CN116682533A - Renal patient nutrition management method and system based on machine learning - Google Patents
Renal patient nutrition management method and system based on machine learning Download PDFInfo
- Publication number
- CN116682533A CN116682533A CN202310968224.0A CN202310968224A CN116682533A CN 116682533 A CN116682533 A CN 116682533A CN 202310968224 A CN202310968224 A CN 202310968224A CN 116682533 A CN116682533 A CN 116682533A
- Authority
- CN
- China
- Prior art keywords
- information
- food
- target object
- nutrition
- intake
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The application provides a kidney disease patient nutrition management method and system based on machine learning, comprising the steps of determining food raw material information and food nutrition ingredient information ingested by a target object based on diet record data and personal information of the target object; according to the food raw material information and the food nutrition ingredient information, combining historical recipe information and pre-acquired standard nutrition intake information of a target object, and determining nutrition balance requirements corresponding to diet record data of the target object through a multi-target optimization algorithm; based on the nutrition balance requirement, setting a personalized recipe and scoring the personalized recipe by combining the dietary taste preference of the target object and the personal information of the target object, and recommending a personalized recipe list with the scoring exceeding a preset recommendation threshold.
Description
Technical Field
The application relates to a nutrition management technology, in particular to a kidney disease patient nutrition management method and system based on machine learning.
Background
The nephrotic syndrome is a group of clinical syndromes caused by various primary and secondary glomerular diseases, and is characterized in that a great amount of proteinuria (24 h urine protein is more than 35g, hypoalbuminemia (plasma albumin is less than 30 g/L), hyperlipidemia and edema are lost from urine due to iron, zinc, copper and key metabolites for regulating calcium metabolism of nephrotic syndrome patients.
Patent publication No. CN113573417B, entitled Intelligent nutrition management System for patients with renal disease, discloses that patient nutrition management information is generated by a renal disease management server and sent to a base station; in response to receiving the patient nutrition management information, transmitting, by the base station, a first downlink control message to the mobile terminal on a first frequency band and a first set of symbols; in response to receiving the first downlink control message sent by the base station, beginning, by the mobile terminal, to receive downlink data sent by the base station on downlink resources indicated by the first downlink control message; allocating, by the base station, a first set of PRACH resources for transmitting a random access preamble to the second mobile terminal in response to starting to transmit downlink data to the mobile terminal; a second downlink control message is transmitted by the base station to the mobile terminal on a second frequency band and prior to the fourth set of time slots in response to allocating the first PRACH resource to the second mobile terminal for transmitting the random access preamble.
The prior art is focused on pushing guidance opinions to patients at regular time by means of wireless communication and a mobile terminal, and realizes whole-course management, detailed management and personalized management, but is not focused on nutrition management related to kidney diseases, and the existing nutrition management is often confusing, and is not focused on the whole nutrition collocation except basic food affecting the kidney diseases.
Disclosure of Invention
The embodiment of the application provides a machine learning-based nutrition management method and system for a kidney disease patient, which can at least solve part of the problems in the prior art, namely that the existing nutrition management is often disordered, and the nutrition management is not concerned with the whole nutrition collocation except basic food which basically influences kidney disease.
In a first aspect of an embodiment of the present application,
provided is a machine learning-based nutrition management method for a renal patient, comprising:
determining food raw material information and food nutrient information ingested by the target object based on the diet recording data and the personal information of the target object;
according to the food raw material information and the food nutrition ingredient information, combining historical recipe information and pre-acquired standard nutrition intake information of a target object, and determining nutrition balance requirements corresponding to diet record data of the target object through a multi-target optimization algorithm;
based on the nutrition balance requirement, setting a personalized recipe and scoring the personalized recipe by combining the dietary taste preference of the target object and the personal information of the target object, and recommending a personalized recipe list with the scoring exceeding a preset recommendation threshold.
In an alternative embodiment of the present application,
the determining, according to the food raw material information and the food nutrition component information, the nutrition balance requirement corresponding to the diet record data of the target object through a multi-target optimization algorithm by combining the history recipe information and the pre-acquired standard nutrition intake information of the target object comprises:
setting a corresponding constraint relation matrix for the food raw material information and the food nutrition ingredient information, taking the history recipe information as a particle group, and initializing the particle group, the speeds and positions of all particles in the particle group and an external file corresponding to the particle group;
constructing an fitness function with the aim of minimizing the nutrition difference between the food nutrition ingredient information contained in the history recipe information and the child standard nutrition intake information;
establishing an adaptive grid space according to the fitness function, equally dividing the adaptive grid space into a plurality of sub-grids, determining target particles through a roulette selection algorithm based on fitness values corresponding to the plurality of sub-grids, and taking the positions and the speeds of the target particles as initial optimal solutions;
and constructing a particle out-of-range set based on the fitness values corresponding to the plurality of sub-grids, merging the particle out-of-range set into the external archive, and iteratively optimizing the speed and the position of particles in the particle swarm until a preset iteration condition is reached, wherein the size of the external archive is limited.
