Infant nutrition recipe intelligent generation system and method

Abstract

The invention discloses a nutrition recipe intelligent generation system and method, and belongs to the technical field of nutrition information processing and intelligent recommendation. The system is suitable for individual nutrition management of individuals of all ages, and particularly, optimization design is carried out for infant stages of 0-72 months. The system intelligently generates a personalized nutrition recipe based on individual basic information, growth indexes, nutrition standards, available food materials, user preferences and feedback and other multi-dimensional data, and supports nutrition intake evaluation and dynamic optimization. The system comprises core modules of data input, nutritional requirement calculation, recipe generation, nutrition analysis and interactive display and the like, and expands and supports functions of sensitization screening, regional and seasonal adaptation, work and rest linkage, family sharing and the like. Through fusion of individual growth data and environmental conditions, the system realizes scientific and high-adaptability nutrition recommendation, improves feeding convenience and intervention accuracy, and has a wide application prospect.

Description

Technical Field

[0001] The present invention relates to the interdisciplinary fields of algorithms, artificial intelligence, nutrition, and health management, and in particular to a nutritional analysis and personalized intelligent recipe generation technology for individuals of all ages, especially for optimizing the feeding needs of infants and young children. Specifically, it relates to a nutritional recipe intelligent generation system and method. Background Art

[0002] With the development of big data and artificial intelligence technologies, personalized recommendation systems are gaining widespread application in areas such as health management and nutritional intervention. Personalized nutritional recipes, especially those tailored to specific populations such as infants, the elderly, and those with chronic diseases, are crucial for improving health and quality of life.

[0003] Most nutrition recommendation systems currently on the market are primarily designed for average adult users. Their recommendations are often based on generic templates or static rules, lacking effective adaptation to individual growth stages, dynamic changes in nutritional needs, and behavioral preferences. This approach struggles to meet the needs of diverse populations, especially infants and young children, who experience rapid growth, high risk of food allergies, and rapidly changing nutritional needs, and who demand greater dynamism, individualization, and food safety.

[0004] Existing technologies haven't yet developed a comprehensive assessment system for nutritional recommendations that encompasses basic information about individuals, health status, nutritional standards, dietary behaviors, and environmental factors. In particular, the lack of a closed-loop mechanism that provides real-time feedback and continuously optimizes recommendation models hinders the development of accurate, dynamic intelligent recipe services. Furthermore, existing systems generally lack a recommendation mechanism that integrates individual data, including that of infants and young children, making it impossible to generate intelligent, dynamic recipes based on multidimensional factors such as growth data, nutritional structure evolution, ingredient sensitivity control, and environmental conditions.

[0005] Therefore, there is an urgent need for an intelligent system that integrates a nutritional science knowledge base, artificial intelligence algorithms, and a user behavior feedback mechanism. It can provide scientific nutrition recommendation plans with dynamic adjustment, personalized generation, and intelligent feedback optimization for people of different age groups, including infants and young children, to improve the scientificity, convenience, and adaptability of nutritional management. Summary of the Invention

[0006] The present invention provides a system and method for intelligently generating nutritional recipes, suitable for individuals of different ages, including infants and young children. Specifically, it provides structural optimization for the growth needs of infants and young children aged 0 to 72 months. This approach aims to address the existing problem of a recommendation mechanism that lacks the ability to dynamically adapt to individual differences, growth rhythms, and feeding risks. By constructing a multi-module collaborative mechanism encompassing feature acquisition, needs assessment, personalized generation, and feedback optimization, the system achieves a closed-loop nutritional recommendation system and has expansion capabilities such as family collaboration, e-commerce procurement, and cross-platform linkage, forming a scientific, safe, and dynamic integrated nutritional management support tool.

