Personalized intelligent catering method and system based on user condition

Abstract

The invention discloses a personalized intelligent catering method and system based on user conditions, and relates to the technical field of personalized nutriology and intelligent foods.The personalized intelligent catering method comprises the steps that user health comprehensive data and multi-dimensional real-time user situation data are collected, and the two kinds of data are preprocessed; constructing a user nutrition demand prediction model by combining a long-short-term memory network model with a multi-layer perceptron model; analyzing the preprocessed multi-dimensional real-time user situation data to obtain an adjustment coefficient influencing the nutritional requirement, and adjusting the personalized nutritional requirement of the user through a linear programming algorithm; and based on the adjusted personalized nutritional requirements of the user and in combination with the taste preference of the user, the matching degree of each food material is analyzed, customized adjustment is performed through a particle swarm optimization algorithm, and a personalized recipe of the user is generated. Food material matching is optimized through the particle swarm optimization algorithm, personalized intelligent catering is successfully achieved, and scientificity, accuracy and user satisfaction of personalized catering are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of personalized nutrition and intelligent food technology, and in particular to a personalized intelligent meal preparation method and system based on user conditions. Background Art

[0002] With the continuous improvement of people's health awareness, personalized nutritional meal matching has gradually become an important research direction in the field of modern nutrition and food technology. In recent years, personalized nutrition has gradually developed, and related technologies have gradually shifted from static calculations based on traditional nutritional demand models to dynamic optimization using advanced technologies such as big data, artificial intelligence and machine learning. In particular, the emergence of long short-term memory (LSTM) network models has provided a new method for processing and analyzing user health data and contextual data. Through deep learning of time series data, the LSTM model can accurately predict changes in users' nutritional needs, thereby providing more accurate guidance for personalized nutritional meal matching. At the same time, the introduction of multi-dimensional real-time user contextual data can better reflect the fluctuations in users' nutritional needs in different contexts. With the application of big data technology, users' health data and behavior data can be collected and processed in real time, providing strong support for personalized intelligent meal matching.

[0003] However, the existing technologies still have some shortcomings in the realization of personalized meal matching. First, most of the existing technologies rely on single health data or static nutritional demand prediction models, and fail to fully consider the fluctuations in nutritional needs of users in different situations, which makes it difficult to achieve the best personalized meal matching effect. Secondly, although some technologies try to make personalized recommendations based on user taste preferences, they lack in-depth analysis of ingredient matching and matching, resulting in the generated recipes having a poor balance in nutrition and taste, which makes it difficult to meet the personalized needs of users. Therefore, in terms of combining health data with situational data, accurately predicting nutritional needs, and dynamically adjusting personalized recipes, the existing technologies still have certain limitations. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a personalized intelligent meal matching method based on user conditions to solve the problems of insufficient consideration of user context data and lack of food matching analysis.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a personalized intelligent meal preparation method based on user conditions, which includes collecting user comprehensive health data and multi-dimensional real-time user situational data, and preprocessing the two types of data; constructing a user nutritional demand prediction model based on a long short-term memory network model, using the preprocessed user comprehensive health data to obtain the user nutritional demand value, and generate the user's personalized nutritional demand; analyzing the preprocessed multi-dimensional real-time user situational data to obtain the adjustment coefficient affecting the nutritional demand, and adjusting the user's personalized nutritional demand through a linear programming algorithm; based on the adjusted user's personalized nutritional demand and combined with the user's taste preferences, analyzing the matching degree of each ingredient, and making customized adjustments through a particle swarm optimization algorithm to generate a user's personalized recipe.

[0007] As a preferred solution of the personalized intelligent meal preparation method based on user conditions of the present invention, wherein: the comprehensive health data of the user includes basic information of the user, physiological health data of the user and activity level of the user; The multi-dimensional real-time user context data includes real-time activity time data, real-time activity location data, and real-time activity environment data; The preprocessing includes data cleaning and standardization.

[0008] As a preferred solution of the personalized intelligent meal preparation method based on user conditions of the present invention, wherein: the user nutritional demand prediction model is constructed based on the long short-term memory network model, and the specific steps are as follows: The pre-processed historical user health comprehensive data is converted into a time series format through a sliding window method; The model is based on the long short-term memory network model; The input layer receives the pre-processed historical user health comprehensive data converted into time series format; LSTM layer uses memory cells and gate mechanisms. The LSTM model selectively remembers and forgets information, capturing long-term dependencies in time series. The fully connected layer extracts features based on the output of the LSTM layer and reduces the dimension, outputting preliminary features; Combined with the multi-layer perceptron model, the input layer receives the preliminary features output by the LSTM model; The concatenation layer concatenates the preliminary features with other relevant factors; The MLP layer inputs the concatenated vectors into a single-layer MLP for further feature combination and transformation; The output layer outputs the final result; Finally, a user nutritional demand prediction model is obtained.

