Weight loss method and system for obese patient

Abstract

The invention discloses a weight loss method and system for an obese patient. The method comprises the steps that body information and physical sign monitoring information of the patient are acquired; inputting the body information and the physical sign monitoring information into a pre-trained obesity evaluation model to obtain an evaluation report of the patient; according to the evaluation report and a pre-constructed knowledge graph, a personalized weight reduction scheme is determined, and the personalized weight reduction scheme is fed back to the patient; collecting weight loss data executed by the patient based on the personalized weight loss scheme; according to the weight reduction data, whether the personalized weight reduction scheme is valid or not is judged, and / or whether the patient executes the personalized weight reduction scheme or not is judged; and if not, determining that the personalized weight reduction scheme is invalid, and returning to execute the personalized weight reduction scheme determined according to the evaluation report until the personalized weight reduction scheme is valid, and / or the patient executes the personalized weight reduction scheme. Therefore, by means of the deep analysis capability of the DeepSeek large model on the data and in combination with the knowledge graph, the problem of insufficient weight reduction accuracy can be solved.

Description

Technical Field

[0001] The present application relates to the interdisciplinary field of artificial intelligence and medical health, and in particular to a weight loss method and system for obese patients. Background Art

[0002] With the growing global obesity problem, obesity and its related diseases (such as diabetes, hypertension, and cardiovascular disease) have become a major public health challenge. Obesity not only affects people's quality of life but also significantly increases the medical burden. Therefore, scientific and effective weight loss management has become an urgent issue that needs to be addressed.

[0003] Existing weight loss methods mainly rely on the doctor's experience and the development of individualized plans. That is, the doctor designs a weight loss plan based on the patient's specific situation (such as height, weight, basal metabolic rate, etc.) and gives diet and exercise advice.

[0004] However, during the weight loss process, patients may not achieve good results and lack accuracy due to untimely behavioral monitoring and program adjustments. Summary of the Invention

[0005] Based on the above-mentioned deficiencies of the prior art, the present application provides a weight loss method and system for obese patients to solve the problem of insufficient weight loss accuracy brought about by the prior art.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] The first aspect of the present application provides a method for weight loss in obese patients, comprising:

[0008] Obtain patient's physical information and vital sign monitoring information;

[0009] Inputting the physical information and the vital sign monitoring information into a pre-trained obesity assessment model to obtain an assessment report for the patient; wherein the obesity assessment model is pre-trained based on the DeepSeek large model; and the DeepSeek large model is pre-trained based on sample physical information and sample vital sign monitoring information of sample patients;

[0010] Determining a personalized weight loss plan based on the assessment report and the pre-built knowledge graph, and feeding the personalized weight loss plan back to the patient;

[0011] collecting weight loss data of the patient based on the personalized weight loss plan;

[0012] Determining, based on the weight loss data, whether the personalized weight loss program is effective, and / or whether the patient is implementing the personalized weight loss program;

[0013] If the personalized weight loss plan is not effective, and\or the patient does not implement the personalized weight loss plan, the personalized weight loss plan is determined to be invalid, and the process of determining the personalized weight loss plan based on the evaluation report is returned to execution until the personalized weight loss plan is effective, and\or the patient implements the personalized weight loss plan.

[0014] Optionally, in the above-mentioned weight loss method for obese patients, inputting the body information and the physical sign monitoring information into a pre-trained obesity assessment model to obtain an assessment report of the patient includes:

[0015] Performing deep feature extraction on the body information and the vital sign monitoring information using the obesity assessment model to obtain patient data features;

[0016] Matching the patient data features with the obesity types in the pre-built knowledge graph using the obesity assessment model to obtain the patient's obesity level;

[0017] Retrieving risk knowledge corresponding to the patient data features from the knowledge graph through the obesity assessment model, and performing a fusion analysis on the patient data features and the risk knowledge to obtain the patient's weight loss risk;

[0018] An assessment report for the patient is generated by the obesity assessment model according to a preset report template, based on the obesity level and the weight loss risk.

[0019] Optionally, in the above-mentioned weight loss method for obese patients, the method for constructing the knowledge graph includes:

[0020] Acquiring multi-source data on obesity and preprocessing the multi-source data to obtain target multi-source data;

[0021] Using the large model to perform entity recognition on the target multi-source data, and obtain multiple entities corresponding to the target multi-source data;

[0022] For each entity, using the large model to analyze the semantic relationship between the entity and the remaining entities other than the entity, to obtain multiple entity relationships corresponding to the entity;

[0023] A knowledge graph is constructed based on all the entities and their corresponding multiple entities; wherein each node in the knowledge graph represents one of the entities, and an edge between any two connected nodes represents one of the entity relationships.

[0024] Optionally, in the above-mentioned weight loss method for obese patients, determining a personalized weight loss plan based on the assessment report and the pre-constructed knowledge graph includes:

[0025] Extract key signs from assessment reports using large models;

[0026] Retrieving dietary knowledge corresponding to the key features and the physical information from a pre-built knowledge graph using the large model, and formulating a diet plan for the patient based on the dietary knowledge, the key features, and the physical information;

[0027] Retrieving exercise knowledge corresponding to the key features and the body information from a pre-built knowledge graph using the large model, and formulating an exercise plan based on the exercise knowledge, the key features, and the body information;

[0028] Retrieving life knowledge corresponding to the vital sign monitoring information from a pre-built knowledge graph using the large model, and formulating a life plan based on the life knowledge and the vital sign monitoring information;

[0029] Utilizing the large model to formulate weight loss goals and expected results based on the physical information and the knowledge of weight loss patterns and weight loss effects in the knowledge graph;

[0030] The diet plan, the exercise plan, the exercise plan, and the weight loss goal are integrated with the expected effect to obtain a weight loss plan, and the weight loss plan is determined as a personalized weight loss plan.

[0031] Optionally, the above-mentioned weight loss method for obese patients further comprises:

[0032] In response to a query operation in a menu bar of the triggering interactive platform, displaying an interactive interface;

[0033] According to the query instruction typed in the interactive interface, query information corresponding to the query instruction is displayed to the patient.

[0034] Optionally, the above-mentioned weight loss method for obese patients further comprises:

[0035] When receiving the question information typed by the patient, pre-processing the question information to obtain target question information;

[0036] Use the big model to conduct in-depth analysis on the target question information to obtain the core intent corresponding to the target question information;

[0037] Retrieving target knowledge corresponding to the core intent from the knowledge graph;

[0038] The large model is used to generate a simple answer based on the target knowledge, and the simple answer is fed back to the patient.

