A personalized medicinal tea drink configuration system and method based on physical constitution recognition

Abstract

The application discloses a kind of individualized medicinal tea drink configuration system and method based on physique identification, specifically related to medicinal tea drink configuration technical field, for solving the problem that individualized medicinal tea drink configuration is not clear;The present application generates user state vector by fusing the static physique data and real-time physiological data of user, and environmental data, and extracts features, analyzes the relationship between physique and environment, generates individualized medicinal tea drink formula and medicinal material proportion, and then dynamically adjusts the medicinal material proportion in tea drink formula by real-time monitoring of environmental change and physique fluctuation, to adapt to different health status and environmental conditions, at the same time, by collecting user feedback data, combined with incremental learning technology, continuously optimize the formula, so as to adaptively adjust according to the change and feedback of user's demand, improve the individualization and dynamic adaptation ability of medicinal tea drink recommendation, help user obtain tea drink formula more in line with its health needs and taste preference, improve health management experience.

Description

Technical Field

[0001] This invention relates to the field of medicinal tea preparation technology, and more specifically, to a personalized medicinal tea preparation system and method based on constitution identification. Background Technology

[0002] Medicinal tea, as a health care method that combines traditional Chinese medicine theory with modern dietary concepts, combines herbs with tea or other natural ingredients and regulates the body, nourishes health, and prevents diseases according to the theories of yin and yang, cold and heat in traditional Chinese medicine.

[0003] With the fast pace of modern life and the diversification of health needs, people's focus on medicinal teas is no longer limited to taste and effects, but increasingly leans towards personalized customization and targeted conditioning. The medicinal tea market is also constantly developing, gradually moving towards personalization and intelligentization.

[0004] The shortcomings of existing technologies: Most existing medicinal diet and tea recommendation systems rely on static personalized recommendation models, which usually only recommend formulas based on the user's physical characteristics (such as body temperature, heart rate, etc.) or environmental data (such as temperature, humidity, etc.). They lack a real-time response mechanism for environmental changes and fluctuations in the user's physical condition, cannot handle the user's dynamic feedback, and have a long update cycle. As a result, the recommendation results cannot adapt to changes in user needs in a timely manner, nor can they continuously optimize formula recommendations through real-time feedback. This limits the flexibility and accuracy of personalized recommendation systems. The systems usually only provide a single recommended formula, ignoring the user's real-time changes in needs under different health states and environmental conditions. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a personalized medicinal tea preparation system and method based on constitution identification, so as to solve the problem of unclear personalized medicinal tea preparation in the above-mentioned background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for preparing personalized medicinal tea based on body constitution identification includes the following steps:

[0008] Collect users’ static physical condition data and real-time physiological data, acquire environmental data through sensors, fuse physical condition and environmental data through a nonlinear mapping function, and generate a user state vector that includes changes in user physical condition and environment.

[0009] Features are extracted from the fused user state vector, the relationship between physical constitution and environmental data is analyzed, and personalized medicinal diet tea recipes are output and the proportion of medicinal materials is calculated to form customized recipes for users.

[0010] The changes are calculated based on real-time environmental data and fluctuations in the user's physical condition, and the proportions of medicinal herbs in the formula are adjusted to form an optimized medicinal tea formula.

[0011] We obtain data on the taste and health effects of medicinal tea drinks through user feedback, and optimize the tea drink formulas through incremental learning.

[0012] In a preferred embodiment, the user's static physical condition data and real-time physiological data are collected, and environmental data is acquired through sensors. The specific process is as follows:

[0013] Users fill in their physical condition information through health records or questionnaires. The physical condition information includes physical condition type, age, gender, body type, allergy history, and family medical history. The physical condition information is then converted into structured data to form static physical condition data.

[0014] Users provide real-time physiological data through smart bracelets and smartwatches. The real-time physiological data includes heart rate, body temperature, blood pressure, blood sugar, sleep quality, and activity level. The real-time physiological data is stored in the form of time series.

[0015] Environmental data collection is conducted by deploying temperature and humidity sensors and air quality monitoring devices to collect real-time climate conditions of the user's environment, including temperature, humidity, air quality, PM2.5 concentration, and air pressure. Combined with the weather API interface, real-time weather information of the user's location is obtained, including current temperature, humidity, and wind speed.

[0016] In a preferred embodiment, physical condition and environmental data are fused using a nonlinear mapping function to generate a user state representation that includes changes in the user's physical condition and environment. The specific process is as follows:

[0017] Synchronize user physical condition data and environmental data over time. Based on the timestamp of each data point, align the user physical condition data and environmental data using the timestamp to correctly match the physical condition and environmental data at each moment.

[0018] After aligning the physical fitness data and environmental data in time, state alignment is performed to form two time series: a physical fitness data series and an environmental data series. The physical fitness data series and the environmental data series are then merged to form a unified input stream.

