Formula food operation and management device based on human-computer interaction technology

Abstract

The application discloses a formula food operation and management device based on human-computer interaction technology, which comprises a human-computer interaction module, a nutrition analysis module, a formula food screening module and a formula food mixing module which are connected in communication. The detailed health information collected by the human-computer interaction module can provide highly personalized nutrition matching information through the nutrition analysis module, so that the postoperative patient can obtain the most suitable nutrition supply for recovery needs. For special nutritional needs that cannot be met directly through standard full-nutrition formula food, the formula food screening module can be combined with a nutrient component matching prediction model to dynamically generate customized nutrient component mixing matching parameters, thereby solving the problem that it is difficult to configure under the traditional method.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of human-computer interaction data processing, and particularly relates to a formula food operation and management device based on human-computer interaction technology. BACKGROUND

[0002] Formula foods, especially specialized formula foods designed for specific health needs, have become an important part of modern medical and nutritional science. Such foods aim to provide precise and balanced nutritional support for people with special nutritional needs, such as postoperative recovery patients, to facilitate the recovery process. They usually contain a series of carefully formulated nutritional ingredients, such as protein, carbohydrates, fats, vitamins and minerals, and possibly dietary fiber and other functional ingredients. The design of formula foods needs to follow strict nutritional principles and clinical evidence to ensure that it can not only meet the basic energy needs of patients, but also specifically support the recovery of specific physiological functions, such as the immune system, wound healing or muscle reconstruction. With the development of nutritional science and the pursuit of personalized medicine, formula foods are increasingly becoming key elements of customized nutritional solutions, especially playing an indispensable role in chronic disease management and postoperative rehabilitation.

[0003] However, the traditional formula food formulation process often relies on the experience and professional knowledge of nutritionists or doctors, and this process may have certain limitations and challenges. First, manual formulation is difficult to achieve a high degree of individualization and precision, because each patient's physical condition, type of surgery, risk of complications, and nutritional absorption capacity are different, and manual judgment may not fully cover these subtle differences. Second, the traditional formulation method is low in efficiency and time-consuming, especially when facing a large number of patients, one-by-one assessment and customized formula become extremely cumbersome, which limits the efficiency and scalability of medical services. In addition, with the rapid development of nutritional science, new research findings and nutritional standards are emerging, and manual formulation may not keep up with the latest scientific progress, resulting in suboptimal nutritional value and safety of formula foods. Finally, manual formulation may also be affected by subjective judgment and potential human errors, such as calculation errors or information omissions, which not only affect the accuracy and safety of the formula, but also may have adverse effects on the patient's recovery process. SUMMARY

[0004] In order to solve at least one of the technical problems mentioned above, the present application provides a formula food operation and management device and method based on human-computer interaction technology, and the specific technical solutions are as follows:

[0005] In one aspect, the present application provides a formula food operation and management device based on human-computer interaction technology, which comprises:

[0006] An interactive module for obtaining user health information provided by a user, the user health information comprising age, weight, surgery type, nutrition diagnosis information, and allergen information;

[0007] A nutrition analysis module for predicting nutrition ratio information required by the user according to the user health information, the nutrition ratio information comprising nutrition type information and nutrition type ratio information;

[0008] A formula food screening module for matching a complete nutrition formula food meeting the nutrition ratio information from a complete nutrition formula food information base, and inputting the nutrition ratio information into a pre-trained nutrient component ratio prediction model to output nutrient component mixture ratio parameters, the nutrient component mixture ratio parameters comprising nutrient component type information and nutrient component proportion information, if no complete nutrition formula food meeting the nutrition ratio information is matched;

[0009] A formula food mixing module for mixing corresponding nutrient components according to the nutrient component mixture ratio parameters to obtain a ratio food.

