A food regimen generation method and system based on the five elements theory of traditional Chinese medicine

Abstract

This invention discloses a method and system for generating dietary therapy plans based on the Five Elements theory of Traditional Chinese Medicine, relating to the fields of artificial intelligence and health science popularization. By collecting initial user characteristic data and dynamic behavioral characteristic data, a two-dimensional feature fusion model is constructed. An AI model is used to classify and determine the user's constitution, and personalized dietary therapy suggestions, recipe combinations, and health science popularization content are generated based on the constitution profile. The system only provides health science popularization and dietary therapy guidance services and does not involve medical diagnosis or treatment. It can be widely applied in the field of digital health science popularization.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and health science popularization, and in particular to a method and system for generating dietary therapy plans based on the Five Elements theory of Traditional Chinese Medicine. Background Technology

[0002] With the rapid development of artificial intelligence technology, the field of health management is gradually extending from traditional medical diagnosis to personalized health services. Existing health management technologies are mainly divided into two categories: one is based on medical diagnosis systems, such as disease risk assessment through physiological indicator detection; the other is based on exercise monitoring systems, such as recording users' exercise volume, heart rate, and other data through devices like smart bracelets and watches. However, the above technologies have the following shortcomings:

[0003] First, while exercise monitoring systems can record users' behavioral data, they lack in-depth analysis of an individual's initial characteristics (such as innate physical predisposition) and acquired behavioral habits, making it impossible to generate truly personalized health plans. For example, the same amount of exercise may have significantly different health effects on users with different physical predispositions, but current technology cannot identify these differences.

[0004] Secondly, existing health and wellness products (such as diet recommendation apps and health management mini-programs) are mostly based on simple rules or general templates, lacking personalized scientific basis, resulting in poor user experience and low user stickiness. Furthermore, most of these products do not address the concept of constitution classification in traditional Chinese medicine and fail to effectively combine traditional culture with modern artificial intelligence technology.

[0005] Therefore, in order to solve the problems existing in the current technology, there is a need for a method and system for generating dietary therapy plans based on the Five Elements theory of Traditional Chinese Medicine. Summary of the Invention

[0006] The present invention aims to at least partially solve one of the technical problems in the above-mentioned technologies.

[0007] To achieve the above objectives, the first aspect of this invention proposes a method for generating a dietary therapy plan based on the Five Elements theory of Traditional Chinese Medicine, comprising the following steps:

[0008] S1: Obtain initial user feature data and dynamic behavioral feature data. This method does not involve disease diagnosis, treatment, or prevention; it is only used for health education and dietary guidance. S2: Generate a two-dimensional feature fusion model for the user based on preset coding rules and a weighted algorithm. The preset coding rules are: using birth date and birth time as input features, and generating one of five feature labels through a preset temporal coding algorithm. The temporal coding algorithm includes: converting birth date into year, month, and day values, and birth time into hour and minute values; generating a multi-dimensional feature vector through weighted summation and normalization; the weight coefficients for the weighted summation are preset weight coefficients; mapping the multi-dimensional feature vector to one of five cluster centers, which are formed by K-means clustering of a predetermined number of samples. The training process involves several steps: S1: The samples are obtained from publicly available demographic datasets or anonymized user data collected within the company in compliance with regulations. The collection time is within a predetermined time range, and the sample distribution covers different years and seasons. The samples undergo representativeness testing to ensure coverage of different regions, ages, and genders. S2: The two-dimensional feature vector is input into a pre-trained decision tree model or K-means clustering model. The model parameters are optimized through K-fold cross-validation to reduce the risk of overfitting on user behavior data, improve classification stability, and output a user's physical condition profile. S3: Based on the physical condition profile, a pre-stored dietary knowledge base is called to generate non-medical personalized dietary advice and health science content. S4: Execution feedback is collected through the user interaction module to dynamically update the physical condition profile.

[0009] In addition, the dietary therapy method based on the Five Elements theory of Traditional Chinese Medicine proposed above according to the present invention may also have the following additional technical features:

[0010] Furthermore, in S1, the initial feature data includes the user's date and time of birth, and the dynamic behavioral feature data includes diet records, sleep records, exercise records, water intake records, mood scores, and physiological state feedback data.