[0019] For each pair of user physical and environmental data, a nonlinear function is used to calculate the interaction effect, and the interaction features are generated and represented by the nonlinear function: ,in, These are physical fitness data at time t; It is the environmental data at time t; , is the model's learning weight, representing the relative influence of physical condition and environmental data; b is the bias term, used to adjust the model's output; It is an activation function;

[0020] All interactive features By aggregation, we obtain the user state vector U(t), which is represented as: ,in, It represents the interaction features between each pair of physical condition and environmental data, where n is the number of feature pairs.

[0021] In a preferred embodiment, features are extracted based on the fused user state vector, the relationship between physical constitution and environmental data is analyzed, and a personalized medicinal tea recipe is output and the proportion of medicinal materials is calculated to form a user-customized recipe. The specific steps are as follows:

[0022] Chi-square test was used to select features that have a significant impact on tea beverage recipe recommendations.

[0023] Chi-square test was used to select physical and environmental characteristics that were positively correlated with the effects of medicinal tea formulas;

[0024] Based on the feature selection results, the features most relevant to the generation of medicinal tea recipes are selected.

[0025] The selected features, after being standardized and normalized, are passed to the recommendation engine;

[0026] Based on the requirements of the task, choose a deep neural network model or a gradient boosting decision tree as the recommendation model;

[0027] Based on the features of the given input, a personalized medicinal tea recipe is obtained through a recommendation model. For each medicinal herb, the recommended proportion is calculated based on the user's physical condition and environmental conditions, and the recommended proportion of each medicinal herb is used as the initial recipe output.

[0028] The initial formula is used, and user feedback on the initial formula is used to generate a user-customized formula.

[0029] In a preferred embodiment, the preliminary formula is used, and a user-customized formula is generated based on user feedback on the preliminary formula. The specific process is as follows:

[0030] After generating the initial recipe, feedback is provided based on the user's taste preferences and actual needs, and the parameters of the recommendation model are dynamically adjusted through feedback-based reinforcement learning.

[0031] The reinforcement learning update formula is: ,in, In the state Take action below The current value; For learning rate, Discount factor; For instant rewards; For the next state The optimal value under the given conditions;

[0032] During the formula generation process, the recommended medicinal ingredients are adjusted based on real-time feedback. The adaptive adjustment formula is as follows: ,in, P(t) represents the generated tea beverage formula, which is an adjustment factor for the formula based on changes in the current environment or the user's physical condition. This indicates the adaptively adjusted medicinal tea recipe;

[0033] Real-time feedback includes the user's health goals, preferences, and current changes in physical condition;

[0034] The final recommended medicinal tea recipes will be provided to users again as customized recipes.

[0035] In a preferred embodiment, the change is calculated based on real-time environmental data and fluctuations in the user's physical condition, and the proportions of the medicinal herbs in the formula are adjusted to form an optimized medicinal tea formula. The specific process is as follows:

[0036] Environmental change monitoring is conducted by acquiring real-time environmental change data through sensors installed in the user's environment; the degree of change in the current environment is determined by analyzing the environmental change data.

[0037] By acquiring real-time environmental change data, the proportion of medicinal herbs in medicinal teas can be adjusted according to the degree of change in the current environment.

[0038] Analyze and adjust the fluctuations in the user's physical condition in real time, and adjust the formula of the medicinal tea drink according to the immediate changes in the user's physical condition;

[0039] Analyze users' physical condition change data, identify trends in physical condition changes, determine the amount of physical condition changes, and adjust the proportion of medicinal materials in the medicinal diet tea formula according to the amount of physical condition changes;

[0040] The formula for comprehensively adjusting medicinal tea recipes involves combining changes in environmental factors and changes in physical constitution. The formula for comprehensive fine-tuning of medicinal tea recipes is as follows: Wherein, P(t) is the initially generated medicinal tea recipe; This is the final adjusted formula. It is an adjustment factor for the fine-tuning of the formula in response to environmental changes. It is an adjusting factor that allows for minor adjustments to the formula based on changes in body constitution. This represents the change in physical condition. This refers to environmental changes.

[0041] A personalized medicinal tea preparation system based on body constitution identification, used to implement the aforementioned personalized medicinal tea preparation method based on body constitution identification, includes:

[0042] The physical condition and environment analysis module is used to collect users’ static physical condition data and real-time physiological data, acquire environmental data through sensors, and fuse physical condition and environmental data through a nonlinear mapping function to generate a user state vector that includes changes in user physical condition and environment.

[0043] The initial formula generation module is used to extract features based on the fused user state vector, analyze the relationship between physical constitution and environmental data, and output personalized medicinal diet tea formulas and calculate the proportion of medicinal materials to form user-customized formulas.

[0044] The formula adjustment module is used to calculate the amount of change based on real-time environmental data and fluctuations in the user's physical condition, and adjust the proportion of medicinal materials in the formula to form an optimized medicinal tea formula;

[0045] The feedback optimization module is used to obtain data on the taste and health effects of medicinal tea drinks through user feedback, and to optimize the tea drink formula through incremental learning.

