Personalized nutritional meal intelligent recommendation method and device, equipment and storage medium
By acquiring users' basic health information and referenced preferred recipes, and combining knowledge graphs for multi-dimensional mining and multi-layer rule verification, the problems of low accuracy and poor scientific adaptability in GDM dietary recommendations have been solved, thus achieving scientific and accurate dietary recommendations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- FUDAN UNIVERSITY
- Filing Date
- 2026-03-17
- Publication Date
- 2026-06-05
Smart Images
Figure CN122157989A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information processing technology, and in particular to a method, apparatus, device, and storage medium for personalized intelligent nutritional diet recommendation. Background Technology
[0002] Gestational diabetes mellitus (GDM) is a disorder of glucose metabolism that occurs during pregnancy. It has a significant impact on maternal and infant health and requires scientific management to reduce risks. Recommending scientific and personalized dietary plans for patients with GDM is crucial, as it is not only the core means of controlling blood sugar but also a key line of defense for protecting maternal and infant health.
[0003] Existing GDM dietary recommendations typically rely solely on simple label matching, making it difficult to accurately infer potential preferences. This can easily lead to an imbalance between blood glucose control needs and user preferences, resulting in low recommendation accuracy and poor scientific adaptability. Summary of the Invention
[0004] This invention provides a personalized nutritional diet intelligent recommendation method, device, equipment, and storage medium. Starting with the user's clearly defined preferred recipes, it combines knowledge graphs to mine related recipes from multiple dimensions and adds multi-layer rule verification steps. This can effectively solve the problems of low recommendation accuracy and poor scientific adaptability, and achieve scientific and accurate recommendation of diets for pregnant women with GDM.
[0005] According to one aspect of the present invention, a personalized intelligent nutritional diet recommendation method is provided, the method comprising:
[0006] Obtain basic health information, preferred recipes, and blood glucose monitoring data of the target individual; wherein, the basic health information includes body mass index, gestational age, and food restrictions;
[0007] The target associated entities of the reference preferred recipes are determined based on a pre-established ingredient database and recipe database; wherein, the associated entities include associated ingredients, associated components, and associated features;
[0008] Based on the pre-established food-recipe-ingredient knowledge graph and the target associated entity, candidate preferred recipes are obtained by multi-dimensional reasoning of the reference preferred recipes;
[0009] Based on the correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the compatibility between the candidate preferred recipes and the food taboos, the candidate preferred recipes are screened to obtain potential preferred recipes.
[0010] Based on the basic health information and the blood glucose monitoring data, a multi-level rule verification is performed on the potential preferred recipes, and a daily dietary plan is recommended for the target subject according to the verification results; wherein, the multi-level rules are determined based on the gestational contraindications, energy matching and blood glucose requirements of gestational diabetes.
[0011] According to another aspect of the present invention, a personalized nutrition and diet intelligent recommendation device is provided, the device comprising:
[0012] The information acquisition module is used to acquire the target object's basic health information, preferred recipes, and blood glucose monitoring data; wherein, the basic health information includes body mass index, gestational age, and food taboos;
[0013] The associated entity determination module is used to determine the target associated entities of the reference preferred recipes based on a pre-established ingredient library and recipe library; wherein, the associated entities include associated ingredients, associated components, and associated features;
[0014] The reasoning module is used to perform multi-dimensional reasoning on the reference preferred recipes based on a pre-established food-recipe-ingredient knowledge graph and the target associated entity to obtain candidate preferred recipes;
[0015] The filtering module is used to filter the candidate preferred recipes to obtain potential preferred recipes based on the correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the adaptability of the candidate preferred recipes to the food taboos.
[0016] The verification module is used to perform multi-level rule verification on the potential preferred recipes based on the basic health information and the blood glucose monitoring data, and recommend a daily dietary plan for the target object according to the verification results; wherein, the multi-level rules are determined based on the gestational contraindications, energy matching and blood glucose requirements of gestational diabetes.
[0017] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:
[0018] At least one processor; and,
[0019] A memory communicatively connected to the at least one processor; wherein,
[0020] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the personalized nutrition and diet intelligent recommendation method according to any embodiment of the present invention.
[0021] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the personalized nutritional diet intelligent recommendation method according to any embodiment of the present invention.