In an alternative embodiment of the present application,
the constructing the fitness function targeting the minimization of the nutritional difference between the food nutritional ingredient information contained in the historical recipe information and the child's standard nutritional intake information comprises:
the fitness function is constructed according to the following formula:
；
wherein f (x) represents a fitness value, U represents the number of non-inferior solutions in a set of non-inferior solutions used to solve the fitness function, A represents a constraint relation matrix, e j Represents the distance between the j-th adjacent non-inferior solutions, E represents the non-inferior solution set, Q represents the number of nutritional ingredients, f i (p) represents a set of food nutrient information contained in the history recipe information, f i (q) represents a standard set of nutritional intake information for children.
In an alternative embodiment of the present application,
the setting a personalized recipe based on the nutritional balancing requirement in combination with the dietary taste preference of the target object and the personal information of the target object comprises:
acquiring the catering information of the target object and the total content of nutrient elements contained in the catering information, and determining recommended intake by combining the personal information of the target object;
determining a thermal energy content in a unit sample of the meal allocation information and a weight of food intake in the meal allocation information according to the meal allocation information of the target object, and determining a thermal energy target value of the target object based on the thermal energy content and the weight of food intake;
a personalized recipe is set based on the recommended intake, the thermal energy target value, and the dietary taste preference of the target subject.
In an alternative embodiment of the present application,
the setting a personalized recipe based on the recommended intake, the thermal energy target value, and the dietary taste preference of the target subject comprises:
the personalized recipe is set according to the following formula:
；
wherein Z represents the personalized recipe, pre represents the dietary taste preference of the target object, m and n represent the number of nutrient elements and the number of meals, e j Representing the heat energy content, X, in a unit sample of the jth meal information j Representing the weight, k, of food intake in the j-th meal information ij Indicating the content of the ith nutrient element in the jth meal information,representing the recommended intake,/->Representing the thermal energy target value.
In a second aspect of an embodiment of the present application,
provided is a machine learning based kidney disease patient nutrition management system comprising:
a first unit for determining food raw material information and food nutrient information ingested by the target object based on diet recording data and personal information of the target object;
a second unit, configured to determine, according to the food raw material information and the food nutrient component information, a nutrition balance requirement corresponding to diet recording data of a target object through a multi-target optimization algorithm in combination with historical recipe information and pre-acquired standard nutrition intake information of the target object;
and the third unit is used for setting personalized recipes and scoring the personalized recipes based on the nutrition balance requirements and combining the dietary taste preference of the target object and the personal information of the target object, and recommending personalized recipe lists with scores exceeding a preset recommendation threshold.
In an alternative embodiment of the present application,
the second unit is further configured to:
setting a corresponding constraint relation matrix for the food raw material information and the food nutrition ingredient information, taking the history recipe information as a particle group, and initializing the particle group, the speeds and positions of all particles in the particle group and an external file corresponding to the particle group;
constructing an fitness function with the aim of minimizing the nutrition difference between the food nutrition ingredient information contained in the history recipe information and the child standard nutrition intake information;
establishing an adaptive grid space according to the fitness function, equally dividing the adaptive grid space into a plurality of sub-grids, determining target particles through a roulette selection algorithm based on fitness values corresponding to the plurality of sub-grids, and taking the positions and the speeds of the target particles as initial optimal solutions;
and constructing a particle out-of-range set based on the fitness values corresponding to the plurality of sub-grids, merging the particle out-of-range set into the external archive, and iteratively optimizing the speed and the position of particles in the particle swarm until a preset iteration condition is reached, wherein the size of the external archive is limited.
In an alternative embodiment of the present application,
the second unit is further configured to:
the fitness function is constructed according to the following formula:
；
wherein f (x) represents a fitness value, U represents the number of non-inferior solutions in a set of non-inferior solutions used to solve the fitness function, A represents a constraint relation matrix, e j Represents the distance between the j-th adjacent non-inferior solutions, E represents the non-inferior solution set, Q represents the number of nutritional ingredients, f i (p) represents a set of food nutrient information contained in the history recipe information, f i (q) represents a standard set of nutritional intake information for children.
In an alternative embodiment of the present application,
the third unit is further configured to:
acquiring the catering information of the target object and the total content of nutrient elements contained in the catering information, and determining recommended intake by combining the personal information of the target object;
determining a thermal energy content in a unit sample of the meal allocation information and a weight of food intake in the meal allocation information according to the meal allocation information of the target object, and determining a thermal energy target value of the target object based on the thermal energy content and the weight of food intake;
a personalized recipe is set based on the recommended intake, the thermal energy target value, and the dietary taste preference of the target subject.
In an alternative embodiment of the present application,
the third unit is further configured to:
the personalized recipe is set according to the following formula:
；
wherein Z represents the personalized recipe, pre represents the dietary taste preference of the target object, and m and n represent the number of nutrient elements and the number of meals respectively，e j Representing the heat energy content, X, in a unit sample of the jth meal information j Representing the weight, k, of food intake in the j-th meal information ij Indicating the content of the ith nutrient element in the jth meal information,representing the recommended intake,/->Representing the thermal energy target value.
The beneficial effects of the scheme of the application can refer to the corresponding parts of technical features in the specific embodiment, and are not repeated here.
