Method and system for predicting diet weight loss intervention effect by using AI technology

Abstract

The invention provides a method and system for predicting a diet weight loss intervention effect by using an AI technology, and relates to the technical field of AI.The method comprises the steps that a multi-modal data collection device is constructed, biosensing data and diet behavior data are collected, multi-modal data preprocessing is carried out, and a corresponding time sequence database is obtained; based on a KEGG knowledge graph technology, external biological pathway data is obtained, a dynamic metabolism digital twinborn model is constructed according to the time sequence database and the external biological pathway data, and a corresponding metabolism parameter flow is obtained; according to the metabolic parameter flow, constructing a multi-modal deep learning prediction model, and obtaining a multi-dimensional prediction result set; and performing learning optimization analysis on the multi-dimensional prediction result set to obtain an optimization instruction set, performing diet weight loss intervention on the user according to the optimization instruction set, acquiring biosensing data and diet behavior data of the user after diet weight loss intervention in real time, and correcting diet weight loss of the user.

Description

Technical Field

[0001] The present invention relates to the field of AI technology, and specifically to a method and system for predicting the effect of dietary weight loss intervention using AI technology. Background Art

[0002] Traditional dietary intervention plans rely on nutritionists' experience or general models (such as BMI formulas) and lack personalized factors (metabolic rate, genes, intestinal flora, etc.), resulting in large differences in effects.

[0003] Disadvantages: Static models cannot dynamically adapt to individual physiological changes (such as metabolic adaptation during the plateau period) and are prone to rebound.

[0004] At the same time, traditional dietary intervention programs rely on manual recording of diet and exercise (such as paper diaries), with a single data dimension (calories only), a high error rate, and a lack of real-time physiological indicators (blood sugar and hormone levels).

[0005] Disadvantages: Data lags lead to untimely intervention adjustments and inability to capture hidden factors (such as emotional eating).

[0006] Existing prediction tools are mostly based on linear regression or simple decision trees, which cannot handle high-dimensional nonlinear relationships (such as the metabolic synergistic effects of food combinations).

[0007] Disadvantages: The confidence level of the predictions is low, and the long-term risks of different intervention strategies (such as the probability of muscle loss) cannot be quantified.

[0008] Adjustments to traditional dietary intervention programs rely on periodic manual assessments (such as weekly follow-up visits) and lack real-time feedback and automated optimization mechanisms.

[0009] Disadvantages: Slow response speed and inability to correct execution deviations in a timely manner (such as overeating during holidays).

[0010] Therefore, a method and system for predicting the effect of dietary weight loss intervention using AI technology is now provided. Summary of the Invention

[0011] In order to solve the above technical problems, the purpose of the present invention is to provide a method and system for predicting the effect of dietary weight loss intervention using AI technology.

[0012] In order to achieve the above object, the present invention provides the following technical solutions: According to one preferred embodiment of the present invention, a method for predicting the effect of dietary weight loss intervention using AI technology includes: Construct a multimodal data acquisition device to collect biosensor data and dietary behavior data, and perform multimodal data preprocessing to obtain the corresponding time series database; Based on the KEGG knowledge graph technology, external biological pathway data is obtained, and a dynamic metabolic digital twin model is constructed according to the time series database and the external biological pathway data, and the corresponding metabolic parameter flow is obtained; Constructing a multimodal deep learning prediction model based on the metabolic parameter flow and obtaining a multidimensional prediction result set; The multidimensional prediction result set is subjected to learning optimization analysis to obtain an optimized instruction set. According to the optimized instruction set, a dietary weight loss intervention is performed on the user, and the user's biosensor data and dietary behavior data after the dietary weight loss intervention are collected in real time to correct the user's dietary weight loss.

[0013] According to one preferred embodiment of the present invention, a method for predicting the effect of dietary weight loss intervention using AI technology, the process of collecting biosensor data and dietary behavior data includes: Several biosensor units and image acquisition units are provided to collect multimodal data of the person to be evaluated during the acquisition period; the multimodal data includes biosensor data and dietary behavior data; The biosensor data includes basic physiological index data, metabolism-related data and exercise-related data; the dietary behavior data includes dietary time data, food intake data and dietary behavior habit data; The basic physiological indicator data include weight, body fat ratio, sleep quality and stress level; the metabolic-related data include resting metabolic rate, blood sugar level and blood lipid index; the exercise-related data include exercise heart rate, exercise mileage and exercise speed average; the eating time data include meal time and meal frequency; the food intake data include food type, food portion, food calorie value and nutritional content; the eating behavior habit data includes eating rate and emotional eating markers.