[0039] A second aspect of the present application provides a weight loss system for obese patients, comprising:

[0040] An information acquisition unit, used to obtain the patient's physical information and vital sign monitoring information;

[0041] An input unit, configured to input the physical information and the vital sign monitoring information into a pre-trained obesity assessment model to obtain an assessment report for the patient; wherein the obesity assessment model is pre-trained based on a DeepSeek large model; and the DeepSeek large model is pre-trained based on sample physical information and sample vital sign monitoring information of sample patients;

[0042] a plan determination unit, configured to determine a personalized weight loss plan based on the assessment report and a pre-built knowledge graph, and feed the personalized weight loss plan back to the patient;

[0043] a data collection unit, configured to collect weight loss data of the patient based on the personalized weight loss plan;

[0044] a judgment unit, configured to judge whether the personalized weight loss plan is effective, and / or whether the patient is implementing the personalized weight loss plan, based on the weight loss data;

[0045] The return execution unit is used to determine that the personalized weight loss plan is invalid if the personalized weight loss plan is not effective and\or the patient does not implement the personalized weight loss plan, and return to the execution of determining the personalized weight loss plan based on the evaluation report until the personalized weight loss plan is effective and\or the patient implements the personalized weight loss plan.

[0046] Optionally, in the above-mentioned weight loss system for obese patients, the input unit includes:

[0047] a feature extraction unit, configured to perform deep feature extraction on the body information and the vital sign monitoring information using the obesity assessment model to obtain patient data features;

[0048] a matching unit, configured to match the patient data features with the obesity types in the pre-built knowledge graph using the obesity assessment model to obtain the patient's obesity level;

[0049] a fusion analysis unit, configured to retrieve risk knowledge corresponding to the patient data features from the knowledge graph through the obesity assessment model, and perform a fusion analysis on the patient data features and the risk knowledge to obtain the patient's weight loss risk;

[0050] A report generating unit is used to generate an assessment report for the patient according to the obesity assessment model and a preset report template, based on the obesity level and the weight loss risk.

[0051] Optionally, the above-mentioned weight loss system for obese patients further includes:

[0052] a preprocessing unit, configured to obtain multi-source data on obesity and preprocess the multi-source data to obtain target multi-source data;

[0053] An entity recognition unit, configured to perform entity recognition on the target multi-source data using a large model to obtain multiple entities corresponding to the target multi-source data;

[0054] an analyzing unit, configured to analyze, for each entity, the semantic relationship between the entity and the remaining entities other than the entity using the large model, to obtain a plurality of entity relationships corresponding to the entity;

[0055] A construction unit is used to construct a knowledge graph based on all the entities and their corresponding multiple entities; wherein each node in the knowledge graph represents one of the entities, and an edge between any two connected nodes represents one of the entity relationships.

[0056] Optionally, in the above-mentioned weight loss system for obese patients, the plan determination unit includes:

[0057] An extraction unit, used to extract key signs from the assessment report using a large model;

[0058] a first retrieval unit, configured to use the large model to retrieve dietary knowledge corresponding to the key features and the physical information from a pre-constructed knowledge graph, and formulate a diet plan for the patient based on the dietary knowledge, the key features, and the physical information;

[0059] a second retrieval unit, configured to use the large model to retrieve movement knowledge corresponding to the key features and the body information from a pre-built knowledge graph, and formulate an exercise plan based on the movement knowledge, the key features, and the body information;

[0060] A third retrieval unit is configured to use the large model to retrieve life knowledge corresponding to the vital sign monitoring information from a pre-built knowledge graph, and formulate a life plan based on the life knowledge and the vital sign monitoring information;

[0061] a formulation unit, configured to formulate a weight loss goal and expected effect by using the large model based on the body information and the knowledge of weight loss rules and weight loss effects in the knowledge graph;

[0062] The program determination subunit is used to integrate the diet plan, the exercise plan, the exercise plan, and the weight loss goal with the expected effect to obtain a weight loss program, and determine the weight loss program as a personalized weight loss program.

[0063] Optionally, the above-mentioned weight loss system for obese patients further includes:

[0064] A display unit, configured to display an interactive interface in response to a query operation in a menu bar of the triggering interactive platform;

[0065] A display unit is used to display query information corresponding to the query instruction to the patient according to the query instruction typed in the interactive interface.

[0066] Optionally, the above-mentioned weight loss system for obese patients further includes:

[0067] a receiving unit, configured to, upon receiving the question information typed by the patient, pre-process the question information to obtain target question information;

[0068] A deep analysis unit, configured to perform a deep analysis on the target question information using a large model to obtain the core intent corresponding to the target question information;

[0069] A fourth retrieval unit, configured to retrieve target knowledge corresponding to the core intent from the knowledge graph;

[0070] An answer generation unit is used to generate a simple answer based on the target knowledge using the large model, and to feed back the simple answer to the patient.

[0071] The present application provides a weight loss method for obese patients, which obtains the patient's physical information and physical sign monitoring information, and then inputs the physical information and physical sign monitoring information into a pre-trained obesity assessment model to obtain the patient's assessment report, wherein the obesity assessment model is pre-trained based on the DeepSeek large model, and the DeepSeek large model is pre-trained based on sample physical information and sample physical sign monitoring information of sample patients. Then, based on the assessment report and the pre-constructed knowledge graph, a personalized weight loss plan is determined, and the personalized weight loss plan is fed back to the patient. Then, the patient's weight loss data based on the personalized weight loss plan is collected. Finally, based on the weight loss data, it is judged whether the personalized weight loss plan is effective, and\or whether the patient has implemented the personalized weight loss plan. If the personalized weight loss plan is not effective, and\or the patient has not implemented the personalized weight loss plan, it is determined that the personalized weight loss plan is invalid, and the process returns to the personalized weight loss plan determined according to the assessment report until the personalized weight loss plan is effective, and\or the patient implements the personalized weight loss plan. By leveraging the DeepSeek model's ability to deeply analyze data and combining it with knowledge graphs, we can comprehensively and accurately assess patients' obesity status and develop personalized weight loss plans, thereby solving the problem of insufficient weight loss accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without any creative work.

[0073] Figure 1 A flow chart of a weight loss method for obese patients provided in an embodiment of the present application;

[0074] Figure 2 A flowchart of a method for generating an evaluation report provided in another embodiment of the present application;

[0075] Figure 3 A schematic diagram of a flow chart of a method for constructing a knowledge graph provided in another embodiment of the present application;

[0076] Figure 4 A flowchart of a method for determining a weight loss plan provided in another embodiment of the present application;

[0077] Figure 5 A flowchart of a method for querying patient information provided in another embodiment of the present application;

[0078] Figure 6 A flowchart of an interaction method provided in another embodiment of the present application;

[0079] Figure 7 This is a general architecture diagram of a weight reduction system provided in an embodiment of the present application;

[0080] Figure 8 This is a structural schematic diagram of a weight loss system for obese patients provided in another embodiment of the present application. DETAILED DESCRIPTION

[0081] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0082] In this application, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

[0083] The present invention provides a method for reducing weight in obese patients. Figure 1 As shown, the specific steps include:

[0084] S101. Obtain the patient's physical information and vital sign monitoring information.

