Construction method for predicting glycolipid metabolic disease model based on artificial intelligence

Abstract

The invention relates to the technical field of medical information management, and discloses a method for constructing a model for predicting glycolipid metabolic disease based on artificial intelligence, which comprises the following steps: identifying the syndrome type and pathological product of a patient with glycolipid metabolic disease, and constructing a risk identification model of the patient with glycolipid metabolic disease according to the syndrome type and pathological product; identifying the disease site, syndrome and disease factors of the glycolipid metabolic disease patient, and analyzing the disease course distribution condition of the glycolipid metabolic disease patient; according to the disease course distribution condition, determining the distinguishing capability of the risk identification model, and executing model optimization processing of the risk identification model to obtain an optimization model; when the optimization processing result meets a preset effect, outputting a feedback result of the glycolipid metabolic disease patient, and setting a result paraphrase of the feedback result; and in combination with the risk identification model, the optimization model and the result paraphrasing, performing model construction processing of the glycolipid metabolic disease patient to obtain a model construction result. According to the invention, the accuracy and reliability of glycolipid metabolic disease prediction can be improved.

Description

Technical Field

[0001] The present invention relates to a method for constructing an artificial intelligence-based prediction model for glycolipid metabolism diseases, belonging to the technical field of medical information management. Background Art

[0002] Glycolipid metabolism diseases refer to a group of diseases caused by abnormal metabolism of sugar and fat, including abnormal blood sugar and obesity, atherosclerosis, etc. They are common and frequently-occurring diseases in clinical practice, and are likely to cause chronic lesions, functional decline or even failure of tissues and organs such as the heart and blood vessels. They are one of the diseases that pose a major threat to human health. With the change of lifestyle and the increase of the aging population, the incidence of glycolipid metabolism diseases has been increasing year by year, becoming a major public health problem globally. Therefore, it is of great significance to develop an effective prediction model to achieve early detection and prevention of glycolipid metabolism diseases.

[0003] Traditional prediction of glycolipid metabolism diseases mainly uses the four diagnostic methods of inspection, auscultation and olfaction, inquiry and palpation. By observing information such as the tongue image, smell, eating habits, and pulse condition of the patient, to understand the physical condition and potential causes of the patient. However, this method may lack objective quantitative indicators, such as the degree of dull complexion, the degree of stringy and slippery pulse, etc., making it difficult for traditional Chinese medicine to give specific numerical ranges and risk levels when predicting glycolipid metabolism diseases, thus affecting the accuracy and reliability of the prediction.

[0004] Therefore, there is an urgent need for a solution to improve the accuracy and reliability of the prediction of glycolipid metabolism diseases. Summary of the Invention

[0005] The present invention provides a method for constructing an artificial intelligence-based prediction model for glycolipid metabolism diseases, and its main purpose is to improve the accuracy and reliability of the prediction of glycolipid metabolism diseases.

[0006] To achieve the above object, a method for constructing an artificial intelligence-based prediction model for glycolipid metabolism diseases provided by the present invention includes: Select patients with glycolipid metabolism diseases, collect the tongue image, pulse signal and inquiry record of the patients with glycolipid metabolism diseases, and extract the key syndrome characteristics of the patients with glycolipid metabolism diseases according to the tongue image, the pulse signal and the inquiry record; According to the key syndrome characteristics, identify the syndrome types of the patients with glycolipid metabolism diseases. Based on the syndrome types, extract the pathological products of the patients with glycolipid metabolism diseases, and combine the syndrome types and the pathological products to construct a risk identification model for the patients with glycolipid metabolism diseases; Extract the distribution characteristics and manifestation types of the pathological products, identify the disease location syndrome elements and disease nature elements of the patients with glycolipid metabolism diseases according to the distribution characteristics and the manifestation types, and analyze the course distribution of the patients with glycolipid metabolism diseases based on the disease location syndrome elements and the disease nature elements; According to the course distribution, determine the discrimination ability of the risk identification model. Based on the discrimination ability, perform model optimization processing on the risk identification model to obtain an optimized model; When the optimization processing result meets the preset effect, use the optimized model to identify the risk factors of the glycolipid metabolism disease patient. According to the risk factors, output the feedback result of the glycolipid metabolism disease patient, and set the result interpretation of the feedback result; Combining the risk identification model, the optimized model, and the result interpretation, perform model construction processing on the glycolipid metabolism disease patient to obtain a model construction result.

[0007] Optionally, extracting the key syndrome characteristics of the glycolipid metabolism disease patient according to the tongue image, the pulse signal, and the interview record includes: Extract the tongue image characteristics of the glycolipid metabolism disease patient according to the tongue image; Based on the tongue image characteristics, analyze the variation law of the tongue image of the glycolipid metabolism disease patient; Identify the floating and sinking degree and pulse rate of the glycolipid metabolism disease patient according to the pulse signal; Based on the floating and sinking degree and the pulse rate, determine the pulse syndrome type of the glycolipid metabolism disease patient; Extract the symptom and sign record of the glycolipid metabolism disease patient from the interview record, and identify the symptom nature of the glycolipid metabolism disease patient according to the symptom and sign record; Combine the variation law of the tongue image, the pulse syndrome type, and the symptom nature to extract the key syndrome characteristics of the glycolipid metabolism disease patient.

[0008] Optionally, identifying the syndrome type of the glycolipid metabolism disease patient according to the key syndrome characteristics includes: Determine the disease nature type of the glycolipid metabolism disease patient according to the key syndrome characteristics; Identify the symptom manifestations of the glycolipid metabolism disease patient, and based on the symptom manifestations, analyze the disease nature of the glycolipid metabolism disease patient; Locate the lesion position of the glycolipid metabolism disease patient according to the disease nature and the disease nature type; Combine the lesion position, the disease nature, and the disease nature type to identify the pathogenic factors of the glycolipid metabolism disease patient; Based on the pathogenic factors, the lesion position, and the disease nature type, analyze the disease development trend of the glycolipid metabolism disease patient; Identify the syndrome type of the glycolipid metabolism disease patient according to the disease nature type, the lesion position, the pathogenic factors, and the disease development trend.

[0009] Optionally, extracting the pathological products of the patient with glycolipid metabolism disease based on the syndrome type includes: Identifying the pathogenic factors of the patient with glycolipid metabolism disease based on the syndrome type; Determining the abnormal body surface parts of the patient with glycolipid metabolism disease according to the pathogenic factors; Analyzing the pathological manifestations of the patient with glycolipid metabolism disease based on the pathogenic factors and the abnormal body surface parts; Identifying the visceral function state of the patient with glycolipid metabolism disease under the pathological manifestations; Identifying the metabolism of qi, blood, body fluid of the patient with glycolipid metabolism disease according to the visceral function state; Extracting the pathological products of the patient with glycolipid metabolism disease based on the visceral function state and the metabolism of qi, blood, body fluid.

