Child nutrition personalized formula recommendation method and system

Abstract

The invention relates to a child nutrition personalized formula recommendation method and system, and belongs to the field of medical care information. The method comprises the following steps: intelligently predicting a weight difference value before and after a target child takes a set health-care formula for a predicted number of days by adopting an A I formula prediction model according to the taking days, item-by-item associated information of the target child and multi-item formula data of the set health-care formula; when the weight difference value is a positive number and a single-day weight difference value obtained by dividing the weight difference value by the predicted number of days exceeds a limit, judging that the set health-care formula is a recommended health-care formula of the target child; according to the invention, aiming at the technical problem that different personalized formula recommendation schemes cannot be analyzed for different children in the prior art, the A I formula prediction model with a customized structure can be adopted to intelligently predict the body weight change before and after different children take different health-care formulas according to multiple items of basic data screened in a targeted manner; therefore, the technical problem is solved.

Description

Technical Field

[0001] The present invention relates to the field of medical care information, and in particular to a method and system for recommending personalized nutrition formulas for children. Background Art

[0002] Personalized nutrition formulas for children can be customized according to factors such as the child's age, gender, and health status to ensure comprehensive and balanced nutrition. In general, the basic principles of children's nutritional needs are as follows: Protein: Children need sufficient protein to support growth and development. High-quality protein sources include meat, fish, soy products, and dairy products. Carbohydrates: Provide energy to ensure that children have enough physical strength to carry out daily activities and learn. They mainly come from whole grains, potatoes, and fruits. Fats: Healthy fats are essential for brain development. Foods rich in unsaturated fats should be selected, such as nuts, fish, and olive oil. Vitamins and minerals: Ensure adequate intake of vitamins A, C, D and minerals such as calcium and iron. This can be achieved by eating more vegetables and fruits. Water: Maintain adequate water intake and drink at least 8 glasses of water a day.

[0003] For example, Chinese invention patent publication CN104522652A proposes a children's nutritional liquid formula and preparation method thereof, comprising the following steps: taking wolfberry, Rehmannia root, red ginseng, ginseng, Achyranthes bidentata, Ziziphus jujuba seeds, Astragalus root, Poria cocos, Chinese angelica root, and mother clove, crushing and grinding into powder, filtering through a 200-mesh sieve, adding the filtered powder to a distilling flask, adding clean water in a ratio of 1:3, and distilling at 60-70°C until the water evaporates to dryness, collecting the steam, mixing it with fruit juice, and adding rock sugar. The beneficial effects of the present invention are: the nutritional liquid prepared by this method is rich in nutrients, does not contain any toxins or additives, has a pure taste without a noticeable medicinal taste, is easy for children to drink, has a good effect, enhances children's resistance and immunity, and better helps children's physical development.

[0004] For example, Chinese invention patent publication CN105532919A proposes a complete nutritional formula for children and its preparation method. Its components include: 20% to 30% whole milk powder, 12% to 19% skim milk powder, 18% to 21% desalted whey powder, 12% to 18% carbohydrates, 6% to 12% vegetable oils and fats, 5% to 10% concentrated whey protein powder, 2% to 6% prebiotic composition, 1.5% to 4.5% lipid composition, 0.15% to 0.25% multivitamins, and 0.45% to 0.55% multiminerals. The complete nutritional formula for children contains high-quality whey protein and all the minerals, vitamins, and other nutrients necessary for children's bodies, and can be used as a single source of nutrition for people, including malnourished children. The preparation method of the present invention uses a wet process to add the lipid composition, reducing the loss of DHA and ARA during the wet addition process.

[0005] However, the above technical solutions are only limited to the material selection and preparation mode of children's nutritional formula, which has universal use significance for children of a certain age. It does not take into account that even different children of the same age may have different nutritional supply effects and physical development promotion effects due to differences in physical development status, eating habits, etc. When using the same nutritional formula, it is also possible. In other words, the various technical solutions in the existing technology do not take into account the provision of personalized formula recommendations for different children, resulting in the inability to improve the flexibility and effectiveness of children's nutritional rationing. Summary of the Invention

[0006] In order to solve the technical problems in the prior art, the present invention provides a method and system for recommending personalized nutrition formulas for children. Based on targeted screening of multiple comprehensive basic data including item-by-item related information related to the latest status of the target child, each formula data associated with the set health formula, and the target child's three-meal dosage of the set health formula, a customized AI formula prediction model is introduced to intelligently predict the weight difference of the target child before and after taking the set health formula in a selected taking mode, and based on the intelligent prediction results, it is determined whether the set health formula belongs to the recommended health formula for the target child, thereby providing valuable reference information for the analysis of different personalized formula recommendation plans for different children, thereby improving the flexibility and effectiveness of children's nutrition rationing.

[0007] According to a first aspect of the present invention, a method for recommending personalized nutritional formulas for children is provided, the method comprising:

[0008] Obtaining the target child's age information, height information, weight information, gender information, and recent illness type coding information to output as item-by-item associated information of the target child;

[0009] Obtaining the respective proportions of various ingredients of the set health care formula for the target child, the respective portion type coding data of various ingredients, and the three dosages corresponding to the three meals of the target child, and outputting the data as multiple formula data of the set health care formula for the target child;

[0010] Performing multiple training on the radial basis neural network to obtain a radial basis neural network after the multiple trainings and outputting the obtained radial basis neural network as an AI recipe prediction model;

[0011] The AI formula prediction model is used to intelligently predict the weight difference before and after the target child takes the set health formula after the predicted number of days the target child takes the set health formula, the target child's item-by-item related information, and multiple formula data of the set health formula taken by the target child;

[0012] When the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing it by the predicted number of days exceeds or equals the set weight threshold, the set health care formula is determined to be the recommended health care formula for the target child; otherwise, the set health care formula is determined to be a non-recommended health care formula for the target child;

[0013] Among them, performing multiple training on the radial basis neural network to obtain the radial basis neural network after multiple training and outputting it as the AI formula prediction model includes: the number of training times performed by the radial basis neural network is positively correlated with the value of the predicted days.