[0085] Specifically, physical information may include height, weight, BMI (body mass index), body fat percentage, waist circumference, waist-to-hip ratio, blood pressure, blood sugar, blood lipids, and lifestyle data such as eating habits, exercise frequency, sleep quality, and psychological state. Vital sign monitoring information is collected from patients using intelligent weight scales, exercise equipment, and other vital sign monitoring tools. This information includes, but is not limited to, height, weight, and BMI (body mass index).

[0086] Optionally, when a patient needs to lose weight, the patient can send physical information and vital sign monitoring information through the data transmission interface provided by the system, or obtain the patient's physical information through a physical examination in the hospital and obtain vital sign monitoring information through the hospital's vital sign monitoring tools.

[0087] It should be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the patient or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0088] S102: Input the body information and physical sign monitoring information into a pre-trained obesity assessment model to obtain an assessment report of the patient.

[0089] Among them, the obesity assessment model is pre-trained based on the DeepSeek large model, and the DeepSeek large model is pre-trained based on the sample body information and sample vital sign monitoring information of sample patients.

[0090] It should be noted that the obesity assessment model is trained based on deep learning and combined with knowledge in a pre-built knowledge graph. The trained obesity assessment model is used to comprehensively assess the patient's obesity level, obesity type, and weight loss risk, and output an assessment report.

[0091] Optionally, in another embodiment of the present application, a specific implementation of step S102 is as follows: Figure 2 As shown, the specific steps include:

[0092] S201. Perform deep feature extraction on body information and physical sign monitoring information through an obesity assessment model to obtain patient data features.

[0093] Specifically, before executing step S201, the physical information and vital sign monitoring information are preprocessed. Numerical physical indicator data, such as height, weight, BMI, and body fat percentage, are normalized and standardized to eliminate the impact of dimensional differences on subsequent analysis. Non-numeric lifestyle data, such as dietary habits and exercise frequency, are encoded and converted into numerical form that the model can recognize.

[0094] Next, the DeepSeek model, also known as the obesity assessment model, uses deep learning algorithms to perform deep feature extraction on the preprocessed body information and vital sign monitoring information. This involves using the multi-layered neural network structure within the obesity assessment model to extract hidden features and patterns from the preprocessed body information and vital sign monitoring information, thereby obtaining patient data features with these hidden features and patterns. For example, the frequency of high-calorie and high-fat intake can be extracted from a patient's long-term dietary habit data, while features such as exercise intensity and exercise type can be extracted from exercise frequency data.

[0095] S202. Match the patient's data features with the obesity types in the pre-built knowledge graph through the obesity assessment model to obtain the patient's obesity level.

[0096] It should be noted that in order to determine the patient's obesity level and type, in this embodiment of the present application, the patient data features extracted in step S201 are associated with the weight management knowledge in the pre-constructed knowledge graph for retrieval. That is, the patient data features are used to retrieve obesity level classification knowledge related to the patient data features in the knowledge graph to determine the patient's obesity level (e.g., mild obesity, moderate obesity, severe obesity, etc.). Then, based on the body fat percentage and body composition data such as waist circumference and waist-to-hip ratio in the physical information, the knowledge graph is searched for matching obesity type knowledge to determine whether the patient has abdominal obesity, systemic obesity, or other types of obesity to determine the patient's obesity type. Therefore, after matching the obesity level and type using the patient data features and physical information with the weight management knowledge in the knowledge graph, the matched obesity level and type are used as the patient's obesity level. The pre-constructed knowledge graph includes, but is not limited to, weight management knowledge.

[0097] S203. Retrieve risk knowledge corresponding to patient data features from the knowledge graph through the obesity assessment model, and perform a fusion analysis of the patient data features and risk knowledge to obtain the patient's weight loss risk.

[0098] Specifically, after determining the patient's obesity level, the model comprehensively assesses the difficulty and potential risks of weight loss for different obesity types, as well as the impact of lifestyle factors on weight loss outcomes and health. Specifically, the obesity assessment model uses patient data features such as age, gender, and physical condition to match individual characteristics with weight loss plans in the knowledge graph. It then analyzes the risks that patients may face during weight loss based on this data and this adaptation knowledge.

[0099] For example, for older patients with a history of cardiovascular disease, the knowledge graph's association between exercise and cardiovascular risk can be used to assess the risk of high-intensity exercise triggering cardiovascular disease during weight loss. For obese patients with diabetes, the knowledge graph's relationship between diet and blood sugar control can be used to analyze the risk of blood sugar fluctuations caused by an unreasonable diet during weight loss.

[0100] S204. Generate an assessment report for the patient based on the obesity assessment model and the preset report template, according to the obesity level and weight loss risk.

[0101] It is understandable that after obtaining the obesity level and weight loss risk, the obesity assessment model will generate an assessment report for the obesity level and weight loss risk according to the preset report template. The assessment report will detail the patient's obesity level and obesity type, list the data features and relevant knowledge in the knowledge graph based on the assessment, and describe the patient's weight loss risk in a graded manner, such as low risk, medium risk, and high risk, and specifically point out the possible risk types and causes. Finally, based on the knowledge graph's knowledge of coping strategies for different obesity conditions and risks, the model provides patients with personalized weight loss recommendations, including dietary adjustment directions (such as recommended food types, daily calorie intake range), exercise plans (such as suitable sports, exercise intensity and frequency), lifestyle improvement recommendations (such as sleep duration, stress management methods), etc., so that patients can clearly understand their own obesity status and subsequent weight loss direction.

[0102] S103. Determine a personalized weight loss plan based on the assessment report and the pre-built knowledge graph, and provide the personalized weight loss plan to the patient.

[0103] Specifically, in an embodiment of the present application, the decision-making reasoning ability of the DeepSeek large model will be used to generate a personalized weight loss plan for the patient based on the evaluation results output by the obesity assessment model and the knowledge in the knowledge graph. The weight loss plan includes a diet plan, an exercise plan, and suggestions for adjusting lifestyle habits. For example, the diet plan will formulate daily food types, intakes, and matching plans based on the patient's nutritional needs and taste preferences. The exercise plan will design the appropriate type, intensity, frequency, and duration of exercise based on the patient's physical condition and exercise goals, and the lifestyle adjustment suggestions will cover aspects such as sleep and stress management. At the same time, the system will also provide patients with an explanation of their weight loss goals and expected results. Finally, the generated personalized weight loss plan will be fed back to the patient so that the patient can perform weight loss operations according to the personalized weight loss plan.