[0022] The technical solution of this invention first obtains the target object's basic health information, reference preferred recipes, and blood glucose monitoring data. The basic health information includes body mass index, gestational age, and food contraindications. Then, based on a pre-established food and recipe database, target associated entities for the reference preferred recipes are determined. These associated entities include associated ingredients, associated components, and associated features. Next, based on a pre-established food-recipe-component knowledge graph and the target associated entities, multi-dimensional reasoning is performed on the reference preferred recipes to obtain candidate preferred recipes. Then, based on the correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the adaptability of the candidate preferred recipes to food contraindications, the candidate preferred recipes are screened to obtain potential preferred recipes. Finally, based on the basic health information and blood glucose monitoring data, multi-level rule verification is performed on the potential preferred recipes, and a daily dietary plan is recommended to the target object based on the verification results. The multi-level rules are determined based on gestational diabetes contraindications, energy matching, and blood glucose requirements. This technical solution starts with the user's clearly defined preferred recipes and combines knowledge graphs to mine related recipes from multiple dimensions. It also adds multiple layers of rule verification, which can effectively solve the problems of low recommendation accuracy and poor scientific adaptability, and achieve scientific and accurate recommendations for GDM pregnant women's diets.
[0023] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description
[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0025] Figure 1 This is a flowchart of a personalized intelligent nutrition and diet recommendation method provided by an embodiment of the present invention;
[0026] Figure 2 This is a flowchart of another personalized nutrition and diet intelligent recommendation method provided by an embodiment of the present invention;
[0027] Figure 3This is a schematic diagram of a personalized intelligent nutrition and diet recommendation method provided according to an embodiment of the present invention;
[0028] Figure 4 This is a schematic diagram of the structure of a personalized nutrition and diet intelligent recommendation device provided in an embodiment of the present invention;
[0029] Figure 5 This is a schematic diagram of the structure of an electronic device that implements a personalized nutrition and diet intelligent recommendation method according to an embodiment of the present invention. Detailed Implementation
[0030] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.
[0031] It should be noted that the terms "first," "second," "target," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. 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 comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0032] Example 1
[0033] Figure 1 This is a flowchart of a personalized intelligent dietary recommendation method provided in Embodiment 1 of the present invention. This embodiment is applicable to the situation of providing scientific and precise dietary recommendations for pregnant women with GDM. This method can be executed by a personalized intelligent dietary recommendation device, which can be implemented in hardware and / or software. This personalized intelligent dietary recommendation device can be configured in an electronic device with data processing capabilities. Figure 1 As shown, the method includes:
[0034] S110: Obtain basic health information, reference recipes, and blood glucose monitoring data of the target subject.
[0035] The target group can be pregnant women with gestational diabetes who require dietary recommendations. The recommended recipes can be recipes explicitly marked as preferred by the target group, and may include one or more recipes, such as steamed sea bass or multigrain rice. Basic health information includes Body Mass Index (BMI), gestational week, and food restrictions. For example, food restrictions may include disliked foods and / or food allergies. Blood glucose monitoring data can be used to reflect blood glucose changes at different times.
[0036] In this embodiment, the target user can actively input their basic health information (including BMI, gestational age, and food restrictions), refer to preferred recipes, and blood glucose monitoring data through a client. For example, the client can be a mobile phone, computer, or other similar device.
[0037] S120, determine the target associated entity for reference to preferred recipes based on the pre-established ingredient library and recipe library.
[0038] The ingredient library stores the GI value, nutritional components, and energy content of ingredients. The recipe library stores the ingredient composition, cooking method, suitable meal size, and associated ingredients / nutrients of recipes. GI (Glycemic Index) measures the degree to which food raises blood sugar levels, reflecting the rate and ability of a food to raise blood sugar compared to glucose. Target associated entities refer to entities corresponding to preferred recipes, specifically including associated ingredients, associated components, and associated features. Associated ingredients can be determined based on the core ingredient composition of the recipe. Associated components can be determined based on the core nutritional components (such as protein) of the core ingredients. Associated features can include cooking characteristics and nutritional characteristics; for example, cooking characteristics can include cooking methods (such as steaming or frying), and nutritional characteristics can include GI characteristics (such as low GI).
[0039] In this embodiment, a general ingredient library and recipe library need to be established in advance and stored in the management terminal for unified management. After obtaining the target object's preferred recipe through the client, the ingredient library and recipe library stored in the management terminal can be called to find the associated entities (including associated ingredients, associated components, and associated features) of the preferred recipe as the target associated entities. For example, assuming the preferred recipe is steamed sea bass, its core ingredient is sea bass, its core nutrient is protein, its cooking method is steaming, and its GI characteristic is low GI, then sea bass can be used as the target associated ingredient, protein as the target associated component, and steaming and low GI as the target associated features.
[0040] S130, based on the pre-established food-recipe-ingredient knowledge graph and target associated entities, perform multi-dimensional reasoning on the reference preferred recipes to obtain candidate preferred recipes.
[0041] The ingredient-recipe-ingredient knowledge graph can be used to describe the relationships between ingredients, recipes, and nutritional components, specifically including entities, entity attributes, and entity relationships. For example, the ingredient-recipe-ingredient knowledge graph includes relationships such as "steamed sea bass - sea bass - protein" and "sea bass - shrimp (seafood)".