[0010] Optionally, constructing a risk identification model for the patient with glycolipid metabolism disease by combining the syndrome type and the pathological products includes: Collecting the four diagnostic data of traditional Chinese medicine of the patient with glycolipid metabolism disease by combining the syndrome type and the pathological products; Extracting the disease evaluation indexes of the patient with glycolipid metabolism disease from the four diagnostic data of traditional Chinese medicine; Identifying the type of glycolipid metabolism disease of the patient with glycolipid metabolism disease by using the disease evaluation indexes, and extracting the independent risk factors corresponding to the type of glycolipid metabolism disease; Determining the category of the demand model of the patient with glycolipid metabolism disease according to the independent risk factors; Constructing a risk identification model for the patient with glycolipid metabolism disease based on the independent risk factors and the category of the demand model.

[0011] Optionally, identifying the syndrome elements of the disease location and the elements of the disease nature of the patient with glycolipid metabolism disease according to the distribution characteristics and the manifestation types includes: Extracting the pathological products corresponding to the distribution characteristics and the manifestation types; Identifying the disease development of the patient with glycolipid metabolism disease according to the pathological products; Analyzing the disease manifestations of the patient with glycolipid metabolism disease based on the disease development; Extracting the associated organs of the patient with glycolipid metabolism disease according to the disease manifestations, and analyzing the degree of disorder of the associated organs; Determining the lesion site of the patient with glycolipid metabolism disease based on the degree of disorder and the disease manifestations; Extracting the pathological features of the patient with glycolipid metabolism disease according to the lesion site and the pathological products; Combined with the diseased site and the pathological features, identify the syndrome elements of the diseased site and the pathogenic factors of the patient with glycolipid metabolism disorder.

[0012] Optionally, based on the syndrome elements of the diseased site and the pathogenic factors, analyze the course distribution of the patient with glycolipid metabolism disorder, including: Based on the syndrome elements of the diseased site and the pathogenic factors, collect the historical consultation data of the patient with glycolipid metabolism disorder; Extract the initial symptoms and current symptoms of the patient with glycolipid metabolism disorder from the historical consultation data; According to the initial symptoms and the current symptoms, identify the symptom change trajectory of the patient with glycolipid metabolism disorder; Identify the symptom turning points of the patient with glycolipid metabolism disorder under the symptom change trajectory; Based on the symptom turning points, define the classified course stages of the patient with glycolipid metabolism disorder; According to the symptom change trajectory and the historical consultation data, construct the probability distribution graph of the pathogenic factors; Based on the probability distribution graph, identify the key pathogenic factors of the patient with glycolipid metabolism disorder; According to the key pathogenic factors, the classified course stages and the symptom change trajectory, analyze the course distribution of the patient with glycolipid metabolism disorder.

[0013] Optionally, according to the course distribution, determine the discrimination ability of the risk identification model, including: Collect the disease distribution data corresponding to the course distribution, and extract the characteristic factors in the disease distribution data; Based on the characteristic factors, set the input features of the risk identification model; According to the input features, determine the target variable of the risk identification model; Use the input features and the target variable to split the course distribution data into the training set and the test set of the risk identification model; Use the training set to train the risk identification model to obtain a trained model; Based on the trained model, perform data analysis and processing on the test set to generate model analysis data; Calculate the data accuracy rate of the model analysis data; According to the data accuracy rate, determine the discrimination ability of the risk identification model.

[0014] Optionally, according to the risk factors, output the feedback result of the patient with glycolipid metabolism disorder, including: According to the risk factors, analyze the disease types of the patient with glycolipid metabolism disorder; Define the health evaluation indicators for the patients with glycolipid metabolism diseases based on the disease types; Calculate the risk scores for the patients with glycolipid metabolism diseases according to the health evaluation indicators; Identify the risk levels of the patients with glycolipid metabolism diseases according to the risk scores; Set the personalized guidance suggestions for the patients with glycolipid metabolism diseases based on the risk classification and the health evaluation indicators; Output the feedback results of the patients with glycolipid metabolism diseases by combining the disease types, the risk levels and the personalized guidance suggestions.

[0015] Compared with the problems described in the background art, in the embodiments of the present invention, by extracting the key syndrome characteristics of the patients with glycolipid metabolism diseases according to the tongue image, the pulse signal and the inquiry record, comprehensive input data can be provided for the prediction model, which helps to improve the accuracy and reliability of the model. Further, in the embodiments of the present invention, by identifying the syndrome types of the patients with glycolipid metabolism diseases according to the key syndrome characteristics, and extracting the pathological products of the patients with glycolipid metabolism diseases based on the syndrome types, a risk identification model of the patients with glycolipid metabolism diseases is constructed, which can more comprehensively reflect the pathological state of the patients with glycolipid metabolism diseases, provide richer information for the risk identification model, and thus realize personalized risk prediction. In the embodiments of the present invention, by extracting the distribution characteristics and manifestation types of the pathological products, the specific location and nature of the patient's lesion can be clarified, which helps to analyze the course distribution of the patients with glycolipid metabolism diseases, clarify the TCM syndrome characteristics and pathological mechanisms at different stages, help to predict the occurrence risk of glycolipid metabolism diseases, and realize early intervention. Further, in the embodiments of the present invention, by determining the discrimination ability of the risk identification model according to the course distribution, the interpretability of the model to the data and the prediction accuracy can be evaluated, so as to more accurately identify the risk of glycolipid metabolism diseases. Further, in the embodiments of the present invention, by outputting the feedback results of the patients with glycolipid metabolism diseases according to the risk factors and setting the result interpretation of the feedback results, the possible complications of the patients with glycolipid metabolism diseases can be predicted in advance, which helps the patients to better understand their own conditions and risk factors, and thus improves the self-management ability of the patients. In the embodiments of the present invention, by combining the risk identification model, the optimization model and the result interpretation, the model construction process of the patients with glycolipid metabolism diseases is executed, and the model construction result is obtained, which can clarify which factors are the most critical for risk prediction, significantly improve the accuracy and generalization ability of the model, enable it to work effectively in different patient groups, improve the acceptability and practicality of the model, help TCM doctors to identify high-risk populations, intervene in advance to prevent the occurrence of diseases, and at the same time can be used as a health management tool to provide personalized guidance suggestions for each patient to help the patients better manage their health status. Therefore, a method for constructing an artificial intelligence-based prediction model for glycolipid metabolism diseases provided by the embodiments of the present invention can improve the accuracy and reliability of glycolipid metabolism disease prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 FIG. is a schematic flowchart of a method for constructing an artificial intelligence-based prediction model for glycolipid metabolism diseases provided by an embodiment of the present invention; Figure 2 FIG. is a schematic block diagram of modules for implementing the method for constructing an artificial intelligence-based prediction model for glycolipid metabolism diseases provided by an embodiment of the present invention.

[0017] The realization of the object, functional features and advantages of the present invention will be further described in conjunction with the embodiments with reference to the accompanying drawings. Detailed implementation manners

[0018] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] The embodiment of the present application provides a construction method for an artificial intelligence-based prediction model for glycolipid metabolic diseases. The execution subject of the construction method for an artificial intelligence-based prediction model for glycolipid metabolic diseases includes, but is not limited to, at least one of electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the construction method for an artificial intelligence-based prediction model for glycolipid metabolic diseases can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0020] Embodiment 1: Refer to Figure 1 As shown, it is a schematic flowchart of a construction method for an artificial intelligence-based prediction model for glycolipid metabolic diseases provided by an embodiment of the present invention. In this embodiment, the construction method for an artificial intelligence-based prediction model for glycolipid metabolic diseases includes: S1. Select patients with glycolipid metabolic diseases, collect the tongue image, pulse signal and medical interview record of the patients with glycolipid metabolic diseases, and extract the key syndrome characteristics of the patients with glycolipid metabolic diseases according to the tongue image, the pulse signal and the medical interview record.