[0104] Optionally, the present application embodiment provides a method for constructing a knowledge graph, such as Figure 3 As shown, the following steps are included:

[0105] S301: Obtain multi-source data on obesity, and pre-process the multi-source data to obtain target multi-source data.

[0106] It should be noted that in order for the knowledge graph to encompass all relevant knowledge about obesity, it is necessary to integrate multiple sources of data, including obesity-related medical knowledge, nutrition knowledge, kinesiology knowledge, and clinical practice experience, so that a weight management knowledge graph can be successfully constructed. This knowledge graph can cover the causes and pathological mechanisms of obesity, the relationship between different individual characteristics (such as age, gender, body composition, etc.) and weight loss plans, the nutritional content of various foods, and exercise calorie consumption data. It also enables the dynamic updating and knowledge reasoning of the knowledge graph through the semantic understanding and knowledge association capabilities of the DeepSeek large model, providing knowledge support for subsequent obesity assessment and plan development.

[0107] Therefore, it is necessary to collect obesity-related medical knowledge, nutrition knowledge, kinesiology knowledge, and clinical practice experience from a wide range of data sources. This includes but is not limited to medical literature databases, professional books, scientific research reports, authoritative medical websites, and clinical case databases. Preprocessing of this multi-source data, including data cleaning, deduplication, and format conversion, is required to ensure data quality and usability.

[0108] S302: Use the large model to perform entity recognition on the target multi-source data to obtain multiple entities corresponding to the target multi-source data.

[0109] It is understandable that the knowledge graph includes entities and the connections between them. Therefore, to obtain the entities in the knowledge graph, we can use the natural language processing capabilities of the DeepSeek large model to perform entity recognition on the target multi-source data. In other words, the DeepSeek large model can be used to conduct in-depth analysis of the target multi-source data to accurately extract entities with specific meanings.

[0110] For example, the model can identify obesity-related disease names (such as metabolic syndrome and cardiovascular disease), body composition indicators (such as weight, body fat percentage, and muscle mass), food names (such as various grains, vegetables, fruits, and meats), sports names (such as running, swimming, yoga, and strength training), and individual characteristics (such as age, gender, height, and weight). For example, in the sentence "Apples are rich in vitamin C and fiber," the model can accurately identify the entities "apple," "vitamin C," and "fiber."

[0111] S303: For each entity, use the large model to analyze the semantic relationship between the entity and the remaining entities other than the entity, and obtain multiple entity relationships corresponding to the entity.

[0112] Specifically, in order to obtain the semantic relationship between each entity and thus connect each entity, we can use the powerful semantic understanding ability of the DeepSeek large model to determine the semantic relationship between entities, that is, by performing semantic analysis on the location of the target multi-source data where the entity is located, we can determine the association between different entities and then obtain multiple entity relationships corresponding to each entity.

[0113] For example, the causal relationship between the causes of obesity and obesity (for example, the causal link between a high-calorie diet and obesity), the compatibility between individual characteristics and weight loss plans (for example, different exercise and dietary weight loss plans are suitable for individuals of different ages, genders, and body compositions), the inclusion relationship between food and nutrients (for example, milk contains nutrients such as protein and calcium), and the correspondence between exercise and calorie consumption (for example, the number of calories burned per kilometer of running). For the sentence "High blood pressure is a common complication of obesity," the model can extract the "complication" relationship between "high blood pressure" and "obesity."

[0114] S304. Build a knowledge graph based on all entities and their corresponding multiple entities.

[0115] Specifically, the entities and entity relationships obtained are organized and stored in the form of a graph, thereby constructing the preliminary framework of the knowledge graph. In this case, each node in the knowledge graph represents an entity, and the edge between any two connected nodes represents an entity relationship.

[0116] For example, the "obesity" node can be connected to cause nodes such as "high-calorie diet" and "lack of exercise" through the "cause" relationship edge, and to disease nodes such as "hypertension" and "diabetes" through the "complication" relationship edge. "Food" nodes such as "apple" can be connected to nutrient nodes such as "vitamin C" and "fiber" through the "nutrient" relationship edge.

[0117] Optionally, in another embodiment of the present application, in step S103, a specific implementation of the personalized weight loss program is determined based on the evaluation report and the pre-built knowledge graph, such as Figure 4 As shown, the specific steps include:

[0118] S401. Extract key signs from the assessment report using a large model.

[0119] It is understandable that in order to develop a personalized weight loss plan for patients, the DeepSeek model can be used to first conduct an in-depth analysis of the results output by the obesity assessment model, that is, to extract key information from the assessment report, including the degree of obesity, type of obesity, weight loss risk level, and the patient's physical indicators, lifestyle and other data characteristics.

[0120] S402. Use the big model to retrieve dietary knowledge corresponding to key features and physical information from the pre-built knowledge graph, and formulate a diet plan for the patient based on the dietary knowledge, key features and physical information.

[0121] It is understandable that personalized weight loss plans include but are not limited to diet plans, exercise plans, lifestyle adjustment suggestions, etc. Therefore, a diet plan for the patient can be developed first. That is, when developing a diet plan, a daily diet plan should be generated based on nutritional needs and taste preferences. Specifically, the nutritional needs analysis needs to be based on the patient's physical indicator data (such as age, gender, height, weight, body fat percentage, basal metabolic rate, etc.), that is, key features and obesity assessment results. The DeepSeek large model will retrieve the nutritional needs knowledge corresponding to the key features and obesity assessment results in the knowledge graph, and calculate the patient's daily calorie requirements and the reasonable intake of various nutrients such as protein, carbohydrates, and fat.

[0122] For example, for a 35-year-old female patient who is 165cm tall and weighs 75kg with a high body fat percentage and is seeking weight loss, the model uses nutritional knowledge from the knowledge graph to determine their daily calorie deficit, as well as the ratios and specific values ​​for protein, carbohydrate, and fat intake. Taste preference integration analyzes the patient's input taste preferences and searches the knowledge graph for food types and pairings that meet their nutritional needs and preferences. If the patient prefers spicy flavors, the model will filter spicy and nutritious ingredients (such as peppers and onions) from the knowledge graph and combine them with other nutritious foods to create a diverse diet. It also considers factors such as the glycemic index and nutritional complementarity of foods to ensure the scientific and rationality of the diet. Daily diet plan generation allocates the calculated nutritional needs and selected foods according to the timing of meals and snacks to create a detailed daily diet plan. Clearly list the food types, intake and specific cooking methods for each meal. For example, breakfast is 50g of whole wheat bread, 1 egg, and 200ml of milk; lunch is 100g of brown rice, 200g of stir-fried vegetables, and 150g of steamed fish.

[0123] S403. Use the big model to retrieve the exercise knowledge corresponding to the key features and body information from the pre-built knowledge graph, and formulate an exercise plan based on the exercise knowledge, key features and body information.