[0014] According to one preferred embodiment of the present invention, a method for predicting the effect of dietary weight loss intervention using AI technology includes performing multimodal data preprocessing on biosensor data and dietary behavior data, including: Obtaining biosensor data and dietary behavior data of the person to be evaluated during the collection period; The biosensor data and dietary behavior data at the same time in the corresponding acquisition period are recorded as a set of multimodal data sets. The acquisition period includes several acquisition moments. The multimodal data sets corresponding to each acquisition moment are recorded as t=1, 2, 3, ..., n, where n is a natural number; accordingly, the multimodal data set corresponding to the acquisition moment t is recorded as ; Multimodal datasets Classify the data in the data into structured data and unstructured data; Performing data recognition on the unstructured data based on text feature recognition technology, and encoding the recognized unstructured data based on Unicode encoding technology; Perform outlier detection on the structured data, specifically: Based on the IQR (interquartile range) method, outliers in sleep quality, stress level, resting metabolic rate, blood glucose level, blood lipid index, and exercise heart rate in the multimodal dataset were filtered out; The processed unstructured data and structured data are timestamped, and the frequencies of different data at the same collection time are matched and aligned through a sliding window. The data are then stored in the database and marked as a time series database.

[0015] According to one preferred embodiment of the present invention, a method for predicting the effect of dietary weight loss intervention using AI technology, based on KEGG knowledge graph technology, the process of obtaining external biological pathway data includes: Introducing the KEGG database; Parse the KEGG XML file using the lxml library to extract an undirected weighted graph consisting of L nodes and K edges; the undirected weighted graph consists of corresponding nodes and edges, i.e., several groups of nodes and edges; The nodes include KEGG gene ID, KEGG CID and KEGG enzyme ID; the edges include participating reactions, catalytic reactions and regulatory relationships; According to the undirected weighted graph, the external biological pathway data consisting of L nodes and K edges is obtained, which is recorded as: ; Where d is the damping coefficient; M is the total number of nodes in the undirected weighted graph; Pointing to a node All nodes of For nodes The number of outgoing links.

[0016] According to one preferred embodiment of the present invention, a method for predicting the effect of dietary weight loss intervention using AI technology, the process of constructing a dynamic metabolic digital twin model includes: Obtain data from a time series database to generate a comprehensive biological health vitality index and a comprehensive dietary health adaptation index; The dynamic metabolic digital twin model is generated based on external biological pathway data, biological comprehensive health vitality index, and dietary comprehensive health adaptation index: ;in, is the metabolic parameter flow of the corresponding node, It is the comprehensive biological health vitality index, Comprehensive dietary health adaptation index, For external biological pathway data.

[0017] According to one preferred embodiment of the present invention, a method for predicting the effect of dietary weight loss intervention using AI technology, the process of generating a biological comprehensive health vitality index and a dietary comprehensive health adaptation index includes:

[0018] Generate a comprehensive biological health and vitality index based on the biological sensor data with corresponding serial numbers in the time series database: Where W is weight (kg); F is body fat percentage (%); S is sleep quality (1-10 points); St is stress level (1-10 points); RMR is resting metabolic rate (kcal / day); BG is blood glucose (mmol / L); BL is blood lipid index (mmol / L); HR is heart rate (bpm); D is distance traveled (km); V is average speed (km / h). 、 Measurement error for the biosensing unit; Generate a comprehensive dietary health adaptation index based on the dietary behavior data of the corresponding serial number in the time series database: Where T is the time of eating, unit (hours), Fq is the frequency of eating, unit (times / day), N is the type of food, unit (kind), Q is the portion size, unit (g), C is the caloric value of the food, unit (kcal), Nu is the nutritional content, type (1-10 points), Sp is the eating rate, unit (g / min), and Em is the emotional eating marker, type (0 / 1); 、 Measuring error for the image acquisition unit; According to one of the preferred embodiments of the present invention, a method for predicting the effect of dietary weight loss intervention using AI technology obtains several groups of biosensor data and dietary behavior data from historical collection periods; and generates a biological comprehensive health vitality index and a dietary comprehensive health adaptation index corresponding to the historical collection period based on the biosensor data and dietary behavior data from several groups of historical collection periods, thereby obtaining a metabolic parameter flow corresponding to the historical collection period, constructing a multimodal deep learning prediction model, and outputting a multidimensional prediction result set.