[0010] Preferably, after the nutrition ratio information is input into the pre-trained nutrient component ratio prediction model to output the nutrient component mixture ratio parameters, the method further comprises:

[0011] Generating a plurality of first ratio parameters within a preset range based on the nutrition ratio information and the nutrient component mixture ratio parameters;

[0012] Generating second ratio parameters marked with taste information based on the ratio parameters and taste information of the nutrient components;

[0013] Reflecting a plurality of the second ratio parameters to the interactive module for user selection, and determining the user's selection as new nutrient component mixture ratio parameters.

[0014] Preferably, the matching of the complete nutrition formula food meeting the nutrition ratio information from the complete nutrition formula food information base, if no complete nutrition formula food meeting the nutrition ratio information is matched, comprises:

[0015] Screening complete nutrition formula foods containing all the nutrition type information and meeting the nutrition type ratio information from the complete nutrition formula food information base.

[0016] Preferably, the device further comprises:

[0017] A ratio parameter mean value smoothing module for mean value smoothing the nutrient component mixture ratio parameters of a plurality of different users, the mean value smoothing comprising:

[0018] The first classification processing is performed on the multiple nutrient component mixing proportion parameters, and multiple first categories containing the same nutrient component combinations are obtained.

[0019] The second classification processing is performed on the nutrient component mixing proportion parameters in the first categories based on the similarity of the nutrient component proportions, and multiple second categories are obtained.

[0020] The average proportion of each nutrient component type in the second categories is calculated, and the corresponding average proportion is applied to all nutrient component mixing proportion parameters in the second categories.

[0021] Preferably, the device further comprises:

[0022] The nutrient component proportion prediction model training module is configured to input the nutrition type information and the nutrition type proportion information as input data, and output the nutrient component mixing proportion parameters as output data, and train the nutrient component proportion prediction model based on a CNN-GRU network.

[0023] In the CNN-GRU network, the output of the GRU network is used as the input of the CNN network.

[0024] Preferably, the method of predicting the nutrition proportion information required by the user according to the user health information comprises:

[0025] The user health information is input into a pre-trained nutrition proportion prediction model to obtain the nutrition proportion information.

[0026] The training process of the nutrition proportion prediction model is as follows:

[0027] The user health information and the corresponding nutrition proportion information of different populations are collected to obtain a nutrition proportion data set.

[0028] The principal component analysis algorithm is used to reduce the dimensionality of the nutrition proportion data set.

[0029] The GBDT algorithm is used to fit the reduced nutrition proportion data set, and the characteristic values with a correlation greater than a preset threshold are selected according to the comprehensive index after fitting.

[0030] The random forest model is trained based on the characteristic value information to obtain the trained nutrition proportion prediction model.

[0031] Preferably, the nutrient components include protein components, fat components, and carbohydrate components.

[0032] In a second aspect, the present application provides a recipe food operation and management method based on human-computer interaction technology, which comprises:

[0033] obtaining user health information provided by a user, the user health information including body weight, surgery type, nutrition diagnosis information, and allergen information;

[0034] predicting nutrition ratio information required by the user according to the user health information, the nutrition ratio information including nutrition type information and nutrition type ratio information;

[0035] matching a total nutrition formula food meeting the nutrition ratio information from a total nutrition formula food information library, and inputting the nutrition ratio information into a pre-trained nutrient component ratio prediction model to output nutrient component mixture ratio parameters including nutrient component type information and nutrient component proportion information if no total nutrition formula food meeting the nutrition ratio information is matched.

[0036] mixing corresponding nutrient components according to the nutrient component mixture ratio parameters to obtain a ratio food.

[0037] Preferably, after the nutrition ratio information is input into the pre-trained nutrient component ratio prediction model to output the nutrient component mixture ratio parameters, the method further includes:

[0038] generating a plurality of first ratio parameters within a preset range based on the nutrition ratio information and the nutrient component mixture ratio parameters;

[0039] generating second ratio parameters marked with taste information based on the ratio parameters and taste information of the nutrient components;

[0040] echoing the plurality of second ratio parameters to a human-computer interaction module for user selection, and determining the user's selection as new nutrient component mixture ratio parameters.