Drawings
FIG. 1 is a flow chart of a method for machine learning based nutrition management of renal patients in accordance with an embodiment of the present application;
fig. 2 is a schematic structural diagram of a renal patient nutrition management system based on machine learning according to an embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
The technical scheme of the application is described in detail below by specific examples. The following embodiments may be combined with each other, and some embodiments may not be repeated for the same or similar concepts or processes.
Fig. 1 is a flow chart of a machine learning-based nutrition management method for a renal patient according to an embodiment of the present application, as shown in fig. 1, the method includes:
s101, determining food raw material information and food nutrition information ingested by a target object based on diet record data and personal information of the target object;
for example, dietary record data of the target subject may be collected, including daily ingested foods and beverages; personal information of the target object, such as age, sex, height, weight, activity level, etc., is collected. Obtaining food material information using publicly available food databases or APIs, such as USDA food databases, food nutritional ingredient databases, etc.; according to the names and the quantity of foods recorded by the target object, correspondingly matching the food raw material information, and recording;
based on the obtained food raw material information, the nutritional ingredient content of each food, such as energy, protein, carbohydrate, fat, fiber, etc., is calculated. Considering the influence of the processing and cooking modes of the food on the nutrition components, and if relevant data are available, carrying out corresponding correction; the total intake of each nutrient is obtained by multiplying the nutrient of each food by the amount of food ingested by the subject.
Analyzing the nutritional ingredient data of food intake, comparing the difference with the recommended intake, and evaluating the dietary health condition of the target object; based on the analysis results, targeted suggestions and improvements are given, such as increasing or decreasing the intake of certain food, adjusting the dietary structure, etc.; the analysis results may be presented to the target in the form of reports, charts or recommended menus, helping them to better understand their own diet and make adjustments accordingly.
S102, according to the food raw material information and the food nutrition ingredient information, combining historical recipe information and pre-acquired standard nutrition intake information of a target object, and determining nutrition balance requirements corresponding to diet record data of the target object through a multi-target optimization algorithm;
illustratively, the multi-objective optimization algorithm of the embodiments of the present application is built based on an improved particle algorithm for multi-objective optimization. In order to achieve the purpose of balanced nutrition, the intake of each nutrition element of a user is ensured to be in a reasonable range when the edible amount of the food materials is calculated. Therefore, in the practical application process, the solution of nutrition balance can be regarded as a multi-objective optimization problem.
Illustratively, publicly available food databases or APIs, such as the USDA food database, may be used; knowing the range of intake or recommended intake of each nutritional ingredient required by the target subject with reference to standards of a professional institution or health organization, such as guidelines of the World Health Organization (WHO) or national health department;
for ease of understanding, the relevant information is described below in one specific example:
food raw material information and food nutrient composition information:
food name: an apple;
energy: 52 kcal, protein: 0.3 g, carbohydrate: 14 g, fat: 0.2 g, fiber: 2.4 g;
food name: chicken breast meat;
energy: 165 kcal, protein: 31 g, carbohydrate: 0g, fat: 3.6 g, fiber: 0 g;
standard nutritional intake information for the target subject:
energy: 2000 kcal, protein: 55 g, carbohydrate: 300 g, fat: 70 g, fiber: 25 g.
In an alternative embodiment of the present application,
the determining, according to the food raw material information and the food nutrition component information, the nutrition balance requirement corresponding to the diet record data of the target object through a multi-target optimization algorithm by combining the history recipe information and the pre-acquired standard nutrition intake information of the target object comprises:
setting a corresponding constraint relation matrix for the food raw material information and the food nutrition ingredient information, taking the history recipe information as a particle group, and initializing the particle group, the speeds and positions of all particles in the particle group and an external file corresponding to the particle group;
constructing an fitness function with the aim of minimizing the nutrition difference between the food nutrition ingredient information contained in the history recipe information and the child standard nutrition intake information;
establishing an adaptive grid space according to the fitness function, equally dividing the adaptive grid space into a plurality of sub-grids, determining target particles through a roulette selection algorithm based on fitness values corresponding to the plurality of sub-grids, and taking the positions and the speeds of the target particles as initial optimal solutions;
and constructing a particle out-of-range set based on the fitness values corresponding to the plurality of sub-grids, merging the particle out-of-range set into the external archive, and iteratively optimizing the speed and the position of particles in the particle swarm until a preset iteration condition is reached, wherein the size of the external archive is limited.
Illustratively, the food material information and the food nutritional ingredient information may include nutritional ingredient data of various foods, such as energy, protein, carbohydrate, fat, fiber, etc.; the historical recipe information may include past diet records of the target subject, recording food names, intake amounts, and nutritional ingredient information; the range of intake or recommended intake of each nutritional ingredient is determined based on the personal characteristics and needs of the target subject.