[0007] The system of the present invention comprises: 1) User client module, used to collect basic information of individuals including infants and young children (such as date of birth, gender, height, weight, head circumference, allergy history, etc.) as well as daily diet and schedule records; it also supports family member management and multi-child collaboration functions, realizing differentiated cross-individual data collection and sharing; 2) Data collection and transmission module, used to synchronize client data to the cloud processing center in real time, ensuring the continuous accumulation and timely update of feature data; 3) Feature database module: used to store and manage individual user static characteristics, dynamic behaviors, growth and development curves, food nutritional properties, dietary guidelines, etc., supporting data structuring, indexing, and rapid access; 4) Nutritional Assessment and Trend Analysis Module: This module assesses current nutritional status based on individual differences, growth and development standards, nutritional rule optimization algorithms, or artificial intelligence models, identifies potential nutritional excesses or deficiencies, and generates quantifiable nutritional deviation indicators. It also combines historical dietary records, intake, and growth rhythms to dynamically track individual nutritional absorption trends, identify nutrients that continuously deviate from the target range, and automatically generate compensatory recommendations to prevent short-term errors from accumulating into long-term imbalances. 5) Recipe Generation Module: This module combines the ingredient database with the recipe library to generate personalized recipe recommendations based on an individual's current nutritional status, growth stage, nutritional needs, dietary preferences, etc., supporting seasonal substitutions and ingredient sensitivity exclusions. 6) Feedback Collection and Model Optimization Module: This module collects active and passive user feedback (e.g., manual changes, sliding changes, not marking as a favorite), analyzes preference changes, growth rhythms, or model performance, and adjusts the weight parameters of various influencing factors in recipe recommendations in real time to dynamically optimize recommendation strategies. 7) Recommendation model switching module: supports multiple recommendation algorithms based on rule logic, machine learning models, or deep neural network models, allowing switching or combination according to different usage scenarios or system performance requirements; 8) Recommendation output module: Outputs the generated recipe plan to the client interface, supporting graphic and text presentation, stage labels, nutritional information annotation and operation suggestions, and can be linked with e-commerce platforms to realize ingredient purchase recommendations; 9) Multi-child collaboration module: This module manages the information of multiple infants and toddlers under the same account, generating personalized recipes for each infant and toddler, with independent recommendations based on individual characteristics such as age, gender, allergies, and dietary preferences. Based on the commonalities and differences among multiple infants and toddlers, combined with shared family food resources, it automatically generates a comprehensive nutritional recipe plan suitable for multiple infants and toddlers. Users can choose to use independent or collaborative recommendation modes, and the system can dynamically adjust the recipe generation strategy based on user preferences or family configuration rules. 10) Auxiliary function modules: including but not limited to feeding risk reminders, nutritional supplement recommendations, sleep and rest schedule and diet linkage reminders, family member sharing, expert consultation access, and third-party e-commerce interface linkage, to enhance the system's practicality and expansion capabilities;

[0008] If the user does not fully fill in the basic information for the infant or child, the system described in the present invention can also generate preliminary trial recipes based on the partial information already provided (such as only age, gender, or a single physical characteristic data entry). This trial recipe is estimated based on default model parameters or population average data. The system will automatically mark it as "Trial Recommendation" and provide a risk warning and guidance for completing the information, reminding the user of the risk of bias and encouraging them to complete the information before using the personalized recommendation.

[0009] The "season" mentioned in the present invention is a general term and can be accurately specified to monthly or daily food replacement recommendations when necessary.

[0010] The "Multi-Child Collaboration Module" described in this invention is not only applicable to the nutritional management of multiple infants in ordinary families, but is also suitable for centralized infant management scenarios such as kindergartens, early childhood education institutions, and confinement centers. It can improve the efficiency of centralized procurement and production while ensuring precise personalized nutrition matching. Furthermore, this module provides the technical and data foundation for business models such as large-scale customization and intelligent distribution of personalized fresh infant food. It is suitable for flexible meal supply linked to production lines such as community kitchens and central kitchens, and promotes the development of new professions such as on-site chef customization services and family nutrition assistants based on infant nutrition data.