[0041] Preferably, the matching of the total nutrition formula food meeting the nutrition ratio information from the total nutrition formula food information library, if no total nutrition formula food meeting the nutrition ratio information is matched, includes:

[0042] screening total nutrition formula foods containing all the nutrition type information and meeting the nutrition type ratio information from the total nutrition formula food information library.

[0043] The present application has the following advantages:

[0044] (1) The present invention collects detailed health information through the human-computer interaction module and provides highly personalized nutritional ratio information through the nutritional analysis module, ensuring that postoperative patients receive the most suitable nutritional supply for their recovery needs. For special nutritional needs that cannot be met directly by standard complete nutritional formula food, the formula food screening module and the nutrient component ratio prediction model are combined to dynamically generate customized nutrient component mixing ratio parameters, which solves the problem of difficult configuration under traditional methods.

[0045] (2) By introducing a mean smoothing module for the ratio parameters, this invention can perform mean smoothing on the ratio parameters of nutrient components mixed by multiple different users, which greatly improves the efficiency of nutrient formulation. On the one hand, by processing the first and second categories, ratios with similar parameters are integrated, reducing the complexity of formulation design. On the other hand, by calculating the average ratio and applying it to all similar ratio parameters, the consistency and accuracy of the formulation are ensured, avoiding errors and deviations that may occur with manual formulation, and improving the stability and reliability of the formulation.

[0046] (3) By using the CNN-GRU network, this invention can better process the temporal and spatial features of nutritional type information. At the same time, it uses the Attention mechanism to strengthen the weight of important features, thereby more accurately predicting the mixing ratio parameters of nutrient components, meeting the nutritional needs of different users, improving the accuracy and practicality of personalized nutritional advice, and in practical applications, it can quickly generate reasonable nutrient ratio schemes based on the user's nutritional type information, thereby improving the efficiency and effectiveness of nutritional management. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the background art, the accompanying drawings used in the embodiments of the present invention or the background art will be described below.

[0048] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with this disclosure and, together with the specification, serve to illustrate the technical solutions disclosed in this invention.

[0049] Figure 1 A flowchart illustrating a method for operating and managing formulated food based on human-computer interaction technology, provided as an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of the mean smoothing process provided in an embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of a formula food operation and management device based on human-computer interaction technology, provided as an embodiment of the present invention. Detailed Implementation

[0052] To enable those skilled in the art to better understand the present invention, 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.

[0053] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0054] In this document, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Furthermore, the term "at least one" in this document means any combination of at least two of any one or more elements. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0055] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0056] Furthermore, to better illustrate the present invention, numerous specific details are set forth in the following detailed embodiments. Those skilled in the art will understand that the present invention can be practiced without certain specific details. In some instances, methods, means, elements, and circuits well known to those skilled in the art have not been described in detail in order to highlight the spirit of the invention.

[0057] Currently, there is a problem with the inability to accurately formulate special nutritional needs that cannot be met directly by standard complete nutritional formula foods.

[0058] Example 1: This invention provides a formula food operation and management device based on human-computer interaction technology, see reference. Figure 3 The device includes:

[0059] The human-computer interaction module 100 is used to obtain user health information provided by the user, including age, weight, type of surgery, nutritional diagnosis information and allergen information;

[0060] The human-computer interaction module 100 collects user health information through a graphical user interface (GUI), speech recognition technology, and natural language processing (NLP) technology. Users can provide information such as age, weight, type of surgery (e.g., cardiac surgery, orthopedic surgery), nutritional diagnosis information (e.g., protein and vitamin deficiencies), and allergen information (e.g., lactose intolerance, nut allergy) by filling out questionnaires, uploading medical reports, or answering questions verbally. NLP technology allows users to input complex medical histories and preferences via text or voice, while the GUI provides intuitive charts and drop-down menus to simplify the input of basic information such as age, weight, and type of surgery.

[0061] The nutrition analysis module 200 is used to predict the nutritional ratio information required by the user based on the user's health information. The nutritional ratio information includes nutritional type information and nutritional type ratio information.