[0019] According to one preferred embodiment of the present invention, a method for predicting the effect of dietary weight loss intervention using AI technology, performing learning optimization analysis on a multi-dimensional prediction result set to obtain an optimized instruction set, and performing dietary weight loss intervention on a user includes: Obtaining a multidimensional prediction result set output by the multimodal deep learning prediction model; Feature extraction is performed on the multidimensional prediction result set. Based on the TF-IDF extraction technology, feature items corresponding to the multidimensional prediction result set are obtained, and a corresponding optimization instruction set is generated based on the feature items. Biosensor data and dietary behavior data of the person to be evaluated after the optimization instruction set usage cycle are collected, and the above process is repeated for updating and iteration.

[0020] In order to achieve at least one of the above-mentioned invention objects, the present invention further provides a system for predicting the effect of dietary weight loss intervention using AI technology, wherein the system executes the above-mentioned method for predicting the effect of dietary weight loss intervention using AI technology, comprising: a multimodal data acquisition module, a multimodal data preprocessing module, a dynamic metabolic digital twin module, a multimodal deep learning prediction module, and an intelligent optimization module; A multimodal data collection module, used to collect multimodal data of the person to be evaluated; The multimodal data preprocessing module preprocesses the multimodal data of the person to be evaluated to obtain the time series database of the person to be evaluated; Dynamic metabolic digital twin module, used to build a dynamic metabolic digital twin model and obtain the corresponding metabolic parameter flow; Multimodal deep learning prediction module, used to build a multimodal deep learning prediction model and obtain a multidimensional prediction result set; The intelligent optimization module performs learning optimization analysis on the multi-dimensional prediction result set to obtain an optimization instruction set, and modifies the user's diet for weight loss according to the optimization instruction set.

[0021] The present invention further provides a computer-readable storage medium, which stores a computer program. The computer program can be executed by a processor to implement the above-mentioned method of predicting the effect of dietary weight loss intervention using AI technology.

[0022] Compared with the prior art, the present invention has the following beneficial effects: 1. Existing technologies rely on single-dimensional data (such as calorie counts) and general models (such as the BMI formula), lacking consideration of individual factors such as metabolic rate, genetics, and gut microbiota. This invention builds a multimodal data acquisition device that integrates biosensor data with dietary behavior data to form a time-series database. Combined with biological pathway data extracted from the KEGG knowledge graph, a dynamic metabolic digital twin model is constructed. This model can quantify individual metabolic characteristics and accurately predict individual genetic, metabolic, and behavioral traits.

[0023] 2. Existing technologies use static models that are unable to capture dynamic changes in physiological states, and intervention adjustments lag behind. The present invention uses a dynamic metabolic digital twin model to map individual metabolic parameter flows in real time, and utilizes a multimodal deep learning model to process high-dimensional nonlinear relationships. Combined with the optimization instruction set for TF-IDF feature extraction, it enables automated iterative optimization of dietary intervention programs. For example, it dynamically adjusts calorie intake and nutrient ratios based on real-time blood sugar and blood lipid data, addressing the shortcomings of traditional programs such as slow response and easy rebound, thereby improving intervention efficiency and long-term effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction to the drawings required for use in the embodiments will be given below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0025] Figure 1 This is a schematic diagram of a method for using AI technology to predict the effects of dietary weight loss intervention.

[0026] Figure 2 This is a system module diagram for predicting the effects of dietary weight loss intervention using AI technology. DETAILED DESCRIPTION

[0027] The following description is intended to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are for illustrative purposes only, and those skilled in the art will readily appreciate other obvious variations. The basic principles of the present invention defined in the following description may be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the present invention.

[0028] It is to be understood that the term "one" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element may be one, while in another embodiment, the number of the elements may be multiple, and the term "one" should not be understood as a limitation on the quantity.