[0014] According to a second aspect of the present invention, a system for recommending personalized nutritional formulas for children is provided, the system comprising:

[0015] An information input mechanism is used to obtain the target child's age information, height information, weight information, gender information, and recent disease type coding information to output as item-by-item associated information of the target child;

[0016] a formula collection mechanism for obtaining the respective proportions of various ingredients of the set health formula for the target child, the respective type coding data of the various ingredients, and the three dosages corresponding to the three meals of the target child, and outputting the data as multiple formula data of the set health formula for the target child;

[0017] A multiple training mechanism is used to perform multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple training and output it as an AI recipe prediction model;

[0018] an intelligent prediction mechanism, connected to the information input mechanism, the formula collection mechanism, and the multiple training mechanism, respectively, for using an AI formula prediction model to intelligently predict the weight difference of the target child before and after taking the set health formula after the target child has taken the set health formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health formula taken by the target child;

[0019] a recommendation processing mechanism connected to the intelligent prediction mechanism, configured to determine that the set health care formula is a recommended health care formula for the target child when the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing the weight difference value by the predicted number of days exceeds or equals a set weight threshold; otherwise, determine that the set health care formula is a non-recommended health care formula for the target child;

[0020] Among them, performing multiple training on the radial basis neural network to obtain the radial basis neural network after multiple training and outputting it as the AI formula prediction model includes: the number of training times performed by the radial basis neural network is positively correlated with the value of the predicted days.

[0021] Compared with the prior art, the present invention has at least the following key inventive features:

[0022] Firstly, based on the item-by-item associated information related to the target child's latest status, the data of each formula associated with the set health formula, and the target child's dosage for the three meals of the set health formula, an artificial intelligence model is used to intelligently predict the target child's weight difference before and after taking the set health formula for the predicted number of days. When the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing it by the predicted number of days exceeds or equals a set weight threshold, the set health formula is determined to be the recommended health formula for the target child, thereby completing a reliable prediction of the future effects of different health formulas on different children. This provides key basic data for the customization of personalized health formula recommendation strategies for different children without the need for the children to undergo actual multi-day dosage tests.

[0023] Secondly, the artificial intelligence model that performs intelligent prediction of the weight difference between the target child and the child after taking the set health care formula for the predicted number of days is a customized AI formula prediction model. The structural customization of the AI formula prediction model is manifested in that the AI formula prediction model is a radial basis function neural network that has completed multiple training sessions, and the number of training sessions performed by the radial basis function neural network is positively correlated with the value of the predicted number of days, thereby providing AI formula prediction models with different structures for taking health care formulas for different durations, ensuring the effectiveness and stability of the intelligent prediction results.

[0024] Thirdly, a comprehensive set of basic data is introduced to perform intelligent prediction of the weight difference before and after the target child takes the set health care formula for a predicted number of days. The basic data includes item-by-item associated information related to the target child's latest status, the data of each portion of the set health care formula, and the target child's three-meal dosage of the set health care formula. Specifically, the item-by-item associated information related to the target child's latest status includes the target child's age, height, weight, gender, and the coding information of the most recent illness type. The targeted screening of the above basic data further ensures the effectiveness and stability of the intelligent prediction results.

[0025] Finally, in each training session of the radial basis neural network, the difference in weight before and after a child takes a certain health formula for a certain number of days is used as the output of the radial basis neural network, and the health formula, the item-by-item association information of the child, and multiple formula data of the health formula are used as the input of the radial basis neural network to perform this training, thereby ensuring the training effect of each training session of the radial basis neural network. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The embodiments of the present invention will be described below with reference to the accompanying drawings, in which:

[0027] Figure 1 The present invention is a technical flow chart of the method and system for recommending personalized nutrition formulas for children.

[0028] Figure 2 The present invention is a flowchart of the steps of a method for recommending personalized nutritional formulas for children according to Example 1 of the present invention.

[0029] Figure 3 The figure is a flowchart of the steps of the method for recommending personalized nutrition formula for children according to Example 2 of the present invention.

[0030] Figure 4 The present invention is a flowchart of the steps of a method for recommending personalized nutritional formulas for children according to Example 3 of the present invention.

[0031] Figure 5 The present invention is a flowchart of the steps of a method for recommending personalized nutritional formulas for children according to Example 4 of the present invention.

[0032] Figure 6 FIG. 5 is a schematic structural diagram of an electronic device according to Embodiment 5 of the present invention.

[0033] Figure 7 Schematic diagram of the structure of a personalized nutrition formula recommendation system for children according to Example 6 of the present invention. DETAILED DESCRIPTION

[0034] like Figure 1 As shown, a technical flow chart of the method and system for recommending personalized nutrition formula for children according to the present invention is given.

[0035] exist Figure 1 The specific technical process of the present invention is as follows:

[0036] Technical Process A: Design a customized AI formula prediction model to intelligently predict the weight difference between target children before and after taking a set health formula for a predicted number of days.

[0037] For example, the structural customization of the AI recipe prediction model is mainly reflected in the following aspects:

[0038] First aspect: the AI recipe prediction model is a radial basis neural network that has been trained multiple times;

[0039] Second, the number of training sessions performed by the radial basis function neural network is positively correlated with the number of days predicted, thereby providing AI formula prediction models with different structures for health formulas of different durations, ensuring the effectiveness and stability of the intelligent prediction results.