[0011] Furthermore, in S2, the construction process of the dual-dimensional feature fusion model includes: mapping the initial feature data into five initial state labels through a preset encoding mapping table, wherein the encoding mapping table takes birth date and birth time as input and outputs one of the five corresponding feature labels; applying a weighted scoring algorithm to the dynamic behavioral feature data to generate acquired behavioral state tendency values, wherein the weights of each behavioral dimension in the weighted scoring algorithm are dynamically adjusted by the user's initial state labels, specifically: calculating the statistical correlation between the initial state labels and each behavioral dimension based on the Pearson correlation coefficient in the training dataset, and determining the correlation based on the correlation strength. The weights of the corresponding dimensions are automatically adjusted. The training dataset contains a predetermined number of user behavior questionnaire samples, which are collected through an online survey platform. The questionnaire content covers dimensions such as eating frequency, sleep duration, exercise intensity, drinking habits, emotional state, and physiological feedback. The sample selection criteria are: complete and consistent data. The statistical correlation has verifiable statistical significance on the training dataset, with a significance level of p < 0.05. The significance test uses the Pearson correlation coefficient t-test method. The initial state labels are combined with the acquired behavioral state tendency values ​​to form a two-dimensional feature vector.

[0012] Furthermore, in S3, the output format of the physical condition profile includes: the user's initial state label; the acquired behavioral state tendency value includes at least one of the following: damp-heat tendency, qi deficiency tendency, yang deficiency tendency, yin deficiency tendency, and phlegm-dampness tendency; the health maintenance dimensions that currently require the most attention include at least one of the following: drinking water, sleep, diet, and exercise; the maximum depth of the decision tree model is dynamically adjusted according to the features of the training data, and the K value of the K-means clustering model is consistent with the number of initial state labels.

[0013] Furthermore, in S4, the process of generating personalized dietary recommendations includes: calling a pre-stored dietary knowledge base, matching corresponding food recommendations, dietary pairings, and taboo prompts based on the physical condition profile; the dietary knowledge base is a relational database, with the physical condition tag as the primary key, and associated with the food recommendation table, recipe pairing table, and taboo prompt table; and combining intangible cultural heritage elements, presented in the form of text, images, or multimodal interaction.

[0014] Furthermore, in S5, the dynamic feedback and report generation includes: generating weekly behavioral data reports, monthly trend analysis reports, and annual panoramic data reports; and dynamically generating status identifiers based on user execution data.

[0015] Furthermore, the status identifiers are data tags automatically triggered based on user behavior data, including: when a user's water intake data reaches a preset threshold for 7 consecutive days, an "Water Intake Met" identifier is automatically generated; when a user's exercise data reaches a preset threshold for 3 consecutive days, an "Exercise Continues" identifier is automatically generated; when a user with a tendency towards damp heat has normal bowel movements for 5 consecutive days, a "Status Stable" identifier is automatically generated; when a user falls asleep before 11 PM for 7 consecutive days, an "Early to Bed Met" identifier is automatically generated; and when a user has no record of overeating for 5 consecutive days, a "70% Full" identifier is automatically generated.

[0016] The second aspect of this invention proposes a dietary therapy plan generation system based on the Five Elements theory of Traditional Chinese Medicine to implement the first aspect, comprising: a data acquisition module for acquiring initial user feature data and dynamic behavioral feature data; a feature fusion modeling module, executed in the processor of a cloud server, communicating with the data acquisition module through an API interface, for performing user data preprocessing, initial label mapping, and behavioral tendency value calculation; an AI classification decision module, with a built-in trained decision tree model or K-means clustering model, for generating a user's physical condition profile; a plan generation module for retrieving corresponding suggestions from a dietary therapy knowledge base and generating text, images, or multimodal interactive content; a user interaction display module for displaying plans to users, collecting execution feedback, displaying reports and status indicators, and including at least one of the following data interaction modes: periodic data entry interface, custom indicator tracking interface, and natural language question and answer interface; and a data storage module for storing user historical data, model parameters, and knowledge base content.

[0017] Furthermore, the data acquisition module receives input data through a user terminal interface, and the AI ​​classification decision module and the solution generation module interact with each other via a RESTful API.

