a Web application construction tool to modify page codes, and sequentially running in a simulation environment and a formal environment until the demand is met. According to the method, the page optimization efficiency and accuracy are improved, the labor cost is reduced, the user demand is responded quickly, and the user experience and the page performance are improved.

Description

Technical Field

[0001] The present application belongs to the technical field of web design, and specifically relates to a method, system, device and medium for optimizing page layout based on a large model. Background Art

[0002] With the development of the internet, web applications are becoming increasingly widespread. Reasonable page layout design plays a vital role in improving user experience, promoting user interaction, and increasing page conversion rates. However, traditional web design relies on manual experience and follows a traditional process of manual design-development-testing-launch. This traditional web design approach presents the following problems: First, existing web design methods employ a discrete, iterative model. This leads to a long time-consuming and sluggish response from data collection to final deployment. For example, related page optimization methods still require manual A / B testing design, data attribution analysis, and code refactoring, making them unable to adapt to changing user behavior characteristics in real time. Furthermore, each iteration of web design requires data analysis, page design, page development, testing, and online maintenance, placing high demands on human resources. Second, traditional web optimization methods rely on sampling statistics and manual hypothesis verification, which can only identify local interaction issues and cannot fully identify user intent. While relevant patents incorporate machine learning into intelligent optimization systems, these methods are only applied to single-point processes such as click-through rate prediction.

[0003] In summary, traditional web design optimization methods are time-consuming and labor-intensive. Although there are some intelligent page layout optimization methods, most of these methods focus on local steps and lack a global perspective and the ability to proactively plan and connect the entire process, making it difficult to achieve efficient and continuous optimization. Summary of the Invention

[0004] In a first aspect, an embodiment of the present application provides a method for optimizing page layout based on a large model, comprising the following steps: S1. Create an agent for page optimization based on the large model and install server deployment tools, web application building tools, and web testing tools in the agent. S2. The agent responds to the user's page generation request and uses the large model to call the web application construction tool to create the initial page code with embedded points and modify it based on the user's feedback. S3. After the agent responds to the user's request, it calls the server deployment tool through the large model to deploy the page in the simulation environment and the production environment. After the user confirms the page code, it calls the web testing tool to test it in the simulation environment and then runs it in the production environment. S4. When the intelligent agent responds to user optimization needs, it regularly collects user operation data through tracking points, and inputs user operation data and user optimization needs into the large model to generate optimization suggestions. Then, based on the optimization suggestions, it calls the Web application construction tool to modify the page code, and then runs it in the simulation environment and the formal environment in sequence until the user optimization needs are met.

[0005] Furthermore, the specific steps of step S1 are as follows: S11. Create the front-end and back-end of the agent for page optimization, and deploy the large model on the back-end of the agent; S12. Integrate the Kubernetes cluster management API and Docker image builder into a server deployment tool, and install the server deployment tool into the agent backend; S13. Integrate the command line interface of the Vue framework and the CSS-in-JS compiler to generate a web application building tool, and install the web application building tool into the agent; S14. Use the Cypress-based end-to-end test and Lighthouse performance testing tools to generate a web test tool and install the web test tool into the agent.

[0006] Furthermore, the specific steps of step S2 are as follows: S21. The agent responds to user input, and when the large model identifies the user's page generation requirements, it analyzes the page application scenario and page type; S22. The large model calls the web application building tool to generate the initial page code for the corresponding scenario and sets a script to collect user operation records as a tracking point in the page component; S23. The large model calls the web application building tool to compile the initial page code and generate a preview display interface and returns it to the user through the agent; S24. Responding to user opinions through the agent; If the user modifies the page code, proceed to step S25; If the user confirms, proceed to step S3; S25. Modify the page code by calling the Web application construction tool through the large model, and return to step S23.

[0007] Furthermore, the specific steps of step S3 are as follows: S31. The agent responds to user input, and when the large model recognizes the user's request, the server deployment tool is called by the large model to select the target server for deployment in the simulation environment and the formal environment; S32. Generate a page from the page code using the large model and deploy it to a simulation environment for online testing; If the online test fails, return to step S24; If the online test passes, go to step S33; S33. Use the large model to deploy the page to the official environment for operation and release.

[0008] Furthermore, the specific steps of step S4 are as follows: S41. The large model regularly collects user operation data and page operation data after the page is published through tracking points, analyzes the user operation data and page operation data, and generates page optimization suggestions based on user optimization needs; S42. The large model calls the web application building tool based on the page optimization suggestions to modify the page code, generate the page, and deploy it to the simulation environment for online testing. If the online test fails, go to step S43; If the online test passes, proceed to step S34; S43. Modify again and return to step S42; S44. Determine whether all page optimization suggestions have been implemented; If yes, go to step S45; If not, return to step S42; S45. The large model generates a page from the page code and deploys it to the simulation environment for testing. After the simulation test passes, it waits for user confirmation. S46. Response to input by the agent; If the user input is a modification suggestion, go to step S47; If the user input is to confirm going online, go to step S48; S47. The large model calls the web application building tool to modify the page code based on the modification suggestions, generates the page, and deploys it to the simulation environment for testing until the user confirms the online release. S48. The large model deploys the page to the official operating environment for release.