[0009] As a preferred solution of the personalized intelligent meal preparation method based on user conditions described in the present invention, the health energy calculator is combined with enhanced AR technology, and the Harris-Benedict algorithm is used to obtain the user's metabolic rate, and the user's comprehensive health data after preprocessing is combined with the user's nutritional demand prediction model to obtain the user's nutritional demand value, and generate the user's personalized nutritional demand. The specific steps are as follows: Using a health energy calculator combined with enhanced AR technology, the system simulates users performing light activities in a virtual environment, captures individual differences by observing users' physical reactions in different situations, and creates a unique health profile for each user. Through the activity logs provided by users and the use of sports bracelets to collect user activity levels for analysis, the average activity intensity of users is obtained; Based on the average activity intensity of the user and the user's health profile, the personalized activity factor is obtained by adjusting the proportion of the user's time spent on different activity types and their corresponding metabolic equivalent values. Based on the user's health profile, the basal metabolic rate is obtained through the health energy calculator and the Harris-Benedict algorithm, and the user's metabolic rate is obtained by combining the personalized activity factor; Based on the user's metabolic rate and the pre-processed user's comprehensive health data, the user's nutritional demand value is obtained through the user's nutritional demand prediction model.

[0010] As a preferred solution of the personalized intelligent meal preparation method based on user conditions of the present invention, wherein: the personalized nutritional needs of users are generated, and the specific steps are as follows: Based on historical user health data, analyze through cluster analysis algorithm and set health thresholds ; Compare user nutritional requirements with health thresholds Make comparisons and assess user health status; when When the number is 0, it means that the user is in poor health and needs to take in a variety of nutrients; when When the user is in good health, he or she does not need to take in nutrients. Based on the user's health status and the nutrients they need, the user's personalized nutritional needs are generated through genetic algorithms combined with nutrition theory.

[0011] As a preferred solution of the personalized intelligent meal preparation method based on user conditions of the present invention, the pre-processed multi-dimensional real-time user situation data is analyzed to obtain the adjustment coefficient affecting the nutritional needs, and the personalized nutritional needs of the user are adjusted by a linear programming algorithm. The specific steps are as follows: Analyze real-time activity time data through time series analysis methods and combine with kinematic principles to determine the activity intensity of each time period; Use the weight allocation method to convert different activity intensities into adjustment coefficients for the impact of activity intensity on nutritional needs; The real-time activity location data and user physiological health data are analyzed through data mining methods and mapped into adjustment coefficients of the activity location affecting nutritional needs; Combine the real-time activity environment data with the influence of human calorie consumption and human water needs, and calculate the adjustment coefficient of the activity environment's impact on nutritional needs through regression analysis. Set minimum and maximum constraints based on the user's basic physiological needs and health status; The linear programming algorithm uses the adjustment coefficients of activity intensity, activity location and activity environment to adjust nutritional requirements, and adjusts the user's personalized nutritional requirements through minimum and maximum constraints.

[0012] As a preferred solution of the personalized intelligent meal matching method based on user conditions of the present invention, the specific steps are as follows: based on the adjusted personalized nutritional needs of the user and combined with the user's taste preferences, the matching degree of each ingredient is analyzed, and customized adjustments are made through the particle swarm optimization algorithm to generate a personalized recipe for the user. Obtain user taste preferences through user's dietary history, feedback ratings, and taste preference questionnaires; The adjusted personalized nutritional needs of users are combined with their taste preferences, and the particle swarm optimization algorithm is used to gradually search for ingredients and dish combinations that meet the needs to form an ingredient matching plan; The Pearson correlation coefficient is used to analyze the matching degree between the nutritional value of each ingredient combination and the user's taste preference; The matching degree of each ingredient combination is weighted and summed to obtain the ingredient combination evaluation score, and the ingredients that meet the user's personalized needs are screened out through the forward sorting method to form a personalized recipe for the user.