Converting the food raw material information and the food nutrition component information into a constraint relation matrix, wherein the rows of the matrix represent the foods, the columns represent the nutrition components, and the matrix elements represent the corresponding nutrition component content of each food; wherein the constraint relation matrix is shown in the following formula:
；
wherein a is ij The content of the nutrient element j in the food material i of unit weight is shown;
initializing historical diet information as particle groups, wherein each particle represents a historical diet record; initializing the speeds and positions of all particles in the particle swarm, establishing corresponding external files for the particle swarm, and recording the current optimal solution. Constructing an fitness function by taking the nutrition difference between food nutrition ingredient information contained in the history recipe information and standard nutrition intake information of a target object as a target; the fitness function evaluates the fitness of each particle, representing the nutritional balance of the diet record; and establishing an adaptation grid space based on the adaptation function, and uniformly dividing the adaptation value into a plurality of sub-grids. Determining target particles, namely the food records with good fitness according to fitness values corresponding to the subgrid by using a roulette selection algorithm; taking the position and the speed of the target particles as an initial optimal solution, and updating the positions and the speeds of all the particles in the particle swarm; and iteratively optimizing the speed and the position of particles in the particle swarm, and gradually approaching to an optimal solution through cooperation and information communication among the particles.
Taking the historical recipe information as a particle swarm, determining an optimal solution through a particle swarm optimization algorithm, and taking the optimal solution as a nutrition balance requirement corresponding to diet record data;
the historical recipe information is initialized to a group of random particles, namely a group of random solutions is obtained, then an optimal solution is sought through iterative computation of fitness values, in each iterative process, the particles continuously update themselves through tracking individual extremum and population extremum of the history, new positions and speeds of the particles are computed, and the steps are repeated until the maximum iterative times are reached, and the iteration is ended.
On the basis, the application also provides an external file corresponding to the particle swarm, wherein the external file can store the corresponding particle position and speed, and the scale of the external file can be set for limiting the dilemma of local optimization in the algorithm solving process.
To determine the nutritional balance requirement, a fitness function may be constructed that minimizes the nutritional difference between the food nutritional ingredient information contained in the historical recipe information and the child's standard nutritional intake information to a target, wherein the fitness function is represented by the following formula:
；
wherein f (x) represents a fitness value, U represents the number of non-inferior solutions in a set of non-inferior solutions used to solve the fitness function, A represents a constraint relation matrix, e j Represents the distance between the j-th adjacent non-inferior solutions, E represents non-inferior solutionsInferior solution set, Q represents the quantity of nutrient components, f i (p) represents a set of food nutrient information contained in the history recipe information, f i (q) represents a standard set of nutritional intake information for children.
The objective function of the model is divided into two parts, Q is the minimization of energy. The primary goal of a nutritional formula is to achieve an energy intake equal to the recommended amount, but in practice is a goal that is not substantially achieved, so the objective function of energy is defined herein as minimizing the energy intake from the food, while adding to the constraints that the total energy intake is equal to or greater than the recommended amount. Ensuring the supply of the daily diet energy. At the same time, there is a constraint on the proportion of three meals per day. The balance of energy intake of three meals is ensured. The energy intake obtained by the preparation of the model is equal to or greater than the recommended amount, so that the exercise advice instruction is required to be added for the excessive energy part.
The second part is reasonable intake of each nutrient, calculates the intake of 11 nutrients, and allows the intake to be infinitely close to the recommended amount in a mode that the intake and the recommended amount are only different by the minimum absolute value, and simultaneously, the constraint that the total intake of each nutrient is in the range of the reasonable intake is added, so that the intake of each nutrient is infinitely close to the recommended amount while conforming to the reasonable intake interval is ensured.
According to the standard nutrition intake information and the historical recipe data of the target object, the nutrition balance requirement corresponding to the diet record data of the target object is determined through a multi-target optimization algorithm, the food intake can be close to the target nutrition intake information through the optimization result, and the personalized recipe meeting the nutrition balance requirement can be generated according to the individual requirement and the limiting condition, so that the nutrition balance of diet is improved;
through cooperation and information exchange of particle swarms, recipes are optimized and searched rapidly, so that better results are obtained; the method can efficiently generate diet record data suitable for the target object, and reduces the time and labor cost of manual adjustment and trial and error;
the scheme can effectively help a target object to realize the nutrition balance requirement of diet recording data, and provide personalized recipe recommendation according to personal information and taste preference; the method has the technical effects of nutrition balance, individuation, high efficiency, practicability and the like, and can provide better diet guidance and improvement suggestion for a target object.
S103, setting personalized recipes and scoring the personalized recipes based on the nutrition balance requirements and combining the dietary taste preference of the target object and the personal information of the target object, and recommending personalized recipe lists with scores exceeding preset recommendation thresholds.
Illustratively, the intake range or recommended intake of each nutrient component is determined as a requirement for nutrient balance based on the standard nutrient intake information of the target subject; consider the target subject's dietary taste preferences, food contraindications, and personal preferences such as preferred food materials, eating habits, special needs, etc. Based on the nutrition balance requirement and the preference of the diet taste, setting personalized recipes comprising the types, the parts and the proportions of foods of each meal; considering food raw material information and food nutrition ingredient information, food in a recipe can be ensured to meet the nutrition balance requirement.