[0009] Furthermore, the user operation data after the page is published in step S41 includes the user click coordinates, and the page operation data includes the exposure time of the page elements, the loading time of the first screen of the page, and whether a conversion event occurs on the page; Analyzing user operation data and page operation data includes the following steps: Count the click rate of page elements based on the user's click coordinates; Calculate user stay time based on page element exposure time; If the user stay time is less than the lower threshold of the visit time, it is judged as a bounce, and the bounce rate of each page is calculated; The page conversion rate is calculated based on the ratio of the number of conversion events on the page to the total number of user visits; Generating page optimization suggestions based on user optimization needs includes the following steps: Construct a multi-objective optimization function for the page design solution based on the click-through rate of page elements, user stay time, page conversion rate, and page bounce rate as a penalty item; With the goal of maximizing the value of the multi-objective optimization function, the page code is modified, and the differentiated code corresponding to each modification plan is used as a page modification suggestion.

[0010] Furthermore, page modification suggestions include the following: The element that needs to be modified; At least one of the page's style, layout, or copy needs to be modified; The target amount of the quantitative indicator expected to be improved.

[0011] Furthermore, before modifying the page code in step S42, the following steps are also included: Evaluate the extent of page code modifications; When the modification exceeds the set threshold, the original page code is retained as branch A; The page code that executes all page modification suggestions is taken as branch B; The specific steps for deploying to the simulation environment for online testing are as follows: Dynamically switch between branch A and branch B versions through CSS variables; Initially, the first proportion of user traffic is allocated to the B branch version; At each first time interval, the user traffic of branch version B is increased by a set amount, and the conversion rate between branch version A and branch version B is calculated; When the conversion rate difference between branch A and branch B exceeds the set threshold and persists for a second period of time, the page code for branch B is fully released. When the error rate of the B branch version is higher than the second ratio and lasts for a third period of time, the page code of the A branch version is switched back.

[0012] In a second aspect, an embodiment of the present application further provides a system for optimizing page layout based on a large model, comprising: The agent creation module is used to create an agent for page optimization based on the large model, and install server deployment tools, web application building tools, and web testing tools in the agent; The initial page code generation module is used to respond to user page generation requirements in the intelligent agent, call the web application construction tool through the large model to create the initial page code with embedded points, and modify and confirm it according to user opinions; The page generation module is used to respond to user requests when the agent is deployed. The server deployment tool is called by the large model to deploy the simulation environment and the formal environment. After the user confirms the page code, it is tested by calling the Web testing tool in the simulation environment before running it in the formal environment. The page optimization module is used to respond to user optimization needs in the intelligent body, regularly collect user operation data through tracking points, and input user operation data and user optimization needs into the large model to generate optimization suggestions. Then, according to the optimization suggestions, the Web application construction tool is called to modify the page code, and then it is run in the simulation environment and the formal environment in sequence until the user optimization needs are met.

[0013] In a third aspect, an embodiment of the present application further provides an electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the program, the steps of the method for optimizing page layout based on a large model as described in the first aspect are implemented.

[0014] In a fourth aspect, an embodiment of the present application further provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for optimizing page layout based on a large model as described in the first aspect are implemented.

[0015] It can be seen from the above technical solutions that this application has the following advantages: The method, system, device and medium for optimizing page layout based on a large model provided in this application realize automation of the entire process from page generation, testing, deployment to optimization, reduce manual intervention and improve work efficiency; by regularly collecting user operation data and generating optimization suggestions, it can quickly respond to user needs and improve user experience; through multi-objective optimization functions and A / B testing solutions based on large models, it ensures that the optimization suggestions are reasonable and effective; reduces the workload of manual analysis, testing and deployment, and reduces labor costs; through grayscale release and A / B testing, it ensures the stability and reliability of the optimization solution, while supporting continuous iteration and evolution. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for the description. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0017] Figure 1 Schematic diagram of the flow of the method for optimizing page layout based on a large model of the present invention.

[0018] Figure 2 It is a schematic diagram of the system for optimizing page layout based on a large model of the present invention. DETAILED DESCRIPTION

[0019] The specific steps of the method for optimizing page layout based on a large model will be described in detail below, and various embodiments of the present disclosure will be described more fully. The present disclosure can have various embodiments, and adjustments and changes can be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but rather that the present disclosure should be understood to cover all adjustments, equivalents and / or alternatives that fall within the spirit and scope of the various embodiments of the present disclosure.