[0029] Please combine Figure 1 The present invention discloses a method for predicting the effect of dietary weight loss intervention using AI technology, the method comprising the following steps: Constructing a multimodal data acquisition device for collecting multimodal data of the person to be evaluated, preprocessing the multimodal data of the person to be evaluated, and obtaining a time series database of the person to be evaluated; Based on the KEGG knowledge graph technology, external biological pathway data is obtained, and a dynamic metabolic digital twin model is constructed according to the time series database and the external biological pathway data, and the corresponding metabolic parameter flow is obtained; Constructing a multimodal deep learning prediction model based on the metabolic parameter flow and obtaining a multidimensional prediction result set; The multidimensional prediction result set is subjected to learning optimization analysis to obtain an optimized instruction set. According to the optimized instruction set, a dietary weight loss intervention is performed on the user, and the user's biosensor data and dietary behavior data after the dietary weight loss intervention are collected in real time to correct the user's dietary weight loss.

[0030] It should be further explained that, in the specific implementation process, the specific process of collecting biosensor data and dietary behavior data includes: Constructing a multimodal data acquisition device, the multimodal data acquisition device comprising a plurality of biosensor units and an image acquisition unit, for acquiring multimodal data of the person to be evaluated during an acquisition period; the multimodal data comprising biosensor data and dietary behavior data; The biosensor units are respectively used to collect biosensor data; the biosensor data includes basic physiological index data, metabolism-related data and exercise-related data; The image acquisition unit is used to collect dietary behavior data; the dietary behavior data includes dietary time data, food intake data and dietary behavior habit data; It should be further explained that the basic physiological indicator data include weight, body fat ratio, sleep quality and stress level; the metabolic-related data include resting metabolic rate, blood sugar level and blood lipid index; the exercise-related data include exercise heart rate, exercise mileage and exercise speed average; the eating time data includes eating time and eating frequency; the food intake data includes food type, food portion, food calorie value and nutritional content; the eating behavior habit data includes eating rate and emotional eating markers.

[0031] It should be further explained that, in the specific implementation process, the multimodal data preprocessing of biosensor data and dietary behavior data and the specific process of obtaining the corresponding time series database include: Obtaining biosensor data and dietary behavior data of the person to be evaluated during the collection period; The biosensor data and dietary behavior data at the same time in the corresponding acquisition period are recorded as a set of multimodal data sets. The acquisition period includes several acquisition moments. The multimodal data sets corresponding to each acquisition moment are recorded as t=1, 2, 3, ..., n, where n is a natural number; accordingly, the multimodal data set corresponding to the acquisition moment t is recorded as ; Multimodal datasets The data in the data set is classified into structured data and unstructured data. It should be further explained that the unstructured data includes food types and emotional eating tags, and the rest are structured data. The unstructured data is identified based on text feature recognition technology, and the identified unstructured data is encoded based on Unicode encoding technology. It should be further explained that encoding the identified unstructured data is conducive to storage in the database, reducing resource waste; Perform outlier detection on the structured data, specifically: Filter out multimodal data sets based on the IQR (interquartile range) method Abnormal values ​​of sleep quality, stress level, resting metabolic rate, blood sugar level, blood lipid index and exercise heart rate; it should be further explained that the standard for judging the presence of abnormalities in the corresponding data is based on existing and public databases; based on linear interpolation, the filtered data is filled in to achieve data integrity; The processed unstructured data and structured data are timestamped, and the frequencies of different data at the same collection time are matched and aligned through a sliding window, and stored in the database, marked as a time series database; it should be further explained that the time series database is horizontally recorded as serial number, collection time t, weight value, body fat ratio, sleep quality, stress level, resting metabolic rate, blood sugar value, blood lipid index, exercise heart rate, exercise mileage, average exercise speed, eating time, eating frequency, food type, food portion, food calorie value, nutritional content, eating rate and emotional eating mark.

[0032] It should be further explained that, in the specific implementation process, based on the KEGG knowledge graph technology, the specific process of obtaining external biological pathway data includes: The KEGG database is introduced. It should be further explained that the KEGG database includes but is not limited to obesity-related data; Parse the KEGG XML file using the lxml library to extract an undirected weighted graph consisting of L nodes and K edges; the undirected weighted graph consists of corresponding nodes and edges, i.e., several groups of nodes and edges; The nodes include KEGG gene IDs, KEGG CIDs, and KEGG enzyme IDs; the edges include participating reactions, catalytic reactions, and regulatory relationships; it should be further explained that the participating reactions include but are not limited to the enzyme encoded by the FTO gene participating in fat metabolism, the catalytic reactions include but are not limited to trypsin catalyzing protein degradation, and the regulatory relationships include but are not limited to insulin activating GLUT4 gene expression; According to the undirected weighted graph, the external biological pathway data consisting of L nodes and K edges is obtained, which is recorded as: ; Where d is the damping coefficient; M is the total number of nodes in the undirected weighted graph; Pointing to a node All nodes of For nodes The number of outgoing links.