[0021] In the embodiment of the present invention, by selecting patients with glycolipid metabolic diseases, the research object for subsequent analysis can be clarified. The patients with glycolipid metabolic diseases refer to individuals suffering from metabolic diseases, such as diabetic patients.

[0022] Furthermore, in the embodiment of the present invention, by collecting the tongue image, pulse signal and medical interview record of the patients with glycolipid metabolic diseases, data support can be provided for the subsequent construction of a glycolipid metabolism and prediction model. The tongue image refers to an image of the patient's tongue taken by a professional device, including the color, shape, etc. of the tongue body and the tongue coating. The pulse signal refers to the characteristic signals such as the waveform, intensity, rhythm, etc. of the patient's pulse recorded by a pulse detection device. The medical interview record refers to the detailed information obtained by communicating with the patient, including symptoms, medical history, lifestyle, family medical history, etc.

[0023] Optionally, the tongue image of the patients with glycolipid metabolic diseases can be collected by a tongue diagnosis instrument, the pulse diagnosis signal of the patients with glycolipid metabolic diseases can be collected by manual pulse-taking record, and the medical interview record of the patients with glycolipid metabolic diseases can be collected by using an electronic medical record system.

[0024] In the embodiment of the present invention, by extracting the key syndrome characteristics of the patient with glycolipid metabolism disease according to the tongue image, the pulse condition signal and the medical interview record, comprehensive input data can be provided for the prediction model, which helps to improve the accuracy and reliability of the model. The key syndrome characteristics refer to the typical manifestations that can reflect the current pathological state and physical characteristics of the patient, such as obesity, unsmooth defecation, frequent nocturia, etc.

[0025] As an embodiment of the present invention, the extracting of the key syndrome characteristics of the patient with glycolipid metabolism disease according to the tongue image, the pulse condition signal and the medical interview record includes: extracting the tongue image characteristics of the patient with glycolipid metabolism disease according to the tongue image; analyzing the changing rules of the tongue image of the patient with glycolipid metabolism disease based on the tongue image characteristics; identifying the floating and sinking degree and the pulse rate of the pulse condition of the patient with glycolipid metabolism disease according to the pulse condition signal; determining the pulse syndrome type of the patient with glycolipid metabolism disease based on the floating and sinking degree and the pulse rate of the pulse condition; extracting the symptom and sign record of the patient with glycolipid metabolism disease from the medical interview record, and identifying the nature of the symptoms of the patient with glycolipid metabolism disease according to the symptom and sign record; and extracting the key syndrome characteristics of the patient with glycolipid metabolism disease by combining the changing rules of the tongue image, the pulse syndrome type and the nature of the symptoms.

[0026] Among them, the tongue image characteristics refer to the characteristic information obtained by observing the color, shape, movement of the tongue body and the color, thickness, dryness and moistness of the tongue coating, etc. The changing rules of the tongue image refer to the dynamic changing trends shown by the tongue image during the occurrence, development and prognosis of the disease, such as the tongue coating changing from thick to thin, from dry to moist. The floating and sinking degree of the pulse condition refers to the depth of the pulse position felt by feeling the pulse, such as floating pulse, sinking pulse, etc. The pulse rate refers to the frequency of the pulse beating, including slow pulse, rapid pulse. The pulse syndrome type refers to the syndrome type judged according to the comprehensive characteristics such as the floating and sinking, rate, strength, rhythm of the pulse condition, such as floating pulse indicating exterior syndrome, rapid pulse indicating cold syndrome, etc. The symptom and sign record refers to the detailed description of the subjective symptoms and objective signs of the patient obtained through the medical interview, such as the patient's self-report of dry mouth, fatigue and polyuria, etc. The nature of the symptoms refers to the pathological essence obtained after classifying and judging the attributes of the symptoms shown by the patient. For example, when the patient shows symptoms such as dry mouth, excessive thirst and polyuria, it may indicate internal heat due to yin deficiency.

[0027] Optionally, the extraction of the tongue image features of the patient with glycolipid metabolism disease can be obtained through digital image processing technology according to the tongue image. For example, deep features of the tongue image can be extracted using deep learning methods, and shallow features can be extracted in combination with edge feature extraction methods. Based on the tongue image features, the analysis of the change law of the tongue image of the patient with glycolipid metabolism disease can be determined by collecting tongue images multiple times and analyzing the dynamic changes of the tongue image during the disease treatment process. According to the pulse signal, the recognition of the floating and sinking degree and pulse rate of the patient with glycolipid metabolism disease can be determined by a pulse analyzer.

[0028] S2. According to the key syndrome characteristics, identify the syndrome types of the patients with glycolipid metabolism disease. Based on the syndrome types, extract the pathological products of the patients with glycolipid metabolism disease. Combine the syndrome types and the pathological products to construct a risk recognition model for the patients with glycolipid metabolism disease.

[0029] In the embodiment of the present invention, by identifying the syndrome types of the patients with glycolipid metabolism disease according to the key syndrome characteristics, the pathological essence of glycolipid metabolism disease can be determined, and a personalized treatment plan can be formulated accordingly. The syndrome type refers to the category of the disease essence obtained by comprehensive judgment based on various information such as the symptoms, signs, tongue image, and pulse condition of the patient, such as qi stagnation and phlegm turbidity syndrome, phlegm stasis intermingled syndrome, etc.

[0030] As an embodiment of the present invention, the identification of the syndrome types of the patients with glycolipid metabolism disease according to the key syndrome characteristics includes: determining the disease nature type of the patients with glycolipid metabolism disease according to the key syndrome characteristics; identifying the symptom manifestations of the patients with glycolipid metabolism disease, and analyzing the disease nature of the patients with glycolipid metabolism disease based on the symptom manifestations; locating the lesion location of the patients with glycolipid metabolism disease according to the disease nature and the disease nature type; combining the lesion location, the disease nature, and the disease nature type to identify the pathogenic factors of the patients with glycolipid metabolism disease; analyzing the disease development trend of the patients with glycolipid metabolism disease based on the pathogenic factors, the lesion location, and the disease nature type; and identifying the syndrome types of the patients with glycolipid metabolism disease according to the disease nature type, the lesion location, the pathogenic factors, and the disease development trend.