[0020] For example, with the rapid development of the internet, web applications are playing an increasingly important role in our daily lives and work. Reasonable page layout design is crucial for improving user experience, facilitating user interaction, and increasing page conversion rates. However, traditional page design optimization methods rely primarily on manual experience, employing a traditional process of manual design, development, testing, and launch. This model presents numerous problems.

[0021] First, existing web design optimization methods employ a discrete, iterative model, resulting in a long cycle from data collection to final deployment, leading to responsiveness delays. For example, some related page optimization methods still require manual A / B testing design, data attribution analysis, and code refactoring, making them unable to adapt to changes in user behavior in real time. Furthermore, each iteration of web design requires a series of tasks, including data analysis, page design, page development, testing, and online operations and maintenance, placing a high demand on manpower.

[0022] Secondly, traditional web page optimization methods rely primarily on sampling statistics and manual hypothesis verification, which can only identify local interaction issues and fail to fully understand and identify user intent. Although some related patents have introduced intelligent optimization systems and machine learning technologies, these technologies have limited application scope, mostly applied only to single-point links such as click-through rate prediction, lacking the ability to optimize the entire process.

[0023] In summary, traditional web design optimization methods not only consume significant time and labor costs, but also lack a global perspective and proactive planning capabilities, making efficient and continuous optimization difficult to achieve. While some intelligent page layout optimization methods exist, most of these methods focus on specific steps and fail to optimize and automate the entire process.

[0024] To address the above issues, this embodiment provides a method for optimizing page layout based on a large model. By integrating the large model with multiple automation tools through an intelligent agent, the entire process from page generation, deployment to optimization is automated.

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0026] See also Figure 1 FIG. 1 is a flow chart of a method for optimizing page layout based on a large model in a specific embodiment, the method comprising the following steps: S1. Create an agent for page optimization based on the large model and install server deployment tools, web application building tools, and web testing tools in the agent. It should be noted that by integrating a large model and multiple automation tools, the entire process of page optimization is automated; S2. The agent responds to the user's page generation request and uses the large model to call the web application construction tool to create the initial page code with embedded points and modify it based on the user's feedback. It should be noted that the user experience and page generation accuracy are improved through real-time preview and user feedback mechanisms; the initial page code is generated by a large model, which reduces the workload of manual code writing and improves generation efficiency; user operation data is collected through embedded points to provide data support for subsequent optimization; S3. After the agent responds to the user's request, it calls the server deployment tool through the large model to deploy the page in the simulation environment and the production environment. After the user confirms the page code, it calls the web testing tool to test it in the simulation environment and then runs it in the production environment. It should be noted that automated deployment tools can improve the deployment efficiency of simulation and production environments; online testing in simulation environments can ensure the stability and performance of pages in production environments; and rollback in case of test failures can effectively control deployment risks. S4. The intelligent agent responds to user optimization needs by regularly collecting user operation data through tracking points. This data and optimization needs are then input into the large model to generate optimization suggestions. Based on the optimization suggestions, the web application building tool is called to modify the page code. The page code is then run in the simulation environment and the production environment until the user's optimization needs are met. It should be noted that continuous optimization of the page is achieved by regularly collecting user operation data and generating optimization suggestions; the rationality and effectiveness of the optimization suggestions are ensured through multi-objective optimization functions and A / B testing mechanisms; and the traffic proportion of the new version is gradually increased through the grayscale release mechanism to achieve risk control during the optimization process.

[0027] This embodiment not only improves the efficiency and accuracy of page optimization, but also reduces labor costs, can quickly respond to user needs, and continuously improve user experience and page performance.