[0033] It should be further explained that, in the specific implementation process, the specific process of building a dynamic metabolic digital twin model and obtaining the corresponding metabolic parameter flow includes: Get data from the time series database; Generate a comprehensive biological health and vitality index based on the biosensor data with corresponding serial numbers in the time series database. It should be further explained that the comprehensive biological health and vitality index reflects multiple aspects of a person's physical condition, metabolic capacity, and athletic performance. It combines weight and body fat ratio related to body composition, integrates resting metabolic rate, blood glucose level, and blood lipid indexes related to metabolism, and considers the effects of sleep and stress on physical recovery and stress, as well as heart rate, mileage, and speed averages in terms of exercise to construct a comprehensive index. The biological comprehensive health vitality index is recorded as: Where W is weight (kg); F is body fat percentage (%); S is sleep quality (1-10 points); St is stress level (1-10 points); RMR is resting metabolic rate (kcal / day); BG is blood glucose (mmol / L); BL is blood lipid index (mmol / L); HR is heart rate (bpm); D is distance traveled (km); V is average speed (km / h). 、 Measurement error for the biosensing unit; Generate a comprehensive dietary health adaptation index based on the dietary behavior data of the corresponding serial number in the time series database. It should be further explained that the comprehensive dietary health adaptation index covers aspects such as the time pattern, frequency, food diversity, portion size, calories, and nutrition of eating. It combines the time and frequency of eating to consider the dietary rhythm, the food type and portion size to reflect the richness and intake of the diet, the caloric value and nutrient content of the food to reflect the energy and nutrient balance of the diet, and the impact of emotional eating on dietary health to construct a comprehensive index; The comprehensive dietary health adaptation index is recorded as: Where T is the time of eating, unit (hours), Fq is the frequency of eating, unit (times / day), N is the type of food, unit (kind), Q is the portion size, unit (g), C is the caloric value of the food, unit (kcal), Nu is the nutritional content, type (1-10 points), Sp is the eating rate, unit (g / min), and Em is the emotional eating marker, type (0 / 1); 、 Measuring error for the image acquisition unit; The dynamic metabolic digital twin model is generated based on the external biological pathway data, the biological comprehensive health vitality index, and the dietary comprehensive health adaptation index: ;in, is the metabolic parameter flow of the corresponding node.

[0034] It should be further explained that, in the specific implementation process, the specific process of building a multimodal deep learning prediction model and obtaining a multidimensional prediction result set includes: Obtaining several sets of historical collection cycles of biosensor data and dietary behavior data; And based on several groups of biosensor data and dietary behavior data of historical collection periods, a biological comprehensive health vitality index and a dietary comprehensive health adaptation index corresponding to the historical collection period are generated, thereby obtaining a metabolic parameter flow corresponding to the historical collection period; Based on convolutional neural networks and metabolic parameter streams corresponding to historical acquisition cycles, a multimodal deep learning prediction model is constructed, and a multidimensional prediction result set is output; It should be further explained that the format of the multi-dimensional prediction result set is the predicted physical state and the predicted metabolic state.

[0035] It should be further explained that, in the specific implementation process, the specific process of performing learning optimization analysis on the multi-dimensional prediction result set and obtaining the optimized instruction set includes: Obtaining a multidimensional prediction result set output by the multimodal deep learning prediction model; Performing feature extraction on the multidimensional prediction result set, obtaining feature items corresponding to the multidimensional prediction result set based on TF-IDF extraction technology, and generating a corresponding optimization instruction set according to the feature items; It should be further explained that the biosensor data and dietary behavior data of the person to be evaluated after the optimized instruction set usage cycle are collected and the above process is repeated to achieve the update and iteration of the model.

[0036] like Figure 2 As shown, a system for predicting the effect of dietary weight loss intervention using AI technology includes a multimodal data acquisition module, a multimodal data preprocessing module, a dynamic metabolic digital twin module, a multimodal deep learning prediction module, and an intelligent optimization module; A multimodal data collection module, used to collect multimodal data of the person to be evaluated; The multimodal data preprocessing module preprocesses the multimodal data of the person to be evaluated to obtain the time series database of the person to be evaluated; Dynamic metabolic digital twin module, used to build a dynamic metabolic digital twin model and obtain the corresponding metabolic parameter flow; Multimodal deep learning prediction module, used to build a multimodal deep learning prediction model and obtain a multidimensional prediction result set; The intelligent optimization module performs learning optimization analysis on the multi-dimensional prediction result set to obtain an optimization instruction set, and modifies the user's diet for weight loss according to the optimization instruction set.

