User behavior analysis method

Abstract

The invention discloses a user behavior analysis method, which is characterized in that a user behavior sequence is modeled by using an RNN (Recurrent Neural Network) model, and the operation mode and behavior characteristics of a user can be deeply understood, so that the user interaction experience is optimized; the ARIMA model is adopted to model page performance data, so that time sequence characteristics of page performance can be accurately captured, and potential performance bottlenecks and change trends can be identified; an RNN model is introduced into an ARIMA model, page performance data is used as an input of a sequence model and is analyzed together with user behavior data, comprehensive analysis of user behaviors and page performance can be realized, and prediction precision and practicability of the model are improved; the trained model parameters are utilized to predict the optimization scheme and optimization cost of the future time point, performance problems and optimization requirements possibly occurring in the future can be predicted, the optimization scheme is formulated and implemented in advance, and potential performance bottlenecks are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of substations, and in particular to a user behavior analysis method. Background Art

[0002] With the rapid development of the Internet and mobile applications, users have higher and higher requirements for the response speed and interactive experience of websites and applications. Users not only expect the page to load within a few seconds, but also expect to enjoy a smooth and delay-free operation experience during use. This high expectation has brought tremendous pressure to developers, forcing them to constantly look for more efficient performance optimization methods. However, traditional page performance monitoring methods are usually too simple to fully capture the complexity of user behavior and the dynamic changes of page performance, and therefore cannot provide effective decision support and performance optimization solutions.

[0003] Traditional page performance monitoring methods mainly rely on basic statistical data, such as page loading time, resource loading time, etc. These methods usually simply record and analyze basic performance indicators such as page loading time, white screen time, resource request time, etc. Although they can provide some basic performance conditions, they lack in-depth analysis of user behavior patterns and page performance time series characteristics. These simple statistical data cannot reflect the real experience of users in different interaction scenarios, nor can they capture the dynamic characteristics of page performance changing over time.

[0004] In addition, traditional methods often ignore the diversity and complexity of user behavior. When using a website or application, users may perform a variety of operations, such as clicking buttons, scrolling pages, and entering text, which will affect page performance. However, traditional methods usually only focus on a single performance indicator, and fail to comprehensively consider the overall behavior patterns of users and the comprehensive impact of these operations on performance.

[0005] This limitation makes it difficult for traditional performance monitoring methods to provide accurate and comprehensive performance evaluation and optimization suggestions when facing complex user behaviors and dynamic page performance changes. Developers can often only rely on experience and intuition to optimize performance, lacking scientific data support and precise optimization directions, which ultimately leads to unsatisfactory optimization results and limited improvement in user experience.

[0006] To this end, the present invention proposes a user behavior analysis method. Summary of the invention

[0007] In view of this, an embodiment of the present invention hopes to provide a user behavior analysis method, including a sequence model and a time series model, which is implemented through the following steps:

[0008] S1, data preparation: collect user behavior data and page performance data on the website or application;

[0009] S2, sequence model construction: use RNN model to model user behavior sequence;

[0010] S3, time series model construction: use ARIMA model to model page performance data;

[0011] S4, introduce the RNN model into the ARIMA model, introduce the constructed time series model into the sequence model, use the page performance data as an input of the sequence model, and analyze it together with the user behavior data;

[0012] S5, using the trained model parameters to predict the optimization plan and optimization cost at future time points, and according to the analysis results, using the obtained time domain adjustment factors and factors, formulate a specific optimization plan;

[0013] S6, combined with the new optimization cost and optimization scheme, adjusts the ARIMA model and RNN model to produce more accurate analysis results about performance bottlenecks.

[0014] Optionally, the acquisition logic for collecting the user's behavior data on the website or application in S1 includes:

[0015] Get the start time and end time of page loading;

[0016] Use JavaScript event listeners to capture user clicks, scrolling, input and other behaviors on the page, record the timestamp of each interaction, and calculate the corresponding response delay

[0017] Generates and records a user unique identifier (UUID), which is stored in the user's browser via a cookie or localStorage to ensure that the user can be identified across sessions.