[0018] Compared with the prior art, the beneficial effects of the present invention are:

[0019] 1. Two-dimensional feature fusion modeling: fusing initial feature labels with dynamic behavioral tendency values ​​to create a more accurate user status profile;

[0020] 2. AI-driven personalization: Achieving dynamic adjustments based on real behavioral data through decision trees and weighted models;

[0021] 3. Closed-loop feedback mechanism: User behavior data can generate visual reports and status identification systems. Attached Figure Description

[0022] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0023] Figure 1This is a system architecture diagram of a method and system for generating dietary therapy plans based on the Five Elements theory of Traditional Chinese Medicine, according to an embodiment of the present invention.

[0024] Figure 2 This is a flowchart of a two-dimensional feature fusion modeling process for a dietary therapy plan generation method and system based on the Five Elements theory of Traditional Chinese Medicine, according to an embodiment of the present invention.

[0025] Figure 3 This is an AI classification and scheme generation logic block diagram of a dietary therapy scheme generation method and system based on the Five Elements theory of Traditional Chinese Medicine, according to an embodiment of the present invention.

[0026] Figure 4 This is a schematic diagram of the user interface of a method and system for generating dietary therapy plans based on the Five Elements theory of Traditional Chinese Medicine, according to an embodiment of the present invention. Detailed Implementation

[0027] Embodiments of the present invention are described in detail below, with examples of the embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0028] The following describes, with reference to the accompanying drawings, a method and system for generating a dietary therapy plan based on the Five Elements theory of Traditional Chinese Medicine.

[0029] like Figures 1 to 4 As shown, an embodiment of the present invention provides a method for generating a dietary therapy plan based on the Five Elements theory of Traditional Chinese Medicine, comprising the following steps:

[0030] S1: Obtain initial user characteristic data and dynamic behavioral characteristic data. The method does not involve the diagnosis, treatment or prevention of diseases, but is only used for health education and dietary guidance.

[0031] It should be noted that the initial feature data includes the user's date and time of birth, and the dynamic behavioral feature data includes diet records, sleep records, exercise records, water intake records, mood scores, and physiological state feedback data.

[0032] S2: Generate a two-dimensional feature fusion model for users based on preset encoding rules and weighting algorithms.

[0033] like Figure 2As shown, the preset encoding rule is as follows: Birth date and birth time are used as input features. A preset temporal encoding algorithm generates one of five feature labels. The temporal encoding algorithm includes: converting the birth date into year, month, and day values; converting the birth time into hour and minute values; and generating a multi-dimensional feature vector through weighted summation and normalization. The weight coefficients of the weighted summation are determined based on statistical analysis of the distribution characteristics of birth dates and birth times in the sample data. In one specific implementation, these coefficients can be set as follows: year weight 0.2, month weight 0.3, day weight 0.1, hour weight 0.3, and minute weight 0.1. The multi-dimensional feature vector is mapped to one of five cluster centers. These five cluster centers are obtained by training 5000 birth date and time samples using K-means clustering. The sample collection time range is within the past five years, and the sample distribution covers different years and seasons. The samples have also undergone representativeness testing to ensure coverage of different regions, ages, and genders.

[0034] It should be noted that the construction process of the dual-dimensional feature fusion model includes: mapping the initial feature data into five initial state labels through a preset encoding mapping table. The encoding mapping table takes birth date and birth time as input and outputs one of the five corresponding feature labels. A weighted scoring algorithm is applied to the dynamic behavioral feature data to generate acquired behavioral state tendency values. The weight of each behavioral dimension in the weighted scoring algorithm is dynamically adjusted by the user's initial state labels. Specifically, the statistical correlation between the initial state labels and each behavioral dimension is calculated based on the Pearson correlation coefficient in the training dataset, and the weights are adjusted according to the magnitude of the correlation. The weights of the corresponding dimensions are dynamically adjusted. The training dataset contains 500 user behavior questionnaire samples, which were collected through an online survey platform. The questionnaire content covers dimensions such as eating frequency, sleep duration, exercise intensity, drinking habits, emotional state, and physiological feedback. The sample selection criteria are: complete and consistent data. The statistical correlation has verifiable statistical significance on the training dataset, with a significance level of p < 0.05. The significance test uses the Pearson correlation coefficient t-test method, combining the initial state label with the acquired behavioral state tendency value to form a two-dimensional feature vector.