[0124] Understandably, developing a patient's exercise plan requires a comprehensive assessment of their physical condition, along with the selection of exercise type, intensity, frequency, and duration. Therefore, the DeepSeek model is used to retrieve appropriate exercise type knowledge from the knowledge graph based on the patient's physical condition (such as cardiopulmonary function, joint flexibility, and the presence of exercise contraindications) and obesity type.

[0125] For example, for patients with a high body weight and high joint stress, low-impact exercises such as swimming, cycling, and water exercises are preferred. For patients with weak cardiopulmonary function, moderate-intensity, gradually progressive aerobic exercise, such as walking and jogging, should be selected, combined with appropriate resistance training to increase muscle mass and improve basal metabolic rate. Exercise intensity, frequency, and duration are determined based on the patient's exercise goals (such as rapid fat loss, muscle building, and improved health) and current physical fitness. Based on the kinematics knowledge in the knowledge graph, appropriate exercise intensity, frequency, and duration are determined. For patients whose primary goal is fat loss, the model will develop a plan for gradually increasing exercise intensity based on their body fat percentage and physical endurance. For example, starting with low-intensity brisk walking, 4-5 times per week for 30-40 minutes each time, gradually transitioning to moderate-intensity jogging or aerobics as the body adapts. Simultaneously, appropriate strength training should be combined, 2-3 times per week for 20-30 minutes each time. When creating a detailed exercise plan, the selected exercise type, intensity, frequency, and duration are refined into a specific exercise plan, including the specific content and timing of each exercise's warm-up, main training, and cool-down. For example, a running exercise plan might include a 5-minute warm-up with joint mobility and slow walking, a 25-minute main training session with moderate running, and a 10-minute cool-down with stretching and deep breathing exercises.

[0126] S404: Use the big model to retrieve life knowledge corresponding to the vital sign monitoring information from the pre-built knowledge graph, and formulate a life plan based on the life knowledge and vital sign monitoring information.

[0127] Understandably, implementing a patient's life plan includes stress management and sleep recommendations. Sleep management recommendations can be based on the patient's lifestyle and health status (i.e., lifestyle knowledge and vital signs monitoring information). The DeepSeek model searches the knowledge graph for relevant knowledge related to sleep, weight management, and health, and then develops personalized sleep adjustment recommendations for the patient.

[0128] For example, for patients who often stay up late and lack sleep, the model will recommend that they adjust their work and rest schedules to ensure 7-8 hours of high-quality sleep every day, and provide methods to improve sleep quality, such as avoiding the use of electronic devices before bed, keeping the bedroom quiet and dark, and establishing a regular sleep schedule. Stress management recommendations take into account the impact of stress on weight and health. Based on the psychology and health management knowledge in the knowledge graph, the model analyzes the patient's mental state and life stressors and provides patients with stress management strategies. For example, it recommends stress relief methods suitable for patients, including meditation, yoga, deep breathing exercises, and the cultivation of hobbies and interests. Patients are advised to arrange their work and life reasonably, learn time management, and avoid excessive fatigue and stress accumulation.

[0129] S405. Use the big model to formulate weight loss goals and expected results based on body information and the knowledge of weight loss rules and effects in the knowledge graph.

[0130] Specifically, to ensure patients have a targeted weight loss goal, personalized weight loss plans also include weight loss goals and expected results, allowing patients to visualize the effects of their weight loss. Therefore, weight loss goal setting requires finding corresponding weight loss patterns from the knowledge graph based on the patient's initial weight, physical condition, and weight loss needs. Then, using the DeepSeek model, we set a scientifically sound weight loss goal for the patient.

[0131] For example, goals are divided into short-term goals (such as 1-2 weeks) and long-term goals (such as 3-6 months). Short-term goals focus on cultivating healthy living habits and initial weight loss, while long-term goals set reasonable weight loss and body composition improvement targets based on the patient's physical potential and health status. For example, for a patient who is 20kg overweight, the short-term goal is to lose 0.5-1kg per week, and the long-term goal is to lose 10-12kg in 6 months, while also achieving a decrease in body fat percentage and an increase in muscle mass. The expected weight loss effects are combined with the effect knowledge corresponding to different weight loss plans in the knowledge graph to explain in detail to patients the expected effects that can be achieved after following the formulated weight loss plan in terms of weight change, improvement in body composition, improvement in health indicators (such as blood pressure, blood sugar, blood lipids, etc.), and improvement in quality of life. Through specific data and case studies, patients can clearly understand the weight loss process and the ultimate benefits, thereby strengthening their confidence and motivation to implement the weight loss plan.

[0132] S406. Integrate the diet plan, exercise plan, and weight loss goal with the expected effect to obtain a weight loss plan, and determine the weight loss plan as a personalized weight loss plan.

[0133] Specifically, the formulated diet plan, exercise plan, lifestyle adjustment suggestions, weight loss goals and expected results will be integrated, and then a complete personalized weight loss plan will be generated according to the preset plan template. The personalized weight loss plan should be presented to the patient in a clear and easy-to-understand manner to facilitate the patient's understanding and implementation, and at the same time provide comprehensive guidance and reference for the patient's subsequent weight loss process.

[0134] S104. Collect the patient's weight loss data based on the personalized weight loss plan.

[0135] It is understandable that after providing feedback on a personalized weight loss plan to the patient, in order to know whether the personalized weight loss plan is effective for the patient or whether the patient is willing to follow the personalized weight loss plan to lose weight, in the embodiment of the present application, during the patient's weight loss execution phase, the patient's physical sign change data (such as weight change, heart rate, sleep data, etc.), dietary intake data (through image recognition or manual input of food information, analysis of food nutrients) and exercise data (recording exercise type, duration, calories consumed, etc.) can be collected in real time through wearable devices (such as smart bracelets, smart body fat scales, etc.) and patient-initiated input. These data are transmitted to the system in real time as weight loss data for storage and analysis.

[0136] S105. Determine whether the personalized weight loss plan is effective based on the weight loss data, and / or whether the patient is implementing the personalized weight loss plan.

[0137] It should be noted that in order to understand the feasibility and adaptability of the patient to the personalized weight loss plan, the weight loss data can be pre-processed by using the DeepSeek large model, and data features can be extracted from the weight loss data and potential associations can be mined, and then matched with the knowledge graph, and the matched data can be compared with the original body information to determine whether the personalized weight loss plan is effective, and\or whether the patient implements the personalized weight loss plan. Therefore, if it is determined that the personalized weight loss plan is not effective, and\or the patient does not implement the personalized weight loss plan, indicating that the personalized weight loss plan is invalid or it is found that the weight loss effect does not meet expectations or the patient has compliance problems, then step S106 is executed.

[0138] Optionally, after executing step S105, the method further includes:

[0139] If the personalized weight loss plan is effective, and\or the patient implements the personalized weight loss plan, it is determined that the personalized weight loss plan is effective for the patient, and the patient can be reminded to continue to lose weight according to the personalized plan.