[0040] Thirdly, in each training of the radial basis neural network, the difference in weight before and after a child takes a certain health formula for a certain number of days is used as the output of the radial basis neural network, and the health formula, the item-by-item associated information of the child, and multiple pieces of formula data of the health formula are used as the input of the radial basis neural network to perform this training, thereby ensuring the training effect of each training of the radial basis neural network;

[0041] Technical Process B: Introducing comprehensive and comprehensive basic data to perform intelligent prediction of the weight difference before and after the target child takes the set health formula for the predicted number of days;

[0042] Specifically, if Figure 1 As shown, the multiple basic data include item-by-item related information related to the latest status of the target child, each recipe data associated with the set health care formula, and the target child's three-meal dosage for the set health care formula;

[0043] More specifically, the item-by-item associated information related to the target child's latest status includes the target child's age information, height information, weight information, gender information, and the coding information of the most recent illness type;

[0044] More specifically, the target child's most recent illness type coding information is the coding information corresponding to the disease type with the shortest duration of illness for the target child, and the single type coding data corresponding to each ingredient is the coding data of the type corresponding to the ingredient;

[0045] The targeted screening of the above-mentioned multiple basic data further ensures the effectiveness and stability of the intelligent prediction results;

[0046] Technical Process C: The AI formula prediction model, customized by Technical Process A, uses multiple basic data specifically selected by Technical Process B to intelligently predict the weight difference before and after the target child takes the set health formula for the predicted number of days;

[0047] Technical process D: judging whether the set health care formula is the recommended health care formula for the target child based on the intelligent prediction result of technical process D;

[0048] For example, when the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing the weight difference value by the predicted number of days exceeds or equals the set weight threshold, the set health care formula is determined to be the recommended health care formula for the target child; otherwise, the set health care formula is determined to be a non-recommended health care formula for the target child;

[0049] In this way, based on the item-by-item associated information related to the target child's latest status, the formula data associated with the set health formula, and the target child's three-meal dosage of the set health formula, an artificial intelligence model is used to intelligently predict the target child's weight difference before and after taking the set health formula for the predicted number of days. When the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing it by the predicted number of days exceeds or equals the set weight threshold, the set health formula is determined to be the recommended health formula for the target child, thereby completing a reliable prediction of the future effects of different health formulas on different children.

[0050] The present invention can provide key basic data for tailoring personalized health care formula recommendation strategies for different children without the need for children to undergo actual multi-day consumption tests.

[0051] The key points of the present invention are: customization of personalized formula recommendation strategies for different children based on intelligent prediction results of weight change data before and after different children take different health care product formulas, customized structural design of AI formula prediction model, targeted screening of multiple basic data in a comprehensive and comprehensive manner, and targeted design of each training of radial basis neural network.

[0052] The method and system for recommending personalized nutritional formulas for children of the present invention will be described in detail below by way of examples.

[0053] Example 1

[0054] Figure 2 The present invention is a flowchart of the steps of a method for recommending personalized nutritional formulas for children according to Example 1 of the present invention.

[0055] like Figure 2 As shown, the method for recommending personalized nutritional formulas for children includes the following specific steps:

[0056] Step S1: Obtain the target child's age information, height information, weight information, gender information, and recent illness type coding information to output as item-by-item associated information of the target child;

[0057] Specifically, obtaining the target child's age information, height information, weight information, gender information, and recent disease type coding information as the target child's item-by-item associated information output includes: using different information collection components to respectively obtain the target child's age information, height information, weight information, gender information, and recent disease type coding information;

[0058] Step S2: Obtaining the respective proportions of various ingredients of the set health care formula for the target child, the respective type coding data of the various ingredients, and the three dosages corresponding to the three meals of the target child, and outputting them as multiple formula data of the set health care formula for the target child;

[0059] Specifically, the proportions of each serving of various ingredients of the set health care formula taken by the target child, the type coding data of each serving of various ingredients, and three serving dosages corresponding to the three meals of the target child are obtained and output as multiple formula data of the set health care formula taken by the target child, including: the three serving dosages corresponding to the three meals of the target child are the three serving dosages corresponding to the three meals of the target child for breakfast, lunch, and lunch respectively;

[0060] Step S3: Perform multiple training on the radial basis neural network to obtain a radial basis neural network after multiple training and output it as an AI recipe prediction model;

[0061] For example, performing multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple trainings and outputting the radial basis neural network as the AI recipe prediction model includes: selecting a numerical simulation mode to implement testing and simulation of a data processing process of performing multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple trainings and outputting the radial basis neural network as the AI recipe prediction model;

[0062] Step S4: using an AI formula prediction model to intelligently predict the weight difference of the target child before and after taking the set health formula after the target child has taken the set health formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health formula taken by the target child;

[0063] Specifically, an AI formula prediction model is used to intelligently predict the weight difference of the target child before and after taking the set health care formula after the target child takes the set health care formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health care formula taken by the target child. The weight difference includes: the weight difference before and after the target child takes the set health care formula for the predicted number of days, the predicted number of days the target child takes the set health care formula, the target child's item-by-item associated information, and multiple formula data of the set health care formula taken by the target child. The input content and output content of the AI formula prediction model can all be in the form of numerical normalized representation;

[0064] Step S5: When the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing the weight difference value by the predicted number of days exceeds or equals the set weight threshold, the set health care formula is determined to be the recommended health care formula for the target child; otherwise, the set health care formula is determined to be a non-recommended health care formula for the target child;

[0065] The step of performing multiple training on the radial basis neural network to obtain a radial basis neural network after the multiple trainings and outputting the obtained radial basis neural network as the AI recipe prediction model comprises: the number of trainings performed by the radial basis neural network is positively correlated with the value of the predicted days;

[0066] For example, the positive correlation between the number of training times performed by the radial basis neural network and the value of the predicted days includes: when the predicted days are 5 days, the number of training times performed by the radial basis neural network is 150 times; when the predicted days are 10 days, the number of training times performed by the radial basis neural network is 300 times; when the predicted days are 15 days, the number of training times performed by the radial basis neural network is 500 times; when the predicted days are 25 days, the number of training times performed by the radial basis neural network is 700 times, and so on;