[0042] In this embodiment, a general knowledge graph of ingredients, recipes, and components also needs to be pre-established and stored in the management terminal for unified management. After determining the target associated entity of the reference preferred recipe, the knowledge graph of ingredients, recipes, and components stored in the management terminal can be invoked. Starting from the target associated entity, the entity relationships in the knowledge graph are used to perform multi-dimensional reasoning on the reference preferred recipe to obtain candidate preferred recipes. For example, multi-dimensional reasoning on the reference preferred recipe can be implemented based on the RippleNet algorithm to discover recipes that the target object may like. The RippleNet algorithm is a recommendation system model that combines knowledge graphs and automatically discovers the user's hierarchical potential interests through preference propagation technology.
[0043] In this embodiment, optionally, candidate preferred recipes are obtained by performing multi-dimensional reasoning on reference preferred recipes based on a pre-established food-recipe-ingredient knowledge graph and target associated entities, including the following steps A1-A4:
[0044] A1. Based on the ingredient-recipe-ingredient knowledge graph, identify related recipes that are similar to the target ingredient as the first candidate recipes.
[0045] In this embodiment, recipes corresponding to similar ingredients can be inferred from the core ingredients of a reference favorite recipe. For example, assuming the reference favorite recipe is steamed sea bass, its target associated ingredient is sea bass. Based on the association relationship of "sea bass-shrimp-seafood" in the ingredient-recipe-ingredient knowledge graph, the associated recipe is determined to be boiled shrimp. For all reference favorite recipes, inference is performed based on the corresponding target associated ingredients, and all the obtained associated recipes constitute the first candidate recipes.
[0046] A2. Based on the ingredient-recipe-ingredient knowledge graph, identify related recipes containing the target related ingredients as the second candidate recipes.
[0047] In this embodiment, recipes for other ingredients containing the same nutrients can be deduced from the core nutritional components of a preferred recipe. For example, assuming the preferred recipe is steamed sea bass, and its target associated component is protein, the associated recipe is pan-fried chicken breast based on the association between "protein and chicken breast (rich in protein)" in the ingredient-recipe-ingredient knowledge graph. For all preferred recipes, deduction is performed based on the corresponding target associated components, and all resulting associated recipes constitute the second candidate recipes.
[0048] A3. Based on the ingredient-recipe-ingredient knowledge graph, identify the associated recipes containing target-related features as the third candidate recipes.
[0049] In this embodiment, recipes containing the same features can also be inferred from the characteristics of the referenced preferred recipes. For example, assuming the referenced preferred recipe is steamed sea bass, its target association features are steaming and low GI. Based on the association relationship of "steaming-steamed egg custard (same cooking method)" in the ingredient-recipe-ingredient knowledge graph, the associated recipe is determined to be shrimp steamed egg custard. For all referenced preferred recipes, inference is performed based on the corresponding target association features, and all the obtained associated recipes constitute the third candidate recipe.
[0050] A4. Determine the candidate preferred recipes based on the first, second, and third candidate recipes.
[0051] After obtaining the first, second, and third candidate recipes through multi-dimensional reasoning, they can be integrated to obtain the candidate preference recipes. For example, when using the RippleNet algorithm to implement multi-dimensional reasoning from A1 to A3, the number of hops can be set to 3 (to avoid overly shallow reasoning that misses potential preferences or overly deep reasoning that deviates from actual needs, ensuring a balance between reasoning depth and accuracy), and the interest decay factor can be set to 0.8 (the further the reasoning hops, the lower the interest weight, prioritizing the retention of candidate preference recipes closely related to the reference preference recipes).
[0052] S140, based on the correlation and feature matching degree between candidate preferred recipes and reference preferred recipes, as well as the adaptability of candidate preferred recipes to food taboos, the candidate preferred recipes are screened to obtain potential preferred recipes.
[0053] In this embodiment, after determining candidate preferred recipes, the correlation and feature matching degree between the candidate preferred recipes and reference preferred recipes, as well as the adaptability of the candidate preferred recipes to food taboos, can be further determined. Based on this, the candidate preferred recipes are then filtered to obtain potential preferred recipes. Optionally, the process of filtering candidate preferred recipes to obtain potential preferred recipes based on the correlation and feature matching degree between the candidate preferred recipes and reference preferred recipes, as well as the adaptability of the candidate preferred recipes to food taboos, includes the following steps B1-B4:
[0054] B1. Determine the correlation and feature matching degree between candidate favorite recipes and reference favorite recipes based on the entity attribute information in the food-recipe-ingredient knowledge graph.