[0140] S106. Determine that the personalized weight loss plan is ineffective.

[0141] It should be noted that when the personalized weight loss plan is not effective, and\or the patient does not implement the personalized weight loss plan, it means that the personalized weight loss plan is ineffective and the patient's compliance is poor. Therefore, DeepSeek can be used to combine the analysis results with the knowledge graph knowledge to dynamically adjust the diet structure, exercise intensity or increase psychological intervention to ensure that the weight loss goal is achieved, that is, return to step S103 until the personalized weight loss plan is effective, and\or the patient implements the personalized weight loss plan.

[0142] In addition, the embodiment of the present application will automatically analyze the reasons and dynamically adjust the weight loss plan based on the analysis results, such as optimizing the diet structure, adjusting the exercise intensity, increasing psychological intervention, etc., to ensure the achievement of the weight loss goal.

[0143] Optionally, in order to facilitate patients to view their own information, another embodiment of the present application further provides a method for querying patient information, such as Figure 5 As shown, the specific steps include:

[0144] S501: In response to a query operation in a menu bar of a triggering interactive platform, an interactive interface is displayed.

[0145] Specifically, the embodiment of the present application provides an interactive interface for patients through interactive platforms such as mobile phone applications and web pages. Therefore, when the patient wants to query relevant information in the interactive interface, he can trigger the query operation in the menu bar of the interactive platform, and the interactive interface will be displayed to the patient.

[0146] S502. According to the query instruction entered in the interactive interface, query information corresponding to the query instruction is displayed to the patient.

[0147] It is understandable that patients can view their own obesity assessment reports, personalized weight loss plans, weight changes and other information in the interactive interface. That is, they only need to type the corresponding query command in the interactive interface, and then the system will query the command to find the corresponding results from the database and display the results in the interactive interface.

[0148] Optionally, in order to provide psychological support and encouragement to patients, patients can ask questions to the AI ​​weight loss doctor, give feedback on their feelings and situations, so that the system can answer the patient's questions, thereby enhancing the patient's confidence and compliance in weight loss. Therefore, in another embodiment of the present application, an interactive method is also provided, such as Figure 6 As shown, the specific steps include:

[0149] S601. When question information entered by a patient is received, pre-process the question information to obtain target question information.

[0150] Specifically, patients can ask questions to the system through various interactive methods such as text input boxes and voice-to-text conversion. Therefore, when the system obtains the question information typed by the patient, it will use the DeepSeek large model to pre-process the question information first, including removing special symbols and spaces in the text, correcting spelling errors, unifying the text format, etc., so as to obtain the target question information. For example, for the recognition errors that may occur in voice-to-text conversion, the model uses the predictive ability of the language model to correct them. For example, if the patient inputs "My knees hurt when I ran recently, what should I do", the model automatically recognizes and corrects it to "My knees hurt when I ran recently, what should I do", laying the foundation for subsequent processing.

[0151] S602: Use the big model to conduct in-depth analysis on the target question information to obtain the core intent corresponding to the target question information.

[0152] It should be noted that in order to flexibly and effectively identify the intention of patients asking questions, the DeepSeek large model can be used to apply the intention recognition technology in natural language processing to conduct in-depth analysis of the target question information. That is, through multi-layer neural networks and attention mechanisms, the keywords and key phrases in the target question information can be captured to understand the core intention of the patient's question.

[0153] For example, for the question "Which foods help with weight loss?", the model identifies the keywords "food" and "weight loss" and determines that the patient's intention is to obtain dietary knowledge related to weight loss. If the question is "I feel very hungry after exercise. Does this mean that weight loss is effective?", the model can understand that the patient's intention is to understand the relationship between exercise, weight loss, and hunger. At the same time, the model will also combine contextual information to accurately grasp the semantics of complex questions. For example, if a patient repeatedly asks, "I jump rope for 30 minutes every day, but my weight has not changed. Is this normal? You said before that you want to control your diet. How do you control it specifically?", the large model can thus link previous and subsequent questions and understand the patient's doubts about exercise and diet.

[0154] S603. Retrieve target knowledge corresponding to the core intent from the knowledge graph.

[0155] It should be noted that after determining the core intent of the patient's question, the DeepSeek large model will be used to quickly search the weight management knowledge graph and related knowledge bases with the core intent as the index. If the patient asks about weight loss foods, the model will search the knowledge graph for knowledge about the nutritional content, caloric value, and weight loss effects of various foods. If the question is about exercise, the relationship between exercise type, exercise intensity, weight loss effect, and body reaction will be retrieved. At the same time, the model will associate the patient's personal information (such as age, gender, physical condition, current weight loss plan, etc.) to filter out knowledge that is more in line with the patient's actual situation. For example, if a patient with arthritis asks about exercise methods, the model will prioritize retrieving knowledge about low-impact exercises suitable for arthritis patients to ensure that the answers provided are targeted.

[0156] S604. Generate simple answers based on the target knowledge using the large model, and provide the simple answers as feedback to the patient.

[0157] It is understandable that in order to enable patients to receive answers that are easy to understand, the DeepSeek large model will generate structured and easy-to-understand answers based on the retrieved knowledge and combined with natural language generation technology. That is, the DeepSeek large model adopts a sequence-to-sequence (Seq2Seq) architecture, and the encoder encodes the retrieved knowledge into semantic vectors, and the decoder then decodes the semantic vectors into natural language text. In the process of generating answers, the DeepSeek large model will adjust the expression method according to the type of question and the needs of the patient. For more professional questions, it uses easy-to-understand metaphors or cases to explain. For example, when explaining the concept of basal metabolic rate, the analogy is "the energy consumed every day to keep the body, a factory, running." At the same time, the model will also check the grammar, logical coherence and content accuracy of the generated answers, remove repeated or redundant information, and optimize the quality of the answers.

[0158] It should be noted that the embodiment of the present application also provides an overall architecture diagram of a weight loss system, which can be found in Figure 7 The content shown.

[0159] The present application provides a weight loss method for obese patients, which obtains the patient's physical information and physical sign monitoring information, and then inputs the physical information and physical sign monitoring information into a pre-trained obesity assessment model to obtain the patient's assessment report, wherein the obesity assessment model is pre-trained based on the DeepSeek large model, and the DeepSeek large model is pre-trained based on sample physical information and sample physical sign monitoring information of sample patients. Then, based on the assessment report and the pre-constructed knowledge graph, a personalized weight loss plan is determined, and the personalized weight loss plan is fed back to the patient. Then, the patient's weight loss data based on the personalized weight loss plan is collected. Finally, based on the weight loss data, it is judged whether the personalized weight loss plan is effective, and\or whether the patient has implemented the personalized weight loss plan. If the personalized weight loss plan is not effective, and\or the patient has not implemented the personalized weight loss plan, it is determined that the personalized weight loss plan is invalid, and the process returns to the personalized weight loss plan determined according to the assessment report until the personalized weight loss plan is effective, and\or the patient implements the personalized weight loss plan. By leveraging the DeepSeek model's ability to deeply analyze data and combining it with knowledge graphs, we can comprehensively and accurately assess patients' obesity status and develop personalized weight loss plans, thereby solving the problem of insufficient weight loss accuracy.

