User big data management method and system based on artificial intelligence and visualization

Abstract

The present invention relates to the field of intelligent medical technology and provides a method and system for user big data management based on artificial intelligence and visualization. The present invention has the beneficial effects of improving the efficiency and accuracy of user interface design, enhancing the consistency and personalization of the user experience, and increasing the responsiveness and user satisfaction of the user interface during visual interaction. These beneficial effects collectively constitute the significant advantages and innovative value of the present invention in the field of user interface design.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to application number CN202410215894.X, filed on February 27, 2024, entitled “User big data management method and system based on artificial intelligence and visualization,” the contents of which are incorporated herein by reference in their entirety. Technical Field

[0003] The present invention relates to the field of smart medical technology, and in particular to a user big data management method and system based on artificial intelligence and visualization. Background Art

[0004] In the design of medical information systems and patient interaction interfaces, effectively understanding and meeting patients' needs has always been a major challenge. With the development of technology, especially the application of artificial intelligence and machine learning in user interface design, more and more research is focusing on how to improve user experience through data analysis.

[0005] Traditional medical information system interface design is often based on fixed templates and processes, lacking consideration for individual patient needs. This can lead to patients experiencing complex operations and difficulty finding or understanding information, impacting the quality and efficiency of medical services.

[0006] In recent years, although some studies have begun to focus on patient user behavior analysis and demand identification, these methods mostly rely on simple user behavior statistics or questionnaires, and are unable to deeply explore patients' actual needs and potential problems when using medical information systems. Summary of the Invention

[0007] In order to improve the above problems, the present invention provides a user big data management method and system based on artificial intelligence and visualization.

[0008] A first aspect of an embodiment of the present invention provides a user big data management method based on artificial intelligence and visualization, which is applied to a smart medical user big data management system. The method includes:

[0009] Obtaining, based on the target patient user tag and the visual interactive interface jump node, a visual feature relationship network of past interface modules, wherein the visual feature relationship network of past interface modules includes content semantic description knowledge of a plurality of visual interfaces, each piece of content semantic description knowledge including interface module switching data between a corresponding visual interface and the remaining visual interfaces in the plurality of visual interfaces, wherein the plurality of visual interfaces includes a target visual interface corresponding to the target patient user tag;

[0010] Combining the associated visual feature relationship network and the previous interface module visual feature relationship network to mine the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces, to obtain an interface switching behavior description, and performing behavior requirement identification on the interface switching behavior description to obtain an interface requirement identification vector;

[0011] Performing interface display jump feature analysis based on the interface requirement identification vector to obtain an analysis result;

[0012] According to the interface module visual features of the target visual interface under the visual interaction interface jump node in the analysis result, the interface module switching decision features corresponding to the target patient user label under the visual interaction interface jump node are determined.

[0013] Preferably, the method of mining the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces by combining the associated visual feature relationship network and the past interface module visual feature relationship network to obtain an interface switching behavior description, and performing behavior requirement identification on the interface switching behavior description to obtain an interface requirement identification vector includes:

[0014] Through the associated visual feature relationship network of the current residual branch, based on the original past interface module visual feature relationship network and the interface requirement recognition vector generated by the previous residual branch, the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces is mined to obtain an interface switching behavior description, and the behavior requirement recognition is performed on the interface switching behavior description to obtain an interface requirement recognition vector, wherein, if the current residual branch is the first residual branch, the interface requirement recognition vector generated by the previous residual branch is a zero vector;

[0015] The next residual branch of the current residual branch is used as the new current residual branch, and the branch is jumped to the associated visual feature relationship network through the current residual branch. Based on the original past interface module visual feature relationship network and the interface requirement identification vector generated by the previous residual branch, the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces is mined to obtain an interface switching behavior description, and the behavior requirement is identified on the interface switching behavior description to obtain an interface requirement identification vector, until the last residual branch outputs the interface requirement identification vector.

[0016] Preferably, the residual branch includes a first behavior demand identification layer, a second behavior demand identification layer, a third behavior demand identification layer and a behavior demand splicing layer;

[0017] The method comprises mining the switching behavior vectors between the visual features of the interface modules of each visual interface and the remaining visual interfaces through the associated visual feature network of the current residual branch, based on the original visual feature network of the past interface modules and the interface requirement recognition vector generated by the previous residual branch, obtaining an interface switching behavior description, and performing behavior requirement recognition on the interface switching behavior description to obtain an interface requirement recognition vector, including:

[0018] By using the associated visual feature network of the first behavior requirement recognition layer in the current residual branch, based on the original past interface module visual feature network and the interface requirement recognition vector generated by the previous residual branch, the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces is mined to obtain a first interface switching behavior description, and the first behavior requirement recognition processing is performed on the first interface switching behavior description to obtain a first interface requirement recognition vector;

[0019] Through the associated visual feature relationship network of the second behavior requirement recognition layer in the current residual branch, based on the original past interface module visual feature relationship network and the interface requirement recognition vector generated by the previous residual branch, the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces is mined to obtain a second interface switching behavior description, and the second behavior requirement recognition processing is performed on the second interface switching behavior description to obtain a second interface requirement recognition vector;

[0020] Performing vector multiplication on the second interface requirement identification vector and the interface requirement identification vector generated by the previous residual branch through the third behavior requirement identification layer in the current residual branch to obtain a first linkage requirement feature;

[0021] By using the associated visual feature relationship network of the third behavior requirement identification layer, mining the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces based on the original past interface module visual feature relationship network and the first linkage requirement feature, obtaining a third interface switching behavior description, and performing third behavior requirement identification processing on the third interface switching behavior description to obtain a third interface requirement identification vector;

[0022] Through the behavioral requirement splicing layer, determine the second linkage requirement feature of the first interface requirement identification vector and the interface requirement identification vector generated by the previous residual branch, as well as the third linkage requirement feature of the target feature relationship network corresponding to the first interface requirement identification vector and the third interface requirement identification vector, determine the sum of the second linkage requirement feature and the third linkage requirement feature, and obtain the interface requirement identification vector generated by the current residual branch, wherein the target feature relationship network is obtained by subtracting the reference feature relationship network from the first interface requirement identification vector.

[0023] Preferably, the number of the visual feature relationship networks of the past interface modules is at least two, and the visual interfaces corresponding to the visual feature relationship networks of each past interface module are the same, but the interface analysis node ranges are different;

[0024] The combined associated visual feature relationship network and the past interface module visual feature relationship network mine the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces, obtain the interface switching behavior description, perform behavioral requirement identification on the interface switching behavior description, and obtain the interface requirement identification vector, including: mining the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces based on the associated visual feature relationship network of each past interface module visual feature relationship network, obtain the interface switching behavior description corresponding to each past interface module visual feature relationship network, perform behavioral requirement identification on each interface switching behavior description, and obtain the interface requirement identification vector corresponding to each past interface module visual feature relationship network.

[0025] Preferably, performing interface display jump feature analysis based on the interface requirement identification vector to obtain an analysis result includes:

[0026] Weighting the interface requirement recognition vectors corresponding to the visual feature relationship networks of the past interface modules to obtain an interface requirement recognition weighted vector;

[0027] The interface display jump feature is analyzed based on the weighted vector identified by the interface requirement to obtain an analysis result.

[0028] Preferably, performing interface display jump feature analysis based on the interface requirement identification vector to obtain an analysis result includes: performing interface display jump feature analysis based on the interface requirement identification vector to obtain an analysis result for each visual interface in the multiple visual interfaces;

[0029] The method of determining the interface module switching decision feature corresponding to the target patient user tag under the visual interaction interface jump node based on the interface module visual features of the target visual interface under the visual interaction interface jump node in the analysis result includes: obtaining the interface module visual features of the target visual interface under the visual interaction interface jump node based on the analysis result of each visual interface in the multiple visual interfaces, and determining the interface module switching decision feature corresponding to the target patient user tag under the visual interaction interface jump node.

[0030] Preferably, the visual feature relationship network of past interface modules is obtained according to the target patient user tag and the visual interactive interface jump node, and the visual feature relationship network of past interface modules includes content semantic description knowledge of multiple visual interfaces, including:

[0031] Determine, based on a visual interaction interface jump node of the visual features of the interface module to be predicted, a range of at least two interface analysis nodes before the visual interaction interface jump node;

[0032] The analysis node range of each interface is decomposed into multiple operation node sets based on the pre-set node step size;

[0033] According to the target patient user label of the visual feature of the interface module to be predicted, determining the target visual interface corresponding to the target patient user label and the sample visual interface corresponding to the target visual interface as the visual interface to be processed;

[0034] For each visual interface to be processed, obtaining interface module switching data of the visual interface to be processed and the remaining visual interfaces to be processed in each operation node set;

[0035] For each interface analysis node range, the interface module switching data of each visual interface to be processed and the remaining visual interfaces to be processed are concentrated according to the operation nodes of the interface analysis node range, and the content semantic description knowledge of each visual interface in the interface analysis node range is obtained;

[0036] According to the content semantic description knowledge corresponding to each interface analysis node range, the visual feature relationship network of the past interface modules corresponding to each interface analysis node range is obtained.