[0031] Among them, the disease nature type refers to the classification of the pathological nature of a disease, mainly including basic types such as cold, heat, deficiency, and excess. The symptom manifestations refer to various subjective feelings and objective signs shown by the patient during the disease process, such as dizziness, fatigue, dry mouth, etc. The disease condition nature refers to the characteristics and attributes of the symptoms shown by the patient. For example, fever, thirst, and constipation are mostly heat syndromes; fear of cold, loose stools, and fatigue are mostly deficiency syndromes. The lesion location refers to the zang-fu organs, meridians, or body parts where the disease is located. The pathogenic factors refer to the causes that lead to the occurrence of the disease, including exogenous pathogenic factors (such as the six exogenous pathogens) and endogenous pathogenic factors (such as emotions). The disease development trend refers to the dynamic change trend of the disease at the current stage, including the severity of the condition, the struggle between healthy qi and pathogenic factors, and the progress of pathogenic factors.

[0032] Optionally, based on the symptom manifestations, the analysis of the disease condition nature of the patient with glycolipid metabolism disease can be achieved through the method of combining the four diagnostic methods. For example, by observing the tongue manifestation, listening to the voice, asking about symptoms, and feeling the pulse to analyze the disease condition nature of the patient with glycolipid metabolism disease. Based on the disease condition nature and the disease nature type, the location of the lesion of the patient with glycolipid metabolism disease can be determined by using traditional Chinese medicine syndrome differentiation. Based on the pathogenic factors, the lesion location, and the disease nature type, the analysis of the disease development trend of the patient with glycolipid metabolism disease can be determined through an autoregressive integrated moving average model.

[0033] Furthermore, in the embodiment of the present invention, by extracting the pathological products of the patient with glycolipid metabolism disease based on the syndrome type, it can reflect the severity of the condition and the change of the disease trend, and help guide the treatment plan of the patient. The pathological products refer to abnormal substances produced in the human body during the disease process due to the disorder of zang-fu organ functions and the abnormal metabolism of qi, blood, and body fluids, including phlegm-fluid retention, static blood, turbid pathogenic factors, etc.

[0034] As an embodiment of the present invention, the extraction of the pathological products of the patient with glycolipid metabolism disease based on the syndrome type includes: identifying the pathogenic factors of the patient with glycolipid metabolism disease based on the syndrome type; determining the abnormal body surface parts of the patient with glycolipid metabolism disease according to the pathogenic factors; analyzing the pathological manifestations of the patient with glycolipid metabolism disease based on the pathogenic factors and the abnormal body surface parts; identifying the zang-fu organ function states of the patient with glycolipid metabolism disease under the pathological manifestations; identifying the qi, blood, and body fluid metabolism conditions of the patient with glycolipid metabolism disease according to the zang-fu organ function states; and extracting the pathological products of the patient with glycolipid metabolism disease based on the zang-fu organ function states and the qi, blood, and body fluid metabolism conditions.

[0035] Among them, the pathogenic factor refers to the initial cause leading to the occurrence of glycolipid metabolism disease, such as damp pathogen, heat pathogen, etc. The body surface abnormal change site refers to the specific site affected by the pathogenic factor, including the zang-fu organs, meridians or specific regions on the body surface, such as the spleen and stomach, kidneys, etc. The pathological manifestation refers to the symptoms and signs shown by the patient after the pathogenic factor acts on the body surface abnormal change site, such as dizziness, chest tightness, etc. The zang-fu organ function state refers to the performance and operation of the patient's zang-fu organ functions under the pathological manifestation, such as the disorder of zang-fu organ functions. The qi, blood and body fluid metabolism condition refers to the state of the processes of generation, distribution, transformation and excretion of qi, blood and body fluid in the human body, such as the disorder of body fluid metabolism, the obstruction of blood circulation, etc.

[0036] Optionally, the identification of the zang-fu organ function state of the glycolipid metabolism disease patient under the pathological manifestation can be determined by feeling the patient's pulse. According to the zang-fu organ function state, the identification of the qi, blood and body fluid metabolism condition of the glycolipid metabolism disease patient can be realized by observing the external manifestations such as the patient's complexion and tongue manifestation.

[0037] In the embodiment of the present invention, by combining the syndrome type and the pathological product, a risk identification model for the glycolipid metabolism disease patient is constructed, which can more comprehensively reflect the pathological state of the glycolipid metabolism disease patient, provide richer information for the risk identification model, and thus realize personalized risk prediction. The risk identification model refers to a prediction tool constructed based on traditional Chinese medicine theory and modern data analysis technology for evaluating the occurrence or disease progression of a patient.

[0038] As an embodiment of the present invention, the construction of the risk identification model for the glycolipid metabolism disease patient by combining the syndrome type and the pathological product includes: collecting the traditional Chinese medicine four diagnostic data of the glycolipid metabolism disease patient by combining the syndrome type and the pathological product; extracting the disease evaluation indexes of the glycolipid metabolism disease patient from the traditional Chinese medicine four diagnostic data; identifying the type of glycolipid metabolism disease of the glycolipid metabolism disease patient by using the disease evaluation indexes, and extracting the independent risk factors corresponding to the type of glycolipid metabolism disease; determining the demand model category of the glycolipid metabolism disease patient according to the independent risk factors; and constructing the risk identification model for the glycolipid metabolism disease patient based on the independent risk factors and the demand model category.

[0039] Among them, the traditional Chinese medicine four diagnostic data refers to the patient information collected by the methods of inspection, auscultation and olfaction, inquiry and palpation in traditional Chinese medicine. The disease evaluation index refers to the quantitative index extracted from the traditional Chinese medicine four diagnostic data for evaluating the patient's condition, such as the color of the tongue coating, the strength of the pulse condition, etc. The type of glycolipid metabolism disease refers to the specific classification of the glycolipid metabolism disease identified according to the disease evaluation indexes, such as hyperlipidemia. The independent risk factor refers to the factor closely related to the occurrence and development of the glycolipid metabolism disease, such as hyperglycemia, obesity, etc. The demand model category refers to the model category determined according to the independent risk factors.

[0040] Optionally, the extraction of the disease evaluation indexes of the patient with glycolipid metabolism disease can be realized by using the principal component analysis method. The extraction of the independent risk factors corresponding to the type of glycolipid metabolism disease can be determined by a multi-factor regression model. Based on the independent risk factors and the category of the demand model, the construction of the risk identification model for the patient with glycolipid metabolism disease can be realized by using the Cox regression model.

[0041] S3. Extract the distribution characteristics and manifestation types of the pathological products. According to the distribution characteristics and the manifestation types, identify the disease location syndrome elements and disease nature elements of the patient with glycolipid metabolism disease. Based on the disease location syndrome elements and the disease nature elements, analyze the course distribution of the patient with glycolipid metabolism disease.

[0042] In the embodiment of the present invention, by extracting the distribution characteristics and manifestation types of the pathological products, the specific location and nature of the patient's lesion can be clarified. The distribution characteristics refer to the specific location and influence range of the pathological products in the human body. For example, phlegm-dampness obstruction may be manifested as thick and greasy tongue coating and limb heaviness, and its influence range includes the tongue body and limbs. The manifestation types refer to the specific symptoms and signs shown after the pathological products are formed in the human body, such as tongue image changes, pulse condition changes, etc.

[0043] Optionally, the extraction of the distribution characteristics and manifestation types of the pathological products can be determined by asking the patient about specific symptoms. For example, dizziness and vertigo may indicate phlegm-dampness and occur in the chest.