[0160] Another embodiment of the present application provides a weight loss system for obese patients, such as Figure 8 As shown, it includes the following units:

[0161] The information acquisition unit 801 is used to acquire the patient's physical information and vital sign monitoring information.

[0162] Input unit 802 is used to input physical information and vital sign monitoring information into a pre-trained obesity assessment model to obtain a patient assessment report. The obesity assessment model is pre-trained based on the DeepSeek large model. The DeepSeek large model is pre-trained based on sample physical information and sample vital sign monitoring information of sample patients.

[0163] The plan determination unit 803 is used to determine a personalized weight loss plan based on the evaluation report and the pre-built knowledge graph, and feed back the personalized weight loss plan to the patient.

[0164] The data collection unit 804 is used to collect the weight loss data of the patient based on the personalized weight loss plan.

[0165] The judgment unit 805 is used to judge whether the personalized weight loss plan is effective and / or whether the patient is implementing the personalized weight loss plan based on the weight loss data.

[0166] Return to execution unit 806, which is used to determine that the personalized weight loss plan is invalid if the personalized weight loss plan is not effective and\or the patient does not implement the personalized weight loss plan, and return to execution to determine the personalized weight loss plan based on the evaluation report until the personalized weight loss plan is effective and\or the patient implements the personalized weight loss plan.

[0167] It should be noted that the specific working process of the above modules in the embodiment of the present application can refer to steps S101 to S106 in the above method embodiment, and will not be repeated here.

[0168] Optionally, in another embodiment of the present application, in a weight loss system for obese patients, the input unit 802 includes:

[0169] The feature extraction unit is used to perform deep feature extraction on body information and vital sign monitoring information through the obesity assessment model to obtain patient data features.

[0170] The matching unit is used to match the patient's data features with the obesity types in the pre-built knowledge graph through the obesity assessment model to obtain the patient's obesity level.

[0171] The fusion analysis unit is used to retrieve the risk knowledge corresponding to the patient data characteristics from the knowledge graph through the obesity assessment model, and perform fusion analysis on the patient data characteristics and risk knowledge to obtain the patient's weight loss risk.

[0172] The report generation unit is used to generate an assessment report for a patient according to the obesity level and weight loss risk through an obesity assessment model in accordance with a preset report template.

[0173] Optionally, another embodiment of the present application provides a weight loss system for obese patients, further comprising:

[0174] The preprocessing unit is used to obtain multi-source data on obesity and preprocess the multi-source data to obtain target multi-source data.

[0175] The entity recognition unit is used to use the large model to perform entity recognition on the target multi-source data to obtain multiple entities corresponding to the target multi-source data.

[0176] The analysis unit is used to analyze the semantic relationship between each entity and the remaining entities other than the entity using the large model to obtain multiple entity relationships corresponding to the entity.

[0177] The construction unit is used to build a knowledge graph based on all entities and their corresponding multiple entities. Each node in the knowledge graph represents an entity, and the edge between any two connected nodes represents an entity relationship.

[0178] Optionally, in a weight loss system for obese patients provided in another embodiment of the present application, the plan determination unit 803 includes:

[0179] The extraction unit is used to extract key signs from the assessment report using the large model.

[0180] The first retrieval unit is used to use a large model to retrieve dietary knowledge corresponding to key features and physical information from a pre-built knowledge graph, and to formulate a diet plan for the patient based on the dietary knowledge, key features and physical information.

[0181] The second retrieval unit is used to use the large model to retrieve exercise knowledge corresponding to key features and body information from a pre-built knowledge graph, and to formulate an exercise plan based on the exercise knowledge, key features and body information.

[0182] The third retrieval unit is used to use the big model to retrieve the life knowledge corresponding to the vital sign monitoring information from the pre-built knowledge graph, and formulate a life plan based on the life knowledge and vital sign monitoring information.

[0183] The formulation unit is used to use the big model to formulate weight loss goals and expected results based on body information and the knowledge of weight loss rules and weight loss effects in the knowledge graph.

[0184] The program determination subunit is used to integrate the diet plan, exercise plan, exercise plan and weight loss goals with the expected effects to obtain a weight loss program, and determine the weight loss program as a personalized weight loss program.

[0185] Optionally, another embodiment of the present application provides a weight loss system for obese patients, further comprising:

[0186] The display unit is configured to display an interactive interface in response to a query operation in a menu bar that triggers the interactive platform.

[0187] The display unit is used to display the query information corresponding to the query instruction to the patient according to the query instruction typed in the interactive interface.

[0188] Optionally, another embodiment of the present application provides a weight loss system for obese patients, further comprising:

[0189] The receiving unit is used to pre-process the question information when receiving the question information typed by the patient to obtain target question information.

[0190] The deep analysis unit is used to use the large model to conduct in-depth analysis of the target problem information to obtain the core intent corresponding to the target problem information.

[0191] The fourth retrieval unit is used to retrieve target knowledge corresponding to the core intent from the knowledge graph.

[0192] The answer generation unit is used to generate simple answers based on target knowledge using a large model and to provide the simple answers as feedback to the patient.

[0193] It should be noted that the specific working processes of the various modules provided in the above embodiments of the present application can refer to the corresponding steps in the above method embodiments, and will not be repeated here.

[0194] It should also be noted that the embodiment of the present application provides a conversational structured information collection system, which has the technical effects of any of the above embodiments, and the embodiment of the present application will not be described in detail here.

[0195] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0196] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for reducing weight in obese patients, characterized in that: include: Obtain patient's physical information and vital sign monitoring information; Inputting the physical information and the vital sign monitoring information into a pre-trained obesity assessment model to obtain an assessment report for the patient; wherein the obesity assessment model is pre-trained based on the DeepSeek large model; and the DeepSeek large model is pre-trained based on sample physical information and sample vital sign monitoring information of sample patients; Determining a personalized weight loss plan based on the assessment report and the pre-built knowledge graph, and feeding the personalized weight loss plan back to the patient; collecting weight loss data of the patient based on the personalized weight loss plan; Determining, based on the weight loss data, whether the personalized weight loss program is effective, and / or whether the patient is implementing the personalized weight loss program; If the personalized weight loss plan is not effective, and\or the patient does not implement the personalized weight loss plan, the personalized weight loss plan is determined to be invalid, and the process of determining the personalized weight loss plan based on the evaluation report is returned to execution until the personalized weight loss plan is effective, and\or the patient implements the personalized weight loss plan.