[0013] In a second aspect, the present invention provides a personalized intelligent meal distribution system based on user conditions, including a data collection module, a model construction module, a model prediction module, a nutritional demand adjustment module and a recipe formulation module; the data collection module is used to collect user health comprehensive data and multi-dimensional real-time user situation data, and pre-process the two types of data; the model construction module constructs a user nutritional demand prediction model by combining a long short-term memory network model with a multi-layer perceptron model; the model prediction module is used to use a health energy calculator in combination with enhanced AR technology and a Harris-Benedict algorithm to obtain the user's metabolic rate, and in combination with the pre-processed user health comprehensive data, obtain the user's nutritional demand value through the user nutritional demand prediction model, and generate the user's personalized nutritional demand; the nutritional demand adjustment module is used to analyze the pre-processed multi-dimensional real-time user situation data to obtain the adjustment coefficient affecting the nutritional demand, and adjust the user's personalized nutritional demand through a linear programming algorithm; the recipe formulation module analyzes the matching degree of each ingredient based on the adjusted user's personalized nutritional needs and combined with the user's taste preferences, and customizes the adjustment through a particle swarm optimization algorithm to generate a user's personalized recipe.

[0014] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the personalized intelligent meal preparation method based on user conditions as described in the first aspect of the present invention is implemented.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the personalized intelligent meal preparation method based on user conditions as described in the first aspect of the present invention is implemented.

[0016] The beneficial effects of the present invention are: accurate prediction of user nutritional needs through the long short-term memory network (LSTM) model combined with the multi-layer perceptron model, and optimization of food ingredients by the particle swarm optimization algorithm, successfully realizing personalized intelligent meal matching. Its beneficial effects are that it can dynamically adjust nutritional needs based on user health status and real-time situational data, optimize food ingredients combination based on user taste preferences, thereby generating recipes that meet both nutritional needs and personalized tastes, significantly improving the scientificity, accuracy and user satisfaction of personalized meal matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0018] Figure 1 This is a flow chart of the personalized intelligent meal preparation method based on user conditions in Example 1.

[0019] Figure 2 Schematic diagram of the personalized intelligent meal preparation system based on user conditions in Example 1. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0023] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and provides a personalized intelligent meal matching method based on user conditions, comprising the following steps: S1. Comprehensive user health data includes basic user information, user physiological health data, and user activity level; It should be noted that the above content has been obtained with the user's consent and is used for legitimate purposes.

[0024] S1.1. Multi-dimensional real-time user context data includes real-time activity time data, real-time activity location data, and real-time activity environment data.

[0025] S1.2. Preprocessing includes data cleaning and standardization.

[0026] It should be noted that the specific steps for data cleaning and standardization of user health comprehensive data and multi-dimensional real-time user context data are as follows: Data cleaning: First, conduct a comprehensive data review to identify and handle missing values. This usually involves analyzing the missing pattern to decide whether to delete records with missing values ​​or use appropriate filling methods, such as using the mean, median, or prediction model to fill them. Then, correct logical errors and outliers. Set a reasonable threshold range to detect and correct extreme values ​​that do not conform to physiological common sense or domain knowledge to ensure data consistency. In particular, when data comes from multiple sources, it is necessary to check whether the date format, measurement unit, etc. are consistent, and resolve any inconsistencies found. Finally, remove completely duplicate or logically redundant records; Standardization: The Z-score standardization method is used, that is, the difference between each data point and the overall mean is calculated and then divided by the standard deviation to convert the data into a standard normal distribution with a mean of 0 and a standard deviation of 1. This method is particularly suitable for situations where features have different magnitudes.

[0027] S2. Build a user nutritional demand prediction model based on the long short-term memory network model.

[0028] S2.1. The preprocessed historical user health comprehensive data is converted into time series format through the sliding window method.

[0029] It should be noted that first, a fixed-size time window is determined, and the window will slide gradually according to the set step size on the preprocessed historical user health comprehensive dataset, capturing the data of all time points in the window as a sample each time it slides; then, the data in each time window is sorted and converted into a time series format suitable for input into the long short-term memory network (LSTM) model, that is, each sample contains data points arranged in sequence from the earliest to the latest time point; in this way, by continuously moving the window, a series of time series samples covering the entire time range can be created.

[0030] S2.2, based on the long short-term memory network model; The input layer receives the pre-processed historical user health comprehensive data converted into time series format; LSTM layer uses memory cells and gate mechanisms. The LSTM model selectively remembers and forgets information, capturing long-term dependencies in time series. The fully connected layer extracts features based on the output of the LSTM layer and reduces the dimension, outputting preliminary features; Combined with the multi-layer perceptron model, the input layer receives the preliminary features output by the LSTM model; The concatenation layer concatenates the preliminary features with other relevant factors; The MLP layer inputs the concatenated vectors into a single-layer MLP for further feature combination and transformation; The output layer outputs the final result; Finally, a user nutritional demand prediction model is obtained.