Scoring the personalized recipes according to the set evaluation indexes and weights, and evaluating factors such as nutrition balance, taste and feasibility of the personalized recipes; scoring may be by quantitative methods such as calculating the degree of deviation of each nutrient component, or qualitative methods using expert assessment. Setting a preset recommendation threshold, and screening out personalized recipes with scores exceeding the threshold; and recommending a personalized recipe list meeting the conditions according to the preference and personal information of the target object.
In an alternative embodiment of the present application,
the setting a personalized recipe based on the nutritional balancing requirement in combination with the dietary taste preference of the target object and the personal information of the target object comprises:
acquiring the catering information of the target object and the total content of nutrient elements contained in the catering information, and determining recommended intake by combining the personal information of the target object;
determining a thermal energy content in a unit sample of the meal allocation information and a weight of food intake in the meal allocation information according to the meal allocation information of the target object, and determining a thermal energy target value of the target object based on the thermal energy content and the weight of food intake;
a personalized recipe is set based on the recommended intake, the thermal energy target value, and the dietary taste preference of the target subject.
Illustratively, the meal allocation information of the target object is acquired, wherein the meal allocation information comprises the type, the quantity and the proportion of food of each meal; the sum of the contents of the various nutritional elements in the meal information is calculated, such as energy, protein, carbohydrate, fat, fiber, etc. Determining a recommended intake of each nutrient element according to the advice or the nutrient guidelines of the professional in combination with the personal information of the target object; and according to personal information such as age, sex, physical condition, activity level and the like of the target object, personalized recommended intake calculation is carried out.
The target thermal energy value of the target object is calculated according to the weight of food intake in the meal information and the thermal energy content in each unit sample, wherein the unit sample refers to the content of nutrient elements in each 100 g or each serving of food. Determining an intake range or recommended intake of each nutrient element in each meal based on the recommended intake; controlling total caloric intake of each meal according to the thermal energy target value to ensure that the total energy meets the target requirement; the proper food type, cooking method and ratio are selected according to taste preference.
Illustratively, assuming that the target object is a 25 year old female, the recommended intake of standard nutritional elements that are required for daily intake are as follows:
energy: 2000 kcal, protein: 60 g, carbohydrate: 250 g, fat: 70 g, fiber: 25 g.
Meal information is as follows:
breakfast: egg (1), whole wheat bread (2 slices), milk (200 ml);
lunch: chicken breast (150 g), rice (1 bowl), vegetable salad (lettuce, tomato, cucumber);
dinner: roasted cod (150 g), brown rice (1 bowl), vegetable salad (green vegetables, carrot, onion).
Calculating a thermal energy target value of the target object according to the weight of food intake in the catering information and the thermal energy content in each unit sample; assuming that the heat energy content of the eggs in the unit sample is 80 kcal, the heat energy content of the chicken breast is 110 kcal.
In an alternative embodiment of the present application,
the setting a personalized recipe based on the recommended intake, the thermal energy target value, and the dietary taste preference of the target subject comprises:
the personalized recipe is set according to the following formula:
；
wherein Z represents the personalized recipe, pre represents the dietary taste preference of the target object, m and n represent the number of nutrient elements and the number of meals, e j Representing the heat energy content, X, in a unit sample of the jth meal information j Representing the weight, k, of food intake in the j-th meal information ij Indicating the content of the ith nutrient element in the jth meal information,representing the recommended intake,/->Representing the thermal energy target value.
Illustratively, said scoring said personalized recipe comprises:
collecting relevant data of personalized recipes, including at least one of food raw material information, food nutrition ingredient data, recipe proportion and user feedback;
processing and integrating the collected data, converting the data into a form which can be used for evaluation, and constructing at least one of a food-nutrition relation matrix and normalized data;
calculating, for each evaluation index, a respective score from food intake and nutritional ingredient data in the recipe;
and carrying out weighted summation on the scores of all the evaluation indexes according to the set weights to obtain the final recipe evaluation score.
The evaluation function can be constructed by combining knowledge and experience of a field expert and data fed back by a user. Based on the set evaluation index and weight, an evaluation function is constructed to quantify the quality degree of the personalized recipe. For each evaluation index, different methods may be used for quantification, such as calculating the degree of deviation of the nutritional ingredients from the target values, user satisfaction surveys.
According to the recommended intake and the nutrient content, ensuring that the intake of each nutrient in the recipe meets the requirements of a target object, thereby promoting the health and the nutrition balance; according to personal preference, eating habit, special requirement and the like of the target object, the variety, the amount and the proportion of the food materials of the recipe are adjusted so as to be more in line with the personal preference and the requirement of the target object; by determining the caloric target value, the total caloric intake in the recipe is controlled. Determining a target thermal energy value of the target object according to the weight of food intake in the meal allocation information and the thermal energy content in each unit sample, thereby controlling the total thermal intake of the recipe, helping to control the weight and maintain the energy balance; providing a practical and feasible personalized recipe based on the catering information, the personal information and the taste preference of the target object; combines the nutrition balance requirement, the personal preference and the nutrition intake target, provides a dietary scheme which meets the actual situation for the target object, and has actual operability and feasibility.