[0067] The step of training the radial basis neural network multiple times to obtain a radial basis neural network after the multiple trainings and outputting the trained network as the AI formula prediction model further includes: in each training of the radial basis neural network, using a known weight difference between a child taking a certain health formula for a certain number of days before and after taking the formula as output content of the radial basis neural network, using the health formula, item-by-item association information of the child, and multiple pieces of formula data of the health formula as input content of the radial basis neural network, and performing the training;

[0068] The step of obtaining the target child's age information, height information, weight information, gender information, and recent illness type coding information as the target child's item-by-item associated information output includes: the target child's recent illness type coding information is coding information corresponding to the disease type with the shortest duration of illness of the target child;

[0069] For example, the coding information of the target child's most recent illness type is the coding information corresponding to the disease type with the shortest illness duration of the target child, including: the coding information corresponding to the disease type with the shortest illness duration of the target child is the ASCII code value of the disease type name corresponding to the disease type with the shortest illness duration of the target child;

[0070] The output of the multiple formula data of the set health care formula for the target child includes: the single-serving type code data corresponding to each ingredient is the code data of the type corresponding to the ingredient;

[0071] Wherein, when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing the weight difference by the predicted number of days exceeds or equals a set weight threshold, the set health care formula is judged to be a recommended health care formula for the target child; otherwise, judging the set health care formula to be a non-recommended health care formula for the target child includes: when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing the weight difference by the predicted number of days is less than a set weight threshold, judging the set health care formula to be a non-recommended health care formula for the target child;

[0072] And wherein, when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing it by the predicted number of days exceeds or equals the set weight threshold, the set health care formula is judged to be the recommended health care formula for the target child; otherwise, judging that the set health care formula is a non-recommended health care formula for the target child also includes: when the intelligently predicted weight difference is a negative number, judging that the set health care formula is a non-recommended health care formula for the target child.

[0073] Example 2

[0074] Figure 3 The figure is a flowchart of the steps of the method for recommending personalized nutrition formula for children according to Example 2 of the present invention.

[0075] like Figure 3 As shown, Figure 2 Unlike the embodiment in , before obtaining the target child's age information, height information, weight information, gender information, and the latest disease type code information as the target child's item-by-item associated information output, that is, before step S1, the method further includes:

[0076] Step S6: Identify the target child using the identity recognition mode, and analyze the target child's age information, height information, weight information, gender information, and recent illness type coding information based on the identity recognition result;

[0077] Specifically, an identity recognition mode is used to identify the target child, and the target child's age information, height information, weight information, gender information and recent illness type coding information are parsed based on the identity recognition result, including: an identity recognition mode based on a visual recognition mechanism is used to identify the target child, and the target child's age information, height information, weight information, gender information and recent illness type coding information are parsed based on the identity recognition result.

[0078] Example 3

[0079] Figure 4 The present invention is a flowchart of the steps of a method for recommending personalized nutritional formulas for children according to Example 3 of the present invention.

[0080] like Figure 4 As shown, Figure 2 Unlike the embodiment in , before obtaining the respective portion proportions of various ingredients of the set health care formula for the target child, the respective portion type coding data of various ingredients, and the three serving dosages corresponding to the three meals of the target child and outputting them as the multiple formula data of the set health care formula for the target child, before step S2, the method further includes:

[0081] Step S7: Accessing the recipe management server through the network transmission interface to obtain the respective proportions of various ingredients of the set health care recipe for the target child and the respective type coding data of various ingredients;

[0082] Specifically, the network transmission interface is used to access the formula management server to obtain the proportions of each portion of various ingredients of the set health formula for the target child and the type coding data of each portion corresponding to each ingredient, including: the formula management server is a big data management node.

[0083] Example 4

[0084] Figure 5 The present invention is a flowchart of the steps of a method for recommending personalized nutritional formulas for children according to Example 4 of the present invention.

[0085] like Figure 5 As shown, Figure 2 Unlike the embodiment in , when the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing the weight difference value by the predicted number of days exceeds or equals the set weight threshold, it is determined that the set health care formula is the recommended health care formula for the target child; otherwise, after determining that the set health care formula is the non-recommended health care formula for the target child, that is, after step S5, the method further includes:

[0086] Step S8: After determining that the set health formula is the recommended health formula for the target child, the name information of the set health formula and the identification result of the target child are packaged into the same network data packet, and the network data packet is wirelessly transmitted to a remote child nutrition management network element using a wireless communication link;

[0087] wherein, after determining that the set health formula is the recommended health formula for the target child, packaging the name information of the set health formula and the identification result of the target child into the same network data packet, and then wirelessly transmitting the network data packet to a remote child nutrition management network element using a wireless communication link includes: the wireless communication link being a frequency division duplex communication link or a time division duplex communication link;

[0088] For example, after determining that the set health care formula is the recommended health care formula for the target child, the name information of the set health care formula and the identity identification result of the target child are packaged into the same network data packet, and then the network data packet is wirelessly sent to the remote child nutrition management network element using a wireless communication link, which also includes: the network data packet is an IP data packet.

[0089] Next, various method embodiments of the present invention will be described in detail.