[0011] Through the collaborative work of these modules, this invention enables intelligent, personalized generation of nutritional recipes for individuals of all ages, including sensitive groups like infants and young children. This approach offers the advantages of high dynamism, timely feedback, and high security, contributing to the establishment of a new model of intelligent health management that is family-led and system-assisted. It also provides convenience for centralized management scenarios like kindergartens and makes it feasible for commercial customization and delivery of infant and young children's meals.

[0012] Compared with the existing technology, the present invention has the following technical innovations and beneficial effects: (1) Automated growth identification and dynamic nutrition adaptation: The system can automatically update the age and recommended nutrition structure based on the date of birth and gender, and dynamically adjust it in combination with the standard growth curve; (2) Matching personalized nutrition recommendations with food availability: Introducing multidimensional factors such as region, season, and allergy level to optimize the recipe structure and improve the feasibility and safety of recommendations; (3) Nutritional compensation and trend tracking mechanism: The system supports the cumulative analysis of multi-day intake data, generates trend reports and periodic nutritional assessments, and links supplement recommendations; (4) Intelligent feedback replacement mechanism: Adjust preference weights based on user feedback, dynamically generate alternative recommended recipes, and improve acceptance and user experience; (5) Scalable and upgradeable AI model path: The system is compatible with collaborative filtering and shallow neural networks, and will be expanded to deep recommendation models such as Wide & Deep, GNN, and Transformer in the future. It will also explore the introduction of causal modeling mechanisms to optimize feeding logic; (6) Family sharing and multi-child collaboration support: Supports multi-role account management and collaborative optimization of multiple infant recipes to improve family parenting efficiency; (7) One-click procurement linkage: The system can directly connect the recommended ingredient list to mainstream e-commerce platforms to facilitate rapid procurement execution; (8) Risk warning and expert support module: Provides tips and recommended observation periods for first-time food makers, and builds a community and expert consulting support system.

[0013] The "recommendation logic" described in this invention can be implemented in a variety of ways, including but not limited to rule-based matching algorithms, weighted scoring models, heuristic optimization strategies, collaborative filtering, and deep learning models. The system can select traditional algorithms or introduce artificial intelligence for optimization based on actual deployment requirements, and possesses good scalability and flexible deployment capabilities.

[0014] In addition to forming a closed-loop coordinated operation, each module of the present invention has independent use value and can be deployed independently according to application requirements, or combined with any two or more modules for operation. The resulting system architecture falls within the scope of protection of the present invention.

[0015] Furthermore, the modules described in this invention can be packaged and deployed as standalone tools, plug-ins, or microservice interfaces, allowing for invocation within the system or by third-party systems. The recipe generation module, in particular, can independently provide personalized recommendations or integrate with modules such as nutrition calculation, nutrition analysis, and feedback optimization to form a closed-loop recommendation logic, adapting to flexible integration and deployment across diverse platforms and hierarchical systems.

[0016] The system supports local deployment and offline operation on user clients, including but not limited to mobile apps, desktop programs, web pages, mini-programs, and wearable device interfaces, to accommodate the needs of different platforms. Furthermore, the system supports deployment in conjunction with cloud modules for model synchronization, data updates, and multi-terminal information sharing, creating a flexible local-cloud hybrid architecture.

[0017] Although the present invention is specifically optimized for the nutritional characteristics of infants and young children aged 0 to 72 months in a preferred embodiment, the system structure and nutrition recommendation method are also applicable to individuals of other age groups including children, adolescents, adults and the elderly.