[0055] Specifically, the similarity between candidate and reference preferred recipes in terms of ingredients, nutritional components, and features can be calculated based on entity attribute information in the ingredient-recipe-ingredient knowledge graph. For example, both boiled shrimp and steamed sea bass contain seafood, thus their correlation is high. Furthermore, the similarity can be determined by the number of identical features between candidate and reference preferred recipes based on entity attribute information in the ingredient-recipe-ingredient knowledge graph. The higher the number of identical features, the higher the feature matching degree. For example, for the associated recipes of steamed sea bass, steamed egg custard and fried eggs, the features of steamed egg custard are steaming and low GI, thus its feature matching degree with steamed sea bass is higher.
[0056] B2. Determine whether the candidate favorite recipes trigger the target object's food taboos, and determine the compatibility between the candidate favorite recipes and food taboos based on the judgment results.
[0057] In this embodiment, it is also necessary to determine whether the candidate preferred recipes trigger the target object's food taboos. If they do, it indicates that the compatibility between the candidate preferred recipes and the food taboos is low; if they do not, it indicates that the compatibility between the candidate preferred recipes and the food taboos is high. For example, the compatibility corresponding to the triggering condition can be set to a negative target value, and the compatibility corresponding to the non-triggered condition can be set to a positive target value.
[0058] B3. The correlation and feature matching between candidate favorite recipes and reference favorite recipes, as well as the fit between candidate favorite recipes and food taboos, are weighted and summed to obtain the target score for each candidate favorite recipe.
[0059] For example, for each candidate favorite recipe, the corresponding target score can be calculated based on the degree of relevance (70%), the degree of compatibility with food contraindications (20%), and the degree of feature matching (10%).
[0060] B4. Sort the target scores in descending order, and determine the top preset number of candidate favorite recipes as potential favorite recipes based on the sorting results.
[0061] All target scores are sorted from highest to lowest, and the top 50 (or 50) candidate favorite recipes in the sorting results are identified as potential favorite recipes.
[0062] S150 performs multi-level rule verification on potential preferred recipes based on basic health information and blood glucose monitoring data, and recommends daily dietary plans for target individuals based on the verification results.
[0063] The multi-level rules are determined based on gestational diabetes mellitus (GDM) contraindications, energy matching, and glycemic requirements. In this embodiment, after determining potential preferred recipes, multi-level rule verification can be performed on the potential preferred recipes based on basic health information and glycemic monitoring data. Daily dietary plans are recommended to the target user based on the verified recipes, ensuring that the recommendations both meet the user's potential preferences and the health requirements of GDM. Pregnancy contraindications include foods prohibited during pregnancy (such as hawthorn and Job's tears) and energy restrictions during pregnancy (such as GI > 70). Energy matching can include daily total energy matching and energy matching per meal. Glycemic requirements can be set based on a preset upper limit for glycemic levels, meaning that the actual monitored blood glucose level does not exceed the preset upper limit. Optionally, multi-level rule verification of potential preferred recipes based on basic health information and glycemic monitoring data includes the following steps C1-C3:
[0064] C1. Eliminate recipes that trigger pregnancy taboos from the potential favorite recipes to obtain the first potential recipe.
[0065] First, determine if there are any potential favorite recipes that trigger pregnancy taboos. If so, remove the recipes that trigger pregnancy taboos from the potential favorite recipes and keep the remaining recipes as the first potential recipes.
[0066] C2. Determine the target subject's total daily energy requirement based on their body mass index and gestational age. Then, select energy-matching recipes from the first potential recipe list as the second potential recipe list based on the target total daily energy requirement and the target number of meals.
[0067] In this embodiment, the target individual's daily total energy requirement can be calculated based on nutritional knowledge, taking into account their BMI and gestational age. This total daily energy requirement is then allocated according to target meal times (e.g., 3 main meals and 2 snacks or 3 main meals and 3 snacks) to obtain the energy requirement for each meal. The target meal times can be input by the target individual via a client (high priority) or a default meal times (low priority). Based on the energy characteristics of the recipes, recipes that simultaneously meet both the daily total energy requirement and the energy requirement per meal are selected as second potential recipes from a first pool of potential recipes. For example, a low-energy shrimp steamed egg custard could be chosen for breakfast, and a medium-energy pan-fried chicken breast could be chosen for lunch.
[0068] C3. Based on the blood glucose monitoring data, determine whether the blood glucose requirements for gestational diabetes are met. Based on the determination results, the second potential recipe is screened to obtain the third potential recipe.