[0035] Specifically, in one implementation, the K-means clustering training is performed 100 times, the initial cluster centers are randomly selected, and the convergence condition is that the change in cluster centers is less than 0.001.

[0036] Specifically, those skilled in the art can set the aforementioned preset weighting coefficients according to the actual application scenario.

[0037] S3: As Figure 3As shown, the two-dimensional feature vector is input into a pre-trained decision tree model or K-means clustering model. The model optimizes the model parameters through K-fold cross-validation to reduce the risk of overfitting the model on user behavior data, improve classification stability, and output a user physical condition profile.

[0038] It should be noted that the output format of the physical condition profile includes: the user's initial state label, the acquired behavioral state tendency value including at least one of the following: damp-heat tendency, qi deficiency tendency, yang deficiency tendency, yin deficiency tendency, and phlegm-dampness tendency, and the health maintenance dimensions that currently require the most attention include at least one of the following: drinking water, sleep, diet, and exercise. The maximum depth of the decision tree model is dynamically adjusted according to the features of the training data, and the K value of the K-means clustering model is consistent with the number of initial state labels.

[0039] S4: Based on the physical condition profile, call the pre-stored dietary knowledge base to generate non-medical personalized dietary advice and health science content.

[0040] It should be noted that the process of generating personalized dietary recommendations includes: calling a pre-stored dietary knowledge base, matching corresponding food recommendations, dietary pairings, and taboo reminders based on the physical condition profile. The dietary knowledge base is a relational database, with the physical condition tag as the primary key, and is associated with a food recommendation table, a recipe pairing table, and a taboo reminder table. It also incorporates intangible cultural heritage elements and is presented in the form of text, images, or multimodal interactive formats.

[0041] S5: Collect execution feedback through the user interaction module and dynamically update the physical condition profile.

[0042] It should be noted that the dynamic feedback and report generation includes: generating weekly behavioral data reports, monthly trend analysis reports, and annual panoramic data reports, and dynamically generating status identifiers based on user execution data.

[0043] It should be noted that the status identifiers are data tags that are automatically triggered based on user behavior data, including: when a user's water intake data reaches a preset threshold for 7 consecutive days, an "Water Intake Met" identifier is automatically generated; when a user's exercise data reaches a preset threshold for 3 consecutive days, an "Exercise Continues" identifier is automatically generated; when a user with a tendency towards damp heat has normal bowel movements for 5 consecutive days, an "Early to Bed Met" identifier is automatically generated; when a user falls asleep before 11 PM for 7 consecutive days, an "Early to Bed Met" identifier is automatically generated; and when a user has no record of overeating for 5 consecutive days, an "70% Full" identifier is automatically generated.

[0044] Specifically, the preset thresholds are set according to the daily water intake standards recommended by the World Health Organization. In one specific implementation, the preset threshold for water intake is 1,500 ml per day, the preset threshold for exercise is 6,000 steps or equivalent exercise per day, and the preset threshold for sleep is falling asleep before 11 p.m. per day.

[0045] like Figures 1 to 4 As shown in the figure, an embodiment of the present invention provides a dietary therapy plan generation system based on the Five Elements theory of Traditional Chinese Medicine, comprising: a data acquisition module for acquiring initial user feature data and dynamic behavioral feature data; a feature fusion modeling module, executed in the processor of a cloud server, communicating with the data acquisition module through an API interface, for performing user data preprocessing, initial label mapping, and behavioral tendency value calculation; an AI classification decision module, with a built-in trained decision tree model or K-means clustering model, for generating a user's physical condition profile; a plan generation module for retrieving corresponding suggestions from a dietary therapy knowledge base and generating text, images, or multimodal interactive content; a user interaction display module for displaying plans to users, collecting execution feedback, displaying reports and status indicators, and including at least one of the following data interaction modes: periodic data entry interface, custom indicator tracking interface, natural language question and answer interface; and a data storage module for storing user historical data, model parameters, and knowledge base content.

[0046] It should be noted that the data acquisition module receives input data through the user terminal interface, and the AI ​​classification decision module and the solution generation module interact with each other via a RESTful API.

[0047] Specifically, user terminals include, but are not limited to, mobile terminals, personal computers, and smart devices.