[0018] Optionally, the acquisition logic for collecting the user's behavior data on the website or application in S1 includes:

[0019] Get the loading time of each resource and record the start loading time and end loading time of each resource;

[0020] Get the time of first paint (FP) and first content paint (FCP) and record the time point of first paint and first content paint of the page;

[0021] The page white screen time is obtained by calculating the difference between the time when the browser starts to initiate the HTTP request and the time when the DOM tree is fully constructed;

[0022] Report the collected user behavior data and page performance data to the backend log server via HTTP requests;

[0023] On the server side, create a log library to store reported data and record the original logs of user behavior data and page performance data;

[0024] Calculate page performance indicators within a specific time period, such as the average white screen time and the average resource loading time.

[0025] Optionally, the steps of using the RNN model to model the user behavior sequence are:

[0026] Collect user operation sequence data from user behavior logs and convert user operations into training sets and test sets;

[0027] Use the training set data to train the RNN model, set the appropriate batch size and number of training rounds, and use the test set data to evaluate the performance of the model, and calculate the model's accuracy, precision, recall and other indicators;

[0028] Deploy the trained RNN model to the production environment, obtain user behavior data in real time, and input it into the RNN model for prediction;

[0029] Based on the prediction results, dynamically adjust page content, recommendation system, etc. to optimize the user experience.

[0030] Optionally, the steps to model page performance data using the ARIMA model are:

[0031] Collect key performance indicators from websites or applications and store the data in a database, recording them in chronological order to form time series data;

[0032] Determine the model's differencing order, autoregressive term, and moving average term, and use the determined differencing order, autoregressive term, and moving average term to build an ARIMA model;

[0033] Perform model prediction on the training set and compare it with the actual data to evaluate the model's prediction performance.

[0034] Optionally, introduce the RNN model into the ARIMA model, introduce the constructed time series model into the sequence model, use the page performance data as an input of the sequence model, and analyze it together with the user behavior data as follows:

[0035] The user behavior data is concatenated with the ARIMA residual to form new multidimensional time series data and the multidimensional time series data is generated into subsequences of fixed length, which are suitable for RNN model input;

[0036] Deploy the trained ARIMA-RNN hybrid model to the production environment for real-time prediction of page performance and user behavior;

[0037] Regularly obtain new page performance data and user behavior data, input them into the ARIMA-RNN model for prediction, and predict performance and behavior in the future;

[0038] Regularly update the model, retrain, and optimize it based on new page performance and user behavior data.

[0039] Optionally, use the trained model parameters to predict the optimization plan and optimization cost at future time points. Based on the analysis results, use the obtained time domain adjustment factors and factors to formulate specific optimization plan steps:

[0040] Collect expected page performance data and expected user behavior data at future points in time and data on external factors that are likely to affect page performance and user behavior;

[0041] Input the expected data at future time points into the trained ARIMA-RNN hybrid model to obtain the page performance indicators and user behavior prediction results at future time points;

[0042] Analyze and predict indicators such as page loading time, white screen time, and interaction delay to identify possible performance bottlenecks. Based on the analysis results, set specific optimization goals, including reducing page loading time, reducing white screen time, and increasing interaction speed. Determine optimization priorities based on the impact and urgency of the goals.

[0043] Based on the predicted changes in performance indicators, determine the time period that needs to be adjusted. Based on the predicted changes in performance indicators, determine the time period that needs to be adjusted and related factors.

[0044] Optionally, the steps to adjust the ARIMA model and RNN model in combination with the new optimization cost and optimization scheme to produce more accurate analysis results about performance bottlenecks are:

[0045] Record and collect implemented optimization measures and corresponding optimization costs; integrate optimization plans, optimization costs, page performance data, and user behavior data to form a new multi-dimensional time series data set;

[0046] Integrate new user behavior data with the residual data of the ARIMA model to generate new sequence data;

[0047] Input the new optimization plan and user behavior data into the updated model for prediction, and adjust the time period and related factors that need to be optimized based on the new prediction results.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. By using the RNN model to model the user behavior sequence, we can deeply understand the user's operation mode and behavior characteristics, thereby optimizing the user interaction experience; using the ARIMA model to model the page performance data can accurately capture the time series characteristics of the page performance and identify potential performance bottlenecks and change trends; introducing the RNN model into the ARIMA model, taking the page performance data as an input of the sequence model, and analyzing it together with the user behavior data can achieve a comprehensive analysis of user behavior and page performance, and improve the prediction accuracy and practicality of the model; using the trained model parameters to predict the optimization plan and optimization cost at future time points can foresee possible performance problems and optimization needs in the future, formulate and implement optimization plans in advance, and avoid potential performance bottlenecks; according to the analysis results, use the obtained time domain adjustment factors and factors to formulate specific optimization plans to ensure the accuracy and effectiveness of the optimization measures.