[0044] Furthermore, in the embodiment of the present invention, by identifying the disease location syndrome elements and disease nature elements of the patient with glycolipid metabolism disease according to the distribution characteristics and the manifestation types, a comprehensive condition analysis can be provided for Chinese medicine practitioners, so as to achieve precise treatment of the patient with glycolipid metabolism disease. The disease location syndrome elements refer to the specific location where the lesion of the patient with glycolipid metabolism disease occurs. The disease nature elements refer to the core pathological characteristics and mechanisms shown in the pathophysiological process of glycolipid metabolism diseases, such as the confusion of clear and turbid, blood stasis blockage, etc.

[0045] As an embodiment of the present invention, identifying the disease location syndrome element and disease nature element of the glycolipid metabolism disorder patient according to the distribution characteristics and the manifestation type includes: extracting the pathological products corresponding to the distribution characteristics and the manifestation type; identifying the disease development condition of the glycolipid metabolism disorder patient according to the pathological products; analyzing the disease manifestations of the glycolipid metabolism disorder patient based on the disease development condition; extracting the associated organs of the glycolipid metabolism disorder patient according to the disease manifestations and analyzing the degree of disorder of the associated organs; determining the lesion site of the glycolipid metabolism disorder patient based on the degree of disorder and the disease manifestations; extracting the pathological characteristics of the glycolipid metabolism disorder patient according to the lesion site and the pathological products; and identifying the disease location syndrome element and disease nature element of the glycolipid metabolism disorder patient by combining the lesion site and the pathological characteristics.

[0046] Among them, the disease development condition refers to the stage and severity of glycolipid metabolism disorder in the patient's body, the disease manifestations refer to the specific symptoms and signs shown by the patient during the disease process, such as heavy limbs, the associated organs refer to the zang-fu organs closely related to the occurrence and development of the disease, such as the kidney, liver and gallbladder, etc., the degree of disorder refers to the severity of the dysfunction of the associated organs, the lesion site refers to the main zang-fu organ or tissue with lesions, such as blood vessels, and the pathological characteristics refer to the specific pathological states of the lesion site, such as yin deficiency and internal heat, damp-heat accumulation in the interior, etc.

[0047] Optionally, the identification of the disease development condition of the glycolipid metabolism disorder patient according to the pathological products can be determined by understanding the pulse condition changes of the patient through pulse diagnosis. The analysis of the disease manifestations of the glycolipid metabolism disorder patient based on the disease development condition can be achieved by observing the patient's complexion through inspection. For example, sallow complexion may indicate spleen deficiency. The analysis of the degree of disorder of the associated organs can be obtained through a symptom scoring method. For example, occasional fatigue and mild dry mouth, and no obvious abnormalities in tongue manifestation and pulse condition are regarded as mild disorder.

[0048] By analyzing the course distribution of the glycolipid metabolism disorder patient based on the disease location syndrome element and the disease nature element, the embodiment of the present invention can clarify the TCM syndrome characteristics and pathological mechanisms at different stages, help predict the occurrence risk of glycolipid metabolism disorder, and achieve early intervention. The course distribution refers to the manifestations and changes of the disease at different stages. For example, patients in the prodromal stage are prone to manifestations such as fatigue, frequent sighing, and emotional depression.

[0049] As an embodiment of the present invention, analyzing the course distribution of the glycolipid metabolism disease patients based on the disease location syndrome elements and the disease nature elements includes: collecting the historical medical record data of the glycolipid metabolism disease patients based on the disease location syndrome elements and the disease nature elements; extracting the initial symptoms and current symptoms of the glycolipid metabolism disease patients from the historical medical record data; identifying the symptom change trajectory of the glycolipid metabolism disease patients according to the initial symptoms and the current symptoms; identifying the symptom turning points of the glycolipid metabolism disease patients under the symptom change trajectory; defining the classified course stages of the glycolipid metabolism disease patients based on the symptom turning points; constructing the probability distribution diagram of the disease nature elements according to the symptom change trajectory and the historical medical record data; identifying the key disease nature elements of the glycolipid metabolism disease patients based on the probability distribution diagram; and analyzing the course distribution of the glycolipid metabolism disease patients according to the key disease nature elements, the classified course stages and the symptom change trajectory.

[0050] Among them, the historical medical record data refers to the four diagnostic records accumulated by the patients during previous medical treatments, including inspection records, auscultation and olfaction records, interrogation records, palpation records, etc. The initial symptoms refer to the symptoms shown by the patients during the first medical treatment or in the early stage of the disease. For example, the symptoms of the patient at the first visit are fatigue, loss of appetite, etc. The current symptoms refer to the symptoms shown by the patients during the most recent medical treatment or follow-up. The symptom change trajectory refers to the evolution process of the patients from the initial symptoms to the current symptoms. The symptom turning points refer to the key change points that appear in the symptom change trajectory. For example, the patient suddenly shows symptoms such as chest tightness and palpitation during a certain follow-up. The classified course stages refer to dividing the course of the patients into different stages according to the symptom turning points and the symptom change trajectory, including the latent period, the prodromal period, the obvious symptom period, the recovery period, etc. The probability distribution diagram refers to the probability distribution of the disease nature elements appearing in different stages constructed based on the symptom change trajectory and the historical medical record data. The key disease nature elements refer to the disease nature characteristics that play a leading role in the development of the disease in different stages of the disease. For example, in the early stage of the glycolipid metabolism disease, phlegm-dampness may be the key disease nature element, while in the late stage, blood stasis may be the key disease nature element.

[0051] Optionally, the identification of the symptom turning points of the glycolipid metabolism disease patients under the symptom change trajectory can be realized by using the latent class trajectory analysis model. The construction of the probability distribution diagram of the disease nature elements according to the symptom change trajectory and the historical medical record data can be achieved by drawing a histogram. The identification of the key disease nature elements of the glycolipid metabolism disease patients based on the probability distribution diagram can be determined by using the probability density curve in the probability distribution diagram.

[0052] S4. Determine the discrimination ability of the risk identification model according to the course distribution, and perform model optimization processing on the risk identification model based on the discrimination ability to obtain an optimized model.

[0053] In the embodiment of the present invention, by determining the discrimination ability of the risk identification model according to the course distribution, the ability of the model to interpret data and prediction accuracy can be evaluated, so as to more accurately identify the risk of glycolipid metabolism disease. The discrimination ability refers to the ability of the risk identification model to interpret and predict the course distribution.

[0054] As an embodiment of the present invention, the determining the discrimination ability of the risk identification model according to the course distribution includes: collecting the disease distribution data corresponding to the course distribution, and extracting the characteristic factors from the disease distribution data; setting the input features of the risk identification model based on the characteristic factors; determining the target variable of the risk identification model according to the input features; splitting the course distribution data into a training set and a test set of the risk identification model by using the input features and the target variable; training the risk identification model by using the training set to obtain a trained model; performing data analysis processing on the test set based on the trained model to generate model analysis data; calculating the data accuracy rate of the model analysis data; and determining the discrimination ability of the risk identification model according to the data accuracy rate.