[0062] Deep neural networks (DNNs), especially long short-term memory networks (LSTM) or Transformer architectures, are used to predict users' nutritional needs. These models can capture patterns in complex time-series data and are suitable for processing health information that changes over time. Based on the above network models, a detailed nutritional database is built, covering the nutritional components of various foods and nutritional supplements, so that the model can accurately predict nutritional type information and nutritional type ratio information.

[0063] The formula food screening module 300 is used to match complete nutritional formula foods that meet the nutritional ratio information from the complete nutritional formula food information database. If no complete nutritional formula food that meets the nutritional ratio information is matched, the nutritional ratio information is input into the pre-trained nutrient component ratio prediction model and the nutrient component mixing ratio parameters are output. The nutrient component mixing ratio parameters include nutrient component type information and nutrient component ratio information.

[0064] The module uses SQL and fuzzy matching algorithms to select products that best meet the user's nutritional needs from a complete nutritional formula food information database. If no suitable product is found, the module will input the nutritional ratio information into a pre-trained nutrient component ratio prediction model. The nutrient component ratio prediction model is based on a deep neural network (DNN) and can generate the optimal nutrient component mixing ratio parameters based on the nutrient type information and nutrient type ratio information.

[0065] The formula food mixing module 400 is used to mix the corresponding nutrient components according to the nutrient component mixing ratio parameters to obtain the formula food.

[0066] The formulation food mixing module 400 includes automated ingredient dispensing and mixing equipment, which precisely controls the addition amount of various nutrient components according to the mixing ratio parameters of the nutrient components. The formulation food mixing module 400 may also include precision weighing sensors, fluid control valves and stirring devices to ensure that each nutrient is accurately mixed in proportion.

[0067] In one possible embodiment, during the postoperative recovery phase, Patient A underwent gastric surgery and required a high-protein, low-fat diet to promote wound healing. Through the human-computer interaction module 100, the device collected Patient A's age (50 years old), weight (70 kg), type of surgery (gastric surgery), nutritional diagnostic information (needs increased protein intake), and allergen information (lactose intolerance). Based on this information, the nutritional analysis module 200 predicted the required high-protein, low-fat, lactose-free nutritional profile for Patient A. The formulated food screening module 300, unable to find a perfectly matching product in the existing complete nutritional formula food database, input the nutritional profile information into a pre-trained nutrient component ratio prediction model, generating personalized nutrient component mixing ratio parameters. Finally, the formulated food mixing module 400 mixed specific nutrient components according to these parameters to prepare a formulated food suitable for Patient A.

[0068] In this embodiment, the detailed health information collected by the human-computer interaction module enables the nutrition analysis module to provide highly personalized nutritional ratio information, ensuring that postoperative patients receive the most suitable nutritional supply for their recovery needs. For special nutritional needs that cannot be met directly by standard complete nutritional formula foods, the formula food screening module, combined with the nutrient component ratio prediction model, can dynamically generate customized nutrient component mixing ratio parameters, solving the problem of difficulty in configuration under traditional methods.

[0069] Preferably, after inputting the nutrient ratio information into the pre-trained nutrient component ratio prediction model and outputting the nutrient component mixing ratio parameters, the method further includes:

[0070] Based on nutrient ratio information and nutrient component mixing ratio parameters, multiple first ratio parameters are generated within a preset range;

[0071] Based on the proportioning parameters and the flavor information of the nutrient components, a second proportioning parameter labeled with flavor information is generated;

[0072] Multiple secondary ratio parameters are displayed to the human-computer interaction module for the user to select, and the user's selection is determined as the new nutrient component mixing ratio parameters.

[0073] In this embodiment, a genetic algorithm is used to generate multiple first ratio parameters within a preset nutrient ratio range (such as the proportions of protein, carbohydrates, fats, vitamins, and minerals). The genetic algorithm can explore the nutrient ratio space and find the optimal or near-optimal solution. Furthermore, upper and lower limits for nutrient intake are set according to clinical nutrition guidelines and individual patient differences to ensure that the generated first ratio parameters meet safety and health standards.