[0018] This invention not only realizes phased and personalized nutrition recommendations for individuals including infants and young children, but also constructs a closed-loop mechanism of nutrition intake-feedback evaluation-model optimization. It has good intelligent scalability and commercial implementation potential, and is suitable for various application scenarios such as maternal and child apps, smart terminals, and home assistants. It has good intelligent scalability and commercial implementation potential. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solution of the present invention, the accompanying drawings schematically illustrate the structure as follows:

[0020] The same reference numerals in the accompanying drawings represent the same functional modules: Figure 1 : Schematic diagram of the overall architecture of the intelligent generation system for infant nutrition recipes of the present invention; Figure 2 : Flowchart of the personalized recipe generation and online ingredient acquisition of the present invention; Figure 3 : Schematic diagram of the nutritional feedback closed loop and model dynamic optimization of the present invention; DETAILED DESCRIPTION

[0021] In order to more clearly illustrate the specific implementation path of the "recommendation optimization module" in the system of the present invention, the working mechanism and technical details of this module in different system stages are further explained in combination with different applicable scenarios of rule algorithms and artificial intelligence models as follows: In a preferred embodiment of the present invention, the recommendation optimization module can be implemented based on a preset rule algorithm or an artificial intelligence model, thereby taking into account the actual needs of lightweight and fast response and deep personalized recommendation. Figure 3 The schematic diagram shows the following two methods: One approach is to use a recommendation optimization method based on rules and traditional algorithms, which is suitable for situations where the amount of user data is small or the system computing power is limited. This approach can be implemented using the following algorithm: (1) Rule matching algorithm: weight factors are set according to the age group, nutritional needs gap, dietary preferences and allergen information of infants and young children, and the optimal recipe combination is selected based on the scoring and ranking mechanism; (2) Heuristic search algorithm: Considering the recipes that meet the nutritional goals as the solution space, a feasible ingredient combination scheme is generated through greedy strategy, backtracking algorithm or dynamic programming method; (3) Linear programming or integer programming model: Convert the stage-specific nutritional needs of infants and young children into objective functions and constraints, and use a mathematical programming solver to calculate the optimal nutritional ratio to achieve refined recommendations.

[0022] The above rules have the advantages of simple implementation, fast response, and low computational overhead compared to traditional algorithms, and are suitable for the system cold start phase or scenarios with clear recommendation targets. Figure 3 As shown in , this system supports generating basic trial recommendation plans based on any available information (such as only age, gender, or allergy items) when the user's basic information is incomplete, and includes risk warnings and feeding precautions in the results to ensure the system's usability and user experience in low-information states.

[0023] Another approach is to use recommendation optimization methods based on artificial intelligence and machine learning models, which is suitable for scenarios where the system has accumulated a large amount of user behavior and feedback data. This approach can be implemented using the following models: (1) Collaborative filtering model: By identifying the similarities in preferences among infant and toddler users, personalized recipe recommendations based on neighborhood are achieved; (2) Shallow neural network or Wide & Deep model: Infant and child characteristics, user behavior data, and food attributes are used as input to train an end-to-end deep learning model to output recommendation results; (3) Graph Neural Network (GNN) model: Builds a graph structure containing the relationships between ingredients, recipes, and users, and mines complex ingredient combinations and preference patterns; (4) Transformer model: used to process user historical behavior sequences and model the temporal dependency of recommendation results; (5) Causal modeling model: By identifying the causal relationship between variables and eliminating pseudo-correlation factors, the stability and interpretability of recommendation results are improved.

[0024] The above artificial intelligence models can be dynamically trained and optimized based on the user interaction data continuously collected by the system, building a closed-loop mechanism of "data input - personalized recommendation - user feedback - model iteration". Figure 3 As shown, the present invention continuously adjusts the personalized recommendation results through the user feedback module, realizes the individualized continuous evolution of the nutrition plan, and effectively improves the recommendation accuracy and system adaptability.

[0025] In actual deployment, the recommendation optimization module can also adopt a hybrid recommendation mechanism. The system can flexibly switch between traditional algorithms and artificial intelligence models according to the operating environment or user preferences, or adopt a weighted fusion strategy to integrate the output results of the two types of models. For example, when a user uses it for the first time, a rule matching strategy is used to generate a basic recommendation plan. As user behavior data gradually accumulates, a machine learning model is introduced to optimize and adjust the recommendation weight, thereby achieving a dynamic balance between system response efficiency and recommendation personalization level. The dynamic switching and fusion logic of the above recommendation paths can also be referred to Figure 3 Understand it.