[0055] Among them, the disease distribution data refers to the distribution information data of the traditional Chinese medicine syndromes, symptoms, signs, etc. of patients at different disease course stages, including the main complaint symptoms, tongue manifestations, pulse conditions, traditional Chinese medicine syndrome types, etc. of patients. The characteristic factors refer to various factors related to the incidence risk of glycolipid metabolism disease extracted from the disease distribution data, such as older age, irregular lifestyle, etc. The input features refer to the input variables of the risk identification model after the characteristic factors are screened and processed. The target variable refers to the result variable that the risk identification model needs to predict, such as whether the patient gets sick and the risk degree of getting sick. The training set refers to a part of the data divided from the disease distribution data for training the risk identification model. The test set refers to another part of the data divided from the disease distribution data for testing and verifying the trained risk identification model. The trained model refers to the model obtained after training the risk identification model by using the training set. The model analysis data refers to the prediction result data generated by the model when the test set data is input into the trained model. The data accuracy rate refers to the proportion of the number of correctly predicted samples in the model analysis data to the total number of samples.

[0056] Optionally, the determination of the target variable of the risk identification model based on the input features can be implemented using scikit-learn tools. The splitting of the disease course distribution data into the training set and test set of the risk identification model using the input features and the target variable can be determined by the simple random splitting method. The calculation of the data accuracy rate of the model analysis data can be implemented using the recall rate formula.

[0057] Furthermore, in the embodiments of the present invention, by performing model optimization processing on the risk identification model based on the discrimination ability to obtain an optimized model, the prediction accuracy of the model for risks can be improved, thereby more effectively identifying high-risk individuals or events. At the same time, the interpretability of the model can also be improved, making its decision-making process more transparent, facilitating understanding and trust. The optimized model refers to an enhanced model obtained by performing a series of improvements and adjustments on the existing risk identification model.

[0058] Optionally, the model optimization processing of the risk identification model can be achieved through hyperparameter optimization methods, such as Bayesian optimization method.

[0059] S5. When the optimization processing result meets the preset effect, use the optimized model to identify the risk factors of the patient with glycolipid metabolism disease. According to the risk factors, output the feedback result of the patient with glycolipid metabolism disease, and set the result interpretation of the feedback result.

[0060] In the embodiments of the present invention, by using the optimized model to identify the risk factors of the patient with glycolipid metabolism disease when the optimization processing result meets the preset effect, the risk patients of glycolipid metabolism disease can be identified early or the risk of complications occurring in patients with glycolipid metabolism disease can be predicted, and truncated intervention can be taken in a timely manner. The risk factors refer to the elements that can reflect the essence of the disease onset and have a warning effect on the occurrence of the disease, such as blood stasis, phlegm turbidity, etc.

[0061] Optionally, the identification of the risk factors of the patient with glycolipid metabolism disease can be determined by calculating the contribution degree of each risk factor to the model prediction.

[0062] Furthermore, in the embodiments of the present invention, by outputting the feedback result of the patient with glycolipid metabolism disease according to the risk factors and setting the result interpretation of the feedback result, the possible complications of the patient with glycolipid metabolism disease can be pre-warned, helping the patient better understand their own condition and risk factors, thereby improving the patient's self-management ability. The feedback result refers to the prediction result generated by the risk identification model based on the input characteristic factors, such as predicting the possible traditional Chinese medicine syndrome types that the patient may have in the future. The result interpretation refers to the detailed explanation and description of the feedback result, such as explaining that the manifestations of qi deficiency syndrome are fatigue, shortness of breath, pale tongue and weak pulse, etc.

[0063] As an embodiment of the present invention, outputting the feedback result of the patient with glycolipid metabolism disease according to the risk factors includes: analyzing the disease type of the patient with glycolipid metabolism disease according to the risk factors; defining the health evaluation indicators of the patient with glycolipid metabolism disease based on the disease type; calculating the risk score of the patient with glycolipid metabolism disease according to the health evaluation indicators; identifying the risk level of the patient with glycolipid metabolism disease according to the risk score; setting personalized guidance suggestions for the patient with glycolipid metabolism disease based on the risk classification and the health evaluation indicators; and outputting the feedback result of the patient with glycolipid metabolism disease by combining the disease type, the risk level and the personalized guidance suggestions.

[0064] Among them, the disease type refers to the classification of the condition of the patient with glycolipid metabolism disease obtained according to the principles of traditional Chinese medicine syndrome differentiation and treatment, combined with information such as the patient's symptoms, signs, tongue manifestation, and pulse condition. The health evaluation indicators refer to a series of indicators used to evaluate the health status of the patient with glycolipid metabolism disease, such as tongue manifestation, pulse condition, etc. The risk score refers to the score calculated by a certain algorithm or model according to the patient's health evaluation indicators to indicate the likelihood of the patient developing a disease or complication. The risk level refers to dividing the patient into different risk categories according to the risk score result. The personalized guidance suggestions refer to targeted treatment and lifestyle adjustment suggestions provided for the patient according to the patient's disease type, risk level and health evaluation indicators, such as traditional Chinese medicine prescriptions, acupuncture, massage, etc.

[0065] Optionally, the analysis of the disease type of the patient with glycolipid metabolism disease according to the risk factors can be achieved through a machine learning classification algorithm, such as a neural network algorithm. The identification of the risk level of the patient with glycolipid metabolism disease according to the risk score can be divided according to the range of the risk score. For example, low risk: risk score < 30, medium risk: 30 risk score < 70, high risk: risk score 70.

[0066] In an optional embodiment of the present invention, according to the health evaluation indicators, the risk score of the patient with glycolipid metabolism disease is calculated using the following formula: Among them, S represents the risk score of the patient with glycolipid metabolism disease, represents the original value of the jth health evaluation indicator, represents the mean value of the jth health evaluation indicator, represents the standard deviation of the jth health evaluation indicator, n represents the number of health evaluation indicators, and j represents the index number of the health evaluation indicator.

[0067] S6. Combine the risk identification model, the optimization model, and the result interpretation to perform model construction processing for the patients with glycolipid metabolism diseases, and obtain the model construction result.

[0068] In the embodiment of the present invention, by combining the risk identification model, the optimization model, and the result interpretation, performing the model construction processing for the patients with glycolipid metabolism diseases, and obtaining the model construction result, it is possible to clarify which factors are most critical for risk prediction, significantly improve the accuracy and generalization ability of the model, enable it to work effectively in different patient groups, improve the acceptability and practicality of the model, help traditional Chinese medicine identify high-risk populations, intervene in advance, and prevent the occurrence of diseases. At the same time, it can be used as a health management tool to provide personalized guidance and suggestions for each patient, helping patients better manage their health status. The model construction processing refers to the entire process from data collection, preprocessing, feature engineering, model selection, training to optimization. The model construction result refers to the final output obtained after the model training and optimization are completed, including the performance indicators and prediction results of the model.