[0050] 2. The present invention adjusts the ARIMA model and the RNN model by combining new optimization costs and optimization schemes to produce more accurate analysis results about performance bottlenecks, which can dynamically adapt to changing user behaviors and page performance, and continuously optimize user experience and system performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or technical descriptions will be briefly introduced below. 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 paying creative work.

[0052] Figure 1 It is a schematic diagram of the present invention. DETAILED DESCRIPTION

[0053] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below;

[0054] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0055] This is accomplished by the following steps:

[0056] S1, data preparation: collect user behavior data and page performance data on the website or application;

[0057] S2, sequence model construction: use RNN model to model user behavior sequence;

[0058] S3, time series model construction: use ARIMA model to model page performance data;

[0059] S4, introduce the RNN model into the ARIMA model, introduce the constructed time series model into the sequence model, use the page performance data as an input of the sequence model, and analyze it together with the user behavior data;

[0060] S5, using the trained model parameters to predict the optimization plan and optimization cost at future time points, and according to the analysis results, using the obtained time domain adjustment factors and factors, formulate a specific optimization plan;

[0061] S6, combined with the new optimization cost and optimization scheme, adjusts the ARIMA model and RNN model to produce more accurate analysis results about performance bottlenecks.

[0062] It should be noted that: through systematic and step-by-step steps, user behavior data and page performance data are fully integrated to achieve accurate prediction and optimization of user experience and page performance; first, the data preparation stage ensures the comprehensiveness and accuracy of basic data, laying a solid foundation for subsequent modeling; by using the RNN model to model user behavior sequences, we can deeply understand the user's operation mode and behavior characteristics, thereby optimizing the user interaction experience; using the ARIMA model to model page performance data can accurately capture the time series characteristics of page performance and identify potential performance bottlenecks and change trends; introducing the RNN model into the ARIMA model, using page performance data as an input to the sequence model together with user behavior data Through analysis, we can achieve a comprehensive analysis of user behavior and page performance, and improve the prediction accuracy and practicality of the model; use the trained model parameters to predict the optimization plan and optimization cost at future time points, foresee possible performance problems and optimization needs in the future, formulate and implement optimization plans in advance, and avoid potential performance bottlenecks; according to the analysis results, use the obtained time domain adjustment factors and factors to formulate specific optimization plans to ensure the accuracy and effectiveness of the optimization measures; finally, in combination with the new optimization cost and optimization plan, adjust the ARIMA model and RNN model to produce more accurate analysis results about performance bottlenecks, dynamically adapt to changing user behavior and page performance, and continuously optimize user experience and system performance.

[0063] In this embodiment, the acquisition logic for collecting the user's behavior data on the website or application in S1 includes:

[0064] Get the start time and end time of page loading;

[0065] Use JavaScript event listeners to capture user clicks, scrolling, input and other behaviors on the page, record the timestamp of each interaction, and calculate the corresponding response delay

[0066] Generates and records a user unique identifier (UUID), which is stored in the user's browser via a cookie or localStorage to ensure that the user can be identified across sessions.

[0067] It should be noted that: by obtaining the start time and end time of page loading, the page loading speed and performance can be accurately measured, providing accurate data support for optimizing page loading time; by using JavaScript event listeners to capture user clicks, scrolling, input and other behaviors on the page and recording the timestamp of each interaction, the user's interactive behavior can be tracked in real time, and the user's operating habits and preferences on the page can be understood in detail, so as to calculate the corresponding response delay, evaluate the interactive performance of the page, and provide a scientific basis for optimizing user experience; generate and record user unique identifiers (UUIDs), and store them in the user's browser through cookies or localStorage to ensure cross-session identification of users, and continuously track the user's behavior and preferences in different sessions to form a complete user behavior portrait, which is helpful to provide personalized services and recommendations and improve user satisfaction; the overall benefit of this method is that through the systematic collection and recording of performance data and user behavior data, not only can the page loading and interactive performance be fully understood, performance bottlenecks and optimization space be identified, but also users can be identified across sessions to achieve long-term tracking and analysis of user behavior, and provide more accurate and personalized services, thereby significantly improving user experience and satisfaction, helping enterprises maintain technological leadership and user loyalty in the fiercely competitive market, and improving operational efficiency and business success rate.

[0068] In this embodiment, the acquisition logic for collecting the user's behavior data on the website or application in S1 includes:

[0069] Get the loading time of each resource and record the start loading time and end loading time of each resource;

[0070] Get the time of first paint (FP) and first content paint (FCP) and record the time point of first paint and first content paint of the page;

[0071] The page white screen time is obtained by calculating the difference between the time when the browser starts to initiate the HTTP request and the time when the DOM tree is fully constructed;

[0072] Report the collected user behavior data and page performance data to the backend log server via HTTP requests;

[0073] On the server side, create a log library to store reported data and record the original logs of user behavior data and page performance data;

[0074] Calculate page performance indicators within a specific time period, such as the average white screen time and the average resource loading time.

[0075] It should be noted that: by obtaining the loading time of each resource and recording the start loading time and end loading time of each resource, the loading efficiency and bottlenecks of each resource on the page can be fully tracked and analyzed, providing accurate data support for subsequent performance optimization; obtaining the time of the first paint (FP) and the first content paint (FCP) and recording the time point of the first painting of the page and the time point of the first content painting can specifically measure the moment when the page is initially displayed and the content is visible, which helps to improve the user's perceived loading speed; by calculating the difference between the time when the browser starts to initiate an HTTP request and the time when the DOM tree is fully constructed, the page blank time can be obtained, which can accurately evaluate the time when the user sees a blank screen during the page loading process, further optimizing the user experience; reporting the collected user behavior data and page performance data to the backend log server through HTTP requests can ensure the timeliness of the data. The system can ensure the integrity of the page performance and support real-time monitoring and analysis. On the server side, a log library is created to store reported data, record the original logs of user behavior data and page performance data, and systematically manage and save a large amount of data, supporting multi-dimensional analysis and mining. The system can calculate page performance indicators within a specific time period, such as the average scrolling time and the average resource loading time, which can fully reflect the performance of the page in different time periods and help identify and solve performance problems. The overall benefit of this method is that through systematic data collection, refined records and scientific data storage, page performance can be fully and accurately tracked and analyzed, thereby effectively identifying performance bottlenecks and optimization space, providing scientific decision-making basis, and ultimately achieving continuous optimization of page performance and significant improvement of user experience, helping enterprises maintain their technological leadership and user satisfaction advantages in the fierce market competition, and improving operational efficiency and business success rate.

[0076] In this embodiment, the steps of using the RNN model to model the user behavior sequence are:

[0077] Collect user operation sequence data from user behavior logs and convert user operations into training sets and test sets;

[0078] Use the training set data to train the RNN model, set the appropriate batch size and number of training rounds, and use the test set data to evaluate the performance of the model, and calculate the model's accuracy, precision, recall and other indicators;

[0079] Deploy the trained RNN model to the production environment, obtain user behavior data in real time, and input it into the RNN model for prediction;

[0080] Based on the prediction results, dynamically adjust page content, recommendation system, etc. to optimize the user experience.

[0081] It should be noted that: by systematically collecting user operation sequence data from user behavior logs and converting these data into training sets and test sets, the diversity and complexity of user behavior can be fully covered, ensuring the representativeness and accuracy of model training and test data; using training set data to train the RNN model and setting appropriate batch size and number of training rounds can ensure that the model fully understands the temporal characteristics and laws of user behavior during the learning process, improving the learning efficiency and generalization ability of the model; in the process of using test set data to evaluate model performance, calculating indicators such as model accuracy, precision, and recall rate can comprehensively and objectively evaluate the model's prediction ability and performance, timely discover and improve the model's deficiencies, and ensure the effectiveness and reliability of the model in practical applications; deploying the trained RNN model to the production environment can achieve real-time monitoring and prediction of user behavior. By obtaining user behavior data in real time and inputting it into the RNN model for prediction, it can dynamically track and analyze user operation behavior and timely capture changes in user needs and preferences; according to the prediction results, dynamically adjust page content, recommendation system, etc. to optimize user experience, and can achieve personalized and intelligent user services, improve user satisfaction and stickiness, and enhance user experience.

[0082] In this embodiment, the steps of using the ARIMA model to model page performance data are as follows:

[0083] Collect key performance indicators from websites or applications and store the data in a database, recording them in chronological order to form time series data;

[0084] Determine the model's differencing order, autoregressive term, and moving average term, and use the determined differencing order, autoregressive term, and moving average term to build an ARIMA model;

[0085] Perform model prediction on the training set and compare it with the actual data to evaluate the model's prediction performance.

[0086] It should be noted that by systematically collecting key performance indicators from websites or applications and storing them in a database, and recording them in chronological order to form time series data, the integrity, continuity and accuracy of the data can be ensured, providing a solid data foundation for subsequent analysis and modeling; determining the difference order, autoregressive term, and moving average term of the model can accurately capture the trend, seasonality and random fluctuations in the time series according to the characteristics and change patterns of the data, and using these parameters to build an ARIMA model is helpful to effectively analyze and predict the future trend of time series data; performing model prediction on the training set and comparing it with the actual data to evaluate the model's prediction performance can not only verify the accuracy and reliability of the model, but also identify the advantages and disadvantages of the model, providing a reference for further optimization and adjustment of the model;

[0087] The overall benefit of this method is that by systematically collecting and processing performance indicator data, ensuring data quality and integrity, and combining the determination of parameters such as differences, autoregression and moving average terms, an ARIMA model suitable for specific data characteristics is constructed, which can accurately capture the inherent laws in the time series and improve the accuracy and reliability of predictions. Through model prediction and performance evaluation on the training set, model problems and room for improvement can be discovered in a timely manner, and the model's predictive ability can be continuously optimized and improved. This not only helps to improve the performance monitoring and prediction capabilities of websites or applications, but also provides a scientific basis for performance optimization and decision-making, ultimately improving user experience, optimizing resource allocation, reducing operating costs, and enhancing system stability and sustainable development capabilities, helping companies maintain their leading position in technological advantages and operational efficiency in the fierce market competition.

[0088] 6. A user behavior analysis method according to claim 1, characterized in that: introducing the RNN model into the ARIMA model, introducing the constructed time series model into the sequence model, taking the page performance data as an input of the sequence model, and analyzing it together with the user behavior data in the following steps:

[0089] The user behavior data is concatenated with the ARIMA residual to form new multidimensional time series data and the multidimensional time series data is generated into subsequences of fixed length, which are suitable for RNN model input;

[0090] Deploy the trained ARIMA-RNN hybrid model to the production environment for real-time prediction of page performance and user behavior;

[0091] Regularly obtain new page performance data and user behavior data, input them into the ARIMA-RNN model for prediction, and predict performance and behavior in the future;

[0092] Regularly update the model, retrain, and optimize it based on new page performance and user behavior data.

[0093] It should be noted that: by splicing user behavior data with ARIMA residuals to form new multidimensional time series data and generating fixed-length subsequences suitable for RNN model input, the advantages of time series analysis and deep learning can be effectively combined to improve the accuracy of data processing and the prediction ability of the model; deploying the trained ARIMA-RNN hybrid model to the production environment can provide efficient and accurate solutions for real-time prediction of page performance and user behavior, which helps to respond to user needs and market changes immediately; regularly obtaining new page performance data and user behavior data, and inputting these data into the ARIMA-RNN model for prediction, can continuously update the prediction of future performance and behavior, ensuring the real-time and accuracy of the prediction; by regularly updating the new page performance data and user behavior data, retraining and optimizing the model, it can not only keep the model up-to-date and effective, but also continuously improve the performance and prediction accuracy of the model, ensuring that it always has efficient adaptability in the ever-changing actual environment;

[0094] The overall benefit of this approach is that it combines the powerful analytical capabilities of the ARIMA model for time series data with the deep learning capabilities of the RNN model for sequence data, achieving comprehensive, dynamic, and accurate prediction and analysis of page performance and user behavior, thereby effectively identifying potential problems and optimization space, formulating timely and effective optimization measures, continuously improving user experience and system performance, reducing operating costs, and improving the scientific nature and agility of business decisions, ultimately helping companies maintain their technological advantages and user satisfaction in market competition and achieve significant improvements in operational efficiency and resource utilization.

[0095] In this embodiment, the trained model parameters are used to predict the optimization scheme and optimization cost at future time points, and according to the analysis results, the obtained time domain adjustment factors and factors are used to formulate the steps of a specific optimization scheme:

[0096] Collect expected page performance data and expected user behavior data at future points in time and data on external factors that are likely to affect page performance and user behavior;

[0097] Input the expected data at future time points into the trained ARIMA-RNN hybrid model to obtain the page performance indicators and user behavior prediction results at future time points;

[0098] Analyze and predict indicators such as page loading time, white screen time, and interaction delay to identify possible performance bottlenecks. Based on the analysis results, set specific optimization goals, including reducing page loading time, reducing white screen time, and increasing interaction speed. Determine optimization priorities based on the impact and urgency of the goals.

[0099] Based on the predicted changes in performance indicators, determine the time period that needs to be adjusted. Based on the predicted changes in performance indicators, determine the time period that needs to be adjusted and related factors.

[0100] It should be noted that this comprehensive step can systematically integrate and utilize existing model parameters and data prediction results to improve the accuracy and efficiency of page performance optimization at future time points. By collecting expected page performance data, user behavior data, and external factor data that may affect page performance and user behavior at future time points, the comprehensiveness and relevance of the prediction model input data can be ensured, thereby improving the accuracy of model predictions. These expected data can be input into the trained ARIMA-RNN hybrid model to give full play to the advantages of the ARIMA model in time series analysis and the powerful performance of the RNN model in processing sequence data, thereby obtaining page performance indicators and user behavior prediction results at future time points. Based on these prediction results, in-depth analysis of page loading time, white space, and other factors can be performed. It can accurately identify possible performance bottlenecks and provide comprehensive performance insights based on key performance indicators such as screen time and interaction latency. Based on these analysis results, it can set specific optimization goals, including reducing page loading time, reducing white screen time and increasing interaction speed, and reasonably determine the optimization priority based on the impact and urgency of the goals to ensure efficient use of resources and time. At the same time, based on the predicted changes in performance indicators, it can determine the time period and related factors that need to be adjusted, dynamically adapt to changes in performance and behavior, and provide accurate time domain adjustment suggestions, so as to formulate more specific and actionable optimization plans, ultimately helping enterprises improve page performance, optimize user experience, reduce operating costs, achieve continuous improvement and stable operation of system performance, and maintain technical advantages and user satisfaction in market competition.

[0101] In this embodiment, the steps of adjusting the ARIMA model and the RNN model in combination with the new optimization cost and optimization scheme to generate more accurate analysis results about the performance bottleneck are as follows:

[0102] Record and collect implemented optimization measures and corresponding optimization costs; integrate optimization plans, optimization costs, page performance data, and user behavior data to form a new multi-dimensional time series data set;

[0103] Integrate new user behavior data with the residual data of the ARIMA model to generate new sequence data;

[0104] New optimization plans and user behavior data are input into the updated model for prediction, and the time period and related factors that need to be optimized are adjusted according to the new prediction results. Not only can the existing optimization results and related cost information be effectively utilized to improve the accuracy and adaptability of model predictions, but also a comprehensive insight into user behavior and page performance can be achieved through the integration and analysis of multi-dimensional data, thereby identifying and locating more detailed performance bottlenecks, thereby formulating more accurate and effective optimization measures; in addition, by combining the time series analysis capabilities of the ARIMA model with the powerful performance of the RNN model in processing sequence data, this method can dynamically adapt to changes in page performance and user behavior, provide real-time performance optimization suggestions and decision support, and ultimately help enterprises improve user experience in complex network environments, optimize resource allocation, reduce operating costs, and improve the overall performance and stability of the system.

Claims

1. A user behavior analysis method, including a sequence model and a time series model, characterized in that: This is accomplished by the following steps: S1, data preparation: collect user behavior data and page performance data on the website or application; S2, sequence model construction: use RNN model to model user behavior sequence; S3, time series model construction: use ARIMA model to model page performance data; S4, introduce the RNN model into the ARIMA model, introduce the constructed time series model into the sequence model, use the page performance data as an input of the sequence model, and analyze it together with the user behavior data; S5, using the trained model parameters to predict the optimization plan and optimization cost at future time points, and according to the analysis results, using the obtained time domain adjustment factors and factors, formulate a specific optimization plan; S6, combined with the new optimization cost and optimization scheme, adjusts the ARIMA model and RNN model to produce more accurate analysis results about performance bottlenecks.

2. A user behavior analysis method according to claim 1, characterized in that: The acquisition logic of collecting the user's behavior data on the website or application in S1 includes: Get the start time and end time of page loading; Use JavaScript event listeners to capture user clicks, scrolling, input, and other behaviors on the page, record the timestamp of each interaction, and calculate the corresponding response delay Generates and records a user unique identifier (UUID), which is stored in the user's browser via a cookie or localStorage to ensure that the user can be identified across sessions.

3. A user behavior analysis method according to claim 1, characterized in that: The acquisition logic of collecting the user's behavior data on the website or application in S1 includes: Get the loading time of each resource and record the start loading time and end loading time of each resource; Get the time of first paint (FP) and first content paint (FCP) and record the time point of first paint and first content paint of the page; The page white screen time is obtained by calculating the difference between the time when the browser starts to initiate the HTTP request and the time when the DOM tree is fully constructed; Report the collected user behavior data and page performance data to the backend log server via HTTP requests; On the server side, create a log library to store reported data and record the original logs of user behavior data and page performance data; Calculate page performance indicators within a specific time period, such as the average white screen time and the average resource loading time.

4. A user behavior analysis method according to claim 1, characterized in that: The steps to use the RNN model to model user behavior sequences are: Collect user operation sequence data from user behavior logs and convert user operations into training sets and test sets; Use the training set data to train the RNN model, set the appropriate batch size and number of training rounds, and use the test set data to evaluate the performance of the model, and calculate the model's accuracy, precision, recall and other indicators; Deploy the trained RNN model to the production environment, obtain user behavior data in real time, and input it into the RNN model for prediction; Based on the prediction results, dynamically adjust page content, recommendation system, etc. to optimize the user experience.

5. A user behavior analysis method according to claim 1, characterized in that: The steps to model page performance data using the ARIMA model are: Collect key performance indicators from websites or applications and store the data in a database, recording them in chronological order to form time series data; Determine the model's differencing order, autoregressive term, and moving average term, and use the determined differencing order, autoregressive term, and moving average term to build an ARIMA model; Perform model prediction on the training set and compare it with the actual data to evaluate the model's prediction performance.

6. A user behavior analysis method according to claim 1, characterized in that: Introduce the RNN model into the ARIMA model, introduce the constructed time series model into the sequence model, use the page performance data as an input of the sequence model, and analyze it together with the user behavior data as follows: The user behavior data is concatenated with the ARIMA residual to form new multidimensional time series data and the multidimensional time series data is generated into subsequences of fixed length, which are suitable for RNN model input; Deploy the trained ARIMA-RNN hybrid model to the production environment for real-time prediction of page performance and user behavior; Regularly obtain new page performance data and user behavior data, input them into the ARIMA-RNN model for prediction, and predict performance and behavior in the future; Regularly update the model, retrain, and optimize it based on new page performance and user behavior data.

7. A user behavior analysis method according to claim 1, characterized in that: Use the trained model parameters to predict the optimization plan and optimization cost at future time points. According to the analysis results, use the obtained time domain adjustment factors and factors to formulate specific optimization plan steps: Collect expected page performance data and expected user behavior data at future points in time and data on external factors that are likely to affect page performance and user behavior; Input the expected data at future time points into the trained ARIMA-RNN hybrid model to obtain the page performance indicators and user behavior prediction results at future time points; Analyze and predict indicators such as page loading time, white screen time, and interaction delay to identify possible performance bottlenecks. Based on the analysis results, set specific optimization goals, including reducing page loading time, reducing white screen time, and increasing interaction speed. Determine optimization priorities based on the impact and urgency of the goals. Based on the predicted changes in performance indicators, determine the time period that needs to be adjusted. Based on the predicted changes in performance indicators, determine the time period that needs to be adjusted and related factors.

8. A user behavior analysis method according to claim 1, characterized in that: The steps to adjust the ARIMA model and RNN model to produce more accurate analysis results about performance bottlenecks in combination with the new optimization cost and optimization scheme are: Record and collect implemented optimization measures and corresponding optimization costs; integrate optimization plans, optimization costs, page performance data, and user behavior data to form a new multi-dimensional time series data set; Integrate new user behavior data with the residual data of the ARIMA model to generate new sequence data; Input the new optimization plan and user behavior data into the updated model for prediction, and adjust the time period and related factors that need to be optimized based on the new prediction results.

Images