[0037] The user data and other data involved in this application are obtained with full consent and authorization, and the collection, use and processing of relevant information comply with the relevant laws, regulations and standards of relevant countries and regions.

[0038] The embodiments disclosed in the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. The embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part, and / or installed from a removable medium. When the computer program is executed by the central processing unit (CPU), the above-mentioned functions defined in the method of the present application are executed. It should be noted that the computer-readable medium mentioned above in the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, a system, device or component of an electrical, magnetic, optical, electromagnetic, infrared segment, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more wire segments, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, electrical wire, optical fiber cable, RF, etc., or any suitable combination thereof.

[0039] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments. Without departing from the principles, the implementation methods of the present invention may be subject to any deformation or modification.

Claims

1. A method for predicting the effect of dietary weight loss intervention using AI technology, characterized in that: include: Construct a multimodal data acquisition device to collect biosensor data and dietary behavior data, and perform multimodal data preprocessing to obtain the corresponding time series database; Based on the KEGG knowledge graph technology, external biological pathway data is obtained, and a dynamic metabolic digital twin model is constructed according to the time series database and the external biological pathway data, and the corresponding metabolic parameter flow is obtained; Constructing a multimodal deep learning prediction model based on the metabolic parameter flow and obtaining a multidimensional prediction result set; The multidimensional prediction result set is subjected to learning optimization analysis to obtain an optimized instruction set. According to the optimized instruction set, a dietary weight loss intervention is performed on the user, and the user's biosensor data and dietary behavior data after the dietary weight loss intervention are collected in real time to correct the user's dietary weight loss.

2. The method for predicting the effect of dietary weight loss intervention using AI technology according to claim 1, characterized in that: The process of collecting biosensor data and dietary behavior data includes: Several biosensor units and image acquisition units are provided to collect multimodal data of the person to be evaluated during the acquisition period; the multimodal data includes biosensor data and dietary behavior data; The biosensor data includes basic physiological index data, metabolism-related data and exercise-related data; the dietary behavior data includes dietary time data, food intake data and dietary behavior habit data; The basic physiological indicator data include weight, body fat ratio, sleep quality and stress level; the metabolic-related data include resting metabolic rate, blood sugar level and blood lipid index; the exercise-related data include exercise heart rate, exercise mileage and exercise speed average; the eating time data include meal time and meal frequency; the food intake data include food type, food portion, food calorie value and nutritional content; the eating behavior habit data includes eating rate and emotional eating markers.

3. The method for predicting the effect of dietary weight loss intervention using AI technology according to claim 2, characterized in that: The process of multimodal data preprocessing of biosensor data and dietary behavior data includes: Obtaining biosensor data and dietary behavior data of the person to be evaluated during the collection period; The biosensor data and dietary behavior data at the same time in the corresponding acquisition period are recorded as a set of multimodal data sets. The acquisition period includes several acquisition moments. The multimodal data sets corresponding to each acquisition moment are recorded as t=1, 2, 3, ..., n, where n is a natural number; accordingly, the multimodal data set corresponding to the acquisition moment t is recorded as ; Multimodal datasets Classify the data in the data into structured data and unstructured data; Performing data recognition on the unstructured data based on text feature recognition technology, and encoding the recognized unstructured data based on Unicode encoding technology; Perform outlier detection on the structured data, specifically: Based on the IQR method, abnormal values ​​of sleep quality, stress level, resting metabolic rate, blood glucose level, blood lipid index and exercise heart rate in the multimodal dataset were filtered out; The processed unstructured data and structured data are timestamped, and the frequencies of different data at the same collection time are matched and aligned through a sliding window. The data are then stored in the database and marked as a time series database.