2. The method according to claim 1, characterized in that The step of inputting the body information and the physical sign monitoring information into a pre-trained obesity assessment model to obtain an assessment report of the patient includes: Performing deep feature extraction on the body information and the vital sign monitoring information using the obesity assessment model to obtain patient data features; Matching the patient data features with the obesity types in the pre-built knowledge graph using the obesity assessment model to obtain the patient's obesity level; Retrieving risk knowledge corresponding to the patient data features from the knowledge graph through the obesity assessment model, and performing a fusion analysis on the patient data features and the risk knowledge to obtain the patient's weight loss risk; An assessment report for the patient is generated by the obesity assessment model according to a preset report template, based on the obesity level and the weight loss risk.

3. The method according to claim 1, characterized in that The method for constructing the knowledge graph includes: Acquiring multi-source data on obesity and preprocessing the multi-source data to obtain target multi-source data; Using the large model to perform entity recognition on the target multi-source data, and obtain multiple entities corresponding to the target multi-source data; For each entity, using the large model to analyze the semantic relationship between the entity and the remaining entities other than the entity, to obtain multiple entity relationships corresponding to the entity; A knowledge graph is constructed based on all the entities and their corresponding multiple entities; wherein each node in the knowledge graph represents one of the entities, and an edge between any two connected nodes represents one of the entity relationships.

4. The method according to claim 1, wherein Determining a personalized weight loss plan based on the assessment report and the pre-built knowledge graph includes: Extract key signs from assessment reports using large models; Retrieving dietary knowledge corresponding to the key features and the physical information from a pre-built knowledge graph using the large model, and formulating a diet plan for the patient based on the dietary knowledge, the key features, and the physical information; Retrieving exercise knowledge corresponding to the key features and the body information from a pre-built knowledge graph using the large model, and formulating an exercise plan based on the exercise knowledge, the key features, and the body information; Retrieving life knowledge corresponding to the vital sign monitoring information from a pre-built knowledge graph using the large model, and formulating a life plan based on the life knowledge and the vital sign monitoring information; Utilizing the large model to formulate weight loss goals and expected results based on the physical information and the knowledge of weight loss patterns and weight loss effects in the knowledge graph; The diet plan, the exercise plan, the exercise plan, and the weight loss goal are integrated with the expected effect to obtain a weight loss plan, and the weight loss plan is determined as a personalized weight loss plan.

5. The method according to claim 1, characterized in that Also includes: In response to a query operation in a menu bar of the triggering interactive platform, displaying an interactive interface; According to the query instruction typed in the interactive interface, query information corresponding to the query instruction is displayed to the patient.

6. The method according to claim 1, characterized in that Also includes: When receiving the question information typed by the patient, pre-processing the question information to obtain target question information; Use the big model to conduct in-depth analysis on the target question information to obtain the core intent corresponding to the target question information; Retrieving target knowledge corresponding to the core intent from the knowledge graph; The large model is used to generate a simple answer based on the target knowledge, and the simple answer is fed back to the patient.

7. A weight loss system for obese patients, characterized in that: include: An information acquisition unit, used to obtain the patient's physical information and vital sign monitoring information; An input unit, configured to input the physical information and the vital sign monitoring information into a pre-trained obesity assessment model to obtain an assessment report for the patient; wherein the obesity assessment model is pre-trained based on a DeepSeek large model; and the DeepSeek large model is pre-trained based on sample physical information and sample vital sign monitoring information of sample patients; a plan determination unit, configured to determine a personalized weight loss plan based on the assessment report and a pre-built knowledge graph, and feed the personalized weight loss plan back to the patient; a data collection unit, configured to collect weight loss data of the patient based on the personalized weight loss plan; a judgment unit, configured to judge whether the personalized weight loss plan is effective, and / or whether the patient is implementing the personalized weight loss plan, based on the weight loss data; The return execution unit is used to determine that the personalized weight loss plan is invalid if the personalized weight loss plan is not effective and\or the patient does not implement the personalized weight loss plan, and return to the execution of determining the personalized weight loss plan based on the evaluation report until the personalized weight loss plan is effective and\or the patient implements the personalized weight loss plan.

8. The system according to claim 7, characterized in that The input unit includes: a feature extraction unit, configured to perform deep feature extraction on the body information and the vital sign monitoring information using the obesity assessment model to obtain patient data features; a matching unit, configured to match the patient data features with the obesity types in the pre-built knowledge graph using the obesity assessment model to obtain the patient's obesity level; a fusion analysis unit, configured to retrieve risk knowledge corresponding to the patient data features from the knowledge graph through the obesity assessment model, and perform a fusion analysis on the patient data features and the risk knowledge to obtain the patient's weight loss risk; A report generating unit is used to generate an assessment report for the patient according to the obesity assessment model and a preset report template, based on the obesity level and the weight loss risk.

9. The system according to claim 7, wherein: Also includes: a preprocessing unit, configured to obtain multi-source data on obesity and preprocess the multi-source data to obtain target multi-source data; An entity recognition unit, configured to perform entity recognition on the target multi-source data using a large model to obtain multiple entities corresponding to the target multi-source data; an analyzing unit, configured to analyze, for each entity, the semantic relationship between the entity and the remaining entities other than the entity using the large model, to obtain a plurality of entity relationships corresponding to the entity; A construction unit is used to construct a knowledge graph based on all the entities and their corresponding multiple entities; wherein each node in the knowledge graph represents one of the entities, and an edge between any two connected nodes represents one of the entity relationships.

10. The system according to claim 7, wherein: The scheme determination unit includes: An extraction unit, used to extract key signs from the assessment report using a large model; a first retrieval unit, configured to use the large model to retrieve dietary knowledge corresponding to the key features and the physical information from a pre-constructed knowledge graph, and formulate a diet plan for the patient based on the dietary knowledge, the key features, and the physical information; a second retrieval unit, configured to use the large model to retrieve movement knowledge corresponding to the key features and the body information from a pre-built knowledge graph, and formulate an exercise plan based on the movement knowledge, the key features, and the body information; A third retrieval unit is configured to use the large model to retrieve life knowledge corresponding to the vital sign monitoring information from a pre-built knowledge graph, and formulate a life plan based on the life knowledge and the vital sign monitoring information; a formulation unit, configured to formulate a weight loss goal and expected effect by using the large model based on the body information and the knowledge of weight loss rules and weight loss effects in the knowledge graph; The program determination subunit is used to integrate the diet plan, the exercise plan, the exercise plan, and the weight loss goal with the expected effect to obtain a weight loss program, and determine the weight loss program as a personalized weight loss program.

Images