[0074] A nutrient component flavor database is constructed to record the taste characteristics of each nutrient component (such as sweetness, acidity, bitterness, etc.) so as to combine taste characteristics with nutritional ratio parameters to generate a second ratio parameter that takes into account both nutrition and taste.

[0075] The human-computer interaction module 100 displays a list of second ratio parameters with flavor labels, allowing users to select according to their personal preferences. Users can select their preferred second ratio parameter by simply clicking or swiping. The user's final selection is used as the new nutrient component mixing ratio parameter. If the user does not select a preferred second ratio parameter within a preset time, the nutrient component mixing ratio parameter is used as the default selection.

[0076] In this embodiment, while ensuring nutritional balance, the proportion of nutrients is finely adjusted to provide a variety of flavored formula food options, which can meet the dietary preferences of different users, better adapt to individual differences, and achieve more precise nutritional supply.

[0077] Preferably, a complete nutritional formula food matching the nutritional ratio information is selected from the complete nutritional formula food information database. If no complete nutritional formula food matching the nutritional ratio information is found, it includes:

[0078] We selected complete nutritional formula foods from the complete nutritional formula food information database that contain information on all nutritional types and meet the nutritional type ratio information.

[0079] A comprehensive nutritional formula food information database will be constructed, including detailed nutritional data such as protein, fat, carbohydrates, vitamins, and minerals, as well as the food's formulation ratios and flavor information. Each entry will be associated with metadata such as food ID, origin, production date, and shelf life.

[0080] Based on nutrient type ratio information and fuzzy matching algorithm, we screen out complete nutritional formula foods (complete nutritional formula foods or specific complete nutritional formula foods) that contain all nutrient type information and meet the nutrient type ratio information from the complete nutritional formula food information database.

[0081] Preferably, the device further includes:

[0082] The ratio parameter mean smoothing module is used to perform mean smoothing on the nutrient component mixing ratio parameters of multiple different users. (See also...) Figure 2 Mean smoothing includes:

[0083] S410, Perform a first classification process on the mixing ratio parameters of multiple nutrient components to obtain multiple first categories containing the same combination of nutrient components;

[0084] S420, based on the similarity of the nutrient component ratios, the nutrient component mixing ratio parameters under the first category are subjected to a second classification process to obtain multiple second categories;

[0085] S430, calculate the average proportion of each nutrient component type in the second category, and apply the corresponding average proportion to the mixing ratio parameters of all nutrient components in the second category.

[0086] In this embodiment, the collected nutrient component mixing ratio parameters from different users are denoised and outlier detected to ensure data quality. A pattern recognition algorithm is used to identify the combination type of nutrient components, grouping parameters with the same nutrient component combination into the same first category. Based on the similarity matrix of nutrient component ratios, algorithms such as K-means or hierarchical clustering are used to further subdivide the parameters in the first category into multiple second categories. Similarity can be calculated using Euclidean distance, cosine similarity, or other applicable similarity metrics.

[0087] The average proportion of nutrient component types in each second category is calculated to obtain the average proportion of that category. The calculated average proportion is then applied to the mixing ratio parameters of all nutrient components in that second category to smooth and standardize the parameter distribution.

[0088] In one possible embodiment, the hyperparameters of the classification algorithm are tuned using methods such as cross-validation to obtain the best classification results.

[0089] In this embodiment, by mixing and smoothing the ratio parameters of nutrient components with similar parameters, the needs of formula foods with similar mixing and ratio parameters of nutrient components can be uniformly adjusted, which significantly improves the preparation speed and overall efficiency.

[0090] Preferably, see below. Figure 3 The device also includes:

[0091] The nutrient component ratio prediction model training module 500 is used to take nutrient type information and nutrient type ratio information as input data and nutrient component mixing ratio parameters as output data, and train the nutrient component ratio prediction model based on CNN-GRU network.

[0092] In the CNN-GRU network, the output of the GRU network serves as the input to the CNN network.

[0093] In this embodiment, a large amount of historical data is acquired as a sample set. The data includes, but is not limited to, user identity information, nutritional type information, and ratio information. The sample set is cleaned to remove outliers and missing values. Furthermore, the data is normalized to ensure that the input data are on the same scale, which is beneficial for model learning.

[0094] Nutrient type and nutrient ratio information are input into the hidden layer of the GRU network, which outputs a feature vector. The GRU network is trained using a two-layer GRU network to balance prediction performance and operational efficiency. The output feature vector of the GRU network is then input into the Attention mechanism layer to calculate the feature weights of different feature vectors and perform weighted summation. The weighted feature vector is then input into the CNN network, where convolutional layers capture local features, and finally, fully connected layers predict the final ratio parameters.

[0095] The model's prediction accuracy is calculated using a loss function. If the prediction accuracy does not meet the preset accuracy, iterative training continues until the preset accuracy is met. The evaluation metrics are also the mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE).

[0096] In this embodiment, by using a CNN-GRU network, the temporal and spatial features of nutrient type information can be better processed. Simultaneously, the Attention mechanism is used to strengthen the weights of important features, thereby more accurately predicting the mixing ratio parameters of nutrient components. This meets the nutritional needs of different users and improves the accuracy and practicality of personalized nutrition recommendations. In practical applications, it can quickly generate reasonable nutrient ratio schemes based on the user's nutrient type information, improving the efficiency and effectiveness of nutrition management.

[0097] Preferably, predicting the required nutritional ratios for a user based on their health information includes:

[0098] The user's health information is input into a pre-trained nutrition ratio prediction model to obtain nutrition ratio information;

[0099] The training process for the nutrient ratio prediction model is as follows:

[0100] Collect user health information and corresponding nutritional information from different groups of people to obtain a nutritional ratio dataset;

[0101] Principal component analysis algorithm is used to perform feature dimensionality reduction on the nutrient ratio dataset;

[0102] The GBDT algorithm was used to fit the dimensionality-reduced nutrient ratio dataset, and feature values ​​with a correlation greater than a preset threshold were selected based on the comprehensive index of the fitted dataset.

[0103] A random forest model is trained based on eigenvalue information to obtain a well-trained nutrient ratio prediction model.

[0104] In this embodiment, a large amount of user health information and corresponding nutritional ratio information are collected to form a nutritional ratio dataset. User health information may include, but is not limited to, age, gender, weight, physical activity level, medical history, genetic factors, etc. In order to ensure the quality of the data, data cleaning is performed to remove outliers and missing values, and numerical data is standardized to ensure that all features are on the same order of magnitude.

[0105] Principal Component Analysis (PCA) is used to reduce the dimensionality of the collected nutrient ratio dataset. PCA transforms the original high-dimensional features into a new set of orthogonal variables, namely principal components. The principal components are sorted according to the variance they explain. By retaining the first few principal components, the number of features is reduced, while keeping the variance of the data as constant as possible, thereby reducing the complexity of the model and the risk of overfitting.

[0106] The gradient boosting decision tree (GBDT) algorithm is used to fit the dimensionality-reduced nutrient ratio dataset. GBDT constructs multiple decision trees and adds them to the model one by one. The training objective of each tree is to fit a regression model on the residual of the previous tree, thereby gradually reducing the prediction error. Based on GBDT and according to the comprehensive index after fitting, feature values ​​with a correlation greater than a preset threshold with the nutrient ratio are selected to further optimize the feature set and remove redundant information.

[0107] Based on feature value information, a random forest (RF) model is trained. Random forest is an ensemble learning algorithm based on decision trees. It predicts by constructing multiple decision trees and using majority voting. During training, each tree is trained independently on a subset of features and a subset of samples, increasing the model's diversity and generalization ability. Through training, we obtain a well-trained nutritional ratio prediction model. When a user uploads their health information, this information is input into the trained nutritional ratio prediction model, which then outputs the user's required nutritional ratio information, including recommended intake proportions of various nutrients (such as protein, carbohydrates, fats, vitamins, and minerals).