[0028] Furthermore, as a refinement and extension of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process in this embodiment, another method for optimizing page layout based on a large model is provided, which includes the following steps: S1. Create an agent for page optimization based on the large model and install server deployment tools, web application building tools, and web testing tools in the agent. The specific steps of step S1 are as follows: S11. Create the front-end and back-end of the agent for page optimization, and deploy the large model on the back-end of the agent; For example, a front-end interface is built using Vue.js, through which users can input page optimization requirements, such as "improve the conversion rate of the product page." A large model based on the Transformer architecture, such as GPT-4, is deployed on the client to process these requirements and generate optimization suggestions. S12. Integrate the Kubernetes cluster management API and Docker image builder into a server deployment tool, and install the server deployment tool into the agent backend; Specifically, Kubernetes is used to manage server clusters, and the Docker image builder is used to quickly deploy web applications. For example, when a new web page version needs to be deployed, the Docker image builder generates an image based on the web page code, and Kubernetes is responsible for scheduling and managing the operation of the image in the cluster. S13. Integrate the command line interface of the Vue framework and the CSS-in-JS compiler to generate a web application building tool, and install the web application building tool into the agent; For example, use Vue CLI to create a project structure and combine it with a CSS-in-JS compiler (such as StyledComponents) to dynamically generate page styles. For example, based on the optimization suggestions generated by the large model, the style code of page elements such as button color and font size can be automatically adjusted. S14. Use the Cypress-based end-to-end test and Lighthouse performance testing tools to generate a web test tool and install the web test tool on the agent. For example, Cypress is used to test page functionality, such as form submission and page navigation; Lighthouse is used to test page performance, such as first-screen load time and interactivity. For example, if Lighthouse detects that the first-screen load time of a page is too long, the large model will generate optimization suggestions, such as compressing image size and optimizing CSS code. It should be noted that basic deployment tools, application building tools, and testing tools can be used to improve deployment speed, while Lighthouse performance testing can reduce the first screen loading time; S2. The agent responds to the user's page generation request and uses the large model to call the web application construction tool to create the initial page code with embedded points and modify and confirm it based on the user's feedback. The specific steps of step S2 are as follows: S21. The agent responds to user input, and when the large model identifies the user's page generation requirements, it analyzes the page application scenario and page type; For example, the user inputs "create an e-commerce product details page", and the large model parses the application scenario as "e-commerce" and the page type as "product details page; S22. The large model calls the web application building tool to generate the initial page code for the corresponding scenario and sets a script to collect user operation records as a tracking point in the page component; For example, the generated initial page code includes components such as product images, prices, and purchase buttons, and embedded scripts are set on these components to record user clicks, stays, and other operations; S23. The large model calls the web application building tool to compile the initial page code and generate a preview display interface and returns it to the user through the agent; It should be noted that users can see the generated page preview on the front-end interface of the agent and can intuitively view the page layout and functions; S24. Responding to user opinions through the agent; If the user modifies the page code, proceed to step S25; If the user confirms, proceed to step S3; For example, the user proposes a modification suggestion of "changing the color of the purchase button to red". The agent passes this suggestion to the big model, and the big model calls the web application construction tool to modify the page code and change the color of the purchase button to red. S25. Modify the page code by calling the Web application building tool through the large model, and return to step S23; Specifically, the user directly modifies the page layout in the preview interface, and the agent passes the modified layout information to the large model. The large model generates new page code and recompiles it to generate a preview display interface; It should be noted that the intelligent agent calls the large model and various tools to generate pages, ensuring the accuracy and efficiency of page generation; and through real-time preview and user feedback, improving user experience and page generation accuracy; S3. After the agent responds to the user's request, it calls the server deployment tool through the large model to deploy the simulation environment and the production environment. After the user confirms the page code, it calls the web testing tool to test it in the simulation environment and then runs it in the production environment. The specific steps of step S3 are as follows: S31. The agent responds to user input, and when the large model recognizes the user's request, the server deployment tool is called by the large model to select the target server for deployment in the simulation environment and the formal environment; For example, a user requests to deploy a page to a simulation environment for testing. The agent calls the server deployment tool, creates a namespace for the simulation environment in the Kubernetes cluster, and deploys the page application. S32. Generate a page from the page code using the large model and deploy it to a simulation environment for online testing; If the online test fails, return to step S24; If the online test passes, go to step S33; For example, the web application runs in a simulation environment and performs end-to-end testing using Cypress to check whether the page functions normally; for example, testing whether users can normally add items to the shopping cart; S33. Deploy the page to the official environment through the large model for operation and release; For example, after the test passes, the web application switches traffic using the blue-green deployment strategy, gradually switching user traffic from the old version to the new version to ensure the stability and performance of the new version in the production environment; It should be noted that the deployment process is stable and reliable through deployment in the simulation environment and the online testing in the simulation environment is used to ensure the stability and performance of the page in the formal environment. S4. The agent responds to user optimization needs by regularly collecting user operation data through tracking points. This data and optimization needs are then input into the large model to generate optimization suggestions. Based on the optimization suggestions, the agent then calls the web application building tool to modify the page code. The application is then run in the simulation environment and then in the production environment until the user's optimization needs are met. The specific steps of step S4 are as follows: S41. The large model regularly collects user operation data and page operation data after the page is published through tracking points, analyzes the user operation data and page operation data, and generates page optimization suggestions based on user optimization needs; Specifically, the tracking script records data such as the user's click coordinates on the page and the exposure time of page elements. For example, it records the number of times a user clicks the "Buy" button on a product details page and the duration of their stay. S42. The large model calls the web application building tool based on the page optimization suggestions to modify the page code, generate the page, and deploy it to the simulation environment for online testing. If the online test fails, go to step S43; If the online test passes, proceed to step S34; For example, based on user operation data, the large model generates optimization suggestions, such as "increasing the exposure time of the product recommendation area"; The web application builder modifies the page code based on the suggestion, increases the display area of ​​the product recommendation area, and redeploys it to the simulation environment for testing. S43. Modify again and return to step S42; For example, if the test finds that the modified page recommendation area loads too slowly, the large model will regenerate optimization suggestions, such as "Optimize the recommendation algorithm to reduce loading time," and modify the page code again for testing; S44. Determine whether all page optimization suggestions have been implemented; If yes, go to step S45; If not, return to step S42; For example, the large model generates multiple optimization suggestions, such as "adjust page layout" and "optimize image loading speed." These suggestions are executed and tested one by one until all suggestions are implemented. S45. The large model generates a page from the page code and deploys it to the simulation environment for testing. After the simulation test passes, it waits for user confirmation. Specifically, after all optimization suggestions have been executed and the page runs stably in the simulation environment, the agent notifies the user for confirmation; S46. Response to input by the agent; If the user input is a modification suggestion, go to step S47; If the user input is to confirm going online, go to step S48; Specifically, the user confirms that the page optimization effect is good and enters "Confirm Online"; S47. The large model calls the web application building tool to modify the page code based on the modification suggestions, generates the page, and deploys it to the simulation environment for testing until the user confirms the online release. For example, a user proposes a modification suggestion, "Change the font color of the product recommendation area to blue." The large model calls the Web application building tool to modify the page code and redeploys it to the simulation environment for testing. S48. The large model deploys the page to the official operating environment for release; Specifically, after the user confirms that everything is correct, the page switches traffic using the blue-green deployment strategy and is officially launched. It should be noted that page optimization ensures the rationality of the optimization process; by regularly collecting user operation data and generating optimization suggestions, continuous optimization of the page can be achieved.