[0031] It should be noted that the user nutritional demand prediction model is constructed based on the long short-term memory network (LSTM) model combined with the multi-layer perceptron (MLP) model. This method effectively captures the time series characteristics and long-term dependencies in the user's comprehensive health data through the LSTM layer, especially for health parameters that change over time. It provides more accurate modeling capabilities; then, the preliminary features of the LSTM output are further combined and transformed using MLP, and optimized after being spliced ​​with other relevant factors, thereby achieving in-depth mining and integration of complex, multi-dimensional health data. Compared with existing technologies, this method not only enhances the understanding and prediction accuracy of individual differences and dynamic living habits, but also improves the comprehensiveness and accuracy of personalized nutritional demand prediction by introducing additional relevant factors, providing users with more scientific and reasonable dietary advice.

[0032] S3. Use the health energy calculator combined with enhanced AR technology and the Harris-Benedict algorithm to obtain the user's metabolic rate. Combined with the pre-processed user health comprehensive data, the user's nutritional needs prediction model is used to obtain the user's nutritional needs value and generate the user's personalized nutritional needs.

[0033] S3.1. Use the health energy calculator in combination with enhanced AR technology to simulate users performing lightweight activities in a virtual environment, capture individual differences by observing users' physical reactions in different situations, and create a unique health profile for each user.

[0034] It should be noted that enhanced AR technology is used to create personalized virtual activity scenes, allowing users to perform lightweight activities in these scenes; the system monitors and records changes in the user's heart rate, blood pressure and other physiological indicators in real time, so as to analyze their energy consumption and body response in specific activities; finally, based on the collected data, a health record containing detailed physiological information and activity responses is built for each user.

[0035] S3.2. Through the activity log provided by the user and the use of a sports bracelet to collect user activity levels for analysis, the average activity intensity of the user is obtained.

[0036] It should be noted that the user's daily activity information is first recorded, such as activity type, duration and intensity; then, the sports bracelet monitors and transmits the user's physiological and exercise data in real time, such as number of steps, heart rate and calories consumed; then, these data are integrated, and comprehensive statistical analysis and machine learning algorithms are applied to calculate the average activity intensity of different activity types, reflecting the user's overall activity level over a period of time; finally, based on these analysis results, the user's average activity intensity is determined.

[0037] S3.3. Based on the user's average activity intensity and combined with the user's health record, adjustments are made according to the proportion of the user's time on different activity types and their corresponding metabolic equivalent values ​​to obtain a personalized activity factor.

[0038] It should be noted that we first analyze the proportion of users' time spent on different activity types, and perform a weighted calculation based on the metabolic equivalents (METs) value of each activity type; then, we adjust the user's average activity intensity according to the time proportion of these activity types and their corresponding METs values, thereby deriving a personalized activity factor that reflects personal living habits and health status.

[0039] S3.4. Based on the user's health record, the basal metabolic rate is obtained through the health energy calculator and the Harris-Benedict algorithm, and the user's metabolic rate is obtained by combining the personalized activity factor.

[0040] It should be noted that the Harris-Benedict algorithm is first used by the Health Energy Calculator to calculate the basal metabolic rate (BMR) based on the user's age, gender, weight and height. Then, combined with the personalized activity factor calculated based on the user's time proportion in different activity types and their corresponding metabolic equivalent values, the basal metabolic rate is adjusted to reflect the user's actual daily activity level, thereby deriving a user metabolic rate that accurately reflects personal lifestyle and health status.

[0041] S3.5. Based on the user's metabolic rate and the pre-processed user health comprehensive data, the user's nutritional demand value is obtained through the user's nutritional demand prediction model, and the expression is: ; in, It's time point The nutritional requirement value of the user, is the weight parameter of the user's nutritional requirement value, is a function symbol, Indicates The model is processed. is the user's metabolic rate, is a fully connected layer operation, yes About time point Preprocessed user health comprehensive data The processing result, It's time point Preprocessed comprehensive user health data, is the weight parameter for adjusting the user's metabolic rate, It is the adjustment coefficient that controls the intensity of the influence of the periodic component in the denominator on the output result. represents the effect of the user's metabolic rate on the periodic factor, is the angular frequency parameter, is the phase offset.

[0042] It should be noted that the above expression processes time series data by combining the long short-term memory network (LSTM) and the multi-layer perceptron (MLP) model, and incorporates the user metabolic rate ( ) and its personalized activity factors, while considering the regulatory effect of cyclical factors on nutritional needs. Compared with existing technologies, this method can not only capture the long-term dependencies in the user's comprehensive health data, but also introduce the metabolic rate and cyclical component adjustment coefficient ( ), achieves a more detailed and personalized assessment of nutritional needs, emphasizes the importance of dynamic changes and individual differences, and provides more accurate and customized user nutritional needs values.

[0043] S4, and generate personalized nutritional needs of users.

[0044] S4.1. Based on historical user health data, analyze through cluster analysis algorithm and set health thresholds .

[0045] It should be noted that the historical user health data of users is first collected and sorted, and then a cluster analysis algorithm (such as K-means or hierarchical clustering) is applied to group the historical user health data to identify user groups with similar health characteristics; then, the statistical summary of the core health indicators of each cluster (such as mean value, standard deviation) is calculated, and the corresponding health threshold is set for each cluster based on these statistical results. .

[0046] S4.2. Compare the user's nutritional needs with the health threshold Make comparisons and assess user health status; when When the number is 0, it means that the user is in poor health and needs to take in a variety of nutrients; when , it indicates that the user is in good health and does not need to take in nutrients.

[0047] It should be noted that when , it indicates that the user's current nutritional requirement value exceeds the set health threshold, which usually indicates that the user's health may be problematic or in a sub-healthy state, and it is necessary to increase the intake of multiple nutrients in a targeted manner to improve the health. At this time, foods rich in specific vitamins, minerals and other essential nutrients will be recommended to help the user return to an ideal health level; when , it means that the user's nutritional requirement value is below the health threshold, reflecting that his current health status is relatively ideal and there is no need to increase the intake of additional nutrients. This means that the user's body can obtain sufficient nutritional support under the existing diet structure to maintain normal physiological functions and good health. Therefore, it is recommended to maintain the existing balanced diet without special adjustments to the diet plan or additional nutritional supplements.

[0048] S4.3. Based on the user's health status and the nutrients they need, the user's personalized nutritional needs are generated through genetic algorithms combined with nutrition theory.

[0049] It should be explained that, first, based on the user's health data and nutritional needs assessment results, the optimization goals and constraints are defined, such as the intake range of essential nutrients, calorie restriction, etc.; then, a genetic algorithm is used to encode, select, crossover and mutate multiple nutritional combinations, simulating the natural selection process to search for the optimal solution space; in this process, nutritional theory is combined to ensure that the generated nutritional needs plan complies with dietary guidelines and nutritional balance principles; and finally a set of personalized nutritional needs that meet the user's specific health needs and optimize the nutrient ratio is output.

[0050] S5. Analyze the pre-processed multi-dimensional real-time user situational data to obtain adjustment coefficients that affect nutritional needs, and adjust the user's personalized nutritional needs through a linear programming algorithm.

[0051] S5.1. Analyze real-time activity time data through time series analysis methods and combine them with kinematic principles to determine the activity intensity of each time period.

[0052] It should be stated that time series analysis techniques (such as ARIMA model or exponential smoothing) are applied to identify activity patterns and trends; then kinematic principles are used, such as parameters such as speed and acceleration, to quantify the intensity levels of different activity types; and then these physical activity indicators are mapped to pre-defined activity intensity levels to determine the specific activity intensity of users in different time periods.

[0053] S5.2. Use the weight allocation method to convert different activity intensities into adjustment coefficients for the impact of activity intensity on nutritional needs.

[0054] It should be explained that first, the basic activity intensity level is determined based on the physical characteristics of the activity type and metabolic equivalents (METs); then, corresponding weights are set for each basic activity intensity level, and these weights reflect the relative impact of the activity intensity on nutritional needs; then, based on the user's actual activity data, an adjustment coefficient that affects nutritional needs is calculated by weighted average method.

[0055] S5.3. Analyze the real-time activity location data and user physiological health data through data mining methods, and map them into adjustment coefficients for the impact of activity locations on nutritional needs.

[0056] It should be explained that first, the user's real-time activity location data and corresponding physiological health indicators (such as heart rate, blood pressure, etc.) are integrated, and data mining techniques (such as cluster analysis or regression models) are used to identify the association patterns between different locations and specific physiological reactions; then the actual impact of the user's activities in different locations under these association patterns on human energy consumption and nutrient absorption is evaluated; then, based on the analysis results, these impacts are quantified and converted into specific adjustment coefficients.

[0057] S5.4. Combine the real-time activity environment data with the influence of human calorie consumption and human water needs, and calculate the adjustment coefficient of the activity environment's impact on nutritional needs through regression analysis.

[0058] It should be explained that the user's real-time activity environment data (such as temperature, humidity, etc.) and the corresponding human calorie consumption and water demand data are first collected; then the regression analysis method is used to explore the relationship between environmental factors and calorie consumption and water demand; based on the results, the specific impact of different environmental conditions on human nutritional needs is quantified, such as the increase in water and electrolyte requirements in high temperature environments; and finally these impacts are converted into specific adjustment coefficients; It should also be explained that the influence of human calorie consumption is mainly reflected in the basal metabolic rate (BMR), daily activity level and the thermal effect of food. Basal metabolic rate refers to the energy required for the body to maintain basic physiological functions in a resting state; daily activity level includes physical activity and exercise, which significantly increase total calorie consumption; the thermal effect of food refers to the energy consumed in the process of digesting, absorbing and metabolizing food. These data are usually estimated by indirect calorimetry, direct measurement or using specific formulas; The influence of human water demand is based on many factors, including ambient temperature, humidity, personal activity intensity and individual health status. In high temperature and high humidity environments, the human body will increase heat dissipation through sweat evaporation, which will lead to increased water loss; high-intensity exercise will also accelerate fluid loss. In addition, certain health conditions (such as fever, vomiting or diarrhea) will also increase water demand. The data sources of these laws are extensive and can be obtained through laboratory studies on water balance, epidemiological surveys and personalized health monitoring equipment.

[0059] S5.5. Set minimum and maximum constraints based on the user’s basic physiological needs and health status.

[0060] It should be stated that first, the user's basic information such as age, gender, weight, height, activity level, etc. is evaluated, and combined with their health status (such as whether they have chronic diseases, allergies or special nutritional needs) to determine basic physiological needs such as daily calories, macronutrients and recommended intake of micronutrients; then, based on medical guidelines and nutritional standards, safe and effective minimum and maximum ranges are set for these nutritional indicators.

[0061] S5.6. The linear programming algorithm uses the adjustment coefficient of the activity intensity affecting nutritional demand, the adjustment coefficient of the activity location affecting nutritional demand, and the adjustment coefficient of the activity environment affecting nutritional demand, and adjusts the user's personalized nutritional demand through minimum and maximum value constraints.

[0062] It should be explained that first, the adjustment coefficients are determined, which quantify the impact of different factors (such as activity intensity, location, and environment) on nutritional needs; then, the minimum and maximum constraints on nutrient intake are set based on the user's physiological needs and health status; then a linear programming model is constructed, the above adjustment coefficients are input as variables, and the objective function is set to optimize nutrient distribution while ensuring that all set constraints are met; by solving the linear programming model, the most suitable nutrient intake under given conditions is calculated, so as to make precise adjustments to the user's personalized nutritional needs.

[0063] S6. Based on the adjusted user's personalized nutritional needs and combined with the user's taste preferences, the matching degree of each ingredient is analyzed, and customized adjustments are made through the particle swarm optimization algorithm to generate a personalized recipe for the user.

[0064] S6.1. Obtain the user's taste preference through the user's dietary history, feedback score and taste preference questionnaire.

[0065] S6.2. Combine the adjusted personalized nutritional needs of users with their taste preferences, and gradually search for ingredients and dish combinations that meet the needs through the particle swarm optimization algorithm to form an ingredient matching plan.

[0066] It should be explained that first, the user's personalized nutritional needs and taste preferences are clarified, such as the need for specific nutrients or the preference for certain foods; then an objective function is defined, which comprehensively considers nutritional matching and taste satisfaction, and uses the particle swarm optimization algorithm to search in the space of possible ingredient and dish combinations; each "particle" in the algorithm represents a potential meal plan, and they update their positions according to the individual optimal solution and the social optimal solution. In the process, each plan is continuously evaluated and adjusted to ensure that it meets both nutritional needs and taste preferences, and finally the ingredient meal plan is output.

[0067] S6.3. Use the Pearson correlation coefficient to analyze the matching degree between the nutritional value of each ingredient combination and the user's taste preference.

[0068] It should be explained that, first, for each meal plan, all the ingredients contained therein and their corresponding nutritional value indicators (such as protein content, vitamin level, etc.) and indicators reflecting user taste preferences (such as the degree of liking for a certain ingredient or preference for cooking method) are determined; then the Pearson correlation coefficient between these nutritional value indicators and taste preference indicators is calculated to quantify the strength and direction of the linear relationship between the two; based on the obtained Pearson correlation coefficient, the degree of matching of ingredient combinations with user taste preferences while meeting nutritional needs is evaluated.

[0069] S6.4. Perform a weighted summation of the matching degree of each ingredient combination to obtain an ingredient combination evaluation score, and select ingredients that meet the user's personalized needs through a positive order sorting method to form a user's personalized recipe.

[0070] It should be explained that first, a score is calculated for each ingredient combination based on the degree of match between its nutritional value and the user's taste preferences, and weights are assigned to these scores based on the importance of the nutrients or the user's specific preferences; then a weighted sum is taken of all the ingredient combination scores in each meal plan to obtain the overall evaluation score of the plan; then all meal plans are sorted in ascending order according to the evaluation scores; finally, the top-ranked plans are selected, that is, those ingredient combinations that meet the user's nutritional needs and their taste preferences, to form the final personalized recipe.

[0071] This embodiment also provides a personalized intelligent meal preparation system based on user conditions, including: a data collection module, a model building module, a model prediction module, a nutritional demand adjustment module and a recipe formulation module; The data collection module is used to collect comprehensive user health data and multi-dimensional real-time user context data, and pre-process the two types of data; Model building module, which builds a user nutrition demand prediction model by combining the long short-term memory network model with the multi-layer perceptron model; The model prediction module is used to use the health energy calculator combined with enhanced AR technology and the Harris-Benedict algorithm to obtain the user's metabolic rate, combine the pre-processed user's comprehensive health data, obtain the user's nutritional demand value through the user's nutritional demand prediction model, and generate the user's personalized nutritional demand; Nutritional demand adjustment module, used to analyze the pre-processed multi-dimensional real-time user situation data, obtain the adjustment coefficients affecting nutritional demand, and adjust the user's personalized nutritional demand through a linear programming algorithm; The recipe formulation module analyzes the matching degree of each ingredient based on the adjusted user's personalized nutritional needs and combined with the user's taste preferences, and generates a personalized recipe for the user through customized adjustments through the particle swarm optimization algorithm.

[0072] This embodiment also provides a computer device, which is suitable for the personalized intelligent meal preparation method based on user conditions, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the personalized intelligent meal preparation method based on user conditions as proposed in the above embodiment.

[0073] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0074] The present embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the personalized intelligent meal preparation method based on user conditions as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0075] In summary, the present invention successfully realizes personalized intelligent meal matching by combining the long short-term memory network (LSTM) model with the multi-layer perceptron model to accurately predict the nutritional needs of users, and optimizing the food combination with the particle swarm optimization algorithm. Its beneficial effect is that it can dynamically adjust the nutritional needs based on the user's health status and real-time situational data, and optimize the food combination based on the user's taste preferences, thereby generating recipes that meet both nutritional needs and personalized tastes, significantly improving the scientificity, accuracy and user satisfaction of personalized meal matching.

[0076] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A personalized intelligent meal matching method based on user conditions, characterized by: include, Collect comprehensive user health data and multi-dimensional real-time user context data, and pre-process the two types of data; By combining the long short-term memory network model with the multi-layer perceptron model, a user nutrition demand prediction model is constructed; The health energy calculator is combined with enhanced AR technology, and the Harris-Benedict algorithm is used to obtain the user's metabolic rate. Combined with the pre-processed user health comprehensive data, the user's nutritional demand value is obtained through the user's nutritional demand prediction model, and the user's personalized nutritional demand is generated; Analyze the pre-processed multi-dimensional real-time user situational data to obtain the adjustment coefficients that affect nutritional needs, and adjust the user's personalized nutritional needs through a linear programming algorithm; Based on the adjusted user's personalized nutritional needs and combined with the user's taste preferences, the matching degree of each ingredient is analyzed, and customized adjustments are made through the particle swarm optimization algorithm to generate a personalized recipe for the user.

2. The personalized intelligent meal matching method based on user status according to claim 1, characterized in that: The user health comprehensive data includes user basic information, user physiological health data and user activity level; The multi-dimensional real-time user context data includes real-time activity time data, real-time activity location data, and real-time activity environment data; The preprocessing includes data cleaning and standardization.

3. The personalized intelligent meal matching method based on user status according to claim 2, characterized in that: The specific steps of constructing the user nutritional demand prediction model are as follows: The pre-processed historical user health comprehensive data is converted into a time series format through a sliding window method; The model is based on the long short-term memory network model; The input layer receives the pre-processed historical user health comprehensive data converted into time series format; LSTM layer uses memory cells and gate mechanisms. The LSTM model selectively remembers and forgets information, capturing long-term dependencies in time series. The fully connected layer extracts features based on the output of the LSTM layer and reduces the dimension, outputting preliminary features; Combined with the multi-layer perceptron model, the input layer receives the preliminary features output by the LSTM model; The concatenation layer concatenates the preliminary features with other relevant factors; The MLP layer inputs the concatenated vectors into a single-layer MLP for further feature combination and transformation; The output layer outputs the final result; Finally, a user nutritional demand prediction model is obtained.

4. The personalized intelligent meal matching method based on user status according to claim 3, characterized in that: The method of using the health energy calculator in combination with enhanced AR technology and applying the Harris-Benedict algorithm to obtain the user's metabolic rate, combining the pre-processed user health comprehensive data, and using the user's nutritional demand prediction model to obtain the user's nutritional demand value, and generate the user's personalized nutritional demand, is as follows: Using a health energy calculator combined with enhanced AR technology, the system simulates users performing light activities in a virtual environment, captures individual differences by observing users' physical reactions in different situations, and creates a unique health profile for each user. Through the activity logs provided by users and the use of sports bracelets to collect user activity levels for analysis, the average activity intensity of users is obtained; Based on the average activity intensity of the user and the user's health profile, the personalized activity factor is obtained by adjusting the proportion of the user's time spent on different activity types and their corresponding metabolic equivalent values. Based on the user's health profile, the basal metabolic rate is obtained through the health energy calculator and the Harris-Benedict algorithm, and the user's metabolic rate is obtained by combining the personalized activity factor; Based on the user's metabolic rate and the pre-processed user's comprehensive health data, the user's nutritional demand value is obtained through the user's nutritional demand prediction model.

5. The personalized intelligent meal matching method based on user status according to claim 4, characterized in that: The specific steps of generating the personalized nutritional needs of users are as follows: Based on historical user health data, analyze through cluster analysis algorithm and set health thresholds ; Compare user nutritional requirements with health thresholds Make comparisons and assess user health status; when When the number is 0, it means that the user is in poor health and needs to take in a variety of nutrients; when When the user is in good health, he or she does not need to take in nutrients. Based on the user's health status and the nutrients they need, the user's personalized nutritional needs are generated through genetic algorithms combined with nutrition theory.

6. The personalized intelligent meal matching method based on user status according to claim 5, characterized in that: The pre-processed multi-dimensional real-time user situation data is analyzed to obtain the adjustment coefficient affecting the nutritional needs, and the user's personalized nutritional needs are adjusted through a linear programming algorithm. The specific steps are as follows: Analyze real-time activity time data through time series analysis methods and combine with kinematic principles to determine the activity intensity of each time period; Use the weight allocation method to convert different activity intensities into adjustment coefficients for the impact of activity intensity on nutritional needs; The real-time activity location data and user physiological health data are analyzed through data mining methods and mapped into adjustment coefficients of the activity location affecting nutritional needs; Combine the real-time activity environment data with the influence of human calorie consumption and human water needs, and calculate the adjustment coefficient of the activity environment's impact on nutritional needs through regression analysis. Set minimum and maximum constraints based on the user's basic physiological needs and health status; The linear programming algorithm uses the adjustment coefficients of activity intensity, activity location and activity environment to adjust nutritional requirements, and adjusts the user's personalized nutritional requirements through minimum and maximum constraints.

7. The personalized intelligent meal matching method based on user status according to claim 6, characterized in that: Based on the adjusted personalized nutritional needs of the user and combined with the user's taste preferences, the matching degree of each ingredient is analyzed, and customized adjustments are made through the particle swarm optimization algorithm to generate a personalized recipe for the user. The specific steps are as follows: Obtain user taste preferences through user's dietary history, feedback ratings, and taste preference questionnaires; The adjusted personalized nutritional needs of users are combined with their taste preferences, and the particle swarm optimization algorithm is used to gradually search for ingredients and dish combinations that meet the needs to form an ingredient matching plan; The Pearson correlation coefficient is used to analyze the matching degree between the nutritional value of each ingredient combination and the user's taste preference; The matching degree of each ingredient combination is weighted and summed to obtain the ingredient combination evaluation score, and the ingredients that meet the user's personalized needs are screened out through the forward sorting method to form a personalized recipe for the user.

8. A personalized intelligent meal matching system based on user status, based on the personalized intelligent meal matching method based on user status according to any one of claims 1 to 7, characterized in that: Including data collection module, model building module, model prediction module, nutrition demand adjustment module and recipe formulation module; The data collection module is used to collect comprehensive user health data and multi-dimensional real-time user context data, and pre-process the two types of data; Model building module, which builds a user nutrition demand prediction model by combining the long short-term memory network model with the multi-layer perceptron model; The model prediction module is used to use the health energy calculator combined with enhanced AR technology and the Harris-Benedict algorithm to obtain the user's metabolic rate, combine the pre-processed user's comprehensive health data, obtain the user's nutritional demand value through the user's nutritional demand prediction model, and generate the user's personalized nutritional demand; The nutritional demand adjustment module is used to analyze the pre-processed multi-dimensional real-time user situation data, obtain the adjustment coefficients that affect nutritional demand, and adjust the user's personalized nutritional demand through a linear programming algorithm; The recipe formulation module analyzes the matching degree of each ingredient based on the adjusted user's personalized nutritional needs and combined with the user's taste preferences, and generates a personalized recipe for the user through customized adjustments through the particle swarm optimization algorithm.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the personalized intelligent meal preparation method based on user conditions described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the personalized intelligent meal preparation method based on user conditions described in any one of claims 1 to 7 are implemented.

Images