[0046] The technical effects and advantages of this invention are as follows:

[0047] This invention collects users' static physical condition data and real-time physiological data, and combines them with environmental data. Through a nonlinear mapping function, it fuses the data to generate a unified state vector that includes changes in users' physical condition and environment, providing a data foundation for personalized medicinal diet and tea recipe recommendations.

[0048] Based on the fused user state vector, the system extracts key features and analyzes the relationship between constitution and environment to generate and calculate personalized medicinal tea recipes and ingredient ratios, ensuring that the recommended teas meet the user's health needs. By monitoring environmental changes and fluctuations in user constitution in real time, the system can dynamically adjust the ingredient ratios and optimize the tea recipes in real time to adapt to different environments and constitution changes. Furthermore, by collecting user feedback data and combining incremental learning technology, the system continuously optimizes the tea recipes, ensuring that the recommendation system can adaptively adjust to changes in user needs and feedback. This effectively improves the personalization, accuracy, and dynamic adaptability of medicinal tea recommendations, providing users with tea recipes that better meet their health needs and taste preferences, and improving the user's health management experience. Attached Figure Description

[0049] Figure 1 This is a flowchart of a personalized medicinal tea preparation method based on body constitution identification according to the present invention.

[0050] Figure 2 This is a schematic diagram of the structure of a personalized medicinal tea preparation system based on body constitution identification according to the present invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Example 1: As Figure 1 As shown, a personalized medicinal tea preparation method based on constitution identification includes the following steps:

[0053] Collect users’ static physical condition data and real-time physiological data, acquire environmental data through sensors, fuse physical condition and environmental data through a nonlinear mapping function, and generate a user state vector that includes changes in user physical condition and environment.

[0054] Features are extracted from the fused user state vector, the relationship between physical constitution and environmental data is analyzed, and personalized medicinal diet tea recipes are output and the proportion of medicinal materials is calculated to form customized recipes for users.

[0055] The changes are calculated based on real-time environmental data and fluctuations in the user's physical condition, and the proportions of medicinal herbs in the formula are adjusted to form an optimized medicinal tea formula.

[0056] We obtain data on the taste and health effects of medicinal tea drinks through user feedback, and optimize the tea drink formulas through incremental learning.

[0057] Step 1: Collect and integrate user physical condition and environmental data. In personalized medicinal tea preparation, the user's physical condition and environmental status are key factors in determining the recommended formula. First, it is necessary to collect and integrate the user's physical condition information and changes in the environment in real time. Through this process, we can accurately understand each user's health status and the impact of the external environment on the user's health, thereby generating the most suitable personalized medicinal tea formula. The specific steps are as follows:

[0058] Collect basic physical information of users. Users can fill in their basic physical information through health records or questionnaires. Physical information includes physical type (such as cold, hot, damp, weak, etc.), age, gender, body type, allergy history, family medical history, etc. The definition of physical type can provide a preliminary understanding of the user's health characteristics.

[0059] Transform physical condition information into structured data, namely the user's static physical condition data, such as: body type (thin / medium / obese); allergy history (yes / no); basic health conditions (such as whether or not they have diabetes, high blood pressure, etc.).

[0060] Users provide real-time physiological data through wearable devices (such as smart bracelets and smartwatches), which typically include heart rate, body temperature, blood pressure, blood sugar, sleep quality, and activity level. This physiological data reflects the user's current physical state and helps understand the user's dynamic physical condition. Real-time physiological data is stored in time series format, for example: ,in, This represents data such as heart rate and body temperature at time i.

[0061] Static physical fitness data and real-time physiological data are combined and used as user physical fitness data as input data, and transformed into a multi-dimensional feature vector that contains both static physical fitness information and dynamic physiological data.

[0062] Environmental data is collected by deploying devices such as temperature and humidity sensors and air quality monitors to collect real-time climate conditions of the user's environment, including temperature, humidity, air quality, PM2.5 concentration, and air pressure. At the same time, by combining weather API interfaces, real-time weather information of the user's location can be obtained, such as current temperature, humidity, and wind speed.

[0063] Environmental data is standardized to ensure that data from different devices and sensors are on the same scale. For example, temperature data is standardized to the range of [0, 1]. Filters (such as Kalman filters) are used to smooth environmental data and remove invalid fluctuations caused by noise interference in sensor data.

[0064] Assuming the environmental data is ,in To represent environmental variables (temperature, humidity, air quality, etc.), each Normalization will be performed to ensure that the data for each environmental variable are consistent and comparable;

[0065] Environmental data is viewed as a dynamic input stream, and its impact on users' physical condition is constantly changing, thus requiring real-time data updates.

[0066] The user's physical condition data is integrated with environmental data to generate a unified user status representation. The specific steps are as follows:

[0067] Time synchronization is performed. User physical condition data and environmental data are collected in real time through different devices and sensors and synchronized with time. Each data point is marked with a timestamp 't', for example, user body temperature data. The ambient temperature data was collected at time t. It is also collected at time t; the two data sources are aligned using timestamp t to ensure that the physical condition and environmental data at each time point can be correctly matched;

[0068] After aligning the physical and environmental data in terms of time, two time series are formed: one for physical data and the other for environmental data. and environmental data sequences These time series are merged to form a unified input stream, where n is the total number of time series.