[0029] In an embodiment of the present invention, based on step S41 and step S42, a possible embodiment will be given below to illustrate its specific implementation scheme in a non-limiting manner.

[0030] The user operation data after the page is published in step S41 includes the user click coordinates, and the page operation data includes the exposure time of the page elements, the loading time of the first screen of the page, and whether a conversion event occurs on the page; Analyzing user operation data and page operation data includes the following steps: Count the click rate of page elements based on the user's click coordinates; Calculate user stay time based on page element exposure time; If the user stay time is less than the lower threshold of the visit time, it is judged as a bounce, and the bounce rate of each page is calculated; The page conversion rate is calculated based on the ratio of the number of conversion events on the page to the total number of user visits; Generating page optimization suggestions based on user optimization needs includes the following steps: Construct a multi-objective optimization function F for the page design solution based on the click-through rate of page elements, user stay time, page conversion rate, and page bounce rate as a penalty item; F = α × (click-through rate) + β × (user stay time) + γ × (user stay time) - δ × (page bounce rate); Among them, α, β, γ, and δ are pre-set weight coefficients; With the goal of maximizing the value of the multi-objective optimization function, the page code is modified, and the differential code corresponding to each modification plan is used as a page modification suggestion; It should be noted that reasonable optimization suggestions are generated through analysis of user operation data and page operation data; through multi-objective optimization functions, key indicators such as click-through rate, conversion rate, and loading time are balanced to ensure optimization results; Page modification suggestions include the following: The element that needs to be modified; At least one of the page's style, layout, or copy needs to be modified; The target amount of the quantitative indicator expected to be improved; It should be noted that by determining the specific content of the modification suggestions, we can ensure that the optimization suggestions are effective. By using the quantitative indicators of expected improvement, we can provide clear optimization goals and effect evaluation benchmarks. Before modifying the page code in step S42, the following steps are also included: Evaluate the extent of page code modifications; When the modification exceeds the set threshold, the original page code is retained as branch A; The page code that executes all page modification suggestions is taken as branch B; The specific steps for deploying to the simulation environment for online testing are as follows: Dynamically switch between branch A and branch B versions through CSS variables; Initially, a first proportion of user traffic is allocated to the B branch version. For example, the first proportion is 5%. At each first time interval, the user traffic of the B branch version is increased by a set amount, and the conversion rate between the A branch version and the B branch version is calculated. For example, the first time period is 2 hours. When the conversion rate difference between branch A and branch B exceeds a set threshold and persists for a second period of time, the page code for branch B is fully released. For example, the second period of time is 24 hours. If the error rate of the B branch version exceeds the second ratio and lasts for a third period of time, the page code of the A branch version is switched back. For example, the second ratio is 1% and the third period of time is 5 minutes. For example, for an e-commerce product details page, user action data includes the coordinates of when users clicked the "Buy" button and the length of time they stayed in the product image area. Page performance data includes the page's first screen load time of 3 seconds, 100 conversion events (such as successful orders) on the page, and 1,000 total visits. Data Analysis: Click-through rate statistics: Statistics on the click rate of page elements based on the coordinates of user clicks; for example, the click-through rate of the "Buy" button is 10% (100 clicks / 1000 visits); Dwell time calculation: Calculates the user dwell time based on the exposure time of page elements; for example, the average dwell time of users in the product image area is 15 seconds; Bounce rate statistics: Users whose stay time is less than the lower threshold of the visit duration (e.g. 5 seconds) are considered to have bounced. The bounce rate of each page is 20%; Conversion rate calculation: The page conversion rate is calculated based on the ratio of the number of conversion events on the page to the total number of user visits, which is 10% (100 conversions / 1000 visits); Optimization suggestion generation: Multi-objective optimization function: Based on the above data, a multi-objective optimization function F is constructed: F = α × (click-through rate) + β × (duration) + γ × (conversion rate) − δ × (bounce rate) Taking the weight coefficients as: α=0.3, β=0.2, γ=0.4, δ=0.1 as an example, the value of the optimization function is: F=0.3×10%+0.2×15+0.4×10%−0.1×20%=4.2 Optimization suggestions: Generate page modification suggestions based on the value of the optimization function, such as "increase the size of the product image area to increase dwell time" and "optimize the purchase button copy to increase click-through rate". Suggested page modifications: Example: Suggested changes include: Elements that need to be modified: product image area, purchase button; Modifications: Increased the size of the product image area from 300px to 400px; changed the purchase button text from "Buy" to "Buy Now"; Expected improvement in quantitative indicators: expected dwell time increased to 20 seconds, click-through rate increased to 15%; Page code modification assessment: For example, if the extent of page code modification is evaluated and it is found that the modified page code is significantly different from the original code (e.g., more than 30% of the code has been modified), the original page code is retained as branch A and the modified page code is used as branch B. A / B testing and phased release: Dynamically switch between A and B branches: Use CSS variables to dynamically switch between A and B branch versions; for example, use CSS variables to control the size of the product image area and the copy of the purchase button; Traffic distribution: Initially, 5% of user traffic is allocated to the B branch version. For example, out of 1,000 visits, 50 visits are to the B branch version. Traffic adjustment and conversion rate calculation: Every two hours, increase the user traffic of branch B by a set amount (e.g., 5% each time) and calculate the conversion rate of branch A and branch B. For example, after 24 hours, the conversion rate of branch B is 15% higher than that of branch A, meeting the set threshold (e.g., the conversion rate difference is greater than 10%). Full release or rollback: If the conversion rate difference of branch B exceeds the set threshold (10%) and lasts for 24 hours, the page code of branch B will be fully released; If the error rate of branch B exceeds 1% and persists for 5 minutes, the page code for branch A will be switched back. For example, if the error rate of branch B reaches 2% within a certain period of time during testing, branch A will be switched back immediately.