[0090] In the method for recommending personalized nutritional formulas for children according to various embodiments of the present invention:

[0091] The AI formula prediction model is used to intelligently predict the weight difference of the target child before and after taking the set health care formula after the target child takes the set health care formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health care formula taken by the target child, including: synchronously inputting the predicted number of days the target child takes the set health care formula, the target child's item-by-item associated information, and the multiple formula data of the set health care formula taken by the target child into the AI formula prediction model, and running the AI formula prediction model to obtain the weight difference of the target child before and after taking the set health care formula for the predicted number of days, output by the AI formula prediction model;

[0092] Specifically, the predicted number of days for the target child to take the set health formula, the item-by-item association information of the target child, and multiple formula data of the set health formula taken by the target child are synchronously input into the AI formula prediction model, and the AI formula prediction model is run to obtain the weight difference of the target child before and after taking the set health formula for the predicted number of days, as output by the AI formula prediction model, including: calculating the difference between two weights corresponding to the same time on two days before and after the target child takes the set health formula for the predicted number of days;

[0093] Among them, the predicted number of days for the target child to take the set health care formula, the item-by-item association information of the target child, and the multiple formula data of the set health care formula taken by the target child are synchronously input into the AI formula prediction model, and the AI formula prediction model is run to obtain the weight difference of the target child before and after taking the set health care formula for the predicted number of days output by the AI formula prediction model, which includes: performing binary numerical processing on the predicted number of days for the target child to take the set health care formula, the item-by-item association information of the target child, and the multiple formula data of the set health care formula taken by the target child, and then synchronously inputting them into the AI formula prediction model.

[0094] And in the method for recommending personalized nutrition formula for children according to various method embodiments of the present invention:

[0095] Synchronously inputting the predicted number of days for the target child to take the set health-care formula, the item-by-item associated information of the target child, and multiple formula data of the set health-care formula taken by the target child into the AI formula prediction model, and running the AI formula prediction model to obtain the weight difference of the target child before and after taking the set health-care formula for the predicted number of days output by the AI formula prediction model, including: the weight difference of the target child before and after taking the set health-care formula for the predicted number of days output by the AI formula prediction model is in a binary numerical representation;

[0096] The binary numerical representation of the weight difference before and after the target child takes the set health-care formula for the predicted number of days output by the AI formula prediction model includes: the weight difference before and after the target child takes the set health-care formula for the predicted number of days output by the AI formula prediction model is a binary numerical value with a fixed number of digits, and the first digit of the binary numerical value with a fixed number of digits represents the positive or negative value of the weight difference before and after the target child takes the set health-care formula for the predicted number of days;

[0097] And wherein, the weight difference of the target child before and after taking the set health-care formula for the predicted number of days output by the AI formula prediction model is in a binary numerical representation, further comprising: the weight difference of the target child before and after taking the set health-care formula for the predicted number of days output by the AI formula prediction model is the weight difference obtained by subtracting the weight of the target child before taking the set health-care formula for the predicted number of days from the weight value of the target child after taking the set health-care formula for the predicted number of days.

[0098] Example 5

[0099] Figure 6 FIG. 5 is a schematic diagram of the structure of an electronic device according to Embodiment 5 of the present invention. Figure 6 As shown, the electronic device includes a memory and one or more processors, the memory stores a computer program, and the computer program is configured to be executed by the one or more processors to complete the following steps:

[0100] Step S1: Obtain the target child's age information, height information, weight information, gender information, and recent illness type coding information to output as item-by-item associated information of the target child;

[0101] Specifically, obtaining the target child's age information, height information, weight information, gender information, and recent disease type coding information as the target child's item-by-item associated information output includes: using different information collection components to respectively obtain the target child's age information, height information, weight information, gender information, and recent disease type coding information;

[0102] Step S2: Obtaining the respective proportions of various ingredients of the set health care formula for the target child, the respective type coding data of the various ingredients, and the three dosages corresponding to the three meals of the target child, and outputting them as multiple formula data of the set health care formula for the target child;

[0103] Specifically, the proportions of each serving of various ingredients of the set health care formula taken by the target child, the type coding data of each serving of various ingredients, and three serving dosages corresponding to the three meals of the target child are obtained and output as multiple formula data of the set health care formula taken by the target child, including: the three serving dosages corresponding to the three meals of the target child are the three serving dosages corresponding to the three meals of the target child for breakfast, lunch, and lunch respectively;

[0104] Step S3: Perform multiple training on the radial basis neural network to obtain a radial basis neural network after multiple training and output it as an AI recipe prediction model;

[0105] For example, performing multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple trainings and outputting the radial basis neural network as the AI recipe prediction model includes: selecting a numerical simulation mode to implement testing and simulation of a data processing process of performing multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple trainings and outputting the radial basis neural network as the AI recipe prediction model;

[0106] Step S4: using an AI formula prediction model to intelligently predict the weight difference of the target child before and after taking the set health formula after the target child has taken the set health formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health formula taken by the target child;

[0107] Specifically, an AI formula prediction model is used to intelligently predict the weight difference of the target child before and after taking the set health care formula after the target child takes the set health care formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health care formula taken by the target child. The weight difference includes: the weight difference before and after the target child takes the set health care formula for the predicted number of days, the predicted number of days the target child takes the set health care formula, the target child's item-by-item associated information, and multiple formula data of the set health care formula taken by the target child. The input content and output content of the AI formula prediction model can all be in the form of numerical normalized representation;

[0108] Step S5: When the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing the weight difference value by the predicted number of days exceeds or equals the set weight threshold, the set health care formula is determined to be the recommended health care formula for the target child; otherwise, the set health care formula is determined to be a non-recommended health care formula for the target child;

[0109] The step of performing multiple training on the radial basis neural network to obtain a radial basis neural network after the multiple trainings and outputting the obtained radial basis neural network as the AI recipe prediction model comprises: the number of trainings performed by the radial basis neural network is positively correlated with the value of the predicted days;

[0110] For example, the positive correlation between the number of training times performed by the radial basis neural network and the value of the predicted days includes: when the predicted days are 5 days, the number of training times performed by the radial basis neural network is 150 times; when the predicted days are 10 days, the number of training times performed by the radial basis neural network is 300 times; when the predicted days are 15 days, the number of training times performed by the radial basis neural network is 500 times; when the predicted days are 25 days, the number of training times performed by the radial basis neural network is 700 times, and so on;