[0069] Compared with the problems described in the background art, in the embodiments of the present invention, by extracting the key syndrome characteristics of the patients with glycolipid metabolism diseases according to the tongue image, the pulse condition signal and the inquiry record, comprehensive input data can be provided for the prediction model, which helps to improve the accuracy and reliability of the model. Further, in the embodiments of the present invention, by identifying the syndrome types of the patients with glycolipid metabolism diseases according to the key syndrome characteristics, and extracting the pathological products of the patients with glycolipid metabolism diseases based on the syndrome types, a risk identification model for the patients with glycolipid metabolism diseases is constructed, which can more comprehensively reflect the pathological state of the patients with glycolipid metabolism diseases, provide richer information for the risk identification model, and thus realize personalized risk prediction. In the embodiments of the present invention, by extracting the distribution characteristics and manifestation types of the pathological products, the specific location and nature of the patient's lesions can be clarified, which helps to analyze the course distribution of the patients with glycolipid metabolism diseases, clarify the TCM syndrome characteristics and pathological mechanisms at different stages, help to predict the occurrence risk of glycolipid metabolism diseases, and realize early intervention. Further, in the embodiments of the present invention, by determining the discrimination ability of the risk identification model according to the course distribution, the interpretability of the model for data and the prediction accuracy can be evaluated, so as to more accurately identify the risk of glycolipid metabolism diseases. Further, in the embodiments of the present invention, by outputting the feedback results of the patients with glycolipid metabolism diseases according to the risk factors and setting the result interpretations of the feedback results, possible complications of the patients with glycolipid metabolism diseases can be warned in advance, which helps the patients to better understand their own conditions and risk factors, and thus improves the self-management ability of the patients. In the embodiments of the present invention, by combining the risk identification model, the optimization model and the result interpretations, the model construction process of the patients with glycolipid metabolism diseases is executed, and the model construction results are obtained, which can clarify which factors are most critical for risk prediction, significantly improve the accuracy and generalization ability of the model, enable it to work effectively in different patient groups, improve the acceptability and practicality of the model, help TCM doctors to identify high-risk populations, intervene in advance to prevent the occurrence of diseases, and at the same time can be used as a health management tool to provide personalized guidance suggestions for each patient to help the patients better manage their health status. Therefore, a method for constructing a model for predicting glycolipid metabolism diseases based on artificial intelligence provided by the embodiments of the present invention can improve the accuracy and reliability of glycolipid metabolism disease prediction.

[0070] Embodiment 2: As Figure 2 shown, it is a functional module diagram of a system for constructing a model for predicting glycolipid metabolism diseases based on artificial intelligence according to the present invention.

[0071] The construction system 200 of an artificial intelligence-based prediction model for glycolipid metabolism diseases according to the present invention can be installed in an electronic device. According to the functions achieved, the construction system of an artificial intelligence-based prediction model for glycolipid metabolism diseases may include a data collection module 201, an initial model construction module 202, a disease analysis module 203, an initial model optimization module 204, a model result interpretation module 205, and a final model generation module 206. The modules in the present invention may also be referred to as units, which refer to a series of computer program segments that can be executed by a processor of an electronic device and can complete fixed functions, and are stored in the memory of the electronic device.

[0072] In the embodiments of the present invention, the functions of each module / unit are as follows: The data collection module 201 is used to select patients with glycolipid metabolism diseases, collect tongue image, pulse signal, and medical history records of the patients with glycolipid metabolism diseases, and extract key syndrome characteristics of the patients with glycolipid metabolism diseases according to the tongue image, the pulse signal, and the medical history records. The initial model construction module 202 is used to identify the syndrome types of the patients with glycolipid metabolism diseases according to the key syndrome characteristics, extract pathological products of the patients with glycolipid metabolism diseases based on the syndrome types, and construct a risk identification model for the patients with glycolipid metabolism diseases by combining the syndrome types and the pathological products.

[0073] The disease analysis module 203 is used to extract the distribution characteristics and manifestation types of the pathological products, identify the disease location syndrome elements and disease nature elements of the patients with glycolipid metabolism diseases according to the distribution characteristics and the manifestation types, and analyze the course distribution of the patients with glycolipid metabolism diseases based on the disease location syndrome elements and the disease nature elements. The initial model optimization module 204 is used to determine the discrimination ability of the risk identification model according to the course distribution, and perform model optimization processing on the risk identification model based on the discrimination ability to obtain an optimized model. The model result interpretation module 205 is used to, when the optimization result meets the preset effect, use the optimized model to identify the risk factors of the patients with glycolipid metabolism diseases, output the feedback results of the patients with glycolipid metabolism diseases according to the risk factors, and set the result interpretation of the feedback results. The final model generation module 206 is used to perform model construction processing on the patients with glycolipid metabolism diseases by combining the risk identification model, the optimized model, and the result interpretation to obtain a model construction result.

[0074] Specifically, each module in the construction system 200 of an artificial intelligence-based prediction model for glycolipid metabolism diseases in the embodiments of the present invention adopts the same as the above Figure 1The same technical means as the method for constructing a glycolipid metabolism disease prediction model based on artificial intelligence described in [reference] can be used, and the same technical effects can be achieved, which will not be elaborated here.

[0075] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention.

[0076] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for constructing a model for predicting glucose and lipid metabolic diseases based on artificial intelligence, characterized in that: The method comprises: Selecting patients with glucose and lipid metabolism diseases, collecting tongue images, pulse signals and medical records of the patients with glucose and lipid metabolism diseases, and extracting key syndrome characteristics of the patients with glucose and lipid metabolism diseases based on the tongue images, pulse signals and medical records; According to the key syndrome characteristics, the syndrome type of the patient with the glucolipid metabolism disease is identified, based on the syndrome type, the pathological products of the patient with the glucolipid metabolism disease are extracted, and the risk identification model of the patient with the glucolipid metabolism disease is constructed by combining the syndrome type and the pathological products; Extracting the distribution characteristics and manifestation types of the pathological products, identifying the disease location syndrome factors and disease characteristics elements of the patients with the glucose and lipid metabolism diseases according to the distribution characteristics and the manifestation types, and analyzing the disease course distribution of the patients with the glucose and lipid metabolism diseases based on the disease location syndrome factors and the disease characteristics elements; Determining the distinguishing ability of the risk identification model according to the disease course distribution, and performing model optimization processing of the risk identification model based on the distinguishing ability to obtain an optimized model; When the optimization processing result meets the preset effect, the risk factors of the patient with the glucolipid metabolism disease are identified by using the optimization model, and according to the risk factors, feedback results of the patient with the glucolipid metabolism disease are output, and a result interpretation of the feedback result is set; In combination with the risk identification model, the optimization model and the result interpretation, the model construction process of the patients with glucose and lipid metabolism diseases is performed to obtain a model construction result.

2. The method for constructing a model for predicting glucose and lipid metabolism diseases based on artificial intelligence according to claim 1, characterized in that: Extracting key syndrome characteristics of the patient with glucose and lipid metabolism disease based on the tongue image, the pulse signal and the medical consultation record includes: Extracting tongue features of the patient with glucose and lipid metabolism disease according to the tongue image; Based on the tongue image characteristics, analyzing the changing patterns of the tongue images of the patients with glucose and lipid metabolism diseases; According to the pulse signal, identifying the pulse fluctuation degree and pulse rate of the patient with glucose and lipid metabolism disease; Determining the pulse syndrome type of the patient with glucose and lipid metabolism disease based on the pulse floating and sinking degree and the pulse velocity; Extracting the symptom and sign records of the patient with the glucose and lipid metabolism disease from the medical consultation records, and identifying the nature of the symptoms of the patient with the glucose and lipid metabolism disease according to the symptom and sign records; The key symptom characteristics of the patients with glucose and lipid metabolism diseases are extracted by combining the changing patterns of the tongue image, the pulse syndrome type and the nature of the symptoms.