In a second aspect of an embodiment of the present application,
fig. 2 is a schematic structural diagram of a renal patient nutrition management system based on machine learning according to an embodiment of the present application, including:
a first unit for determining food raw material information and food nutrient information ingested by the target object based on diet recording data and personal information of the target object;
a second unit, configured to determine, according to the food raw material information and the food nutrient component information, a nutrition balance requirement corresponding to diet recording data of a target object through a multi-target optimization algorithm in combination with historical recipe information and pre-acquired standard nutrition intake information of the target object;
and the third unit is used for setting personalized recipes and scoring the personalized recipes based on the nutrition balance requirements and combining the dietary taste preference of the target object and the personal information of the target object, and recommending personalized recipe lists with scores exceeding a preset recommendation threshold.
In an alternative embodiment of the present application,
the second unit is further configured to:
setting a corresponding constraint relation matrix for the food raw material information and the food nutrition ingredient information, taking the history recipe information as a particle group, and initializing the particle group, the speeds and positions of all particles in the particle group and an external file corresponding to the particle group;
constructing an fitness function with the aim of minimizing the nutrition difference between the food nutrition ingredient information contained in the history recipe information and the child standard nutrition intake information;
establishing an adaptive grid space according to the fitness function, equally dividing the adaptive grid space into a plurality of sub-grids, determining target particles through a roulette selection algorithm based on fitness values corresponding to the plurality of sub-grids, and taking the positions and the speeds of the target particles as initial optimal solutions;
and constructing a particle out-of-range set based on the fitness values corresponding to the plurality of sub-grids, merging the particle out-of-range set into the external archive, and iteratively optimizing the speed and the position of particles in the particle swarm until a preset iteration condition is reached, wherein the size of the external archive is limited.
In an alternative embodiment of the present application,
the second unit is further configured to:
the fitness function is constructed according to the following formula:
；
wherein f (x) represents a fitness value, U represents the number of non-inferior solutions in a set of non-inferior solutions used to solve the fitness function, A represents a constraint relation matrix, e j Represents the distance between the j-th adjacent non-inferior solutions, E represents the non-inferior solution set, Q represents the number of nutritional ingredients, f i (p) represents a set of food nutrient information contained in the history recipe information, f i (q) represents a standard set of nutritional intake information for children.
In an alternative embodiment of the present application,
the third unit is further configured to:
acquiring the catering information of the target object and the total content of nutrient elements contained in the catering information, and determining recommended intake by combining the personal information of the target object;
determining a thermal energy content in a unit sample of the meal allocation information and a weight of food intake in the meal allocation information according to the meal allocation information of the target object, and determining a thermal energy target value of the target object based on the thermal energy content and the weight of food intake;
a personalized recipe is set based on the recommended intake, the thermal energy target value, and the dietary taste preference of the target subject.
In an alternative embodiment of the present application,
the third unit is further configured to:
the personalized recipe is set according to the following formula:
；
wherein Z represents the personalized recipe, pre represents the dietary taste preference of the target object, m and n represent the number of nutrient elements and the number of meals, e j Representing the heat energy content, X, in a unit sample of the jth meal information j Represents the j-th ligandWeight, k of food intake in meal information ij Indicating the content of the ith nutrient element in the jth meal information,representing the recommended intake,/->Representing the thermal energy target value.
The present application may be a method, apparatus, system, and/or computer program product. The computer program product may include a computer readable storage medium having computer readable program instructions embodied thereon for performing various aspects of the present application.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the application.
Claims (10)
1. A machine learning based nutrition management method for a renal patient, comprising:
determining food raw material information and food nutrient information ingested by the target object based on the diet recording data and the personal information of the target object;
according to the food raw material information and the food nutrition ingredient information, combining historical recipe information and pre-acquired standard nutrition intake information of a target object, and determining nutrition balance requirements corresponding to diet record data of the target object through a multi-target optimization algorithm;
based on the nutrition balance requirement, setting a personalized recipe and scoring the personalized recipe by combining the dietary taste preference of the target object and the personal information of the target object, and recommending a personalized recipe list with the scoring exceeding a preset recommendation threshold.
2. The method according to claim 1, wherein determining, by a multi-objective optimization algorithm, the nutritional balance requirement corresponding to the dietary record data of the target subject based on the food material information and the food nutritional ingredient information in combination with the historical recipe information and the pre-acquired standard nutritional intake information of the target subject comprises:
setting a corresponding constraint relation matrix for the food raw material information and the food nutrition ingredient information, taking the history recipe information as a particle group, and initializing the particle group, the speeds and positions of all particles in the particle group and an external file corresponding to the particle group;
constructing an fitness function with the aim of minimizing the nutrition difference between the food nutrition ingredient information contained in the history recipe information and the child standard nutrition intake information;
establishing an adaptive grid space according to the fitness function, equally dividing the adaptive grid space into a plurality of sub-grids, determining target particles through a roulette selection algorithm based on fitness values corresponding to the plurality of sub-grids, and taking the positions and the speeds of the target particles as initial optimal solutions;
and constructing a particle out-of-range set based on the fitness values corresponding to the plurality of sub-grids, merging the particle out-of-range set into the external archive, and iteratively optimizing the speed and the position of particles in the particle swarm until a preset iteration condition is reached, wherein the size of the external archive is limited.