[0111] The step of training the radial basis neural network multiple times to obtain a radial basis neural network after the multiple trainings and outputting the trained network as the AI formula prediction model further includes: in each training of the radial basis neural network, using a known weight difference between a child taking a certain health formula for a certain number of days before and after taking the formula as output content of the radial basis neural network, using the health formula, item-by-item association information of the child, and multiple pieces of formula data of the health formula as input content of the radial basis neural network, and performing the training;

[0112] The step of obtaining the target child's age information, height information, weight information, gender information, and recent illness type coding information as the target child's item-by-item associated information output includes: the target child's recent illness type coding information is coding information corresponding to the disease type with the shortest duration of illness of the target child;

[0113] For example, the coding information of the target child's most recent illness type is the coding information corresponding to the disease type with the shortest illness duration of the target child, including: the coding information corresponding to the disease type with the shortest illness duration of the target child is the ASCII code value of the disease type name corresponding to the disease type with the shortest illness duration of the target child;

[0114] The output of the multiple formula data of the set health care formula for the target child includes: the single-serving type code data corresponding to each ingredient is the code data of the type corresponding to the ingredient;

[0115] Wherein, when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing the weight difference by the predicted number of days exceeds or equals a set weight threshold, the set health care formula is judged to be a recommended health care formula for the target child; otherwise, judging the set health care formula to be a non-recommended health care formula for the target child includes: when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing the weight difference by the predicted number of days is less than a set weight threshold, judging the set health care formula to be a non-recommended health care formula for the target child;

[0116] and wherein, when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing the weight difference by the predicted number of days exceeds or equals a set weight threshold, determining that the set health care formula is a recommended health care formula for the target child; otherwise, determining that the set health care formula is a non-recommended health care formula for the target child further comprises: when the intelligently predicted weight difference is a negative number, determining that the set health care formula is a non-recommended health care formula for the target child;

[0117] like Figure 6As shown, illustratively, P processors are provided, where P is a natural number greater than or equal to 1.

[0118] Example 6

[0119] Figure 7 Schematic diagram of the structure of a personalized nutrition formula recommendation system for children according to Example 6 of the present invention.

[0120] like Figure 7 As shown, the children's nutrition personalized formula recommendation system includes the following components:

[0121] An information input mechanism is used to obtain the target child's age information, height information, weight information, gender information, and recent disease type coding information to output as item-by-item associated information of the target child;

[0122] Specifically, obtaining the target child's age information, height information, weight information, gender information, and recent disease type coding information as the target child's item-by-item associated information output includes: using different information collection components to respectively obtain the target child's age information, height information, weight information, gender information, and recent disease type coding information;

[0123] a formula collection mechanism for obtaining the respective proportions of various ingredients of the set health formula for the target child, the respective type coding data of the various ingredients, and the three dosages corresponding to the three meals of the target child, and outputting the data as multiple formula data of the set health formula for the target child;

[0124] Specifically, the proportions of each serving of various ingredients of the set health care formula taken by the target child, the type coding data of each serving of various ingredients, and three serving dosages corresponding to the three meals of the target child are obtained and output as multiple formula data of the set health care formula taken by the target child, including: the three serving dosages corresponding to the three meals of the target child are the three serving dosages corresponding to the three meals of the target child for breakfast, lunch, and lunch respectively;

[0125] A multiple training mechanism is used to perform multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple training and output it as an AI recipe prediction model;

[0126] For example, performing multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple trainings and outputting the radial basis neural network as the AI recipe prediction model includes: selecting a numerical simulation mode to implement testing and simulation of a data processing process of performing multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple trainings and outputting the radial basis neural network as the AI recipe prediction model;

[0127] an intelligent prediction mechanism, connected to the information input mechanism, the formula collection mechanism, and the multiple training mechanism, respectively, for using an AI formula prediction model to intelligently predict the weight difference of the target child before and after taking the set health formula after the target child has taken the set health formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health formula taken by the target child;

[0128] Specifically, an AI formula prediction model is used to intelligently predict the weight difference of the target child before and after taking the set health care formula after the target child takes the set health care formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health care formula taken by the target child. The weight difference includes: the weight difference before and after the target child takes the set health care formula for the predicted number of days, the predicted number of days the target child takes the set health care formula, the target child's item-by-item associated information, and multiple formula data of the set health care formula taken by the target child. The input content and output content of the AI formula prediction model can all be in the form of numerical normalized representation;

[0129] a recommendation processing mechanism connected to the intelligent prediction mechanism, configured to determine that the set health care formula is a recommended health care formula for the target child when the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing the weight difference value by the predicted number of days exceeds or equals a set weight threshold; otherwise, determine that the set health care formula is a non-recommended health care formula for the target child;

[0130] The step of performing multiple training on the radial basis neural network to obtain a radial basis neural network after the multiple trainings and outputting the obtained radial basis neural network as the AI recipe prediction model comprises: the number of trainings performed by the radial basis neural network is positively correlated with the value of the predicted days;

[0131] For example, the positive correlation between the number of training times performed by the radial basis neural network and the value of the predicted days includes: when the predicted days are 5 days, the number of training times performed by the radial basis neural network is 150 times; when the predicted days are 10 days, the number of training times performed by the radial basis neural network is 300 times; when the predicted days are 15 days, the number of training times performed by the radial basis neural network is 500 times; when the predicted days are 25 days, the number of training times performed by the radial basis neural network is 700 times, and so on;

[0132] The step of training the radial basis neural network multiple times to obtain a radial basis neural network after the multiple trainings and outputting the trained network as the AI formula prediction model further includes: in each training of the radial basis neural network, using a known weight difference between a child taking a certain health formula for a certain number of days before and after taking the formula as output content of the radial basis neural network, using the health formula, item-by-item association information of the child, and multiple pieces of formula data of the health formula as input content of the radial basis neural network, and performing the training;