[0108] Preferably, the nutrient components include: a protein component, a fat component, and a carbohydrate component.

[0109] In one possible embodiment, the nutrient component includes a thickening component and an electrolyte component.

[0110] Example 2: This invention provides a method for operating and managing formulated foods based on human-computer interaction technology. (See attached document.) Figure 1 The methods include:

[0111] S100: Obtain user health information provided by the user, including weight, type of surgery, nutritional diagnosis information, and allergen information;

[0112] S200 predicts the nutritional ratio information required by the user based on the user's health information. The nutritional ratio information includes nutritional type information and nutritional type ratio information.

[0113] S300: Match complete nutritional formula foods that meet the nutritional ratio information from the complete nutritional formula food information database. If no complete nutritional formula food that meets the nutritional ratio information is matched, input the nutritional ratio information into the pre-trained nutrient component ratio prediction model and output the nutrient component mixing ratio parameters. The nutrient component mixing ratio parameters include nutrient component type information and nutrient component ratio information.

[0114] S400: Mix the corresponding nutrient components according to the nutrient component mixing ratio parameters to obtain the formulated food.

[0115] Preferably, after inputting the nutrient ratio information into the pre-trained nutrient component ratio prediction model and outputting the nutrient component mixing ratio parameters, the method further includes:

[0116] Based on nutrient ratio information and nutrient component mixing ratio parameters, multiple first ratio parameters are generated within a preset range;

[0117] Based on the proportioning parameters and the flavor information of the nutrient components, a second proportioning parameter labeled with flavor information is generated;

[0118] Multiple secondary ratio parameters are displayed to the human-computer interaction module for the user to select, and the user's selection is determined as the new nutrient component mixing ratio parameters.

[0119] Preferably, a complete nutritional formula food matching the nutritional ratio information is selected from the complete nutritional formula food information database. If no complete nutritional formula food matching the nutritional ratio information is found, it includes:

[0120] We selected complete nutritional formula foods from the complete nutritional formula food information database that contain information on all nutritional types and meet the nutritional type ratio information.

[0121] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A device for operating and managing formulated food based on human-computer interaction technology, characterized in that, The device is used for nutritional management of postoperative patients, and the device includes: The human-computer interaction module is used to obtain user health information provided by the user, including age, weight, type of surgery, nutritional diagnosis information and allergen information. The nutrition analysis module is used to predict the required nutritional ratio information for the user based on the user's health information. The nutritional ratio information includes nutritional type information and nutritional type ratio information. The formulated food screening module is used to match complete nutritional formula foods that meet the nutritional ratio information from the complete nutritional formula food information database. If no complete nutritional formula food that meets the nutritional ratio information is matched, the nutritional ratio information is input into a pre-trained nutrient component ratio prediction model to dynamically generate nutrient component mixing ratio parameters, which are then directly transmitted to the formulated food mixing module for real-time mixing. The nutrient component mixing ratio parameters include nutrient component type information and nutrient component ratio information; the nutrient components include: protein components, fat components, and carbohydrate components. Based on nutrient ratio information and nutrient component mixing ratio parameters, multiple first ratio parameters are generated within a preset range; Construct a flavor database for nutrient components to record the taste characteristics of each nutrient component; Based on the proportioning parameters and the flavor information from the nutrient component flavor database, a second proportioning parameter labeled with flavor information is generated; Multiple secondary ratio parameters are displayed to the human-computer interaction module for the user to select, and the user's selection is determined as the new nutrient component mixing ratio parameters. A mean smoothing module for proportioning parameters is used to perform mean smoothing processing on the mixing proportioning parameters of the nutrient components from multiple different users to improve the consistency of the proportioning, including: A first classification process is performed on the combined proportion parameters of multiple nutrient components to obtain multiple first categories containing the same combination of nutrient components; Based on the similarity of the nutrient component ratios, the nutrient component mixing ratio parameters under the first category are subjected to a second classification process to obtain multiple second categories; Calculate the average proportion of each nutrient component type in the second category, and apply the corresponding average proportion to the mixing ratio parameters of all nutrient components in the second category. The formula food mixing module is used to mix the corresponding nutrient components in real time according to the nutrient component mixing ratio parameters after mean smoothing, so as to obtain a formula food suitable for postoperative patients.