4. The method for predicting the effect of dietary weight loss intervention using AI technology according to claim 3, characterized in that: Based on KEGG knowledge graph technology, the process of obtaining external biological pathway data includes: Introducing the KEGG database; Parse the KEGG XML file using the lxml library to extract an undirected weighted graph consisting of L nodes and K edges; the undirected weighted graph consists of corresponding nodes and edges, i.e., several groups of nodes and edges; The nodes include KEGG gene ID, KEGG CID and KEGG enzyme ID; the edges include participating reactions, catalytic reactions and regulatory relationships; According to the undirected weighted graph, the external biological pathway data consisting of L nodes and K edges is obtained, which is recorded as: ; Where d is the damping coefficient; M is the total number of nodes in the undirected weighted graph; Pointing to a node All nodes of For nodes The number of outgoing links.

5. The method for predicting the effect of dietary weight loss intervention using AI technology according to claim 4, characterized in that: The process of building a dynamic metabolic digital twin model includes: Obtain data from a time series database to generate a comprehensive biological health vitality index and a comprehensive dietary health adaptation index; The dynamic metabolic digital twin model is generated based on external biological pathway data, biological comprehensive health vitality index, and dietary comprehensive health adaptation index: ;in, is the metabolic parameter flow of the corresponding node, It is the comprehensive biological health vitality index, Comprehensive dietary health adaptation index, For external biological pathway data.

6. The method for predicting the effect of dietary weight loss intervention using AI technology according to claim 5, characterized in that: The process of generating the comprehensive biological health vitality index and the comprehensive dietary health adaptation index includes: Generate a comprehensive biological health and vitality index based on the biological sensor data with corresponding serial numbers in the time series database: Where W is weight; F is body fat percentage; S is sleep quality; St is stress level; RMR is resting metabolic rate; BG is blood glucose level; BL is blood lipid index; HR is exercise heart rate; D is exercise mileage; V is average exercise speed; 、 Measurement error for the biosensing unit; Generate a comprehensive dietary health adaptation index based on the dietary behavior data of the corresponding serial number in the time series database: Where T is the time of eating, Fq is the frequency of eating, N is the type of food, Q is the amount of food, C is the caloric value of food, Nu is the nutritional content, Sp is the eating rate, and Em is the emotional eating marker; 、 Measure the error for the image acquisition unit.

7. The method for predicting the effect of dietary weight loss intervention using AI technology according to claim 5, characterized in that: Acquire several sets of biosensor data and dietary behavior data from historical collection periods; and generate a comprehensive biological health vitality index and a comprehensive dietary health adaptation index for the corresponding historical collection period based on the biosensor data and dietary behavior data from the historical collection periods, thereby obtaining the metabolic parameter flow for the corresponding historical collection period, constructing a multimodal deep learning prediction model, and outputting a multidimensional prediction result set.

8. The method for predicting the effect of dietary weight loss intervention using AI technology according to claim 7, characterized in that: The process of conducting learning optimization analysis on the multi-dimensional prediction result set, obtaining the optimized instruction set, and implementing dietary weight loss intervention for users includes: Obtaining a multidimensional prediction result set output by the multimodal deep learning prediction model; Feature extraction is performed on the multidimensional prediction result set. Based on the TF-IDF extraction technology, feature items corresponding to the multidimensional prediction result set are obtained, and a corresponding optimization instruction set is generated based on the feature items. Biosensor data and dietary behavior data of the person to be evaluated after the optimization instruction set usage cycle are collected, and the above process is repeated for updating and iteration.

9. A system for predicting the effect of dietary weight loss intervention using AI technology, wherein the system executes the method for predicting the effect of dietary weight loss intervention using AI technology according to any one of claims 1 to 8, characterized in that: include: Multimodal data acquisition module, multimodal data preprocessing module, dynamic metabolic digital twin module, multimodal deep learning prediction module and intelligent optimization module; A multimodal data collection module, used to collect multimodal data of the person to be evaluated; The multimodal data preprocessing module preprocesses the multimodal data of the person to be evaluated to obtain the time series database of the person to be evaluated; Dynamic metabolic digital twin module, used to build a dynamic metabolic digital twin model and obtain the corresponding metabolic parameter flow; Multimodal deep learning prediction module, used to build a multimodal deep learning prediction model and obtain a multidimensional prediction result set; The intelligent optimization module performs learning optimization analysis on the multi-dimensional prediction result set to obtain an optimization instruction set, and modifies the user's diet for weight loss according to the optimization instruction set.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and the computer program can be executed by a processor to implement a method for predicting the effect of dietary weight loss intervention using AI technology as described in any one of claims 1 to 8.

Images