[0069] By combining user physical condition data and environmental data using high-order feature extraction methods, the complex interaction between the two can be captured. Since the responses of different physical condition types to the environment are non-linear, a non-linear mapping function is used to handle the interactions between the data. The specific steps are as follows:

[0070] For each pair of data related to physical constitution and environment, a nonlinear function is used to calculate the interaction effect. That is, for each pair of user physical constitution data and environmental data, a nonlinear function is used to calculate the interaction effect. The generation of interaction features can be achieved through the following nonlinear function: ,in, These are physical data at time t, such as body temperature and heart rate; It refers to environmental data at time t, such as temperature and humidity. , is the model's learning weight, representing the relative influence of physical condition and environmental data; b is the bias term, used to adjust the model's output; These are activation functions, commonly including ReLU, Sigmoid, or Tanh, which help capture non-linear relationships;

[0071] All interactive features After aggregation, the final user state vector U(t) is obtained, which can be represented as: ,in, It represents the interaction features of each pair of physical constitution and environmental data, where n is the number of feature pairs;

[0072] This final user state representation vector not only contains static information about the user's constitution, but also incorporates the dynamic impact of environmental changes on the constitution. For example, certain constitution types (such as cold constitution) are more sensitive to changes in external temperature. By analyzing the user state representation, the recommended medicinal tea recipes can be adjusted. This complete data flow from the user's constitution to the environment forms an accurate user state vector, which can provide a data foundation for subsequent personalized medicinal tea recommendations.

[0073] Step 2 involves personalized medicinal tea recipe recommendations. This step utilizes AI technology to intelligently analyze the fused data, generating medicinal tea recipes that match the user's constitution and environmental needs. By understanding changes in constitution and environment, the recommendation engine can flexibly adjust the recommended medicinal ingredients, ensuring that the recommended teas not only meet the user's health needs but also have practical efficacy. The specific steps are as follows:

[0074] Extract the most important features for recommendation from the fused user state vector and perform necessary preprocessing, namely the fused user state vector in step 1. The user state vector contains the combined influence of user physical condition data and environmental data, and is a multi-dimensional feature vector. This state vector will be passed as input to the recipe recommendation engine.

[0075] The most influential features for generating medicinal tea recipes are extracted from the fused user state vector. Since user state vectors usually contain multiple features, the purpose of feature selection is to reduce redundant features, remove irrelevant features, and retain features that are crucial to recipe recommendation. The steps are as follows:

[0076] Statistical methods such as information gain and chi-square test are used to select features that significantly influence tea beverage recipe recommendations. For example, the relationship between a user's body temperature and the external temperature has a significant impact on the recommended recipe. The system will select key features based on this information. The chi-square test formula is as follows: ,in, It is the frequency of observed information i. is the frequency of the expected information i, and n is the total number of information items;

[0077] Through chi-square test, select physical and environmental characteristics that are strongly correlated (positively correlated) with the target variable (the effect of medicinal tea formula);

[0078] Based on the feature selection results, the k features most relevant to the generation of medicinal tea recipes were finally selected. These characteristics can be considered the most crucial factors for recommending medicinal teas.

[0079] At the same time, data standardization and normalization are performed. Data standardization and normalization are preprocessing steps in machine learning. The purpose is to eliminate the differences in the units of measurement between features, ensure that the influence of different features on the model is on the same order of magnitude, and avoid certain features from becoming dominant due to their large numerical range.

[0080] Assume the selected eigenvector is The data may have different scales (e.g., body temperature range is 36-38 degrees Celsius, air quality range is 0-100). Standardization transforms each feature into zero mean and unit variance, allowing different features to be compared on the same scale. The standardization formula is: , These are the original eigenvalues; Let be the mean of the i-th feature; Let be the standard deviation of the i-th feature;

[0081] The eigenvalues ​​are restricted to a uniform range (e.g., [0, 1]) and normalized. The normalization formula is: ,in, These are the original eigenvalues; and These are the minimum and maximum values ​​of the i-th feature, respectively;

[0082] Data after standardization and normalization The standardized input features are then passed to the recommendation engine.

[0083] Based on the input user feature vector, generate personalized medicinal tea recipes that best meet the user's needs, and transform the user's physical condition and environmental conditions into specific medicinal recipe recommendations;

[0084] Depending on the task requirements, different models are selected to handle the recommendation problem. Commonly used recommendation models include: Deep Neural Networks (DNN), which learns the complex relationship between physical condition and environmental data through multilayer perceptron networks; and Gradient Boosting Decision Trees (GBDT), which captures the relationship between user status and tea recipes through a tree structure model.

[0085] The recommendation model is trained by training a dataset (historical records of physical condition, environment, and recipes), enabling the model to learn the complex nonlinear relationship between medicinal tea recipes and users' physical condition and environment.

[0086] Given the input feature vector X(t), a personalized medicinal tea recipe P(t) is obtained through the recommendation model M, that is: , where P(t) represents the generated tea recipe, which includes the proportion of each herb.

[0087] For each medicinal herb, the system will calculate a recommended ratio based on the user's physical condition and environmental conditions. This recommended ratio will be used as the initial prescription output.

[0088] The initial formula is used, and user feedback on the initial formula is used to generate a user-customized formula;

[0089] After generating the initial recipe, the system needs to adjust it based on the user's taste preferences and actual needs. Through adaptive algorithms, the recommendation system can dynamically adjust the tea recipe based on user feedback to ensure that the recommended recipe meets both health requirements and taste preferences. The specific steps are as follows:

[0090] User feedback on medicinal tea drinks (such as taste and effects) will help the system learn users' taste preferences. Through feedback-based reinforcement learning, the system can dynamically adjust the parameters of the recommendation model. For example, if users prefer certain herbs (such as goji berries and chrysanthemums), the system will automatically increase the proportion of these herbs.

[0091] The reinforcement learning update formula is: ,in, In the state Take action below The current value; For learning rate, Discount factor; For immediate rewards (such as taste ratings from user feedback); For the next state The optimal value under the given conditions;

[0092] Adaptive adjustment is performed, meaning that during the formula generation process, the system not only considers the user's constitution and environment, but also adjusts the recommended medicinal diet ingredients based on real-time feedback (such as the user's health goals, preferences, and current changes in constitution). For example, when the climate changes, the system may automatically adjust the proportion of warming or dampness-removing medicinal materials. The adaptive adjustment formula is: ,in, The adjustment ratio of medicinal ingredients is determined by the adjustment factors of the formula based on the current environment or changes in the user's physical condition.

[0093] The final recommended medicinal tea recipe will be provided to the user, which includes the specific proportions of each medicinal herb and its function (such as replenishing qi, clearing heat, etc.).

[0094] After using the medicinal tea, users can submit feedback (such as taste, effects, satisfaction, etc.) through the application or interface. This feedback will be transmitted back to the recommendation system as a basis for future formula optimization.

[0095] Step 3: Based on dynamic environmental adaptation and real-time fine-tuning, after generating personalized medicinal tea recipes, real-time fine-tuning is performed according to changes in the environment (such as temperature, humidity, air quality, etc.) and changes in the user's physical condition (such as body temperature, heart rate, etc.). This ensures that the recommended medicinal teas can be flexibly adjusted according to the current environment and the user's immediate health status, and can continuously optimize the recommendations to better meet the user's health needs and actual situation. The specific steps are as follows:

[0096] Environmental change monitoring is conducted by acquiring real-time external environmental data through sensors installed in the user's environment (such as temperature and humidity sensors, air quality monitors, etc.). Let's assume the environmental data is... ,in, The current temperature; Current humidity; Current humidity;

[0097] By analyzing environmental change data, the system can determine the degree of environmental change. For example, when the external temperature drops sharply, the system may need to add warming and nourishing herbs (such as ginger and goji berries); when the humidity is too high, the system may recommend more dampness-removing herbs (such as Job's tears and red beans).

[0098] Hypothetical factors influencing environmental change The magnitude of environmental change at time t (the amount of environmental change) is given by the formula: Where f is a nonlinear function representing the impact of environmental changes on formula adjustments; For environmental data at time t, The environmental data is for time t−1;

[0099] By acquiring environmental changes in real time The system adjusts the proportion of medicinal herbs in the medicinal tea based on the degree of change in the current environment. The adjustment coefficient for the medicinal herbs... Adjustments can be made based on the degree of environmental change: Wherein, P(t) is the initially generated medicinal tea recipe; It is a formula that has been slightly adjusted according to changes in the environment. It is an adjustment factor for the fine-tuning of the formula due to environmental changes.

[0100] The formula of medicinal tea is fine-tuned based on the user's physical condition fluctuations and real-time adjustments. Since the user's physical condition is dynamic, it is necessary to monitor the user's physical condition in real time and make adaptive adjustments to the formula.

[0101] Body constitution change monitoring: Changes in a user's body constitution are typically reflected in real-time health data (such as heart rate, body temperature, etc.). This data is collected through wearable devices and transmitted to the system. For example, assuming the user's body constitution data is... ,in, This represents the user's body temperature at time t. This represents the user's heart rate at time t.

[0102] The system analyzes users' data on changes in their physical condition to identify trends. For example, when body temperature rises, the system may recommend more herbs with heat-clearing properties (such as chrysanthemum and honeysuckle); when users experience fatigue, the system may recommend herbs with qi-tonifying properties (such as ginseng and astragalus).

[0103] Assuming factors influencing changes in user physical condition The formula for the change in health indicators such as body temperature is: Where gx is a non-linear function representing the effect of changes in physical constitution on the tea beverage formula; This refers to the user's physical condition data at the current moment. This refers to the user's physical condition data from the previous moment;

[0104] Based on the amount of change in body constitution The system will adjust the proportions of medicinal herbs in the herbal tea recipe. The slightly adjusted recipe is as follows: P(t) is the initially generated medicinal tea recipe; It is a formula that has been slightly adjusted based on changes in the user's physical condition; It is the adjustment coefficient of tea beverage formula based on changes in physical condition.

[0105] The environment and physical condition work together to influence the needs of tea recipes. Therefore, by taking into account the influence of both, the proportion of medicinal herbs in tea can be adjusted more precisely. For example, in cold weather and when body temperature is high, the system will increase the proportion of warming and tonifying herbs; while in humid environments and when the body is fatigued, the system will increase the proportion of dampness-removing herbs.

[0106] Taking into account both environmental and physical changes, the formula is fine-tuned to more accurately adapt to the current health status and external conditions. Through a dual adaptation mechanism, the system can provide precise tea formulas in complex and dynamic environments.

[0107] The effects of environmental and physical changes on medicinal tea formulas need to be comprehensively adjusted by combining two adjustment factors. and These represent the effects of environmental changes and changes in physical constitution on formula adjustments, respectively. It is an adjustment factor for the fine-tuning of the formula in response to environmental changes. This is an adjustment factor for formula fine-tuning in response to changes in physical constitution; it can be used to represent the fine-tuning coefficients for changes in environment and physical constitution. The comprehensive fine-tuning formula is: Wherein, P(t) is the initially generated medicinal tea recipe; This is the final adjusted formula.

[0108] Through dynamic environmental adaptation and real-time fine-tuning, the system can flexibly adjust the formula of medicinal tea drinks in response to changes in the environment and the user's physical condition. The combination of environmental change monitoring, physical condition fluctuation monitoring, and a dual adaptation mechanism makes the formula recommendations more accurate and tailored to user needs.

[0109] Step 4: Intelligent optimization based on user feedback. By collecting real-time user feedback, the personalized medicinal tea recommendation system is intelligently optimized to ensure that it can continuously provide the best formulas based on the user's constantly changing physical condition and needs, adjusting and improving the accuracy of recommendations. The specific steps are as follows:

[0110] Collect user feedback after using medicinal tea drinks, understand users' true feelings about the tea drink recipes, and provide new input data for the recommendation engine. Feedback information may include taste, effects, satisfaction, changes in health status, etc.

[0111] Users can provide feedback in various ways (such as apps, smart devices, voice assistants, etc.), and the types of feedback include but are not limited to: taste (sweet, bitter, sour, etc.), physical reaction (whether they feel comfortable, tired, or have indigestion, etc.), and overall satisfaction.

[0112] Each feedback session includes the user ID, feedback type, feedback value, feedback time, etc., forming a structured feedback dataset.

[0113] Assuming user feedback includes information such as taste satisfaction and health effects, sentiment analysis methods can be used to extract sentiment tendencies from the feedback. For example, taste feedback can be used to extract satisfaction ratings from descriptions such as "too sweet" or "too bitter".

[0114] A word vector model (such as Word2Vec) is used to process the user's natural language feedback to obtain a sentiment index, which represents the user's overall satisfaction with the tea drink;

[0115] Extract key features (such as health outcome scores) from the feedback data, and convert these features into numerical form to be passed to the recommendation system as new input features.

[0116] Based on user feedback data, the parameters of the recommendation model are automatically adjusted, enabling the recommendation engine to continuously adapt to user needs. Through adaptive optimization, the system can quickly adjust the recommendation strategy based on feedback from different users, thereby improving the personalization of the recommendations.

[0117] Incremental learning algorithms are used to update the recommendation model based on each user feedback. For example, in deep learning, online learning methods are used to dynamically update the weights of the neural network, allowing the model to adjust the recommendation strategy based on the latest user feedback. The incremental learning formula is: ,in, These are the parameters of the model at time t; The loss function is used to calculate the error between the model's predictions and the user's feedback. The learning rate; ) is the loss function with respect to parameters The gradient represents the impact of feedback on the model update;

[0118] The system compares user feedback with the predicted values ​​output by the current model to generate an adjustment factor. Based on the health effects or satisfaction levels observed in the feedback, the parameters of the recommendation system are adjusted.

[0119] Through incremental learning and feedback-driven adaptive optimization, the recommendation system can adjust its recommendation strategy in real time based on each user's actual feedback. This enables the system to continuously improve its responsiveness to user needs. Such adaptive optimization can greatly improve the accuracy of recommendations and ensure that the system always provides users with tea recipes that meet their health needs.

[0120] In summary, step 4 involves continuously optimizing personalized medicinal tea recipes through user feedback and intelligent optimization mechanisms. By collecting and processing feedback data, incremental learning and online optimization, and a feedback-driven adjustment mechanism, the system can dynamically adjust its recommendation strategy based on users' actual needs, changes in health status, and taste preferences. Ultimately, the system can provide more personalized and accurate tea recommendations and continuously improve recommendation effectiveness and user satisfaction through adaptive optimization.

[0121] It should be noted that the thresholds involved in the embodiments can be determined according to specific scenarios and needs.

[0122] This invention collects users' static physical condition data and real-time physiological data, and combines them with environmental data. Through a nonlinear mapping function, it fuses the data to generate a unified state vector that includes changes in users' physical condition and environment, providing a data foundation for personalized medicinal diet and tea recipe recommendations.

[0123] Based on the fused user state vector, the system extracts key features and analyzes the relationship between constitution and environment to generate and calculate personalized medicinal tea recipes and ingredient ratios, ensuring that the recommended teas meet the user's health needs. By monitoring environmental changes and fluctuations in user constitution in real time, the system can dynamically adjust the ingredient ratios and optimize the tea recipes in real time to adapt to different environments and constitution changes. Furthermore, by collecting user feedback data and combining incremental learning technology, the system continuously optimizes the tea recipes, ensuring that the recommendation system can adaptively adjust to changes in user needs and feedback. This effectively improves the personalization, accuracy, and dynamic adaptability of medicinal tea recommendations, providing users with tea recipes that better meet their health needs and taste preferences, and improving the user's health management experience.

[0124] Example 2: A personalized medicinal tea preparation system based on body constitution identification, such as... Figure 2 As shown, it specifically includes:

[0125] The physical condition and environment analysis module is used to collect users’ static physical condition data and real-time physiological data, acquire environmental data through sensors, and fuse physical condition and environmental data through a nonlinear mapping function to generate a user state vector that includes changes in user physical condition and environment.

[0126] The initial formula generation module is used to extract features based on the fused user state vector, analyze the relationship between physical constitution and environmental data, and output personalized medicinal diet tea formulas and calculate the proportion of medicinal materials to form user-customized formulas.

[0127] The formula adjustment module is used to calculate the amount of change based on real-time environmental data and fluctuations in the user's physical condition, and adjust the proportion of medicinal materials in the formula to form an optimized medicinal tea formula;

[0128] The feedback optimization module is used to obtain data on the taste and health effects of medicinal tea drinks through user feedback, and to optimize the tea drink formula through incremental learning.

[0129] The above formulas are all dimensionless calculations. Dimensionless calculations can be performed using various methods such as standardization, which will not be elaborated here. The formulas are derived from software simulations based on a large amount of collected data, and the preset parameters in the formulas can be set by those skilled in the art according to the actual situation.

[0130] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, ATA hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state ATA hard disk.

[0131] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0132] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0133] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0134] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0135] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0136] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for preparing personalized medicinal tea based on body constitution identification, characterized in that, Includes the following steps: Collect users’ static physical condition data and real-time physiological data, acquire environmental data through sensors, fuse physical condition and environmental data through a nonlinear mapping function, and generate a user state vector that includes changes in user physical condition and environment. Features are extracted from the fused user state vector, the relationship between physical constitution and environmental data is analyzed, and personalized medicinal diet tea recipes are output and the proportion of medicinal materials is calculated to form customized recipes for users. The changes are calculated based on real-time environmental data and fluctuations in the user's physical condition, and the proportions of medicinal herbs in the formula are adjusted to form an optimized medicinal tea formula. We obtain data on the taste and health effects of medicinal tea drinks through user feedback, and optimize the tea drink formulas through incremental learning. By fusing physical condition and environmental data through a nonlinear mapping function, a user state representation that includes changes in the user's physical condition and environment is generated. The specific process is as follows: Synchronize user physical condition data and environmental data over time, align user physical condition data and environmental data using timestamps, and match user physical condition data and environmental data at each moment; After aligning the physical fitness data and environmental data in time, state alignment is performed to form two time series: the physical fitness data series and the environmental data series. For each pair of user physical condition data and environmental data, a nonlinear function is used to calculate the interaction effect and generate interaction features; Aggregate all interaction features to obtain the user state vector; Based on the fused user state vector, features are extracted, the relationship between physical constitution and environmental data is analyzed, and personalized medicinal tea recipes are output and the proportion of medicinal materials is calculated to form user-customized recipes. The specific steps are as follows: Chi-square test was used to select physical and environmental characteristics that were positively correlated with the effects of medicinal tea formulas; Based on the feature selection results, the features most relevant to the generation of medicinal tea recipes are selected. The selected features, after being standardized and normalized, are passed to the recommendation engine; Based on the requirements of the task, choose a deep neural network model or a gradient boosting decision tree as the recommendation model; Based on the features of the given input, a personalized medicinal tea recipe is obtained through a recommendation model. For each medicinal herb, the recommended proportion is calculated based on the user's physical condition and environmental conditions, and the recommended proportion of each medicinal herb is used as the initial recipe output. The initial formula is used, and user feedback on the initial formula is used to generate a user-customized formula; By fusing physical condition and environmental data through a nonlinear mapping function, a user state representation that includes changes in the user's physical condition and environment is generated. The specific process is as follows: Synchronize user physical condition data and environmental data over time. Based on the timestamp of each data point, align the user physical condition data and environmental data using the timestamp to correctly match the physical condition and environmental data at each moment. After aligning the physical fitness data and environmental data in time, state alignment is performed to form two time series: a physical fitness data series and an environmental data series. The physical fitness data series and the environmental data series are then merged to form a unified input stream. For each pair of user physical and environmental data, a nonlinear function is used to calculate the interaction effect, and the interaction features are generated and represented by the nonlinear function: ,in, These are physical fitness data at time t; It is the environmental data at time t; , is the model's learning weight, representing the relative influence of physical condition and environmental data; b is the bias term, used to adjust the model's output; It is an activation function; All interactive features By aggregation, we obtain the user state vector U(t), which is represented as: ,in, It represents the interaction features of each pair of physical constitution and environmental data, where n is the number of feature pairs; The initial formula is used, and user feedback on the initial formula is used to generate a customized formula for the user. The specific process is as follows: After generating the initial recipe, feedback is provided based on the user's taste preferences and actual needs, and the parameters of the recommendation model are dynamically adjusted through feedback-based reinforcement learning. The reinforcement learning update formula is: ,in, In the state Take action below The current value; For learning rate, Discount factor; For instant rewards; For the next state The optimal value under the given conditions; During the formula generation process, the recommended medicinal ingredients are adjusted based on real-time feedback. The adaptive adjustment formula is as follows: ,in, P(t) represents the generated tea beverage formula, which is an adjustment factor for the formula based on changes in the current environment or the user's physical condition. This indicates the adaptively adjusted medicinal tea recipe; Real-time feedback includes the user's health goals, preferences, and current changes in physical condition; The final recommended medicinal tea recipes will be provided to users again as customized recipes. The changes are calculated based on real-time environmental data and fluctuations in the user's physical condition, and the proportions of medicinal herbs in the formula are adjusted to form an optimized medicinal tea formula. The specific process is as follows: Environmental change monitoring is conducted by acquiring real-time environmental change data through sensors installed in the user's environment; the degree of change in the current environment is determined by analyzing the environmental change data. By acquiring real-time environmental change data, the proportion of medicinal herbs in medicinal teas can be adjusted according to the degree of change in the current environment. Analyze and adjust the fluctuations in the user's physical condition in real time, and adjust the formula of medicinal tea drinks according to the immediate changes in the user's physical condition; Analyze users' physical condition change data, identify trends in physical condition changes, determine the amount of physical condition change, and adjust the proportion of medicinal materials in the medicinal diet tea formula according to the amount of physical condition change; The formula for comprehensively adjusting medicinal tea recipes involves combining changes in environmental factors and changes in physical constitution. The formula for comprehensive fine-tuning of medicinal tea recipes is as follows: Wherein, P(t) is the initially generated medicinal tea recipe; This is the final adjusted formula. It is an adjustment factor for the fine-tuning of the formula in response to environmental changes. It is an adjusting factor that allows for minor adjustments to the formula based on changes in body constitution. This represents the change in physical condition. This refers to environmental changes.

2. The method for preparing personalized medicinal tea based on body constitution identification according to claim 1, characterized in that: The process involves collecting users' static physical condition data and real-time physiological data, and acquiring environmental data through sensors. The specific steps are as follows: Users fill in their physical condition information through health records or questionnaires. The physical condition information includes physical condition type, age, gender, body type, allergy history, and family medical history. The physical condition information is then converted into structured data to form static physical condition data. Users provide real-time physiological data through smart bracelets and smartwatches. The real-time physiological data includes heart rate, body temperature, blood pressure, blood sugar, sleep quality, and activity level. The real-time physiological data is stored in the form of time series. Environmental data collection is conducted by deploying temperature and humidity sensors and air quality monitoring devices to collect real-time climate conditions of the user's environment, including temperature, humidity, air quality, PM2.5 concentration, and air pressure. Combined with the weather API interface, real-time weather information of the user's location is obtained, including current temperature, humidity, and wind speed.

3. A personalized medicinal tea preparation system based on constitution identification, used to implement the personalized medicinal tea preparation method based on constitution identification as described in any one of claims 1-2, characterized in that, include: The physical condition and environment analysis module is used to collect users’ static physical condition data and real-time physiological data, acquire environmental data through sensors, and fuse physical condition and environmental data through a nonlinear mapping function to generate a user state vector that includes changes in user physical condition and environment. The initial formula generation module is used to extract features based on the fused user state vector, analyze the relationship between physical constitution and environmental data, and output personalized medicinal tea formulas and calculate the proportion of medicinal materials to form user-customized formulas. The formula adjustment module is used to calculate the amount of change based on real-time environmental data and fluctuations in the user's physical condition, and adjust the proportion of medicinal materials in the formula to form an optimized medicinal tea formula; The feedback optimization module is used to obtain data on the taste and health effects of medicinal tea drinks through user feedback, and to optimize the tea drink formula through incremental learning.

Images