[0069] In this embodiment, blood glucose monitoring data can also be compared with a preset upper limit for blood glucose. If all blood glucose monitoring data do not exceed the preset upper limit for blood glucose, it is determined that the blood glucose requirements for gestational diabetes are met. In this case, all second potential recipes can be retained (i.e., no screening is performed on the second potential recipes), and the second potential recipes can be directly used as the third potential recipes. If any blood glucose monitoring data exceeds the preset upper limit for blood glucose, it is determined that the blood glucose requirements for gestational diabetes are not met. In this case, low-GI, low-carbohydrate recipes (such as retaining mixed grain rice and excluding white rice) need to be screened from the second potential recipes as the third potential recipes. It can be understood that the third potential recipe is the recipe that has passed the multi-layer rule verification.
[0070] In this embodiment, optionally, recommending a daily dietary plan for the target object based on the verification result includes: determining a target recommended recipe from a third potential recipe; combining the target recommended recipe according to the target meal times to obtain a daily dietary plan; and sending the daily dietary plan to the target object.
[0071] Specifically, the process first determines whether the number of potential third-party recipes matches the target number of meals. If the number of potential third-party recipes exceeds the target number of meals, it indicates that multiple recipes correspond to a single meal, and a recipe can be randomly selected as the recommended recipe for that meal. If the number of potential third-party recipes equals the target number of meals, it indicates that only one recipe corresponds to a single meal, and each recipe can be directly used as the recommended recipe for that meal. The set of all recommended recipes constitutes the target recommended recipes. These target recommended recipes are then combined according to the target meals to create personalized daily meal plans, which are then pushed to the target recipients via the client. Furthermore, the daily meal plans for the target recipients can be synchronized to the management system for doctors to view and manually adjust.
[0072] The technical solution of this invention first obtains the target object's basic health information, reference preferred recipes, and blood glucose monitoring data. The basic health information includes body mass index, gestational age, and food contraindications. Then, based on a pre-established food and recipe database, target associated entities for the reference preferred recipes are determined. These associated entities include associated ingredients, associated components, and associated features. Next, based on a pre-established food-recipe-component knowledge graph and the target associated entities, multi-dimensional reasoning is performed on the reference preferred recipes to obtain candidate preferred recipes. Then, based on the correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the adaptability of the candidate preferred recipes to food contraindications, the candidate preferred recipes are screened to obtain potential preferred recipes. Finally, based on the basic health information and blood glucose monitoring data, multi-level rule verification is performed on the potential preferred recipes, and a daily dietary plan is recommended to the target object based on the verification results. The multi-level rules are determined based on gestational diabetes contraindications, energy matching, and blood glucose requirements. This technical solution starts with the user's clearly defined preferred recipes and combines knowledge graphs to mine related recipes from multiple dimensions. It also adds multiple layers of rule verification, which can effectively solve the problems of low recommendation accuracy and poor scientific adaptability, and achieve scientific and accurate recommendations for GDM pregnant women's diets.
[0073] In this embodiment, optionally, after recommending a daily dietary plan for the target based on the verification results, the method further includes: if the blood glucose monitoring data does not meet the blood glucose requirements for gestational diabetes, then sending a blood glucose warning message to the target.
[0074] In this embodiment, when the blood glucose monitoring data of the target individual is detected to be inconsistent with the blood glucose requirements for gestational diabetes, a blood glucose warning message can be sent to the target individual via the client to remind them to pay attention to their blood glucose status. The blood glucose warning message may include the abnormal value, its corresponding measurement time, and the number of abnormal values. For example, the blood glucose warning message can be displayed via pop-up window, SMS, and / or voice playback.
[0075] Example 2
[0076] Figure 2 This is a flowchart of a personalized nutritional diet intelligent recommendation method provided in Embodiment 2 of the present invention. This embodiment is based on the above embodiment and optimized. Specifically, the optimization includes: after recommending a daily diet plan to the target object based on the verification result, it further includes: receiving feedback information from the target object on the daily diet plan; wherein, the feedback information includes response behavior and browsing time, and the response behavior includes saving and rejecting; updating the association weight of multi-dimensional reasoning based on the feedback information, and optimizing the multi-dimensional reasoning process based on the updated association weight.
[0077] like Figure 2 As shown, the method in this embodiment specifically includes the following steps:
[0078] S210: Obtain basic health information, reference recipes, and blood glucose monitoring data of the target subject.
[0079] The basic health information includes body mass index, gestational age, and food restrictions.
[0080] S220, determine the target associated entity for reference to preferred recipes based on the pre-established ingredient library and recipe library.
[0081] Among them, associated entities include associated ingredients, associated components, and associated features.
[0082] S230, based on the pre-established food-recipe-ingredient knowledge graph and target associated entities, perform multi-dimensional reasoning on the reference preferred recipes to obtain candidate preferred recipes.