2. The formula food operation and management device based on human-computer interaction technology according to claim 1, characterized in that, The complete nutritional formula foods that match the nutritional ratio information from the complete nutritional formula food information database include: Complete nutritional formula foods containing all the nutritional types and conforming to the nutritional type ratio information are selected from the complete nutritional formula food information database.

3. The formula food operation and management device based on human-computer interaction technology according to claim 1, characterized in that, The device further includes: The nutrient component ratio prediction model training module is used to take the nutrient type information and the nutrient type ratio information as input data, and the nutrient component mixing ratio parameters as output data, and train the nutrient component ratio prediction model based on the CNN-GRU network. In this CNN-GRU network, the output of the GRU network serves as the input to the CNN network.

4. The formula food operation and management device based on human-computer interaction technology according to claim 1, characterized in that, The step of predicting the required nutritional ratios for the user based on the user's health information includes: The user's health information is input into a pre-trained nutrition ratio prediction model to obtain the nutrition ratio information; The training process of the nutrient ratio prediction model is as follows: Collect the health information of users from different groups and the corresponding nutritional ratio information to obtain a nutritional ratio dataset; Principal component analysis was used to perform feature dimensionality reduction on the nutrient ratio dataset. The GBDT algorithm was used to fit the dimensionality-reduced nutrient ratio dataset, and feature values ​​with a correlation greater than a preset threshold were selected based on the comprehensive index of the fitted dataset. A random forest model is trained based on the feature value information to obtain the trained nutrient ratio prediction model.

5. A method for operating and managing formulated foods based on human-computer interaction technology, characterized in that, The method, applied to a human-computer interaction-based food preparation and management device as described in any one of claims 1-4, comprises: Obtain user health information provided by the user, including weight, type of surgery, nutritional diagnosis information, and allergen information; Based on the user's health information, the required nutritional ratio information for the user is predicted, and the nutritional ratio information includes nutritional type information and nutritional type ratio information. The system matches complete nutritional formula foods that match the nutritional ratio information from the complete nutritional formula food information database. If no complete nutritional formula food that matches the nutritional ratio information is matched, the nutritional ratio information is input into the pre-trained nutrient component ratio prediction model, and the nutrient component mixing ratio parameters are output. The nutrient component mixing ratio parameters include nutrient component type information and nutrient component ratio information. The corresponding nutrient components are mixed according to the nutrient component mixing ratio parameters to obtain the formulated food.

6. The method for operating and managing formulated food based on human-computer interaction technology according to claim 5, characterized in that, After inputting the nutrient ratio information into the pre-trained nutrient component ratio prediction model and outputting the nutrient component mixing ratio parameters, the method further includes: Based on the nutrient ratio information and the nutrient component mixing ratio parameters, multiple first ratio parameters are generated within a preset range; Based on the ratio parameters and the flavor information of the nutrient components, a second ratio parameter labeled with flavor information is generated; The multiple second ratio parameters are displayed to the human-computer interaction module for the user to select, and the user's selection is determined as the new mixing ratio parameters of the nutrient components.

7. The method for operating and managing formulated food based on human-computer interaction technology according to claim 5, characterized in that, The complete nutritional formula foods that match the nutritional ratio information from the complete nutritional formula food information database include: Complete nutritional formula foods containing all the nutritional types and conforming to the nutritional type ratio information are selected from the complete nutritional formula food information database.

Images