3. The method for constructing a model for predicting glucose and lipid metabolism diseases based on artificial intelligence according to claim 1, characterized in that: The step of identifying the syndrome type of the patient with the glucose and lipid metabolism disease according to the key syndrome characteristics includes: Determining the disease type of the patient with the glucose and lipid metabolism disease according to the key syndrome characteristics; Identifying the symptoms of the patient with the glucose and lipid metabolism disease, and analyzing the nature of the disease of the patient with the glucose and lipid metabolism disease based on the symptoms; Locating the lesion position of the patient with the glucose and lipid metabolism disease according to the nature of the disease and the type of disease; Combining the lesion location, the nature of the symptoms and the disease type, identifying the pathogenic factors of the patient with the glucose and lipid metabolism disease; Analyze the development trend of the disease of the patient with the glucose and lipid metabolism disease based on the pathogenic factors, the location of the lesion and the type of disease; According to the disease type, the lesion location, the pathogenic factors and the development trend of the symptoms, the syndrome type of the patient with the glycolipid metabolism disease is identified.

4. The method for constructing a model for predicting glucose and lipid metabolism diseases based on artificial intelligence according to claim 1, characterized in that: The step of extracting the pathological products of the patient with glucose and lipid metabolism disease based on the syndrome type includes: Based on the syndrome type, identifying the etiological factors of the patient with the glucose and lipid metabolism disease; Determining the abnormal site on the body surface of the patient with the glucose and lipid metabolism disease according to the pathogenic factors; Analyzing the pathological manifestations of the patient with the glucose and lipid metabolism disease based on the pathogenic factors and the abnormal sites on the body surface; Identifying the functional state of the internal organs of the patient with the glucose and lipid metabolism disease under the pathological manifestations; According to the functional status of the internal organs, identifying the metabolism of qi, blood and body fluids of the patient with the glucose and lipid metabolism disease; Based on the functional status of the internal organs and the metabolism of qi, blood and body fluids, the pathological products of the patient with the glucose and lipid metabolism disease are extracted.

5. The method for constructing a model for predicting glucose and lipid metabolism diseases based on artificial intelligence according to claim 1, characterized in that: The step of combining the syndrome type and the pathological product to construct a risk identification model for patients with glucose and lipid metabolism diseases comprises: In combination with the syndrome type and the pathological product, the TCM four diagnostic data of the patient with the glucose and lipid metabolism disease are collected; Extracting the symptom evaluation index of the patient with glucose and lipid metabolism disease from the four diagnostic data of traditional Chinese medicine; Using the disease evaluation index to identify the type of glucose and lipid metabolism disease of the patient with glucose and lipid metabolism disease, and extracting independent risk factors corresponding to the type of glucose and lipid metabolism disease; Determining the demand model category of the patient with glucose and lipid metabolism disease according to the independent risk factors; Based on the independent risk factors and the demand model categories, a risk identification model for patients with glycolipid metabolism diseases is constructed.

6. The method for constructing a model for predicting glucose and lipid metabolism diseases based on artificial intelligence according to claim 1, characterized in that: The step of identifying the disease location syndrome and disease nature elements of the patient with the glucose and lipid metabolism disease according to the distribution characteristics and the manifestation type includes: extracting the pathological products corresponding to the distribution characteristics and the manifestation types; According to the pathological product, identifying the disease progression of the patient with the glucose and lipid metabolism disease; Analyze the symptoms of the patient with the glucose and lipid metabolism disease based on the disease progression; According to the symptoms, extracting the associated organs of the patient with the glucose and lipid metabolism disease, and analyzing the degree of disorder of the associated organs; Based on the degree of disorder and the symptoms, determining the lesion site of the patient with the glucose and lipid metabolism disease; Extracting the pathological characteristics of the patient with the glucose and lipid metabolism disease according to the lesion site and the pathological product; The disease location and pathological characteristics are combined to identify the disease location syndrome and disease nature elements of the patient with the glucose and lipid metabolism disease.

7. The method for constructing a model for predicting glucose and lipid metabolism diseases based on artificial intelligence according to claim 1, characterized in that: The analyzing the disease course distribution of the patient with the glucose and lipid metabolism disease based on the disease location syndrome and the disease nature elements includes: Based on the disease location syndrome and the disease nature elements, collecting historical medical data of the patient with the glucose and lipid metabolism disease; Extracting the initial symptoms and current symptoms of the patient with glucose and lipid metabolism disease from the historical medical consultation data; According to the initial symptoms and the current symptoms, identifying the symptom change trajectory of the patient with the glucose and lipid metabolism disease; Identify the symptom turning point of the patient with glucose and lipid metabolism disease under the symptom change trajectory; Based on the symptom turning point, defining the classification course stage of the patient with the glucose and lipid metabolism disease; Constructing a probability distribution diagram of the pathogenic factors according to the symptom change trajectory and the historical medical consultation data; Based on the probability distribution graph, identifying key pathogenic factors of the patient with the glucose and lipid metabolism disease; According to the key pathogenic factors, the classified disease stages and the symptom change trajectory, the disease course distribution of patients with glycolipid metabolism diseases is analyzed.

8. The method for constructing a model for predicting glucose and lipid metabolism diseases based on artificial intelligence according to claim 1, characterized in that: Determining the distinguishing ability of the risk identification model according to the disease course distribution includes: Collecting disease distribution data corresponding to the disease course distribution, and extracting characteristic factors from the disease distribution data; Based on the characteristic factors, setting input characteristics of the risk identification model; Determining a target variable of the risk identification model according to the input features; Splitting the disease course distribution data into a training set and a test set of the risk identification model using the input features and the target variable; Using the training set to train the risk identification model to obtain a training model; Based on the training model, performing data analysis processing on the test set to generate model analysis data; Calculating the data accuracy of the model analysis data; The distinguishing ability of the risk identification model is determined according to the data accuracy.

9. The method for constructing a model for predicting glucose and lipid metabolism diseases based on artificial intelligence according to claim 1, characterized in that: Outputting feedback results of the patient with glucose and lipid metabolism disease according to the risk factors includes: Analyzing the symptom type of the patient with glucose and lipid metabolism disease according to the risk factors; Based on the disease type, define the health evaluation index of the patient with glucose and lipid metabolism disease; Calculating the risk score of the patient with glucose and lipid metabolism disease according to the health evaluation index; According to the risk score, identifying the risk level of the patient with glucose and lipid metabolism disease; Based on the risk classification and the health evaluation index, providing personalized guidance suggestions for the patient with the glucose and lipid metabolism disease; In combination with the disease type, the risk level and the personalized guidance suggestions, feedback results of the patient with glucose and lipid metabolism disease are output.

Images