[0026] The intelligent infant nutrition recipe generation system described in this invention boasts strong algorithmic adaptability and engineering scalability, supporting both lightweight and rapid deployment scenarios and large-scale intelligent childcare platforms with strong demands for deeply personalized recommendations. The recommendation optimization module can be deployed via an API, SDK component, or microservice, supporting access via protocols such as RESTful, GraphQL, and RPC, facilitating integration and callable across multiple platforms (e.g., mobile devices, web applications, childcare platforms, or hospital systems).

Claims

1. A nutritional recipe intelligent generation system, characterized by: The system is applicable to the nutritional management of various individuals including infants and young children, among which infants and young children aged 0 to 72 months are the focus of system function optimization and application. The system includes: (1) a data input module for basic information of infants and young children, including date of birth, gender, height, weight, head circumference, allergen information, dietary preferences, and some or all static and dynamic information such as doctor's orders and expert advice; (2) a nutritional demand calculation module for generating phased nutritional target values based on the growth stage of infants and young children and preset nutritional recommendation standards; (3) a recipe generation module for generating personalized recipes that meet the needs according to the nutritional target values based on the currently available food information; (4) a nutritional analysis module for comparing the eating situation corresponding to the generated recipe with the nutritional structure of infants and young children, and outputting a nutritional intake assessment report and compensation suggestions; (5) a display and interaction module for displaying recommended recipes and nutritional analysis results to users, and supporting users to provide feedback.

2. The system according to claim 1, wherein: It is suitable for nutritional recommendations and analysis of individuals of different age groups, preferably for infants and young children. The system can automatically identify the user's age group and load the corresponding nutritional model to achieve phased optimization.

3. The system according to claim 1, wherein: Further including: (6) User feedback collection module, used to collect users' acceptance, execution and subjective evaluation of recipes; (7) Recommendation optimization module, used to dynamically update recommendation strategies based on user feedback data, wherein the recommendation strategies include rule-based adjustment, expert system optimization and adaptive optimization based on artificial intelligence models, and the optimization technologies may include collaborative filtering models, neural networks, Wide & Deep models, graph neural networks (GNNs), Transformer models, causal modeling models, knowledge graphs and logical reasoning algorithms, etc.; the recommendation optimization module supports providing preliminary personalized suggestions based on user tag similarity or default model strategies when complete feedback data has not been collected.

4. The system according to claim 1, wherein: The recipe generation module supports both static template recommendations based on user input data and dynamic intelligent recommendations that combine nutritional needs and ingredient data; static templates can be selected, edited or saved by users as commonly used recipe templates, and dynamic recommendations can be automatically generated based on preset rules, algorithms or artificial intelligence models.

5. The system according to claim 1 or 2, characterized in that Further including: (8) Dynamic timing optimization module, which is used to collect daily routine data of infants and young children, and dynamically adjust the time and content of recommended meals accordingly. It also analyzes the daily, weekly or monthly nutritional intake trends and generates medium- and long-term compensation recommendations.

6. A system according to any preceding claim, characterised in that Further including: (9) A regional and seasonal adaptation module, which is used to screen local low-allergenic ingredients based on the user's geographical location and current season information, and optimize the weight of ingredient recommendations in combination with allergenicity level labels; the module allows users to select a national template to circumvent regional restrictions.

7. A system according to any preceding claim, characterised in that: When the basic information of individuals such as infants and young children is incomplete, the system supports generating preliminary trial recipes based on some available data, and provides usage risk warnings and information completion guidance to prevent users from misusing inappropriate recipes.

8. A system according to any preceding claim, characterised in that Further including: (10) A multi-child collaboration module, which is used to manage the information of multiple infants and young children under the same account. The module supports: generating personalized recipes for each infant and young child, and independently recommending them based on their individual characteristics such as age, gender, allergies, and dietary preferences; Based on the commonalities and differences of multiple infants and young children, combined with the family's shared food resources, a combination of nutritional recipe plans suitable for multiple infants and young children is automatically generated; users can choose to use independent recommendation or collaborative recommendation mode, and the system can dynamically adjust the recipe generation strategy according to user preferences or family configuration rules.

9. A system according to any preceding claim, characterised in that Further including: (11) Intake risk management module, which is used to identify potential missing nutrients based on nutritional analysis results and generate supplement recommendations, set reminder frequency, dosage and time points to avoid repeated or excessive supplementation; and provide feeding precautions when recommending a certain ingredient for the first time, including gradual trial and observation period recommendations, etc.

10. A system according to any preceding claim, characterised in that Further including: (12) An auxiliary execution module, which is used to organize the ingredients required for generating recipes into a shopping list, support online ordering, jump to a third-party e-commerce platform or export to image / PDF / Excel document format, and produce and display tutorial videos for recommended recipe matching. The video matching is based on the similarity of ingredients, nutrition labels or dish types.

11. The system according to claim 1, wherein: It further includes a family sharing module that supports adding family member accounts by scanning codes, invitation codes, etc., and authorizes them to simultaneously view and edit infant and child information and recommended content.

12. A system according to any preceding claim, characterised in that: The data input module is further used to automatically calculate the growth age in days based on the date of birth and age in months, and dynamically update the nutritional target value for the corresponding stage, eliminating the need for users to repeatedly input operations.

13. A system according to any preceding claim, characterised in that: The food material database includes multi-level allergenicity rating labels for optimizing the priority of food material recommendations during recipe generation.

14. According to the system described in any of the preceding claims, each functional module supports deployment as an independent sub-module, or deployment and operation in combination of any two or more modules according to specific needs; the modules can be encapsulated as local components, plug-in tools or microservices, and deployed on different user terminals or third-party platforms, and the various combined structures formed are all within the scope of protection of the present invention.

15. A system according to any preceding claim, characterised in that: The system supports local deployment and offline operation on the user client, which includes mobile APP, web pages, mini-programs, desktop programs or other terminal devices; it can also be operated in conjunction with the functional modules through a cloud server to achieve policy updates and data synchronization.

16. A system according to any preceding claim, characterised in that: Each functional module supports encapsulation and deployment in the form of API interface, SDK component, microservice, integrated compilation or functional service. The module functions can be deployed on the client, server or edge device. The interface structure supports protocols or methods including but not limited to RESTful, GraphQL, RPC, WebSocket, gRPC, message queue communication, etc., for integrated calls by external systems or user terminals.

17. A method for intelligently generating infant nutrition recipes, characterized in that: The process includes the following steps: S1, receiving basic information and dietary preferences of infants and young children; S2, generating phased nutritional target values based on the infants' growth stages and nutritional recommendations; S3, generating personalized recipes that meet nutritional targets based on low-allergenic ingredients available in the current region and season; S4, comparing recommended recipes with nutritional target values to generate a nutritional intake assessment report and compensation recommendations; S5 displays recommended recipes and analysis results, and supports user feedback interaction.

18. The method according to claim 14, characterized in that Further including: S6, dynamically adjust the recommendation strategy based on user feedback information. The feedback includes acceptance, execution status and subjective evaluation. The adjustment methods include rule optimization, artificial intelligence model optimization and doctor and expert experience correction.

19. The method according to claim 14 or 15, characterized in that Further including: S7, analyzes infants’ daily routine data and dynamically adjusts the time and content of recommended meals; S8, analyzes nutritional intake trends and generates mid- to long-term optimization recommendations; S9, prompts risk information and precautions for first-time use of food ingredients, and controls feeding rhythm and trial cycle.

Images