[0031] It should be noted that A / B testing and grayscale release are used to ensure the stability and reliability of the optimization plan; traffic distribution and error rate monitoring are used to control the risks in the optimization process.

[0032] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0033] like Figure 2 As shown, the following is an embodiment of the system for optimizing page layout based on a large model provided by an embodiment of the present disclosure. The system and the method for optimizing page layout based on a large model in the above-mentioned embodiments belong to the same inventive concept. For details not fully described in the embodiment of the system for optimizing page layout based on a large model, please refer to the embodiment of the method for optimizing page layout based on a large model above.

[0034] The system includes: The agent creation module is used to create an agent for page optimization based on the large model, and install server deployment tools, web application building tools, and web testing tools in the agent; The initial page code generation module is used to respond to user page generation requirements in the intelligent agent, call the web application construction tool through the large model to create the initial page code with embedded points, and modify and confirm it according to user opinions; The page generation module is used to respond to user requests when the agent is deployed. The server deployment tool is called by the large model to deploy the simulation environment and the formal environment. After the user confirms the page code, it is tested by calling the Web testing tool in the simulation environment before running it in the formal environment. The page optimization module is used to respond to user optimization needs in the intelligent body, regularly collect user operation data through tracking points, and input user operation data and user optimization needs into the large model to generate optimization suggestions. Then, according to the optimization suggestions, the Web application construction tool is called to modify the page code, and then it is run in the simulation environment and the formal environment in sequence until the user optimization needs are met.

[0035] This embodiment improves the efficiency and accuracy of page optimization, reduces labor costs, responds to user needs quickly, and improves user experience and page performance through the interactive collaboration of the intelligent body creation module, initial page code generation module, page generation module, and page optimization module.

[0036] The method for optimizing page layout based on a large model provided in the embodiment of the present application can be applied to electronic devices. Those skilled in the art will understand that the electronic device structure involved in the embodiment of the present invention does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown, or combine certain components, or arrange components differently. In the embodiment of the present invention, the electronic device includes but is not limited to a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described and / or required herein.

[0037] The electronic device may include a processor, an external memory interface, an internal memory, a universal serial bus (USB) interface, a charging management module, a power management module, a battery, a wireless communication module, an audio module, a speaker, a microphone, a sensor module, a button, a camera, a display, and a SIM card interface, etc.

[0038] It is understood that the structures illustrated in the embodiments of the present application do not constitute specific limitations on the electronic device. In other embodiments of the present application, the electronic device may include more or fewer components than shown, or combine or separate certain components, or arrange the components differently. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0039] A processor may include one or more processing units, such as a central processing unit (CPU), an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, a memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural-network processing unit (NPU). Different processing units may be independent devices or integrated into one or more processors.

[0040] The processor can be the nerve center and command center of the electronic device. The controller can generate operation control signals based on the instruction opcode and timing signal to complete the control of instruction fetching and execution.

[0041] The processor may also include a memory for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. This memory can store instructions or data that the processor has just used or is reusing. If the processor needs to use the instruction or data again, it can directly call it from the memory. This avoids repeated accesses, reduces processor latency, and thus improves system efficiency.

[0042] The above-mentioned electronic device implements the method of optimizing page layout based on a large model of the present application, which creates an intelligent agent for page optimization based on a large model, and installs a server deployment tool, a Web application construction tool and a Web testing tool in the intelligent agent; when the intelligent agent responds to the user's page generation needs, it calls the Web application construction tool through the large model to create an initial page code with embedded points and modifies and confirms it according to the user's opinions; when the intelligent agent responds to the user's implementation request, it calls the server deployment tool through the large model to deploy the simulation environment and the formal environment, and after the user confirms, the page code is first tested by calling the Web testing tool in the simulation environment, and then runs it in the formal environment; when the intelligent agent responds to the user's optimization needs, it regularly collects user operation data through embedded points, and inputs the user operation data and user optimization needs into the large model to generate optimization suggestions, and then calls the Web application construction tool according to the optimization suggestions to modify the page code, and then runs it in the simulation environment and the formal environment in turn until the user's optimization needs are met. The technical solution achieves the integration of large models and multiple automation tools by the intelligent agent to realize the full process automation from page generation, deployment to optimization, which not only improves the efficiency and accuracy of page optimization, but also reduces labor costs, can quickly respond to user needs, and continuously improve user experience and page performance.

[0043] The storage medium provided in the present application stores a program product that can implement a method for optimizing page layout based on a large model.

[0044] The method for optimizing page layout based on a large model includes: creating an intelligent agent for page optimization based on a large model, installing a server deployment tool, a Web application building tool, and a Web testing tool in the intelligent agent; when the intelligent agent responds to user page generation needs, the Web application building tool is called through the large model to create an initial page code with embedded points and is modified and confirmed according to user opinions; when the intelligent agent responds to user implementation requests, the server deployment tool is called through the large model to deploy the simulation environment and the formal environment, and after user confirmation, the page code is first tested by calling the Web testing tool in the simulation environment, and then run in the formal environment; when the intelligent agent responds to user optimization needs, user operation data is regularly collected through embedded points, and the user operation data and user optimization needs are input into the large model to generate optimization suggestions, and then the Web application building tool is called according to the optimization suggestions to modify the page code, and then it is run in the simulation environment and the formal environment in sequence until the user optimization needs are met.

[0045] In some possible implementations, the method for optimizing page layout based on a large model disclosed herein can be implemented in the form of a program product, which includes program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of various exemplary implementations of the present disclosure described in the above "Exemplary Method" section of this specification.

[0046] The storage medium of the present disclosure can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, a system, device or component of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0047] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for optimizing page layout based on a large model, characterized in that: The steps include: S1. Create an agent for page optimization based on the large model and install server deployment tools, web application building tools, and web testing tools in the agent. S2. The agent responds to the user's page generation request and uses the large model to call the web application construction tool to create the initial page code with embedded points and modify it based on the user's feedback. S3. After the agent responds to the user's request, it calls the server deployment tool through the large model to deploy the page in the simulation environment and the production environment. After the user confirms the page code, it calls the web testing tool to test it in the simulation environment and then runs it in the production environment. S4. When the intelligent agent responds to user optimization needs, it regularly collects user operation data through tracking points, and inputs user operation data and user optimization needs into the large model to generate optimization suggestions. Then, based on the optimization suggestions, it calls the Web application construction tool to modify the page code, and then runs it in the simulation environment and the formal environment in sequence until the user optimization needs are met.

2. The method for optimizing page layout based on a large model according to claim 1, characterized in that: The specific steps of step S2 are as follows: S21. The agent responds to user input, and when the large model identifies the user's page generation requirements, it analyzes the page application scenario and page type; S22. The large model calls the web application building tool to generate the initial page code for the corresponding scenario and sets a script to collect user operation records as a tracking point in the page component; S23. The large model calls the web application building tool to compile the initial page code and generate a preview display interface and returns it to the user through the agent; S24. Responding to user opinions through the agent; If the user modifies the page code, proceed to step S25; If the user confirms, proceed to step S3; S25. Modify the page code by calling the Web application construction tool through the large model, and return to step S23.