[0037] Preferably, performing interface display jump feature analysis based on the interface requirement identification vector to obtain an analysis result includes:

[0038] Performing knowledge transformation on the interface requirement identification vector to obtain interface requirement transformation knowledge;

[0039] An interface display jump feature analysis is performed based on the interface requirement transformation knowledge to obtain an analysis result.

[0040] A second aspect of an embodiment of the present invention provides a smart medical user big data management system, comprising: a processor and a memory and a bus connected to the processor; the processor and the memory communicate with each other through the bus; the processor is used to call the computer program in the memory to execute the above-mentioned user big data management method based on artificial intelligence and visualization.

[0041] A third aspect of an embodiment of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the above-mentioned user big data management method based on artificial intelligence and visualization.

[0042] The AI-based and visualization-based user big data management method and system provided by the present invention efficiently captures a network of visual feature relationships across past interface modules by combining target patient user tags with visual interactive interface jump nodes. This network encompasses the semantic descriptions of the content across multiple visual interfaces and the data on the interface module transitions between them. This innovative approach not only enriches the data foundation for user interface design but also improves the efficiency of switching between interfaces and the consistency of the user experience.

[0043] Furthermore, by combining the network of associated visual features with the network of visual features of previous interface modules, this paper further explores the visual feature switching behavior vectors between each visual interface and the remaining visual interfaces. This vectorized description makes interface switching behavior more precise and quantifiable, providing a solid foundation for subsequent behavioral requirement identification.

[0044] By identifying behavioral needs based on the description of interface switching behavior, the present invention can accurately obtain an interface need identification vector. This vector not only reflects the actual needs of users, but also provides strong support for personalized and intelligent interface design.

[0045] Furthermore, the present invention analyzes interface display jump characteristics based on the interface requirement identification vector, thereby obtaining targeted analysis results. This result provides clear guidance for the visual characteristics of the interface modules under the jump node of the visual interactive interface, making the interface design more in line with user expectations and usage habits.

[0046] Ultimately, by determining the interface module switching decision features corresponding to the target patient's user tag, the present invention achieves precise decision-making at the jump node of the visual interactive interface. This decision not only improves the responsiveness and accuracy of the user interface, but also greatly enhances user satisfaction and convenience during use.

[0047] In summary, the present invention achieves significant benefits by improving the efficiency and accuracy of user interface design, enhancing the consistency and personalization of the user experience, and increasing the responsiveness and user satisfaction of the user interface during visual interaction. These benefits collectively constitute the significant advantages and innovative value of the present invention in the field of user interface design. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 This is a flowchart of a user big data management method based on artificial intelligence and visualization provided by an embodiment of the present invention.

[0050] Figure 2 A schematic diagram of the product modules of a smart medical user big data management system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The following describes exemplary embodiments of the present disclosure in more detail with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0052] In order to better understand the above technical solution, the technical solution of the present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0053] See also Figure 1 , which is a flowchart of a user big data management method based on artificial intelligence and visualization provided by an embodiment of the present invention. The method is applied to a smart medical user big data management system. The specific content description of the method includes S110-S140.

[0054] S110. The smart medical user big data management system obtains a visual feature relationship network of past interface modules based on the target patient user tag and the visual interactive interface jump node. The visual feature relationship network of past interface modules includes content semantic description knowledge of multiple visual interfaces. Each content semantic description knowledge includes interface module switching data between the corresponding visual interface and the remaining visual interfaces in the multiple visual interfaces. The multiple visual interfaces include the target visual interface corresponding to the target patient user tag.

[0055] In S110 , target patient user tags are used to describe and categorize specific attributes or characteristics of the patient user. These tags can be defined based on factors such as the patient's age, gender, health status, disease type, and treatment preferences. For an elderly patient with type 2 diabetes, user tags might include "over 65 years old," "male," "type 2 diabetes," and "on oral medication."

[0056] Visual interface transition nodes: In a visual interface (such as a health management system app or website), a user navigates from one interface to another at a specific intersection or decision point. In a health management app, a user might navigate from the "Home" page to the "Health Data" page, and then from "Health Data" to the "Exercise History" page. Here, "Home," "Health Data," and "Exercise History" can all be considered transition nodes.

[0057] The visual feature relationship network of past interface modules: This network records and analyzes the visual features and relationships between different interface modules when users have used visual interactive interfaces in the past. By analyzing historical user data, the system discovered that whenever users view the "blood sugar data" interface, they often view the "dietary suggestions" interface. This sequence and pattern constitutes part of the visual feature relationship network.

[0058] Visual interface: An interface that allows users to interact with the app visually, typically consisting of graphics, text, images, and controls. The interface of a health management app on a smartphone is an example of a visual interface, where users interact with the app by touching buttons, sliders, and other on-screen features.

[0059] Content semantics: This describes and explains the interface's content, elements, and layout, often used to understand and explain the interface's functionality and purpose. For a "health data" interface, this might include the interface title, labels and explanations for each data field, and data visualization methods (e.g., charts or tables).

[0060] Interface module switching data: This records the user's switching from one module to another in the visual interactive interface, including the frequency, duration, and path of the switch. System records show that users switched from the "Health Consultation" module to the "Medication Reminder" module 30 times in the past week, with an average stay of 2 minutes each time.

[0061] Target visual interface: In a visual interactive system, an interface that is associated with a user's specific label or need and requires special attention or optimization. For elderly patients with type 2 diabetes, the target visual interface may be the "blood glucose monitoring" interface, as this interface is directly related to their primary health need (i.e., blood glucose management).

[0062] S120. The smart medical user big data management system combines the associated visual feature relationship network and the past interface module visual feature relationship network to mine the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces, obtains the interface switching behavior description, performs behavioral demand identification on the interface switching behavior description, and obtains the interface demand identification vector.

[0063] In S120, the associated visual feature relationship network is a network structure that describes the visual features (such as layout, color, size, etc.) between different visual elements, interfaces or modules and their mutual relationships. This relationship network helps to understand and predict the user's navigation and behavior in a visual interactive environment. In the associated visual feature relationship network of a health management application, the "Home" button may always be located at the bottom center of the screen, adjacent to the "Settings" and "Help" buttons. This layout relationship forms a stable visual feature pattern, and users can quickly find the required functions based on these features.

[0064] Switching behavior vector: A mathematical representation used to describe a user's behavior when switching from one interface or module to another in a visual interactive interface. This vector can include dimensions such as switching frequency, direction, and duration. A switching behavior vector might indicate that a user switches from the "Health Data" interface to the "Exercise History" interface five times per week, with an average switching duration of three seconds. This vector can help the system understand the user's navigation habits and needs.

[0065] Interface transition behavior description: A detailed description of user transitions within a visual interactive interface, including the transition type, path, and triggering conditions. This description helps understand user behavior patterns and preferences. For example, the interface transition behavior description might indicate that after viewing "blood sugar data," users often switch to the "dietary recommendations" interface to view dietary recommendations. This description can help the system optimize interface layout and navigation flow.

[0066] Behavioral Needs Identification: This system analyzes user behavior data (such as interface transitions, clicks, and input) to identify and understand users' potential needs, expectations, or problems. This identification helps improve the user experience and the system's personalized service capabilities. Through behavioral needs identification, the system discovered that users frequently missed medications in the "Medication Reminder" interface. This suggests that users may need a more prominent reminder function or more flexible reminder settings.

[0067] Interface Requirements Identification Vector: A mathematical representation used to describe user needs, expectations, or improvement suggestions for a specific interface, as derived through the behavioral requirements identification process. This vector can encompass multiple dimensions, such as functional requirements, layout requirements, and interaction requirements. An interface requirements identification vector might indicate that user needs for an "exercise record" interface include: displaying more detailed exercise data (such as steps, distance, and calories burned); providing a comparison function for historical exercise data; and optimizing the interface's layout and color scheme. This vector can provide guidance for subsequent interface design and optimization.

[0068] S130. The smart medical user big data management system performs interface display jump feature analysis based on the interface demand identification vector to obtain an analysis result.

[0069] In S130, interface display jump feature analysis: This is an analysis process that aims to deeply study and understand the characteristics and patterns when a user jumps from one interface display (or module) to another in a visual interactive interface. This analysis usually involves the collection, organization and analysis of data such as the user's navigation path, jump frequency, dwell time, and possible triggering factors. In health management applications, users often jump from the "Health Data" interface to the "Exercise Suggestions" interface. Through the interface display jump feature analysis, the system found that this jump mainly occurs after the user checks his blood sugar or weight data, and the jump frequency is correlated with the changing trend of these data of the user. This indicates that the user may want to adjust his exercise plan based on his health data.