[0048] Example 1: User Initialization and Physical Condition Profile Generation

[0049] Users input their date and time of birth through a terminal device and complete a behavioral habit questionnaire. The system encodes the date and time of birth as initial status labels, outputs a profile of the user's current physical condition through a decision tree model, and marks the current highest priority dimension for improvement.

[0050] In one specific implementation, the weight coefficients of the weighted summation are set in the same way as in S2 above. The cluster centers are obtained by training 5,000 birth date and time samples using K-means clustering. The sample collection time range is within the past 5 years. The training dataset contains 500 user behavior questionnaire samples, and the statistical significance level is set to p<0.05.

[0051] Example 2: Daily Behavior Tracking and Status Report Generation

[0052] Users fill in behavioral data such as diet, sleep, exercise, and water intake every day. The system only records the data for the first 7 days. On the 8th day, the AI ​​model outputs a "behavioral observation report", which includes: a comparative analysis of the user's initial characteristics and current behavioral status, the behavioral dimensions that are currently recommended to be focused on, and specific dietary and behavioral adjustment references.

[0053] Example 3: Custom Data Tracking and Trend Analysis

[0054] Users can customize and record indicators such as "menstrual period date", "weight value", and "mood score". The system generates a line chart of data changes every week, a monthly report every month, and a panoramic data report at the end of the year.

[0055] Example 4: Interactive Health Science Popularization Consultation

[0056] Users can input "I've been waking up at 3 a.m. lately" through the terminal interface. The system will then retrieve corresponding suggestions from the knowledge base based on the user's physical condition profile and output them in a non-medical health science format, such as: "It is recommended that you enter a sleep preparation state before 11 p.m. and reduce screen time before bed. This suggestion is for health science reference only."

[0057] Compliance Statement

[0058] The system and method described in this invention are only for non-medical health education and dietary guidance services. They do not involve any form of medical diagnosis, treatment, intervention or effect evaluation. All output content does not contain medical advice and does not replace the diagnosis and treatment of doctors or medical institutions. If users have physical discomfort or disease symptoms, they should seek medical attention in time and follow professional medical advice.

[0059] In the description of this specification, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0060] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0061] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A food regimen generation method based on the five elements theory of traditional Chinese medicine, characterized in that, Includes the following steps: S1: Obtain initial user characteristic data and dynamic behavioral characteristic data. The method does not involve the diagnosis, treatment or prevention of diseases, but is only used for health education and dietary guidance. S2: Generate a two-dimensional feature fusion model for users based on preset encoding rules and weighting algorithms; The preset encoding rule is as follows: the birth date and birth time are used as input features, and one of five feature labels is generated through a preset time-series encoding algorithm. The time-series encoding algorithm includes: converting the birth date into year, month, and day values, converting the birth time into hour and minute values, and generating a multi-dimensional feature vector through weighted summation and normalization. The weight coefficients of the weighted summation are preset weight coefficients. The multidimensional feature vector is mapped to one of five cluster centers, which are obtained by training a predetermined number of samples using K-means clustering. The samples are from publicly available demographic datasets or anonymous user data collected in compliance with company regulations. The collection time range is within a predetermined time range, and the sample distribution covers different years and seasons. The samples have undergone representativeness testing to ensure coverage of different regions, ages, and genders. S3: Input the two-dimensional feature vector into a pre-trained decision tree model or K-means clustering model. The model optimizes the model parameters through K-fold cross-validation to reduce the risk of overfitting the model on user behavior data, improve classification stability, and output a user physical condition profile. S4: Based on the physical condition profile, call the pre-stored dietary knowledge base to generate non-medical personalized dietary advice and health science content; S5: Collect execution feedback through the user interaction module and dynamically update the physical condition profile. 2.The food regimen generation method based on the five elements theory of traditional Chinese medicine according to claim 1, characterized in that, In S1, the initial feature data includes the user's date and time of birth, and the dynamic behavioral feature data includes diet records, sleep records, exercise records, water intake records, mood scores, and physiological state feedback data. 3.The food regimen generation method based on the five elements theory of traditional Chinese medicine of claim 1, wherein, In S2, the process of constructing the two-dimensional feature fusion model includes: The initial feature data is mapped to five initial state labels through a preset encoding mapping table. The encoding mapping table takes the birth date and birth time as input and outputs one of the five corresponding feature labels. A weighted scoring algorithm is used to generate acquired behavioral state tendency values ​​for dynamic behavioral feature data. The weights of each behavioral dimension in the weighted scoring algorithm are dynamically adjusted by the user's initial state label. Specifically, the statistical correlation between the initial state label and each behavioral dimension is calculated based on the Pearson correlation coefficient in the training dataset. The weights of the corresponding dimensions are automatically adjusted according to the magnitude of the correlation. The training dataset contains a predetermined number of user behavior questionnaire samples, which are collected through an online survey platform. The questionnaire content covers dimensions such as eating frequency, sleep duration, exercise intensity, drinking habits, emotional state, and physiological feedback. The sample selection criteria are: complete and consistent data. The statistical correlation has verifiable statistical significance on the training dataset, with a significance level of p < 0.

05. The significance test uses the t-test method of Pearson correlation coefficient. The initial state label is combined with the acquired behavioral state tendency value to form a two-dimensional feature vector. 4.The food regimen generation method based on the five elements theory of traditional Chinese medicine of claim 1, wherein, In S3, the output format of the physical condition profile includes: the user's initial state label; Acquired behavioral state tendency values ​​include at least one of the following: damp-heat tendency, qi deficiency tendency, yang deficiency tendency, yin deficiency tendency, and phlegm-dampness tendency; The most important dimensions of health maintenance that need to be focused on at present include at least one of the following: drinking water, sleep, diet, and exercise; The maximum depth of the decision tree model is dynamically adjusted based on the features of the training data, and the K value of the K-means clustering model is consistent with the number of initial state labels. 5.The food regimen generation method based on the five elements theory of traditional Chinese medicine according to claim 1, characterized in that, In S4, the process of generating personalized dietary recommendations includes: calling the pre-stored dietary knowledge base and matching corresponding food recommendations, dietary combinations and taboo prompts according to the physical condition profile; The dietary knowledge base is a relational database, with body constitution status tags as the primary key, and linked to food recommendation tables, recipe combination tables, and taboo reminder tables; Incorporating elements of intangible cultural heritage, presented in the form of text, images, or multimodal interactive formats.

6. The method for generating a dietary therapy plan based on the Five Elements theory of Traditional Chinese Medicine according to claim 1, characterized in that, In S5, the dynamic feedback and report generation includes: generating weekly behavioral data reports, monthly trend analysis reports, and annual panoramic data reports; and dynamically generating status identifiers based on user execution data.

7. The food regimen generation method based on the five elements theory of traditional Chinese medicine according to claim 6, characterized in that, The status identifier, which is a data tag automatically triggered based on user behavior data, includes: When a user's drinking water data reaches a preset threshold for 7 consecutive days, an "drinking water meets the standard" label will be automatically generated. When a user's exercise data reaches a preset threshold for 3 consecutive days, an "exercise continuity" label is automatically generated. When a user with a tendency towards damp heat has normal bowel movements for 5 consecutive days, a "stable condition" label will be automatically generated. When a user falls asleep before 11 PM for 7 consecutive days, an "early sleep target met" label is automatically generated. When a user has no record of overeating for 5 consecutive days, a "70% full" label is automatically generated. 8.A food regimen generation system based on a five-element theory of traditional Chinese medicine, characterized by include: The data acquisition module is used to acquire initial user characteristic data and dynamic behavioral characteristic data; The feature fusion modeling module runs on the processor of the cloud server and communicates with the data acquisition module through the API interface. It is used to perform user data preprocessing, initial label mapping and behavioral tendency value calculation. The AI ​​classification and decision-making module has a built-in pre-trained decision tree model or K-means clustering model to generate a user's physical condition profile. The solution generation module is used to retrieve corresponding suggestions from the dietary knowledge base and generate text, images, or multimodal interactive content. The user interaction display module is used to display the plan to the user, collect execution feedback, display reports and status indicators, and includes at least one of the following data interaction modes: periodic data entry interface, custom indicator tracking interface, and natural language question and answer interface. The data storage module is used to store user historical data, model parameters, and knowledge base content. 9.The food regimen generation system based on the five elements theory of traditional Chinese medicine according to claim 8, wherein, The data acquisition module receives input data through the user terminal interface, and the AI ​​classification decision module and the solution generation module interact with each other via a RESTful API.

Images