[0083] S240, based on the correlation and feature matching degree between candidate preferred recipes and reference preferred recipes, as well as the adaptability of candidate preferred recipes to food taboos, the candidate preferred recipes are screened to obtain potential preferred recipes.
[0084] S250 performs multi-level rule verification on potential preferred recipes based on basic health information and blood glucose monitoring data, and recommends daily dietary plans for target individuals based on the verification results.
[0085] The multi-layered rules are determined based on gestational contraindications, energy matching, and blood glucose requirements for gestational diabetes. Furthermore, the specific implementation methods of S210-S250 can be found in the detailed description of Embodiment 1 above, and will not be repeated here.
[0086] S260 receives feedback from the target subject regarding the daily dietary plan.
[0087] In this embodiment, after recommending a daily meal plan to the target object based on the verification results, it also supports receiving feedback information from the target object regarding the daily meal plan, in order to optimize the multi-dimensional reasoning process based on the feedback information. The feedback information includes response behavior (specifically, adding to favorites and rejecting) and browsing time. It can be understood that adding to favorites indicates that the target object likes the recommended daily meal plan, and rejecting indicates that the target object dislikes the recommended daily meal plan. Furthermore, if the target object does not respond, the browsing time of each recipe in the daily meal plan can be used as an auxiliary criterion. Specifically, if the browsing time is greater than or equal to a preset time, it can be assumed that the target object likes the recommended recipe; if the browsing time is less than the preset time, it can be assumed that the target object dislikes the recommended recipe.
[0088] S270, update the association weights of multi-dimensional reasoning based on the feedback information, and optimize the multi-dimensional reasoning process based on the updated association weights.
[0089] For example, if the target user marks the recommended boiled shrimp as a favorite, the association weight of "sea bass-shrimp" is increased, and similar related recipes can be inferred in the future. If the target user rejects the recommended pan-fried chicken breast, the association weight of "protein-chicken breast" is decreased, and similar inferences are reduced in the future.
[0090] Figure 3 This is a schematic diagram of a personalized intelligent dietary recommendation method provided in Embodiment 2 of the present invention. Figure 3 As shown, the user first inputs basic health information (such as BMI, gestational age, and food restrictions) on a mobile device, along with blood glucose data and a preferred recipe (such as steamed sea bass). Then, using the RippleNet algorithm, the target associated entities (including associated ingredients, components, and features) of the preferred recipe are determined based on the ingredient and recipe databases in the basic data. A three-hop reasoning process is then performed based on the ingredient-recipe-component knowledge graph in the basic data to generate potential preferred recipes. These potential preferred recipes undergo multi-layered rule validation, including excluding pregnancy-related restrictions, energy compatibility checks, and blood glucose-linked adjustments, to generate a final dietary plan. This plan is then displayed on the mobile device and simultaneously synchronized with the management system. Furthermore, user feedback on the recommended dietary plan is received, and the RippleNet association weights are updated based on this feedback. Subsequent reasoning is then optimized based on the updated association weights.
[0091] The technical solution of this invention, after recommending a daily dietary plan for the target object based on the verification results, also receives feedback information from the target object regarding the daily dietary plan; wherein, the feedback information includes response behavior and browsing time, and the response behavior includes saving and rejecting; then, the association weights of multi-dimensional reasoning are updated based on the feedback information, and the multi-dimensional reasoning process is optimized based on the updated association weights. This technical solution, starting with the user's clearly defined preferred recipes and combining knowledge graphs for multi-dimensional mining of associated recipes, and adding multi-layer rule verification steps, can effectively solve the problems of low recommendation accuracy and poor scientific adaptability, and achieve scientific and accurate recommendations for GDM pregnant women's diets; at the same time, it can dynamically adjust the multi-dimensional reasoning process based on the user's feedback information on the recommended diets, improving the adaptability of personalized recommendations.
[0092] Example 3
[0093] Figure 4 This is a schematic diagram of a personalized nutrition and diet intelligent recommendation device provided in Embodiment 3 of the present invention. This device can execute the personalized nutrition and diet intelligent recommendation method provided in any embodiment of the present invention, and possesses the corresponding functional modules and beneficial effects of the method. Figure 4 As shown, the device includes:
[0094] The information acquisition module 310 is used to acquire the target object's basic health information, preferred recipes, and blood glucose monitoring data; wherein, the basic health information includes body mass index, gestational age, and food taboos;
[0095] The associated entity determination module 320 is used to determine the target associated entities of the reference preferred recipes based on a pre-established ingredient library and recipe library; wherein, the associated entities include associated ingredients, associated components, and associated features;
[0096] The reasoning module 330 is used to perform multi-dimensional reasoning on the reference preferred recipes based on the pre-established food-recipe-ingredient knowledge graph and the target associated entity to obtain candidate preferred recipes;
[0097] The filtering module 340 is used to filter the candidate preferred recipes to obtain potential preferred recipes based on the correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the adaptability of the candidate preferred recipes to the food taboos.