3. The method of claim 2, wherein said constructing an fitness function targeting a nutrition difference of food nutritional ingredient information contained in said historical recipe information and said child's standard nutrition intake information comprises:
the fitness function is constructed according to the following formula:
；
wherein f (x) represents a fitness value, U represents the number of non-inferior solutions in a set of non-inferior solutions used to solve the fitness function, A represents a constraint relation matrix, e j Represents the distance between the j-th adjacent non-inferior solutions, E represents the non-inferior solution set, Q represents the number of nutritional ingredients, f i (p) represents a set of food nutrient information contained in the history recipe information, f i (q) represents a standard set of nutritional intake information for children.
4. The method of claim 1, wherein setting a personalized recipe based on the nutritional balancing requirements in combination with dietary taste preferences of the target subject and personal information of the target subject comprises:
acquiring the catering information of the target object and the total content of nutrient elements contained in the catering information, and determining recommended intake by combining the personal information of the target object;
determining a thermal energy content in a unit sample of the meal allocation information and a weight of food intake in the meal allocation information according to the meal allocation information of the target object, and determining a thermal energy target value of the target object based on the thermal energy content and the weight of food intake;
a personalized recipe is set based on the recommended intake, the thermal energy target value, and the dietary taste preference of the target subject.
5. The method of claim 4, wherein the setting a personalized recipe based on the recommended intake, the thermal energy target value, and the dietary taste preference of the target subject comprises:
the personalized recipe is set according to the following formula:
；
wherein Z represents the personalized foodSpectrum pre represents the dietary taste preference of the target object, m and n represent the number of nutrient elements and the number of meals, e j Representing the heat energy content, X, in a unit sample of the jth meal information j Representing the weight, k, of food intake in the j-th meal information ij Indicating the content of the ith nutrient element in the jth meal information,representing the recommended intake,/->Representing the thermal energy target value.
6. A machine learning based renal patient nutrition management system, comprising:
a first unit for determining food raw material information and food nutrient information ingested by the target object based on diet recording data and personal information of the target object;
a second unit, configured to determine, according to the food raw material information and the food nutrient component information, a nutrition balance requirement corresponding to diet recording data of a target object through a multi-target optimization algorithm in combination with historical recipe information and pre-acquired standard nutrition intake information of the target object;
and the third unit is used for setting personalized recipes and scoring the personalized recipes based on the nutrition balance requirements and combining the dietary taste preference of the target object and the personal information of the target object, and recommending personalized recipe lists with scores exceeding a preset recommendation threshold.
7. The system of claim 6, wherein the second unit is further configured to:
setting a corresponding constraint relation matrix for the food raw material information and the food nutrition ingredient information, taking the history recipe information as a particle group, and initializing the particle group, the speeds and positions of all particles in the particle group and an external file corresponding to the particle group;
constructing an fitness function with the aim of minimizing the nutrition difference between the food nutrition ingredient information contained in the history recipe information and the child standard nutrition intake information;
establishing an adaptive grid space according to the fitness function, equally dividing the adaptive grid space into a plurality of sub-grids, determining target particles through a roulette selection algorithm based on fitness values corresponding to the plurality of sub-grids, and taking the positions and the speeds of the target particles as initial optimal solutions;
and constructing a particle out-of-range set based on the fitness values corresponding to the plurality of sub-grids, merging the particle out-of-range set into the external archive, and iteratively optimizing the speed and the position of particles in the particle swarm until a preset iteration condition is reached, wherein the size of the external archive is limited.
8. The system of claim 7, wherein the second unit is further configured to:
the fitness function is constructed according to the following formula:
；
wherein f (x) represents a fitness value, U represents the number of non-inferior solutions in a set of non-inferior solutions used to solve the fitness function, A represents a constraint relation matrix, e j Represents the distance between the j-th adjacent non-inferior solutions, E represents the non-inferior solution set, Q represents the number of nutritional ingredients, f i (p) represents a set of food nutrient information contained in the history recipe information, f i (q) represents a standard set of nutritional intake information for children.
9. The system of claim 6, wherein the third unit is further configured to:
acquiring the catering information of the target object and the total content of nutrient elements contained in the catering information, and determining recommended intake by combining the personal information of the target object;
determining a thermal energy content in a unit sample of the meal allocation information and a weight of food intake in the meal allocation information according to the meal allocation information of the target object, and determining a thermal energy target value of the target object based on the thermal energy content and the weight of food intake;
a personalized recipe is set based on the recommended intake, the thermal energy target value, and the dietary taste preference of the target subject.