[0070] The analysis result is a general term for the conclusions or data obtained from a certain analysis or processing process. In the context of interface display jump feature analysis, the analysis result usually refers to a series of findings, patterns, rules or suggestions obtained by analyzing the jump behavior of the user interface. These results can be used to optimize the interface design, enhance the user experience or improve the system function. In the example of the health management application mentioned above, the analysis results may include: "Users tend to seek exercise advice after viewing unsatisfactory health data" and "There is a significant correlation between users' jump behavior and the trend of changes in their health data." Based on these analysis results, the system can adjust the interface design, such as automatically popping up relevant exercise suggestions or health consultation links when users view bad health data, thereby providing a more personalized and responsive user experience.

[0071] S140. The smart medical user big data management system determines the interface module switching decision features corresponding to the target patient user tag under the visual interaction interface jump node based on the interface module visual features of the target visual interface under the visual interaction interface jump node in the analysis result.

[0072] In S140, the interface module switching decision features refer to the set of characteristics and rules used to determine when switching from one interface module to another in a visual interactive system. These characteristics may include user interaction behavior, interface module content attributes, system operating status, etc., while the rules are the logic or algorithms developed based on these characteristics to guide switching decisions. Understanding and applying these decision features is crucial for improving user experience, optimizing system performance, and enabling personalized services.

[0073] In a health management application, the interface module switching decision features may include the following aspects:

[0074] User behavior characteristics: The system may record and analyze user interaction behaviors, such as click frequency, sliding direction, dwell time, etc. For example, if a user frequently clicks on the "Blood Sugar" submodule in the "Health Data" module, the system may prioritize recommending "Blood Sugar"-related content or functions to the user, or automatically switch to the "Blood Sugar" submodule the next time the user opens the app;

[0075] Interface content attribute characteristics: Each interface module has its own specific content attributes, such as importance, urgency, and relevance. These attributes can be captured through content semantic description knowledge. For example, for a diabetic patient, the content of the "Hypoglycemia Warning" module may have higher importance and urgency. Therefore, when the user's blood sugar level falls below the safe range, the system may automatically switch to the "Hypoglycemia Warning" module to remind the user to take timely action;

[0076] System operating status characteristics: The system's operating status also affects the decision-making of interface module switching. For example, if the system detects that the user is performing a task that requires concentration (such as filling out a health questionnaire), it may reduce unnecessary interface switching or interruptions to maintain the user's focus;

[0077] Time and context are also important factors in switching decisions. For example, when a user opens the app in the morning, the system might prioritize the "Morning Health Check" module; in the evening, it might switch to the "Sleep Quality Monitoring" module. Furthermore, if the user is in a specific context, such as a hospital or clinic, the system might present specific interface modules relevant to that context.

[0078] In summary, the interface module switching decision features are multi-dimensional and dynamic, and need to be customized and optimized according to specific user needs and system goals. By rationally applying these features, a more intelligent, efficient, and humanized interface interaction experience can be achieved.

[0079] In this embodiment of the present invention, the smart medical user big data management system performs a series of data analysis and interface optimization decisions based on the specific needs and operating habits of target patients when providing medical information services. The following is a specific application scenario of how this system is implemented.

[0080] First, the smart healthcare user big data management system uses the target patient's tags (such as age, gender, medical history, preferences, etc.) and the jump nodes they use when using the visual interactive interface to obtain a visual feature relationship network of past interface modules. This relationship network contains the content semantic description knowledge of multiple visual interfaces, each of which records the module switching data between the corresponding visual interface and other interfaces in detail. These visual interfaces include the target visual interface corresponding to the target patient's user tag. In other words, the system pays special attention to those interfaces that are highly correlated with the user tag.

[0081] Next, the system combines the currently associated visual feature network with the historical interface module visual feature network to further explore the switching behavior vectors between each visual interface and the remaining visual interfaces in terms of interface module visual features. These vectors describe the user's switching patterns and habits when using different interfaces. By analyzing these switching behavior vectors, the system can generate detailed descriptions of interface switching behaviors and further identify behavioral needs based on these descriptions, thereby obtaining interface need identification vectors. These vectors reflect the user's potential needs and expectations when operating the interface.

[0082] The system then analyzes the interface's transition characteristics based on these interface requirement recognition vectors. This process primarily analyzes the user's potential transition paths and choices when operating the interface, as well as the reasons and motivations behind these choices. The results of this analysis provide the system with important clues for optimizing interface layout and navigation flow.

[0083] Finally, the system determines the interface module switching decision features corresponding to the target patient user tag at the jump node based on the visual features of the target visual interface in the analysis results. These decision features will guide the system in providing users with a more personalized and efficient interface experience, including but not limited to adjusting the interface layout, optimizing the navigation process, and recommending relevant content.

[0084] In general, the smart medical user big data management system continuously optimizes and improves the user experience and satisfaction of medical information services by deeply analyzing users' operating habits and needs, as well as their behaviors and choices when using visual interactive interfaces.

[0085] In another specific application scenario, a smart healthcare user big data management system first extracts a visual feature network of past interface modules based on the user tags of elderly patients with type 2 diabetes and the interactive interface jump nodes they navigate when using the health management system. This network includes content semantic descriptions of multiple visual interfaces, detailing the transition data between each visual interface and other interfaces, particularly the transition data for the target visual interface corresponding to the target patient's user tag.

[0086] Next, the system combines the currently associated visual feature network with the historical interface module visual feature network to further explore the switching behavior vectors between each visual interface and the remaining visual interfaces in terms of interface module visual features. These vectors effectively describe the patient's switching patterns and habits between different interfaces. By analyzing these switching behavior vectors, the system can generate detailed descriptions of interface switching behaviors and further identify behavioral needs based on these descriptions, thereby obtaining interface need identification vectors. These vectors reflect the patient's potential needs and expectations when using the interface.

[0087] The system then analyzes these interface requirement recognition vectors to identify the interface's transition characteristics. This process primarily analyzes the patient's likely transition paths and choices when operating the interface, as well as the reasons and motivations behind these choices. The results of this analysis provide important clues for optimizing the interface layout and navigation process.

[0088] Finally, the system determines the interface module switching decision features corresponding to the target patient's user tag at the jump node based on the visual features of the target visual interface in the analysis results. These decision features will guide the system in providing a more personalized and efficient interface experience for patients, including but not limited to adjusting the interface layout, optimizing the navigation process, and recommending relevant content.

[0089] The application of this smart medical user big data management system can effectively improve blood sugar control, self-management, and quality of life for elderly patients with type 2 diabetes. At the same time, the system also provides important reference and support for medical staff in the practice of intelligent health management for elderly patients with type 2 diabetes outside the hospital.

[0090] In another specific application scenario, the system first identifies the target patient users, namely elderly patients with type 2 diabetes, and determines their behaviors and preferences when using the health management system based on their user tags (such as age, gender, disease course, treatment method, etc.). The system also records and analyzes the jump nodes of these patients between different visual interactive interfaces. These nodes reflect the navigation paths and choices made by the patients during use. By collecting and analyzing this data, the system constructs a visual feature relationship network of past interface modules. This relationship network is a complex network structure that contains content semantic description knowledge of multiple visual interfaces. This knowledge describes the functions, layout, elements of each visual interface, and their relationships and switching modes with other interfaces. In particular, the system focuses on the target visual interface corresponding to the target patient user tag and analyzes the switching data between it and other interfaces.

[0091] The system then leverages a network of associated visual features (which may be predefined or generated based on a machine learning algorithm) and a network of visual features from past interface modules to further explore switching behavior between interfaces. It analyzes the similarities and differences in the visual features of each interface module with the remaining visual interfaces and calculates the switching behavior vectors between them. These vectors not only describe the frequency and pattern of switching between interfaces but also reflect the patient's behavior and preferences during the switching process. Through in-depth analysis of these switching behavior vectors, the system can generate detailed descriptions of interface switching behavior. These descriptions may include the type of switch (such as sequential switching, random switching, fallback switching, etc.), the frequency and duration of switching, and the user behavior during the switching process (such as clicking, sliding, typing, etc.). The system then performs behavioral demand identification on these descriptions and extracts key interface demand identification vectors. These vectors reflect the patient's potential needs, expectations, and problems when using the interface, providing important clues for subsequent interface optimization.

[0092] Next, the system will analyze the interface display jump features based on the interface demand identification vector. It will analyze the possible jump paths and choices of patients when operating the interface, and try to understand the reasons and motivations behind these choices. For example, patients may often jump from the "Health Education" interface to the "Exercise Health" interface, which indicates that they may be interested in the impact of exercise on diabetes and want to learn more about it. By parsing the interface display jump features, the system can gain an in-depth understanding of how patients use and navigate the health management system. These understandings can not only help the system optimize the interface layout and navigation process (such as putting highly related interfaces together or providing more intuitive navigation methods), but also provide a basis for the system to recommend relevant content (such as recommending relevant health education articles or exercise plans based on the patient's interests and needs).

[0093] Finally, the system determines the interface module switching decision features based on the visual features of the target visual interface under the visual interaction interface jump node in the parsed results. These decision features are a set of rules and models that describe how the system should switch and display interface modules in different situations. For example, when a patient switches from the "Basic Information" interface to the "My Blood Sugar" interface, the system may dynamically adjust the display content and layout of the "My Blood Sugar" interface based on the patient's blood sugar data and historical records.

[0094] By applying these interface modules to switch decision-making features, the system can provide patients with a more personalized and efficient interface experience. This not only improves patient satisfaction and compliance (because they can more easily find and use the information and functions they need), but also helps medical staff better understand and manage patients' health conditions (because they can evaluate patients' treatment effectiveness and quality of life through the data collected and analyzed by the system).

[0095] It is important to note that for steps S110-S140, the visual feature network of past interface modules is obtained based on the target patient's user tag and the jump node of the visual interactive interface. This step mainly obtains the interface feature network based on user tags and interactive behaviors and has no direct connection with direct disease diagnosis or treatment.

[0096] By combining the network of associated visual features with the network of visual features of previous interface modules, we can mine the switching behavior vectors between the visual features of each interface module and the remaining visual interfaces. This step involves data mining and analysis and is not directly related to disease diagnosis or treatment.

[0097] Behavioral requirements are identified based on the interface switching behavior description to obtain an interface requirement identification vector. This step is an analysis and understanding of user behavior, aiming to optimize the user interface rather than directly diagnose or treat the disease.

[0098] The interface display and jump features are analyzed based on the interface demand identification vector to obtain the analysis results. This step optimizes the interface display and jump features based on the analysis results of user behavior needs, and also does not directly involve the diagnosis or treatment of the disease.

[0099] Based on the visual features of the target visual interface under the jump node in the visual interaction interface in the analysis results, the interface module switching decision features corresponding to the target patient user tag are determined. This step is the process of determining the user interface decision and does not directly involve the diagnosis or treatment of the disease.

[0100] In summary, the content described primarily involves user interface optimization and user behavior analysis, and does not directly involve disease diagnosis or treatment. These contents are more closely related to user interface design, user experience optimization, and data analysis and processing methods in the computer technology field, and are therefore patentable technical solutions.

[0101] In some exemplary embodiments, the combined associated visual feature relationship network and the past interface module visual feature relationship network mine the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces to obtain an interface switching behavior description, and perform behavioral demand identification on the interface switching behavior description to obtain an interface demand identification vector, including: through the associated visual feature relationship network of the current residual branch, based on the original past interface module visual feature relationship network, and the interface demand identification vector generated by the previous residual branch, mining the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces to obtain an interface switching behavior description, and perform behavioral demand identification on the interface switching behavior description to obtain To the interface requirement identification vector, wherein, if the current residual branch is the first residual branch, the interface requirement identification vector generated by the previous residual branch is a zero vector; the next residual branch of the current residual branch is used as the new current residual branch, and the link is jumped to the associated visual feature relationship network through the current residual branch, based on the original past interface module visual feature relationship network and the interface requirement identification vector generated by the previous residual branch, the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces is mined to obtain an interface switching behavior description, and the interface switching behavior description is subjected to behavioral requirement identification to obtain an interface requirement identification vector, until the last residual branch outputs the interface requirement identification vector.

[0102] In some exemplary embodiments, the system implements a deep technical solution to fine-tune and optimize user navigation behavior in a visual interactive interface. The following is a detailed example explanation of this technical solution.

[0103] First, the system maintains two key relationship networks: the associated visual feature relationship network and the past interface module visual feature relationship network. The associated visual feature relationship network captures the associations and mutual influences between visual features across different visual interfaces, while the past interface module visual feature relationship network records the relationships and transition patterns between visual features of various interface modules during users' past interactions with the interface.

[0104] The system's workflow begins by mining the switching behavior vector between each visual interface and the remaining visual interfaces' interface module visual features, using the associated visual feature network of the current residual branch, combined with the original visual feature network of past interface modules and the interface requirement recognition vector generated by the previous residual branch (for the first residual branch, this vector is zero). This switching behavior vector encodes the user's behavioral patterns, frequency, and trends when switching from one interface module to another.

[0105] Next, the system uses these switching behavior vectors to generate interface switching behavior descriptions. These descriptions detail the user's navigation path between interfaces, dwell time, switching preferences, and other information, providing rich context for the system to understand user behavior.

[0106] The system then performs behavioral demand identification on the interface switching behavior descriptions. By analyzing the user's switching behavior, the system can identify potential needs, expectations, or problems encountered by the user. These needs may include a desire for specific functionality, suggestions for improving the interface layout, or the need to optimize the interaction process. The results of behavioral demand identification are encoded into an interface demand identification vector, which contains the user's suggestions for improvement and optimization of each interface module.

[0107] The system then uses the next residual branch after the current one as the new current residual branch and repeats the above process: mining the switching behavior vector by associating the visual feature relationship network, generating a description of the interface switching behavior, and identifying the behavioral requirements to obtain a new interface requirement identification vector. This process is iterative between residual branches, with each residual branch further refining and optimizing its understanding of user behavior based on the previous branch.

[0108] Finally, when the residual branch at the end outputs its interface requirement recognition vector, the system completes a comprehensive analysis of user behavior. This final interface requirement recognition vector contains the system's in-depth understanding of user behavior requirements, providing strong data support for subsequent interface design optimization and personalized services.

[0109] Through this technical solution, the system can not only accurately capture the user's navigation behavior patterns in the visual interactive interface, but also deeply understand the user's potential needs and expectations, thereby providing users with a more intelligent, efficient and humane interface interaction experience.

[0110] It's important to understand that the above describes a process that uses multiple residual branches to process the visual feature network of interface modules to mine interface switching behavior vectors and identify behavioral needs. This process appears to be a technical solution for user interface optimization and user experience analysis. It analyzes user switching behavior between visual interfaces to identify user needs and optimize interface design accordingly.

[0111] Specifically, this process involves using the associated visual feature network and the visual feature network of past interface modules to mine interface switching behavior vectors, obtaining interface switching behavior descriptions. These descriptions are then used to identify behavioral requirements and generate interface requirement recognition vectors. This process is iteratively performed across multiple residual branches, with each residual branch generating a new interface requirement recognition vector based on the output of the previous branch and the original visual feature network of past interface modules.

[0112] From the description, this process doesn't directly involve disease diagnosis or treatment. It focuses more on user interface design and user experience optimization, analyzing user interface switching behavior to identify user needs and improve interface design accordingly. Therefore, this process does not fall under the category of disease diagnosis and treatment.

[0113] In summary, the content described primarily involves user interface optimization and user experience analysis, and does not directly involve disease diagnosis or treatment. Therefore, it can be determined that the content described does not involve disease diagnosis and treatment. Instead, it is more closely related to user interface design, user experience optimization, and data analysis and processing methods in the field of computer technology, and is therefore a patentable technical solution.

[0114] In some other preferred embodiments, the residual branch includes a first behavior demand identification layer, a second behavior demand identification layer, a third behavior demand identification layer and a behavior demand splicing layer; the associated visual feature relationship network of the current residual branch is based on the original past interface module visual feature relationship network and the interface demand identification vector generated by the previous residual branch to mine the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces to obtain an interface switching behavior description, and the interface switching behavior description is subjected to behavior demand identification to obtain an interface demand identification vector, including: through the associated visual feature relationship network of the first behavior demand identification layer in the current residual branch, based on the original past interface module visual feature relationship network and the interface demand identification vector generated by the previous residual branch Based on the original visual feature relationship network of the past interface modules and the interface requirement identification vector generated by the previous residual branch, the switching behavior vector between the visual features of the interface modules of each visual interface and the remaining visual interfaces is mined to obtain a first interface switching behavior description, and the first behavior requirement identification processing is performed on the first interface switching behavior description to obtain a first interface requirement identification vector; through the associated visual feature relationship network of the second behavior requirement identification layer in the current residual branch, based on the original visual feature relationship network of the past interface modules and the interface requirement identification vector generated by the previous residual branch, the switching behavior vector between the visual features of the interface modules of each visual interface and the remaining visual interfaces is mined to obtain a first interface switching behavior description. The second interface switching behavior description is processed by the second behavior requirement identification process to obtain the second interface requirement identification vector; through the third behavior requirement identification layer in the current residual branch, the second interface requirement identification vector is vector multiplied with the interface requirement identification vector generated by the previous residual branch to obtain the first linkage requirement feature; through the associated visual feature relationship network of the third behavior requirement identification layer, the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces is mined according to the original past interface module visual feature relationship network and the first linkage requirement feature to obtain the third interface switching behavior description, and the first linkage requirement feature is used to identify the first interface switching behavior vector. The three-interface switching behavior description is used to perform a third behavior demand identification process to obtain a third interface demand identification vector; through the behavior demand splicing layer, the second linkage demand feature of the first interface demand identification vector and the interface demand identification vector generated by the previous residual branch, as well as the third linkage demand feature of the target feature relationship network corresponding to the first interface demand identification vector and the third interface demand identification vector are determined, and the sum of the second linkage demand feature and the third linkage demand feature is determined to obtain the interface demand identification vector generated by the current residual branch, wherein the target feature relationship network is obtained by subtracting the reference feature relationship network from the first interface demand identification vector.

[0115] In some preferred embodiments, the system's technical solution further refines the structure and functionality of the residual branch to improve the accuracy of identifying user behavior needs. Specifically, each residual branch now includes four key layers: the first behavior need identification layer, the second behavior need identification layer, the third behavior need identification layer, and the behavior need splicing layer. These layers work together to analyze and understand the user's interface switching behavior in a progressively deeper manner.

[0116] First, the system uses the first behavior requirement recognition layer in the current residual branch, utilizing its associated visual feature network, combined with the original past interface module visual feature network and the interface requirement recognition vector generated by the previous residual branch, to mine the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces. This process reveals the user's initial behavior pattern when switching between interfaces and generates a first interface switching behavior description. Next, the system performs first behavior requirement recognition processing on the first interface switching behavior description, identifies the user's initial needs, and encodes them into a first interface requirement recognition vector.

[0117] The system then enters the second behavioral demand recognition layer, similarly utilizing the network of associated visual features and related information to further mine the interface transition behavior vectors and obtain a second interface transition behavior description. This description provides more details and context about the user's behavior. The system then performs second behavioral demand recognition on the second interface transition behavior description, identifying deeper user needs and encoding them into a second interface demand recognition vector.

[0118] At the third behavioral demand identification layer, the system first performs a special operation: it performs vector multiplication on the second interface demand identification vector and the interface demand identification vector generated by the previous residual branch to obtain the first linkage demand feature. This feature reflects the association and interaction between the current branch and the previous branch in terms of user behavioral needs. Next, the system again uses the associated visual feature relationship network, combined with the original past interface module visual feature relationship network and the first linkage demand feature, to dig deeper into the interface switching behavior vector and generate a third interface switching behavior description. The third interface switching behavior description is processed for third behavioral demand identification, and the system obtains the third interface demand identification vector, which contains higher-level needs and trends of user behavior.

[0119] Finally, the behavioral demand splicing layer is responsible for integrating the demand identification results of each level. It first determines the second linkage demand feature of the first interface demand identification vector and the interface demand identification vector generated by the previous residual branch, which reflects the demand linkage relationship between the current level and the previous level. At the same time, it also determines the target feature relationship network corresponding to the first interface demand identification vector and the third linkage demand feature of the third interface demand identification vector. The target feature relationship network is obtained by taking the difference between the reference feature relationship network and the first interface demand identification vector. It represents the user demand characteristics unique to the current level. The behavioral demand splicing layer adds the second linkage demand feature and the third linkage demand feature to obtain the final interface demand identification vector generated by the current residual branch.

[0120] Through this refined residual branch structure and workflow, the system can more accurately capture and understand users' complex behavioral patterns in visual interactive interfaces, providing stronger support for subsequent interface optimization and personalized services.

[0121] In another application scenario of intelligent health management for elderly patients with type 2 diabetes, Mr. Li, a 65-year-old type 2 diabetes patient, has just registered and logged into a smart healthcare management platform. He hopes to better manage his health through this platform.

[0122] Patient user interface

[0123] Mr. Li uploaded his medical records and personal information, including medical history, drug allergies, and family medical history, to the [Patient User Center Interface]. This information was synchronized to the [Medical Data Information Interface] for medical staff to review and reference.

[0124] Patient-side health education interface

[0125] Mr. Li browsed the [Patient Health Education Interface] and watched videos and articles on dietary guidance, exercise health, and psychological counseling for type 2 diabetes. He completed some knowledge-based questionnaires, and the system awarded him points based on his progress. These points can be redeemed for health-related products in the platform's points mall.

[0126] Patient-side exercise health interface and diet management interface

[0127] Mr. Li checks the exercise guide on the [Patient-side Exercise Health Interface] every day and follows it to perform moderate exercise. His exercise data is monitored in real time via a smart bracelet and uploaded to the platform. He also records his diet on the [Patient-side Diet Management Interface], including the types of food and amounts consumed at each meal. This data is synchronized to the [Care and Nursing Data Information Interface] for analysis and evaluation by medical staff.

[0128] Patient-side online interactive interface

[0129] Mr. Li consulted with medical staff about his condition and medication in the "Cloud Doctor" module of the "Patient Online Interactive Interface" and received a timely response. He also shared his experiences and insights on blood sugar control with other diabetics in the "Diabetes Forum" and actively participated in the discussions.

[0130] Medical data information interface and online consultation interface

[0131] The medical staff reviewed Mr. Li's medical history, personal information, lifestyle data, and health status data on the [Medical Data Information Interface]. Based on this data, they assessed Mr. Li's condition and responded to his questions on the [Medical Online Consultation Interface]. They also participated in discussions on the [Diabetic Forum], providing professional advice and guidance to patients.

[0132] Technical Implementation

[0133] In this smart healthcare management platform, patient behavioral data is collected and analyzed in real time. Using the associated visual feature network of the current residual branch, the system mines the switching behavior vectors between the visual features of each visual interface and the remaining visual interfaces based on the original visual feature network of previous interface modules and the interface requirement recognition vector generated by the previous residual branch. These vectors describe the patient's switching behavior and frequency between different interfaces.

[0134] Through multi-level in-depth analysis and processing, the system obtains descriptions of interface switching behaviors and performs behavioral demand identification on these descriptions, generating interface demand identification vectors. These vectors reflect patients' needs and expectations for interface functionality, layout, interaction flow, and user experience.

[0135] Finally, through the behavioral needs concatenation layer, the system integrates the behavioral needs identification results of the previous layers to obtain the final interface needs identification vector generated by the current residual branch. This vector is a comprehensive description of the patient's basic needs, deep needs, and personalized needs.

[0136] Based on this demand identification vector, the platform can optimize and improve the interface design in a targeted manner, improving patient satisfaction and compliance. At the same time, medical staff can also provide patients with more accurate and personalized health management services based on this data.

[0137] It's important to understand that the above description involves a complex technical process involving multiple behavioral demand identification layers and a behavioral demand splicing layer, which work together to mine and analyze switching behaviors between user interfaces and ultimately generate an interface demand recognition vector. This process involves multiple steps, including using a network of associated visual features and a network of visual features from previous interface modules to mine switching behavior vectors, performing multi-level behavioral demand recognition processing, and linking and integrating various recognition results through various methods (such as vector multiplication and subtraction).

[0138] While this process may appear complex and involves extensive data analysis and processing, its core purpose isn't to diagnose or treat disease. Instead, it focuses on understanding and optimizing user behavior and needs when interacting with visual interfaces. This is achieved by deeply analyzing user switching patterns between interfaces, identifying their behavioral needs, and generating an interface needs identification vector based on this information. This vector can then be used to refine user interface design and provide a better user experience.

[0139] Therefore, judging by its core purpose and actual application scenarios, the content described does not constitute a disease diagnosis or treatment method. It is more of a technical means for user interface optimization and user experience analysis. While it may involve analyzing and processing user behavior and needs, this analysis and processing is not intended to diagnose or treat any disease. Instead, it is intended to improve user interface design and enhance the user experience. In summary, we can conclude that the content described is not a disease diagnosis or treatment method, but rather a technical means for user interface optimization and user experience analysis.

[0140] In other possible embodiments, the number of the visual feature relationship networks of the past interface modules is at least two, the visual interfaces corresponding to each past interface module visual feature relationship network are the same, and the interface analysis node ranges are different; the combined associated visual feature relationship network and the past interface module visual feature relationship network mine the switching behavior vectors between the visual features of the interface modules of each visual interface and the remaining visual interfaces to obtain an interface switching behavior description, and perform behavioral requirement identification on the interface switching behavior description to obtain an interface requirement identification vector, including: mining the switching behavior vectors between the visual features of the interface modules of each visual interface and the remaining visual interfaces based on the visual feature relationship networks of each past interface module through the associated visual feature relationship network to obtain the interface switching behavior description corresponding to each past interface module visual feature relationship network, and perform behavioral requirement identification on each interface switching behavior description to obtain the interface requirement identification vector corresponding to each past interface module visual feature relationship network.

[0141] In some embodiments, the technical solution involves processing multiple visual feature relationship networks of past interface modules. These relationship networks each correspond to the same visual interface, but their interface analysis node scopes are different. This means that although they are all analyzing the same visual interface, the nodes or elements they focus on may be different.

[0142] First, each visual feature relationship network of past interface modules captures the relationships between the visual features of the modules in the visual interface. These relationships may include positional relationships, size relationships, color relationships, and other relationships between modules. Due to the different scopes of analysis nodes, different relationship networks may reveal different visual feature relationships even for the same visual interface.

[0143] Next, the technical solution combines the network of associated visual features with the network of visual features of these previous interface modules to mine the switching behavior vectors of the interface module visual features between each visual interface and the remaining visual interfaces. The "switching behavior vector" here can be understood as a quantitative description of how one interface transforms into another. It may include information such as which modules have changed and how these changes occurred.

[0144] By mining these switching behavior vectors, the technical solution can generate interface switching behavior descriptions. These descriptions detail the changes that occur when switching from one interface to another. Then, behavioral requirements identification is performed on these interface switching behavior descriptions. The goal of this step is to understand the user needs or intentions reflected by these switching behaviors.

[0145] Ultimately, the results of behavioral requirement identification are converted into interface requirement identification vectors. These vectors express user interface requirements in a way that can be understood and processed by computer systems. Each visual feature relationship network of past interface modules corresponds to one such interface requirement identification vector, forming a vector set that provides an important reference for subsequent user interface design or optimization.

[0146] In this process, emerging technical terms such as "switching behavior vector" and "interface requirement identification vector" are key tools for quantifying and describing interface switching behavior and user needs. They enable technical solutions to understand and meet user needs in a more precise and actionable way.

[0147] In the application scenario of intelligent health management for elderly patients with type 2 diabetes, the technical solution can be implemented in detail as follows.

[0148] First, the system constructs at least two visual feature networks of past interface modules. These networks are all associated with the same visual interface, but each network analyzes a different range of interface nodes. For example, one network might focus on analyzing the button layout and color matching of a blood glucose monitoring interface, while another might focus on the font size and contrast of a user's personal information settings interface.

[0149] These visual feature networks of interface modules were constructed by analyzing the interaction behaviors of elderly patients with type 2 diabetes using health management applications. The system recorded information such as the patient's visual focus movements, click events, and dwell time as they navigated different functional interfaces. This data was then used to construct a visual feature network between interface modules.

[0150] Next, the system combines the network of associated visual features with the network of visual features of these previous interface modules to further explore the switching behavior vectors of the interface module visual features between each visual interface and the remaining visual interfaces. The "switching behavior vector" here is a quantitative description that reveals the changes in the patient's visual focus and interaction behavior when switching from one interface to another.

[0151] For example, when a patient switches from a blood glucose monitoring interface to a diet log interface, the system can detect that the patient's focus shifts from a blood glucose chart to a table of food intake and nutritional information. This switching behavior vector includes information such as the distance, speed, and direction of the patient's visual focus.

[0152] By mining these switching behavior vectors, the system can obtain detailed descriptions of interface switching behaviors. These descriptions not only record the patient's behavioral trajectory during the interface switching process, but also reflect the patient's usage habits and preferences.

[0153] The system then identifies behavioral needs based on these interface transition behavior descriptions. This step aims to understand patients' interface usage needs in order to provide them with more personalized and convenient health management services. For example, if the system discovers that patients often get lost or have trouble finding the desired functional modules when switching interfaces, it can optimize the interface design to address these issues, such as adding navigation prompts or adjusting module layouts.

[0154] Ultimately, the results of behavioral need identification are converted into interface need identification vectors. These vectors express the patient's interface needs in a way that can be understood and processed by the computer system. Based on these vectors, the system can adjust the interface design to meet the patient's needs and enhance their user experience.

[0155] Throughout the entire process, the system continuously learns and optimizes, gradually adapting to the usage habits and preferences of different patients, providing them with more accurate and personalized health management services. Furthermore, this technical solution based on a visual feature relationship network also provides a useful reference for other types of intelligent health management applications.

[0156] It should be understood that the above embodiments involve using the visual feature relationship network of past interface modules and the associated visual feature relationship network to mine switching behavior vectors between visual interfaces and obtain interface switching behavior descriptions. These descriptions are then used to identify behavioral requirements to obtain interface requirement identification vectors. This process primarily analyzes and understands user needs based on their interface interaction behaviors in order to optimize user interface design.

[0157] As the description suggests, this process doesn't involve any medical examinations, tests, analyses, or treatments. It focuses primarily on user interface design and user experience optimization, not on disease diagnosis or treatment. While this process may involve analyzing user behavior and needs, the purpose of this analysis is to improve the user interface, not to diagnose or treat a disease.

[0158] In addition, the terms such as "past interface module visual feature relationship network" and "associated visual feature relationship network" mentioned in the content, as well as the mining and identification processes between them, are concepts and methods in the fields of computer technology and user interface design, and have no direct connection with disease diagnosis and treatment in the medical field.

[0159] In summary, we can conclude that the content described here is not a disease diagnosis or treatment method. It is more like a technical means of user interface optimization and user experience analysis, and belongs to the application of computer technology and user interface design.

[0160] In some further exemplary embodiments, the interface display jump feature analysis is performed based on the interface requirement identification vector to obtain an analysis result, including: weighting the interface requirement identification vector corresponding to each of the past interface module visual feature relationship networks to obtain an interface requirement identification weighted vector; and performing interface display jump feature analysis based on the interface requirement identification weighted vector to obtain an analysis result.

[0161] In the application scenario of intelligent health management for elderly patients with type 2 diabetes, the implementation of the technical solution can be further refined as follows.

[0162] The system first processes the interface requirement recognition vectors generated by visual feature networks of multiple past interface modules. These vectors reflect the needs and preferences of elderly patients for different interface elements when using health management applications. Because each network analyzes a different range of interface nodes, the interface requirement recognition vectors generated by each network may also vary.

[0163] To develop a comprehensive interface requirement recognition vector that reflects the patient's overall needs, the system weights the interface requirement recognition vectors corresponding to the visual feature networks of each interface module. Weighting can be based on factors such as the importance and reliability of each network, or the patient's frequency of use of the corresponding interface. This weighting process yields a weighted interface requirement recognition vector that more accurately reflects the patient's actual needs and preferences.

[0164] Next, the system identifies a weighted vector based on this interface requirement and analyzes the interface transition characteristics. This step aims to understand the characteristics of interface display and transitions when patients use the app, and how these characteristics meet their needs. For example, the system can analyze the frequency, switching paths, and dwell time of patients switching between different interfaces to derive interface transition characteristics.

[0165] During the analysis process, the system takes into account factors such as the patient's usage habits, cognitive ability, and health status. For example, for elderly patients with poor eyesight, the system may increase font size and contrast to reduce their reading difficulties; for patients who often forget operation steps, the system may provide more detailed navigation and prompt information to help them complete the operation smoothly.

[0166] Ultimately, the system generates an analysis of the interface's display jump characteristics. This analysis includes specific recommendations for optimizing interface design, improving user experience, and addressing patient needs. For example, the system might suggest adjusting the position or size of a button, adding a shortcut to a function, or optimizing the layout of a specific interface.

[0167] This technical solution enables the system to more accurately understand the needs and preferences of elderly patients with type 2 diabetes, providing them with more personalized and convenient health management services. This data-driven approach also provides a valuable reference for other types of intelligent health management applications.

[0168] In some possible embodiments, the interface display jump feature analysis is performed based on the interface requirement identification vector to obtain the analysis result, including: performing interface display jump feature analysis based on the interface requirement identification vector to obtain the analysis result of each visual interface in the multiple visual interfaces; determining the interface module switching decision feature corresponding to the target patient user tag under the visual interaction interface jump node based on the interface module visual feature of the target visual interface under the visual interaction interface jump node in the analysis result, including: obtaining the interface module visual feature of the target visual interface under the visual interaction interface jump node based on the analysis result of each visual interface in the multiple visual interfaces, and determining the interface module switching decision feature corresponding to the target patient user tag under the visual interaction interface jump node.

[0169] In the application scenario of intelligent health management for elderly patients with type 2 diabetes, the technical solution can be specifically implemented as follows.

[0170] First, the system analyzes the interface display jump characteristics based on the previously generated interface demand recognition vectors. These vectors are generated based on the interface interaction behaviors of elderly patients when using health management applications, reflecting their needs and preferences for different interface elements.

[0171] The parsing process is performed on each of the multiple visual interfaces. The system analyzes the interface requirement recognition vector for each visual interface and extracts features related to interface display and transitions. These features may include interface layout, button location and size, font color and size, use of images, and the patient's switching behavior between these interfaces, such as the frequency, path, and sequence of switching.

[0172] For example, the system may find that elderly patients prefer to use large fonts and high contrast interfaces when viewing blood sugar data, while they prefer to use interfaces with food pictures and detailed nutritional information when recording their diet. These findings are the results of analyzing the jump characteristics of interface display.

[0173] Next, the system uses these analysis results to determine the visual features of the target visual interface module under the visual interaction interface jump node. Here, the "target visual interface" refers to the interface the patient is currently using or will soon use, and the "visual interaction interface jump node" refers to the decision point at which the patient switches between different interfaces.

[0174] The system analyzes the patient's interface needs identification vectors and interface display jump characteristics at these decision points to determine which interface module visual features the target visual interface should have. These features should be able to meet patients' needs, improve their user experience, and guide them to make effective interface transitions.

[0175] For example, when an elderly patient jumps from the blood glucose monitoring interface to the diet record interface, the system may design the food images and nutritional composition table on the diet record interface to be more eye-catching and easy to read based on the previous analysis results, so that the patient can quickly find and understand the relevant information.

[0176] Finally, the system determines the interface module switching decision features corresponding to the target patient user tag at the jump node in the visual interactive interface. The "target patient user tag" here refers to the personalized tags set by the system for elderly patients, such as age, gender, vision status, and cognitive ability. These tags help the system more accurately understand the patient's needs and preferences.

[0177] The system combines the patient's user tags and their behavioral data when switching interfaces to determine the most suitable interface module switching decision features for them. These features should reflect the patient's usage habits and preferences, helping them complete interface switching operations more smoothly and efficiently.

[0178] Through this technical solution, the system can provide more personalized and convenient health management services for elderly patients with type 2 diabetes, improving their user experience and quality of life. At the same time, this data-driven approach also provides a useful reference for other types of intelligent health management applications.

[0179] In some alternative embodiments, the visual feature relationship network of past interface modules is obtained based on the target patient user label and the visual interaction interface jump node, and the visual feature relationship network of past interface modules includes content semantic description knowledge of multiple visual interfaces, including: determining the range of at least two interface analysis nodes before the visual interaction interface jump node based on the visual feature of the interface module to be predicted; decomposing each interface analysis node range into multiple operation node sets based on a pre-set node step size; determining the target visual interface corresponding to the target patient user label and the target interface analysis node set based on the visual feature of the interface module to be predicted; The sample visual interface corresponding to the target visual interface is used as the visual interface to be processed; for each visual interface to be processed, the interface module switching data of the visual interface to be processed and the remaining visual interfaces to be processed in each operation node set is obtained; for each interface analysis node range, the content semantic description knowledge of each visual interface in the interface analysis node range is obtained based on the interface module switching data of each visual interface to be processed and the remaining visual interfaces to be processed in the operation node set of the interface analysis node range; based on the content semantic description knowledge corresponding to each interface analysis node range, the visual feature relationship network of the past interface modules corresponding to each interface analysis node range is obtained.

[0180] In the application scenario of intelligent health management for elderly patients with type 2 diabetes, the implementation of the technical solution can be specifically detailed as follows.

[0181] First, the system uses the target patient's user tag and the current visual interaction interface jump node to obtain a visual feature network of past interface modules. This network contains semantic descriptions of the content of multiple visual interfaces, which provides a deep understanding of the interaction between interface elements and patient behaviors.

[0182] To construct this relationship network, the system first identifies at least two screen analysis node ranges preceding the current jump node. These node ranges represent the sequence of screens the patient traversed before reaching the current jump node. For example, a patient might transition from the home screen to the blood glucose monitoring screen and then to the diet log screen. In this case, the home screen and the blood glucose monitoring screen are the first two analysis node ranges of the diet log screen.

[0183] Next, the system breaks down each interface analysis node into multiple action node sets. These action node sets refer to the specific actions the patient performs on each interface, such as clicking a button or swiping the screen. This breakdown can be based on a pre-set node step size, which can be adjusted based on the patient's operating habits and the complexity of the interface.

[0184] The system then determines the target visual interface and sample visual interface based on the target patient's user tag. The target visual interface is the interface the patient is currently using or will soon use, while the sample visual interface is used as a reference and comparison interface. They can be different interfaces of the same type or different types.

[0185] Next, for each visual interface to be processed, the system obtains its interface module switching data in each operation node set and the remaining visual interfaces to be processed. This data reflects the patient's behavior patterns and preferences when switching between different interfaces.

[0186] Then, for each interface analysis node range, the system switches data based on the interface modules in its operation node set, obtaining content semantic description knowledge of each visual interface within that range. This knowledge provides a deep understanding of interface elements and patient interaction behaviors, which can help the system better understand patients' needs and preferences.

[0187] Finally, the system analyzes the semantic description of the content within each interface node range to derive a visual feature network of past interface modules corresponding to that range. This network represents the system's understanding and memory of the patient's past interactions, helping it more accurately predict future interactions and provide more personalized and convenient health management services.

[0188] In some other replaceable embodiments, interface display jump feature analysis is performed based on the interface requirement identification vector to obtain an analysis result, including: performing knowledge transformation on the interface requirement identification vector to obtain interface requirement transformation knowledge; and performing interface display jump feature analysis based on the interface requirement transformation knowledge to obtain an analysis result.

[0189] In the application scenario of intelligent health management for elderly patients with type 2 diabetes, the technical solution implemented by the system can be explained in detail as follows:

[0190] First, the system performs knowledge transformation on the acquired interface requirement identification vectors. Knowledge transformation is a process that converts the original interface requirement identification vectors into a more understandable and analyzable form, known as interface requirement transformation knowledge. This process may involve operations such as data standardization, normalization, dimensionality reduction, or feature extraction to better reveal the hidden information and patterns within the vectors.

[0191] For example, the system might normalize certain elements in the interface requirement identification vector so that their values ​​fall within a uniform range, eliminating dimensional differences between different elements. Alternatively, the system might use methods such as principal component analysis to reduce the dimensionality of the vector and extract its key features to simplify subsequent analysis.

[0192] After obtaining the interface demand transformation knowledge, the system uses this knowledge to analyze the interface display jump characteristics. This process primarily analyzes the patient's jump behavior between different interfaces and the relationship between these behaviors and interface elements. Based on the information in the interface demand transformation knowledge, the system identifies characteristics such as the patient's dwell time, click frequency, and jump path on different interfaces, as well as the relationship between these characteristics and interface elements such as interface layout, button location, and font size.

[0193] For example, the system may find that when a button on a certain interface is too close to the edge of the screen, it is difficult for elderly patients to accurately click the button due to limited vision or finger dexterity. In this case, the system will record this feature as a reference for optimizing the interface design.

[0194] Ultimately, the system generates an analysis result that comprehensively describes the patient's navigation behavior across different interfaces and the relationships between interface elements. Based on this analysis, the system assesses whether the current interface design meets the patient's needs, identifies areas for improvement, and provides specific optimization recommendations.

[0195] This technical solution enables the system to more deeply understand the interface interactions of elderly patients with type 2 diabetes, identify existing problems and obstacles, and provide them with more personalized and efficient health management services. Furthermore, this data-driven approach can provide valuable insights and references for other types of intelligent health management applications.

[0196] In summary, the present invention's beneficial effect lies in efficiently acquiring a network of visual feature relationships across past interface modules by combining target patient user tags with visual interaction interface jump nodes. This network encompasses the semantic descriptions of the content across multiple visual interfaces and the data on the interface module transitions between them. This innovative approach not only enriches the data foundation for user interface design but also improves the efficiency of switching between interfaces and the consistency of the user experience.

[0197] Furthermore, by combining the network of associated visual features with the network of visual features of previous interface modules, this paper further explores the visual feature switching behavior vectors between each visual interface and the remaining visual interfaces. This vectorized description makes interface switching behavior more precise and quantifiable, providing a solid foundation for subsequent behavioral requirement identification.

[0198] By identifying behavioral needs based on the description of interface switching behavior, the present invention can accurately obtain an interface need identification vector. This vector not only reflects the actual needs of users, but also provides strong support for personalized and intelligent interface design.

[0199] Furthermore, the present invention analyzes interface display jump characteristics based on the interface requirement identification vector, thereby obtaining targeted analysis results. This result provides clear guidance for the visual characteristics of the interface modules under the jump node of the visual interactive interface, making the interface design more in line with user expectations and usage habits.

[0200] Ultimately, by determining the interface module switching decision features corresponding to the target patient's user tag, the present invention achieves precise decision-making at the jump node of the visual interactive interface. This decision not only improves the responsiveness and accuracy of the user interface, but also greatly enhances user satisfaction and convenience during use.

[0201] In summary, the present invention achieves significant benefits by improving the efficiency and accuracy of user interface design, enhancing the consistency and personalization of the user experience, and increasing the responsiveness and user satisfaction of the user interface during visual interaction. These benefits collectively constitute the significant advantages and innovative value of the present invention in the field of user interface design.

[0202] Please refer to Figure 2The embodiment of the present invention further provides a smart medical user big data management system 100, comprising a processor 111, a memory 112, and a bus 113 connected to the processor 111. The processor 111 and the memory 112 communicate with each other via the bus 113. The processor 111 is configured to call program instructions in the memory 112 to execute the aforementioned user big data management method based on artificial intelligence and visualization.

[0203] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or system that includes a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, commodity, or system. In the absence of further limitations, an element defined by the phrase "comprises a..." does not exclude the presence of other identical elements in the process, method, commodity, or system that includes the element.

[0204] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0205] The above are merely embodiments of the present invention and are not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. A user big data management method based on artificial intelligence and visualization, characterized in that: Applied to the smart medical user big data management system, the method includes: Obtaining, based on the target patient user tag and the visual interactive interface jump node, a visual feature relationship network of past interface modules, wherein the visual feature relationship network of past interface modules includes content semantic description knowledge of a plurality of visual interfaces, each piece of content semantic description knowledge including interface module switching data between a corresponding visual interface and the remaining visual interfaces in the plurality of visual interfaces, wherein the plurality of visual interfaces includes a target visual interface corresponding to the target patient user tag; Combining the associated visual feature relationship network and the visual feature relationship network of the past interface modules to mine the switching behavior vectors between the visual features of the interface modules of each visual interface and the remaining visual interfaces, obtaining an interface switching behavior description, and performing behavioral requirement identification on the interface switching behavior description to obtain an interface requirement identification vector, which includes: mining the switching behavior vectors between the visual features of the interface modules of each visual interface and the remaining visual interfaces through the associated visual feature relationship network of the current residual branch, based on the original visual feature relationship network of the past interface modules, and the interface requirement identification vector generated by the previous residual branch, obtaining an interface switching behavior description, performing behavioral requirement identification on the interface switching behavior description, and obtaining an interface requirement identification vector; Performing interface display jump feature analysis based on the interface requirement identification vector to obtain an analysis result; According to the interface module visual features of the target visual interface under the visual interaction interface jump node in the analysis result, the interface module switching decision features corresponding to the target patient user label under the visual interaction interface jump node are determined.

2. The user big data management method based on artificial intelligence and visualization according to claim 1 is characterized in that: If the current residual branch is the first residual branch, the interface requirement identification vector generated by the previous residual branch is a zero vector.

3. The user big data management method based on artificial intelligence and visualization according to claim 2 is characterized in that: The method combines the associated visual feature relationship network and the previous interface module visual feature relationship network to mine the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces to obtain an interface switching behavior description, and performs behavior requirement identification on the interface switching behavior description to obtain an interface requirement identification vector, including: The next residual branch of the current residual branch is used as the new current residual branch, and the branch is jumped to the associated visual feature relationship network through the current residual branch. Based on the original past interface module visual feature relationship network and the interface requirement identification vector generated by the previous residual branch, the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces is mined to obtain an interface switching behavior description, and the behavior requirement is identified on the interface switching behavior description to obtain an interface requirement identification vector, until the last residual branch outputs the interface requirement identification vector.

4. The user big data management method based on artificial intelligence and visualization according to claim 1 is characterized in that: The residual branch includes the first behavior demand recognition layer, the second behavior demand recognition layer, the third behavior demand recognition layer and the behavior demand splicing layer; The method comprises mining the switching behavior vectors between the visual features of the interface modules of each visual interface and the remaining visual interfaces through the associated visual feature network of the current residual branch, based on the original visual feature network of the past interface modules and the interface requirement recognition vector generated by the previous residual branch, obtaining an interface switching behavior description, and performing behavior requirement recognition on the interface switching behavior description to obtain an interface requirement recognition vector, including: By using the associated visual feature network of the first behavior requirement recognition layer in the current residual branch, based on the original past interface module visual feature network and the interface requirement recognition vector generated by the previous residual branch, the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces is mined to obtain a first interface switching behavior description, and the first behavior requirement recognition processing is performed on the first interface switching behavior description to obtain a first interface requirement recognition vector; Through the associated visual feature relationship network of the second behavior requirement recognition layer in the current residual branch, based on the original past interface module visual feature relationship network and the interface requirement recognition vector generated by the previous residual branch, the switching behavior vector between the interface module visual features of each visual interface and the remaining visual interfaces is mined to obtain a second interface switching behavior description, and the second behavior requirement recognition processing is performed on the second interface switching behavior description to obtain a second interface requirement recognition vector; Performing vector multiplication on the second interface requirement identification vector and the interface requirement identification vector generated by the previous residual branch through the third behavior requirement identification layer in the current residual branch to obtain a first linkage requirement feature; By using the associated visual feature relationship network of the third behavior requirement identification layer, mining the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces based on the original past interface module visual feature relationship network and the first linkage requirement feature, obtaining a third interface switching behavior description, and performing third behavior requirement identification processing on the third interface switching behavior description to obtain a third interface requirement identification vector; Through the behavioral requirement splicing layer, determine the second linkage requirement feature of the first interface requirement identification vector and the interface requirement identification vector generated by the previous residual branch, as well as the third linkage requirement feature of the target feature relationship network corresponding to the first interface requirement identification vector and the third interface requirement identification vector, determine the sum of the second linkage requirement feature and the third linkage requirement feature, and obtain the interface requirement identification vector generated by the current residual branch, wherein the target feature relationship network is obtained by subtracting the reference feature relationship network from the first interface requirement identification vector.

5. The user big data management method based on artificial intelligence and visualization according to claim 1 is characterized in that: The number of the visual feature relationship networks of the past interface modules is at least two, and the visual interfaces corresponding to the visual feature relationship networks of each past interface module are the same, but the interface analysis node ranges are different; The combined associated visual feature relationship network and the past interface module visual feature relationship network mine the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces, obtain the interface switching behavior description, perform behavioral requirement identification on the interface switching behavior description, and obtain the interface requirement identification vector, including: mining the switching behavior vectors between the interface module visual features of each visual interface and the remaining visual interfaces based on the associated visual feature relationship network of each past interface module visual feature relationship network, obtain the interface switching behavior description corresponding to each past interface module visual feature relationship network, perform behavioral requirement identification on each interface switching behavior description, and obtain the interface requirement identification vector corresponding to each past interface module visual feature relationship network.

6. The method for user big data management based on artificial intelligence and visualization according to claim 5, characterized in that: The performing of interface display jump feature analysis based on the interface requirement identification vector to obtain an analysis result includes: Weighting the interface requirement recognition vectors corresponding to the visual feature relationship networks of the past interface modules to obtain an interface requirement recognition weighted vector; The interface display jump feature is analyzed based on the weighted vector identified by the interface requirement to obtain an analysis result.

7. The user big data management method based on artificial intelligence and visualization according to claim 1 is characterized in that: The performing interface display jump feature analysis based on the interface requirement identification vector to obtain an analysis result includes: performing interface display jump feature analysis based on the interface requirement identification vector to obtain an analysis result for each visual interface in the multiple visual interfaces; The method of determining the interface module switching decision feature corresponding to the target patient user tag under the visual interaction interface jump node based on the interface module visual features of the target visual interface under the visual interaction interface jump node in the analysis result includes: obtaining the interface module visual features of the target visual interface under the visual interaction interface jump node based on the analysis result of each visual interface in the multiple visual interfaces, and determining the interface module switching decision feature corresponding to the target patient user tag under the visual interaction interface jump node.

8. The method for user big data management based on artificial intelligence and visualization according to claim 5, characterized in that: According to the target patient user label and the visual interactive interface jump node, a visual feature relationship network of past interface modules is obtained, wherein the visual feature relationship network of past interface modules includes content semantic description knowledge of multiple visual interfaces, including: Determine, based on a visual interaction interface jump node of the visual features of the interface module to be predicted, a range of at least two interface analysis nodes before the visual interaction interface jump node; The analysis node range of each interface is decomposed into multiple operation node sets based on the pre-set node step size; According to the target patient user label of the visual feature of the interface module to be predicted, determining the target visual interface corresponding to the target patient user label and the sample visual interface corresponding to the target visual interface as the visual interface to be processed; For each visual interface to be processed, obtaining interface module switching data of the visual interface to be processed and the remaining visual interfaces to be processed in each operation node set; For each interface analysis node range, the interface module switching data of each visual interface to be processed and the remaining visual interfaces to be processed are concentrated according to the operation nodes of the interface analysis node range, and the content semantic description knowledge of each visual interface in the interface analysis node range is obtained; According to the content semantic description knowledge corresponding to each interface analysis node range, the visual feature relationship network of the past interface modules corresponding to each interface analysis node range is obtained.

9. The method for managing user big data based on artificial intelligence and visualization according to any one of claims 1 to 5, characterized in that: Performing interface display jump feature analysis based on the interface requirement identification vector to obtain analysis results, including: Performing knowledge transformation on the interface requirement identification vector to obtain interface requirement transformation knowledge; An interface display jump feature analysis is performed based on the interface requirement transformation knowledge to obtain an analysis result.

10. A smart medical user big data management system, characterized by: It includes a processor, a memory and a bus connected to the processor; the processor and the memory communicate with each other via the bus; The processor is used to call the computer program in the memory to execute the user big data management method based on artificial intelligence and visualization as described in any one of claims 1-9.

11. A computer-readable storage medium, characterized in that A program is stored thereon, which, when executed by a processor, implements the user big data management method based on artificial intelligence and visualization as described in any one of claims 1 to 9.

Images