[0098] The verification module 350 is used to perform multi-level rule verification on the potential preferred recipes based on the basic health information and the blood glucose monitoring data, and recommend a daily dietary plan for the target object according to the verification results; wherein, the multi-level rules are determined based on the gestational contraindications, energy matching and blood glucose requirements of gestational diabetes.
[0099] Optionally, the inference module 330 is specifically used for:
[0100] Based on the food-recipe-ingredient knowledge graph, related recipes of the same type as the target food ingredient are identified as the first candidate recipes;
[0101] Based on the food-recipe-ingredient knowledge graph, related recipes containing the target related ingredients are identified as second candidate recipes;
[0102] Based on the food-recipe-ingredient knowledge graph, the associated recipes containing the target association features are identified as the third candidate recipes;
[0103] Candidate preferred recipes are determined based on the first candidate recipe, the second candidate recipe, and the third candidate recipe.
[0104] Optionally, the filtering module 340 is specifically used for:
[0105] The correlation and feature matching degree between the candidate preferred recipe and the reference preferred recipe are determined based on the entity attribute information in the food-recipe-ingredient knowledge graph.
[0106] Determine whether the candidate preferred recipes trigger the target object's food taboos, and determine the compatibility between the candidate preferred recipes and the food taboos based on the determination result;
[0107] The correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the adaptability of the candidate preferred recipes to the food taboos, are weighted and summed to obtain the target score corresponding to each candidate preferred recipe.
[0108] The target scores are sorted in descending order, and the top preset number of candidate favorite recipes are determined as potential favorite recipes based on the sorting results.
[0109] Optionally, the verification module 350 is used for:
[0110] The first potential recipe is obtained by removing recipes that trigger the pregnancy taboos from the potential preferred recipes;
[0111] The target object's total daily energy requirement is determined based on the target object's body mass index and gestational age. Based on the total daily energy requirement and target meal number, a recipe with matching energy is selected from the first potential recipe as a second potential recipe.
[0112] Based on the blood glucose monitoring data, it is determined whether the blood glucose requirements for gestational diabetes are met. Based on the determination result, the second potential recipe is screened to obtain the third potential recipe.
[0113] Optionally, the verification module 350 is further configured to:
[0114] Determine the target recommended recipe from the third potential recipe;
[0115] The target recommended recipes are combined according to the target meal times to obtain a daily dietary plan;
[0116] The daily meal plan is sent to the target object.
[0117] Optionally, the device further includes: an inference optimization module, used for:
[0118] After recommending a daily diet plan to the target object based on the verification results, the system receives feedback information from the target object regarding the daily diet plan; wherein, the feedback information includes response behavior and browsing time, and the response behavior includes adding to favorites and rejecting;
[0119] The association weights of the multi-dimensional reasoning are updated based on the feedback information, and the multi-dimensional reasoning process is optimized based on the updated association weights.
[0120] Optionally, the device further includes: a blood glucose warning module, used for:
[0121] After recommending a daily dietary plan for the target subject based on the verification results, if the blood glucose monitoring data does not meet the blood glucose requirements for gestational diabetes, a blood glucose warning message is sent to the target subject.
[0122] The personalized nutrition and diet intelligent recommendation device provided in this embodiment of the invention can execute the personalized nutrition and diet intelligent recommendation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method execution.
[0123] Example 4
[0124] Figure 5 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.
[0125] like Figure 5 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.
[0126] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0127] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as personalized nutritional diet intelligent recommendation methods.
[0128] In some embodiments, the personalized nutrition and dietary intelligent recommendation method can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the personalized nutrition and dietary intelligent recommendation method described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the personalized nutrition and dietary intelligent recommendation method by any other suitable means (e.g., by means of firmware).
[0129] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0130] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0131] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.
[0132] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0133] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.
[0134] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.
[0135] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.
[0136] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.
Claims
1. A personalized intelligent dietary recommendation method, characterized in that, The method includes: Obtain basic health information, preferred recipes, and blood glucose monitoring data of the target individual; wherein, the basic health information includes body mass index, gestational age, and food restrictions; The target associated entities of the reference preferred recipes are determined based on a pre-established ingredient database and recipe database; wherein, the associated entities include associated ingredients, associated components, and associated features; Based on the pre-established food-recipe-ingredient knowledge graph and the target associated entity, candidate preferred recipes are obtained by multi-dimensional reasoning of the reference preferred recipes; Based on the correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the compatibility between the candidate preferred recipes and the food taboos, the candidate preferred recipes are screened to obtain potential preferred recipes. Based on the basic health information and the blood glucose monitoring data, a multi-level rule verification is performed on the potential preferred recipes, and a daily dietary plan is recommended for the target subject according to the verification results; wherein, the multi-level rules are determined based on the gestational contraindications, energy matching and blood glucose requirements of gestational diabetes.