[0133] The step of obtaining the target child's age information, height information, weight information, gender information, and recent illness type coding information as the target child's item-by-item associated information output includes: the target child's recent illness type coding information is coding information corresponding to the disease type with the shortest duration of illness of the target child;

[0134] For example, the coding information of the target child's most recent illness type is the coding information corresponding to the disease type with the shortest illness duration of the target child, including: the coding information corresponding to the disease type with the shortest illness duration of the target child is the ASCII code value of the disease type name corresponding to the disease type with the shortest illness duration of the target child;

[0135] The output of the multiple formula data of the set health care formula for the target child includes: the single-serving type code data corresponding to each ingredient is the code data of the type corresponding to the ingredient;

[0136] Wherein, when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing the weight difference by the predicted number of days exceeds or equals a set weight threshold, the set health care formula is judged to be a recommended health care formula for the target child; otherwise, judging the set health care formula to be a non-recommended health care formula for the target child includes: when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing the weight difference by the predicted number of days is less than a set weight threshold, judging the set health care formula to be a non-recommended health care formula for the target child;

[0137] And wherein, when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing it by the predicted number of days exceeds or equals the set weight threshold, the set health care formula is judged to be the recommended health care formula for the target child; otherwise, judging that the set health care formula is a non-recommended health care formula for the target child also includes: when the intelligently predicted weight difference is a negative number, judging that the set health care formula is a non-recommended health care formula for the target child.

[0138] In addition, the present invention may also cite the following technical contents to further demonstrate the outstanding substantial progress of the present invention:

[0139] Synchronously inputting the predicted number of days for the target child to take the set health-care formula, the item-by-item associated information of the target child, and multiple formula data of the set health-care formula taken by the target child into the AI formula prediction model, and running the AI formula prediction model to obtain the weight difference of the target child before and after taking the set health-care formula for the predicted number of days output by the AI formula prediction model, further comprising: using a synchronous control interface to complete synchronous input control of the predicted number of days for the target child to take the set health-care formula, the item-by-item associated information of the target child, and multiple formula data of the set health-care formula taken by the target child into the AI formula prediction model;

[0140] For example, using a synchronous control interface to complete synchronous input control of the predicted number of days for the target child to take the set health care formula, the item-by-item associated information of the target child, and multiple formula data of the set health care formula taken by the target child into the AI formula prediction model includes: optionally using a programmable logic control device to implement the synchronous control interface;

[0141] The positive correlation between the number of training times performed by the radial basis function neural network and the value of the predicted number of days comprises: selecting a numerical conversion formula to express a numerical conversion relationship that positively correlates the number of training times performed by the radial basis function neural network and the value of the predicted number of days;

[0142] And wherein, the numerical conversion relationship in which the number of training times performed by the radial basis neural network and the value of the predicted days are positively correlated by a numerical conversion formula is selected, including: in the numerical conversion formula, the value of the predicted days is the input value of the numerical conversion formula, and the number of training times performed by the radial basis neural network corresponding to the value of the predicted days is the output value of the numerical conversion formula.

[0143] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present disclosure, and they should all be included in the scope of the claims and description of the present disclosure.

Claims

1. A method for recommending personalized nutritional formulas for children, characterized in that: The method comprises: Obtaining the target child's age information, height information, weight information, gender information, and recent illness type coding information to output as item-by-item associated information of the target child; Obtaining the respective proportions of various ingredients of the set health care formula for the target child, the respective portion type coding data of various ingredients, and the three dosages corresponding to the three meals of the target child, and outputting the data as multiple formula data of the set health care formula for the target child; Performing multiple training on the radial basis neural network to obtain a radial basis neural network after the multiple trainings and outputting the obtained radial basis neural network as an AI recipe prediction model; The AI formula prediction model is used to intelligently predict the weight difference before and after the target child takes the set health formula after the predicted number of days the target child takes the set health formula, the target child's item-by-item related information, and multiple formula data of the set health formula taken by the target child; When the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing it by the predicted number of days exceeds or equals the set weight threshold, the set health care formula is determined to be the recommended health care formula for the target child; otherwise, the set health care formula is determined to be a non-recommended health care formula for the target child; Among them, performing multiple training on the radial basis neural network to obtain the radial basis neural network after multiple training and outputting it as the AI formula prediction model includes: the number of training times performed by the radial basis neural network is positively correlated with the value of the predicted days.

2. The method for recommending personalized nutritional formulas for children according to claim 1, wherein: Performing multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple trainings and outputting it as the AI formula prediction model also includes: in each training performed on the radial basis neural network, using the difference in weight before and after a child takes a certain health formula for a certain number of days as the output content of the radial basis neural network, using the health formula, the item-by-item association information of the child, and multiple formula data of the health formula as the input content of the radial basis neural network to perform this training.

3. The method for recommending personalized nutritional formulas for children according to claim 2, wherein: The target child's age information, height information, weight information, gender information, and recent illness type coding information are obtained to output as item-by-item associated information of the target child, including: the target child's recent illness type coding information is the coding information corresponding to the disease type with the shortest duration of illness of the target child; The output of the multiple formula data of the set health care formula for the target child includes: the single-serving type code data corresponding to each ingredient is the code data of the type corresponding to the ingredient; Wherein, when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing the weight difference by the predicted number of days exceeds or equals a set weight threshold, the set health care formula is judged to be a recommended health care formula for the target child; otherwise, judging the set health care formula to be a non-recommended health care formula for the target child includes: when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing the weight difference by the predicted number of days is less than a set weight threshold, judging the set health care formula to be a non-recommended health care formula for the target child; Among them, when the intelligently predicted weight difference is a positive number and the single-day weight difference obtained by dividing it by the predicted number of days exceeds or equals the set weight threshold, the set health care formula is judged to be the recommended health care formula for the target child. Otherwise, judging that the set health care formula is a non-recommended health care formula for the target child also includes: when the intelligently predicted weight difference is a negative number, judging that the set health care formula is a non-recommended health care formula for the target child.