3. The method for optimizing page layout based on a large model according to claim 2, characterized in that: The specific steps of step S3 are as follows: S31. The agent responds to user input, and when the large model recognizes the user's request, the server deployment tool is called by the large model to select the target server for deployment in the simulation environment and the formal environment; S32. Generate a page from the page code using the large model and deploy it to a simulation environment for online testing; If the online test fails, return to step S24; If the online test passes, go to step S33; S33. Use the large model to deploy the page to the official environment for operation and release.

4. The method for optimizing page layout based on a large model according to claim 3, characterized in that: The specific steps of step S4 are as follows: S41. The large model regularly collects user operation data and page operation data after the page is published through tracking points, analyzes the user operation data and page operation data, and generates page optimization suggestions based on user optimization needs; S42. The large model calls the web application building tool based on the page optimization suggestions to modify the page code, generate the page, and deploy it to the simulation environment for online testing. If the online test fails, go to step S43; If the online test passes, proceed to step S34; S43. Modify again and return to step S42; S44. Determine whether all page optimization suggestions have been implemented; If yes, go to step S45; If not, return to step S42; S45. The large model generates a page from the page code and deploys it to the simulation environment for testing. After the simulation test passes, it waits for user confirmation. S46. Response to input by the agent; If the user input is a modification suggestion, go to step S47; If the user input is to confirm going online, go to step S48; S47. The large model calls the web application building tool to modify the page code based on the modification suggestions, generates the page, and deploys it to the simulation environment for testing until the user confirms the online release. S48. The large model deploys the page to the official operating environment for release.

5. The method for optimizing page layout based on a large model according to claim 4, characterized in that: The user operation data after the page is published in step S41 includes the user click coordinates, and the page operation data includes the exposure time of the page elements, the loading time of the first screen of the page, and whether a conversion event occurs on the page; Analyzing user operation data and page operation data includes the following steps: Count the click rate of page elements based on the user's click coordinates; Calculate user stay time based on page element exposure time; If the user stay time is less than the lower threshold of the visit time, it is judged as a bounce, and the bounce rate of each page is calculated; The page conversion rate is calculated based on the ratio of the number of conversion events on the page to the total number of user visits; Generating page optimization suggestions based on user optimization needs includes the following steps: Construct a multi-objective optimization function for the page design solution based on the click-through rate of page elements, user stay time, page conversion rate, and page bounce rate as a penalty item; With the goal of maximizing the value of the multi-objective optimization function, the page code is modified, and the differentiated code corresponding to each modification plan is used as a page modification suggestion.

6. The method for optimizing page layout based on a large model according to claim 5, characterized in that: Page modification suggestions include the following: The element that needs to be modified; At least one of the page's style, layout, or copy needs to be modified; The target amount of the quantitative indicator expected to be improved.

7. The method for optimizing page layout based on a large model according to claim 6, characterized in that: Before modifying the page code in step S42, the following steps are also included: Evaluate the extent of page code modifications; When the modification exceeds the set threshold, the original page code is retained as branch A; The page code that executes all page modification suggestions is taken as branch B; The specific steps for deploying to the simulation environment for online testing are as follows: Dynamically switch between branch A and branch B versions through CSS variables; Initially, the first proportion of user traffic is allocated to the B branch version; At each first time interval, the user traffic of branch version B is increased by a set amount, and the conversion rate between branch version A and branch version B is calculated; When the conversion rate difference between branch A and branch B exceeds the set threshold and persists for a second period of time, the page code for branch B is fully released. When the error rate of the B branch version is higher than the second ratio and lasts for a third period of time, the page code of the A branch version is switched back.

8. A system for optimizing page layout based on a large model, characterized in that: include: The agent creation module is used to create an agent for page optimization based on the large model, and install server deployment tools, web application building tools, and web testing tools in the agent; The initial page code generation module is used to respond to user page generation requirements in the intelligent agent, call the web application construction tool through the large model to create the initial page code with embedded points, and modify and confirm it according to user opinions; The page generation module is used to respond to user requests when the agent is deployed. The server deployment tool is called by the large model to deploy the simulation environment and the formal environment. After the user confirms the page code, it is tested by calling the Web testing tool in the simulation environment before running it in the formal environment. The page optimization module is used to respond to user optimization needs in the intelligent body, regularly collect user operation data through tracking points, and input user operation data and user optimization needs into the large model to generate optimization suggestions. Then, according to the optimization suggestions, the Web application construction tool is called to modify the page code, and then it is run in the simulation environment and the formal environment in sequence until the user optimization needs are met.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method implements the steps of the method for optimizing page layout based on a large model as claimed in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for optimizing page layout based on a large model as claimed in any one of claims 1 to 7 are implemented.

Images