2. The method according to claim 1, characterized in that, Based on a pre-established food-recipe-ingredient knowledge graph and the target associated entities, candidate preferred recipes are obtained through multi-dimensional reasoning of the reference preferred recipes, including: Based on the food-recipe-ingredient knowledge graph, related recipes of the same type as the target food ingredient are identified as the first candidate recipes; Based on the food-recipe-ingredient knowledge graph, related recipes containing the target related ingredients are identified as second candidate recipes; Based on the food-recipe-ingredient knowledge graph, the associated recipes containing the target association features are identified as the third candidate recipes; Candidate preferred recipes are determined based on the first candidate recipe, the second candidate recipe, and the third candidate recipe.
3. The method according to claim 1 or 2, characterized in that, Based on the correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the compatibility between the candidate preferred recipes and the food taboos, the candidate preferred recipes are screened to obtain potential preferred recipes, including: The correlation and feature matching degree between the candidate preferred recipe and the reference preferred recipe are determined based on the entity attribute information in the food-recipe-ingredient knowledge graph. Determine whether the candidate preferred recipes trigger the target object's food taboos, and determine the compatibility between the candidate preferred recipes and the food taboos based on the determination result; The correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the adaptability of the candidate preferred recipes to the food taboos, are weighted and summed to obtain the target score corresponding to each candidate preferred recipe. The target scores are sorted in descending order, and the top preset number of candidate favorite recipes are determined as potential favorite recipes based on the sorting results.
4. The method according to claim 3, characterized in that, Based on the basic health information and the blood glucose monitoring data, multi-level rule validation is performed on the potential preferred recipes, including: The first potential recipe is obtained by removing recipes that trigger the pregnancy taboos from the potential preferred recipes; The target object's total daily energy requirement is determined based on the target object's body mass index and gestational age. Based on the total daily energy requirement and target meal number, a recipe with matching energy is selected from the first potential recipe as a second potential recipe. Based on the blood glucose monitoring data, it is determined whether the blood glucose requirements for gestational diabetes are met. Based on the determination result, the second potential recipe is screened to obtain the third potential recipe.
5. The method according to claim 4, characterized in that, Based on the verification results, a daily dietary plan is recommended for the target object, including: Determine the target recommended recipe from the third potential recipe; The target recommended recipes are combined according to the target meal times to obtain a daily dietary plan; The daily meal plan is sent to the target object.
6. The method according to claim 1, characterized in that, After recommending a daily dietary plan for the target object based on the verification results, the method further includes: Receive feedback information from the target object regarding the daily dietary plan; wherein, the feedback information includes response behavior and browsing time, and the response behavior includes saving and rejecting; The association weights of the multi-dimensional reasoning are updated based on the feedback information, and the multi-dimensional reasoning process is optimized based on the updated association weights.
7. The method according to claim 1, characterized in that, After recommending a daily dietary plan for the target object based on the verification results, the method further includes: If the blood glucose monitoring data does not meet the blood glucose requirements for gestational diabetes, a blood glucose warning message is sent to the target individual.
8. A personalized intelligent nutrition and diet recommendation device, characterized in that, The device includes: The information acquisition module is used to acquire the target object's basic health information, preferred recipes, and blood glucose monitoring data; wherein, the basic health information includes body mass index, gestational age, and food taboos; The associated entity determination module is used to determine the target associated entities of the reference preferred recipes based on a pre-established ingredient library and recipe library; wherein, the associated entities include associated ingredients, associated components, and associated features; The reasoning module is used to perform multi-dimensional reasoning on the reference preferred recipes based on a pre-established food-recipe-ingredient knowledge graph and the target associated entity to obtain candidate preferred recipes; The filtering module is used to filter the candidate preferred recipes to obtain potential preferred recipes based on the correlation and feature matching degree between the candidate preferred recipes and the reference preferred recipes, as well as the adaptability of the candidate preferred recipes to the food taboos. The verification module is used to perform multi-level rule verification on the potential preferred recipes based on the basic health information and the blood glucose monitoring data, and recommend a daily dietary plan for the target object according to the verification results; wherein, the multi-level rules are determined based on the gestational contraindications, energy matching and blood glucose requirements of gestational diabetes.
9. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the personalized nutritional diet intelligent recommendation method according to any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the personalized nutritional diet intelligent recommendation method according to any one of claims 1-7.