4. The method for recommending personalized nutrition formulas for children according to claim 3, wherein: Before obtaining the target child's age information, height information, weight information, gender information, and recent illness type coding information to output as the target child's item-by-item associated information, the method further includes: The target child is identified using an identity recognition model, and the target child's age information, height information, weight information, gender information, and recent illness type coding information are parsed based on the identity recognition results.

5. The method for recommending personalized nutrition formulas for children according to claim 3, wherein: Before obtaining the respective portion proportions of various ingredients of the set health formula for the target child, the respective portion type coding data of the various ingredients, and the three serving dosages corresponding to the three meals of the target child and outputting them as the multiple formula data of the set health formula for the target child, the method further includes: The network transmission interface is used to access the formula management server to obtain the respective proportions of various ingredients of the set health care formula for the target child and the respective type coding data of various ingredients.

6. The method for recommending personalized nutrition formulas for children according to claim 3, wherein: When the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing the weight difference value by the predicted number of days exceeds or equals a set weight threshold, it is determined that the set health care formula is a recommended health care formula for the target child; otherwise, after determining that the set health care formula is a non-recommended health care formula for the target child, the method further includes: After determining that the set health formula is the recommended health formula for the target child, the name information of the set health formula and the identification result of the target child are packaged into the same network data packet, and the network data packet is wirelessly transmitted to a remote child nutrition management network element using a wireless communication link; Among them, after determining that the set health care formula is the recommended health care formula for the target child, the name information of the set health care formula and the identity identification result of the target child are packaged into the same network data packet, and then the network data packet is wirelessly sent to a remote child nutrition management network element using a wireless communication link, including: the wireless communication link is a frequency division duplex communication link or a time division duplex communication link.

7. The method for recommending personalized nutritional formulas for children according to claim 6, wherein: The AI formula prediction model is used to intelligently predict the weight difference of the target child before and after taking the set health care formula after the target child takes the set health care formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health care formula taken by the target child, including: synchronously inputting the predicted number of days the target child takes the set health care formula, the target child's item-by-item associated information, and the multiple formula data of the set health care formula taken by the target child into the AI formula prediction model, and running the AI formula prediction model to obtain the weight difference of the target child before and after taking the set health care formula for the predicted number of days, output by the AI formula prediction model; Among them, the predicted number of days for the target child to take the set health care formula, the item-by-item association information of the target child, and the multiple formula data of the set health care formula taken by the target child are synchronously input into the AI formula prediction model, and the AI formula prediction model is run to obtain the weight difference of the target child before and after taking the set health care formula for the predicted number of days output by the AI formula prediction model, which includes: performing binary numerical processing on the predicted number of days for the target child to take the set health care formula, the item-by-item association information of the target child, and the multiple formula data of the set health care formula taken by the target child, and then synchronously inputting them into the AI formula prediction model.

8. The method for recommending personalized nutritional formulas for children according to claim 7, wherein: Synchronously inputting the predicted number of days for the target child to take the set health-care formula, the item-by-item associated information of the target child, and multiple formula data of the set health-care formula taken by the target child into the AI formula prediction model, and running the AI formula prediction model to obtain the weight difference of the target child before and after taking the set health-care formula for the predicted number of days output by the AI formula prediction model, including: the weight difference of the target child before and after taking the set health-care formula for the predicted number of days output by the AI formula prediction model is in a binary numerical representation; The binary numerical representation of the weight difference before and after the target child takes the set health-care formula for the predicted number of days output by the AI formula prediction model includes: the weight difference before and after the target child takes the set health-care formula for the predicted number of days output by the AI formula prediction model is a binary numerical value with a fixed number of digits, and the first digit of the binary numerical value with a fixed number of digits represents the positive or negative value of the weight difference before and after the target child takes the set health-care formula for the predicted number of days; Among them, the binary numerical representation of the weight difference of the target child before and after taking the set health care formula for the predicted number of days output by the AI formula prediction model also includes: the weight difference of the target child before and after taking the set health care formula for the predicted number of days output by the AI formula prediction model is the weight difference obtained by subtracting the weight value of the target child before taking the set health care formula for the predicted number of days from the weight value of the target child after taking the set health care formula for the predicted number of days.

9. A children's nutrition personalized formula recommendation system, the system implementing the method according to claim 1, characterized in that: The system comprises: An information input mechanism is used to obtain the target child's age information, height information, weight information, gender information, and recent disease type coding information to output as item-by-item associated information of the target child; a formula collection mechanism for obtaining the respective proportions of various ingredients of the set health formula for the target child, the respective type coding data of the various ingredients, and the three dosages corresponding to the three meals of the target child, and outputting the data as multiple formula data of the set health formula for the target child; A multiple training mechanism is used to perform multiple training on the radial basis neural network to obtain the radial basis neural network after the multiple trainings and output it as an AI recipe prediction model; an intelligent prediction mechanism, connected to the information input mechanism, the formula collection mechanism, and the multiple training mechanism, respectively, for using an AI formula prediction model to intelligently predict the weight difference of the target child before and after taking the set health formula after the target child has taken the set health formula for the predicted number of days, the target child's item-by-item associated information, and multiple formula data of the set health formula taken by the target child; a recommendation processing mechanism connected to the intelligent prediction mechanism, configured to determine that the set health care formula is a recommended health care formula for the target child when the weight difference value predicted by the intelligent prediction is a positive number and the single-day weight difference obtained by dividing the weight difference value by the predicted number of days exceeds or equals a set weight threshold; otherwise, determine that the set health care formula is a non-recommended health care formula for the target child; Among them, performing multiple training on the radial basis neural network to obtain the radial basis neural network after multiple training and outputting it as the AI formula prediction model includes: the number of training times performed by the radial basis neural network is positively correlated with the value of the predicted days.

Images