10. The system of claim 9, wherein the third unit is further configured to:
the personalized recipe is set according to the following formula:
；
wherein Z represents the personalized recipe, pre represents the dietary taste preference of the target object, m and n represent the number of nutrient elements and the number of meals, e j Representing the heat energy content, X, in a unit sample of the jth meal information j Representing the weight, k, of food intake in the j-th meal information ij Indicating the content of the ith nutrient element in the jth meal information,representing the recommended intake,/->Representing the thermal energy target value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310968224.0A CN116682533A (en) | 2023-08-03 | 2023-08-03 | Renal patient nutrition management method and system based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310968224.0A CN116682533A (en) | 2023-08-03 | 2023-08-03 | Renal patient nutrition management method and system based on machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116682533A true CN116682533A (en) | 2023-09-01 |
Family
ID=87781324
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310968224.0A Pending CN116682533A (en) | 2023-08-03 | 2023-08-03 | Renal patient nutrition management method and system based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116682533A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117198466A (en) * | 2023-11-08 | 2023-12-08 | 北京四海汇智科技有限公司 | Diet management method and system for kidney disease patients |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100975960B1 (en) * | 2010-01-29 | 2010-08-18 | 이종홍 | School meal managing system for controlling student's nutritive conditions |
| CN111727429A (en) * | 2018-02-08 | 2020-09-29 | 朱威 | Personalized intelligent catering method and system based on user conditions |
| CN112786154A (en) * | 2021-01-18 | 2021-05-11 | 京东方科技集团股份有限公司 | Recipe recommendation method and device, electronic equipment and storage medium |
| CN116434915A (en) * | 2023-06-07 | 2023-07-14 | 北京四海汇智科技有限公司 | Management method and system for guaranteeing balanced dietary nutrition of children |
-
2023
- 2023-08-03 CN CN202310968224.0A patent/CN116682533A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100975960B1 (en) * | 2010-01-29 | 2010-08-18 | 이종홍 | School meal managing system for controlling student's nutritive conditions |
| CN111727429A (en) * | 2018-02-08 | 2020-09-29 | 朱威 | Personalized intelligent catering method and system based on user conditions |
| CN112786154A (en) * | 2021-01-18 | 2021-05-11 | 京东方科技集团股份有限公司 | Recipe recommendation method and device, electronic equipment and storage medium |
| CN116434915A (en) * | 2023-06-07 | 2023-07-14 | 北京四海汇智科技有限公司 | Management method and system for guaranteeing balanced dietary nutrition of children |
Non-Patent Citations (1)
| Title |
|---|
| 邓红: "营养套餐的设计原则与方法", 西南民族大学学报, vol. 33, no. 04, pages 801 - 806 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117198466A (en) * | 2023-11-08 | 2023-12-08 | 北京四海汇智科技有限公司 | Diet management method and system for kidney disease patients |
| CN117198466B (en) * | 2023-11-08 | 2024-01-26 | 北京四海汇智科技有限公司 | Diet management method and system for kidney disease patients |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6725619B2 (en) | System and method for user-specific adjustment of nutritional intake | |
| US20160217420A1 (en) | Conditioning system for nutritional substances | |
| US9902511B2 (en) | Transformation system for optimization of nutritional substances at consumption | |
| KR102221784B1 (en) | Automation Method for supplying customized menu | |
| Maillot et al. | Dietary changes needed to reach nutritional adequacy without increasing diet cost according to income: an analysis among French adults | |
| Slimani et al. | Perspectives of using the EPIC-SOFT programme in the context of pan-European nutritional monitoring surveys: methodological and practical implications | |
| Seljak | Computer-based dietary menu planning | |
| Wilde et al. | Using the thrifty food plan to assess the cost of a nutritious diet | |
| CN116682533A (en) | Renal patient nutrition management method and system based on machine learning | |
| CN116417114B (en) | Student healthy diet management system based on full life cycle | |
| Babb | America’s “thrifty food plan”: hunger, mathematics, and the valuation of nutrition assistance | |
| KR20180073779A (en) | Automation Method for supplying customized menu | |
| CN116343992A (en) | Multi-factor meal arranging device and method based on nutrition and health concepts | |
| US20210374165A1 (en) | Methods and systems for displaying refreshment outlooks | |
| CN113744840A (en) | System and method for realizing personalized instant nutrition evaluation and guidance based on centralized meal supply environment | |
| Moreira et al. | CardNutri: A software of weekly menus nutritional elaboration for scholar feeding applying evolutionary computation | |
| Sevilla-Villanueva et al. | Generating complete menus from nutritional prescriptions by using advanced cbr and real food databases | |
| Vossenaar et al. | An overview of the main pre-survey tasks required for large-scale quantitative 24-hour recall dietary surveys in low-and middle-income countries | |
| Deitchler et al. | Planning and Design Considerations for Quantitative 24-Hour Recall Dietary Surveys in Low-and Middle-Income Countries | |
| Kulturel-Konak et al. | Menu engineering for continuing care senior living facilities with captive dining patrons | |
| CN115359869A (en) | Multi-objective optimization-based dietary nutrition collocation method | |
| CN109493945A (en) | A kind of the intelligent nutrition catering system and dietary management method of health slimming | |
| López-López et al. | Automatic diet generation by artificial bee colony algorithm | |
| Moreira et al. | A multi-objective approach for the menu planning problem: A brazilian case study | |
| La Cruz et al. | Gradient Descent-Based Optimization for Personalized Nutritional Needs and Health/Activity Goals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |