A diabetes data supervision method and system based on real-time collection

Abstract

The present invention provides a method and system for supervising diabetes data based on real-time collection, relating to the technical field of diabetes data analysis, including: collecting static characteristic parameters of a patient, where the static characteristic parameters include the patient's personal information and treatment information; periodically collecting dynamic characteristic parameters of the patient, where the dynamic characteristic parameters include the patient's blood glucose data; constructing a feature vector based on the static characteristic parameters and the dynamic characteristic parameters, where the feature vector includes a prediction correction feature vector and a physical sign feature vector, and the larger the value in the prediction correction feature vector and the physical sign feature vector, the greater the negative impact on diabetes; and performing a risk assessment of nephropathy lesions through a nephropathy prediction module according to the feature vector. The present invention has the advantages of being able to provide a more accurate risk assessment of nephropathy lesions under diabetes that can better adapt to individual differences.

Description

Technical Field

[0001] The present invention relates to the technical field of diabetes data analysis, and more particularly, to a method and system for monitoring diabetes data based on real-time collection. Background Art

[0002] Diabetes is a metabolic disease characterized by chronic hyperglycemia. The long-term hyperglycemic state can damage multiple organs throughout the body, and diabetic nephropathy is one of the most serious complications of diabetic patients.

[0003] Existing methods for assessing the risk of nephropathy usually perform static analysis based on single physiological indicators or data at specific time nodes, ignoring individual differences among patients and the long-term effects of dynamic blood glucose changes in diabetic patients. Blood glucose fluctuations are the key factors leading to the occurrence of diabetic complications. At the same time, in existing blood glucose risk prediction models, the patient data and features extracted are often not fully expressed and utilized. If the existing blood glucose risk prediction models are used to predict nephropathy, it is easy to ignore individual differences and lead to inaccurate assessment of nephropathy risk.

[0004] Therefore, it is necessary to further optimize the existing diabetes data monitoring scheme to achieve a more accurate assessment of the risk of nephropathy under the influence of diabetes that can better adapt to individual differences. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for monitoring diabetes data based on real-time collection, which can provide a more accurate assessment of the risk of nephropathy under the influence of diabetes that can better adapt to individual differences.

[0006] The present invention is achieved through the following technical solutions:

[0007] A method for monitoring diabetes data based on real-time collection includes the following steps:

[0008] Collect static characteristic parameters of the patient, where the static characteristic parameters include the patient's personal information and treatment information;

[0009] Periodically collect dynamic characteristic parameters of the patient, where the dynamic characteristic parameters include the patient's blood glucose data;

[0010] Construct a feature vector based on the static characteristic parameters and the dynamic characteristic parameters. The feature vector includes a prediction correction feature vector and a physical sign feature vector. The larger the value in the prediction correction feature vector and the physical sign feature vector, the greater the negative impact on diabetes;

[0011] Perform a risk assessment of nephropathy lesions through a nephropathy prediction module according to the feature vector.

[0012] Preferably, the personal information includes age, height, weight, and habitual exercise time; the treatment information includes the duration of illness.

[0013] Preferably, the prediction correction feature vector is constructed based on the static feature parameters, and the method includes:

[0014] Obtain the personal correction parameter according to the personal information :

[0015] ;

[0016] where age represents age, w represents weight in kilograms, h represents height in meters, t represents the average daily exercise duration in minutes, and are both adjustment coefficients, b and q are both adjustment biases, and the adjustment coefficients and the adjustment biases are empirical parameters;

[0017] Obtain the treatment correction parameter according to the treatment information :

[0018] ;

[0019] where represents the duration of illness in years;

[0020] Construct the prediction correction feature vector :

[0021] .

[0022] Preferably, the method for obtaining the adjustment coefficients and the adjustment biases is:

[0023] Based on big data, obtain the age, height, weight, and habitual exercise time of multiple diabetes patients;

[0024] Obtain the maximum value and the minimum value of the age among them, and the minimum value of the maximum value and the minimum value of the body mass index;

[0025] Establish equations and solve to obtain the adjustment coefficients and the adjustment biases:

[0026] .

[0027] Preferably, the physical sign feature vector is obtained through the dynamic feature parameters, and the method includes:

[0028] Collect N blood glucose values of the patient within the time period and form a blood glucose time series;

[0029] Set a hyperglycemic threshold and a hypoglycemic threshold ;

[0030] Obtain the number of parameters in the blood glucose time series that are not lower than and record the corresponding blood glucose acquisition time and form a first blood glucose record time series in time sequence;

[0031] Obtain the number of parameters in the blood glucose time series that are not higher than and record the corresponding blood glucose acquisition time and form a second blood glucose record time series in time sequence;

[0032] Obtain the coefficient of variation CV of the blood glucose time series, where the coefficient of variation is the ratio of the standard deviation to the mean;

[0033] Obtain the difference between every two adjacent values in the first blood glucose record time series and take the minimum value

[0034] ; Obtain the difference between every two adjacent values in the second blood glucose record time series and take the minimum value

[0035] ; Obtain the physical sign feature vector

[0036] :

[0037] .

[0038] Preferably, the nephropathy prediction module includes:

[0039] An input layer for inputting the feature vector;

[0040] A feature fusion layer for performing multimodal fusion on the feature vector to obtain a fused feature matrix;

[0040] An output layer for outputting a prediction result of whether nephropathy lesions have occurred according to the fused feature matrix.

[0041] Preferably, the method for performing multimodal fusion on the feature vector is:

[0042] Perform a linear transformation on the prediction correction feature vector and the physical sign feature vector :

[0043] ;

[0044] ;

[0045] , ;

[0046] , ;

[0047] ;

[0048] Among them, and are respectively after linear transformation and , and are linear transformation weights, and are linear transformation biases, is the target dimension of linearization, is 's dimension, is 's dimension, max is the maximum value function, and respectively represent real number matrices with dimensions of and , represents a one-dimensional vector with dimension d;

[0049] Obtain the concatenated feature vector S:

[0050] ;

[0051] Obtain the attention weights :

[0052] ;

[0053] Among them, represents the softmax function, is the attention weight matrix, is the attention bias;

[0054] Obtain the fused feature matrix :

[0055] ;

[0056] Among them, is the dot product operator of matrices.

[0057] Preferably, the method for outputting the prediction result of whether there is kidney disease lesion according to the fused feature vector is:

[0058] Perform a linear transformation on the fusion feature matrix to obtain intermediate parameters :

[0059] ;

[0060] wherein is the transformation weight and the dimension is , is the transformation bias and is a constant;

[0061] Obtain the probability of nephropathy by sigmoid :

[0062] ;

[0063] wherein, e is the natural constant.

[0064] Preferably, when predicting the nephropathy prediction module, the label used is determined by the glomerular filtration rate. When the glomerular filtration rate is lower than the preset threshold, the label is nephropathy, otherwise the label is no nephropathy.

[0065] A diabetes data supervision system based on real-time collection, which is applied to the above-mentioned diabetes data supervision method based on real-time collection, includes:

[0066] A static collection module for collecting static characteristic parameters of a patient, where the static characteristic parameters include the personal information and treatment information of the patient;

[0067] A dynamic collection module for periodically collecting dynamic characteristic parameters of a patient, where the dynamic characteristic parameters include the blood glucose data of the patient;

[0068] A feature construction module for constructing a feature vector based on the static characteristic parameters and the dynamic characteristic parameters. The feature vector includes a prediction correction feature vector and a physical sign feature vector. The larger the value in the prediction correction feature vector and the physical sign feature vector, the greater the negative impact on diabetes.

[0069] An evaluation module for evaluating the risk of nephropathy through the nephropathy prediction module according to the feature vector.

[0070] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0071] The present invention collects the static characteristic parameters and dynamic characteristic parameters of diabetic patients, and constructs a comprehensive feature vector based on various types of data, which helps to more comprehensively reflect the individual health status and disease risk factors;

[0072] The data selected and features constructed in the present invention include both the blood glucose fluctuation situation, the peaks and valleys of blood glucose peaks, and the impacts of individual exercise, obesity, and age, enhancing the comprehensiveness, individual pertinence, and accuracy of feature expression;

[0073] The present invention constructs static features and dynamic features into a prediction correction feature vector and a physical sign feature vector respectively, enabling the quantification of both the blood glucose data features of the patient himself and the degree of influence of individual differences on risks, and further enhancing the reliability of subsequent prediction applied to the model;

[0074] The present invention analyzes the fused feature vector through a nephropathy prediction module, can comprehensively consider the long-term blood glucose control situation and the differences in individual constitutions and living habits, and more accurately output the predicted nephropathy lesion risk situation to assist clinicians in early screening, early diagnosis, and personalized intervention;

[0075] The present invention is reasonably designed, applicable to different types of diabetic patients, can overcome individual differences for targeted data processing, and has a wider applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 is a schematic flowchart of a diabetes data supervision method based on real-time collection provided by Embodiment 1 of the present invention;

[0077] Figure 2 is a schematic principle diagram of a diabetes data supervision system based on real-time collection provided by Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0078] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0079] Embodiment 1

[0080] This embodiment provides a diabetes data supervision method based on real-time collection. Refer to Figure 1 , and includes the following steps:

[0081] Collect static feature parameters of the patient. The static feature parameters include the patient's personal information and treatment information;

[0082] Periodically collect dynamic feature parameters of the patient. The dynamic feature parameters include the patient's blood glucose data;

[0083] Construct a feature vector based on static feature parameters and dynamic feature parameters. The feature vector includes a prediction correction feature vector and a physical sign feature vector. The larger the value in the prediction correction feature vector and the physical sign feature vector, the greater the negative impact on diabetes.

[0084] According to the feature vector, perform a risk assessment of kidney disease lesions through a kidney disease prediction module.

[0085] As a preferred solution of this embodiment, the personal information includes age, height, weight, and habitual exercise time; the treatment information includes the duration of illness.

[0086] In this embodiment, static feature parameters and dynamic feature parameters of diabetic patients are collected. By integrating multi-dimensional and multi-type individual health data, a more representative and comprehensive feature vector is constructed, enabling the kidney disease prediction module in subsequent steps to not only focus on the abnormalities of single indicators when assessing the risk of kidney disease, but also comprehensively analyze the patient's long-term physiological state and disease development trend. In the specific design, the static feature parameters mainly cover the patient's personal information including age, height, weight, and habitual exercise time, and the treatment information mainly includes the duration of illness. The above parameters are usually relatively stable in the short term. The dynamic feature parameters are the blood glucose data continuously recorded by the patient within a certain period, and the blood glucose data usually has certain short-term fluctuations. Through the combination of several types of data in this embodiment, more representative features can be extracted more accurately for each patient with different situations.

[0087] That is to say, based on the solution of this embodiment, the static feature parameters can express the impact of the personal situation of diabetic patients on kidney disease lesions, and the static feature parameters are to detect blood glucose data in the traditional way. Usually, the situation of kidney disease lesions can be predicted through the characteristics of blood glucose data, and in this embodiment, by additionally obtaining dynamic feature parameters, the prediction results obtained only relying on blood glucose data can be further corrected and optimized.

[0088] In summary, based on the joint establishment of features from multi-source data set in this embodiment and prediction through features, it helps to comprehensively reflect the overall health status and potential risk factors of kidney disease of diabetic patients in different situations, and provides a solid data basis for the accurate prediction and management of diseases more specifically and reliably.

[0089] In this embodiment, based on the static feature parameters, the prediction correction feature vector is constructed. The method includes:

[0090] Obtain personal correction parameters according to the personal information :

[0091] ;

[0092] Among them, age represents age, w represents weight in kilograms, h represents height in meters, and t represents the average daily exercise duration in minutes. and are respectively adjustment coefficients, and b and q are respectively adjustment biases. The adjustment coefficients and the adjustment biases are empirical parameters.

[0093] Obtain the treatment correction parameter according to the treatment information :

[0094] ;

[0095] Among them, represents the duration of illness in years;

[0096] Construct the prediction correction feature vector :

[0097] .

[0098] Furthermore, the method for obtaining the adjustment coefficients and the adjustment biases is as follows:

[0099] Based on big data, obtain the ages, heights, weights, and habitual exercise times of multiple diabetic patients.

[0100] Obtain the maximum value and the minimum value of the ages among them, the maximum value and the minimum value of the average daily exercise durations, the maximum value and the minimum value of the body mass indices;

[0101] Establish equations and solve to obtain the adjustment coefficients and the adjustment biases:

[0102] .

[0103] Among the static characteristic parameters selected in this embodiment: as age increases, the basic function of the kidneys gradually declines (such as a decrease in glomerular filtration rate GFR, an increase in glomerulosclerosis, etc.), and the ability of the kidneys of diabetic patients to withstand the damage caused by persistent hyperglycemia weakens; while abnormal BMI (such as obesity with too high BMI) is an independent risk factor for diabetic nephropathy. For example, overweight or obesity will lead to increased insulin resistance, and the state of persistent hyperglycemia increases glomerular hyperfiltration and pressure burden, and is also likely to induce kidney disease lesions; finally, regular exercise helps to improve insulin sensitivity, stabilize blood sugar levels, etc., and is a favorable factor for preventing diabetic nephropathy lesions; the longer the disease duration, that is, the longer the course of the disease, the greater the risk of kidney disease lesions. Therefore, when constructing the prediction correction feature vector based on the static characteristic parameters in this embodiment, the personal correction parameters obtained according to the personal information are mainly that the value of the personal correction parameter will be larger when the age is older, the MBI is larger, and the exercise is less. The settings of the adjustment coefficient and the adjustment bias are to standardize different parameters into a range, so that the obtained based on age, obtained based on MBI, and obtained based on the exercise situation

[0104] have as similar value ranges as possible. Such a design helps to eliminate the dimension difference, avoid feature dominance, and also helps to improve the feature fusion effect subsequently. And the obtained treatment correction parameter is that the longer the disease duration, the larger the value.

[0104] On the other hand, the reason for dividing the feature extraction of the static characteristic parameters into treatment correction parameters and personal correction parameters in this embodiment is that the personal correction parameters mainly describe the individual's static physical signs and behavior habits, that is, the influence on the kidneys is more due to factors such as basic risk, long-term metabolic level, and exercise regulation. And the treatment correction parameter directly reflects the degree of chronic damage caused by diabetes. That is to say, the essential incentives for the influence of these two parameters on kidney disease lesions are different, and grouping for extraction helps the subsequent model to capture features more accurately.

[0105] Next, it is to capture the features related to blood sugar fluctuations through blood sugar data, and obtain the physical sign feature vector through the dynamic characteristic parameters. The method includes:

[0106] Collect N blood sugar values of the patient within the time period and form a blood sugar time series;

[0107] Set the hyperglycemia threshold and the hypoglycemia threshold ;

[0108] Obtain the number of parameters not less than in the blood sugar time series , and record the corresponding blood sugar acquisition time and form the first blood sugar record time series in time sequence;

[0109] Obtain the number of parameters in the blood glucose time series whose values are not higher than and record the corresponding blood glucose acquisition time and form a second blood glucose record time series in time sequence;

[0110] Obtain the coefficient of variation CV of the blood glucose time series, where the coefficient of variation is the ratio of the standard deviation to the average;

[0111] Obtain the difference between every two adjacent values in the first blood glucose record time series and take the minimum value ;

[0112] Obtain the difference between every two adjacent values in the second blood glucose record time series and take the minimum value ;

[0113] Obtain the physical sign feature vector :

[0114] .

[0115] In this embodiment, when obtaining the physical sign feature vector through the dynamic feature parameters, N blood glucose values of the patient collected within the time period T are obtained and a blood glucose time series is formed. Then, the coefficient of variation of the blood glucose during this period, the number of times of high blood glucose and low blood glucose, and the time difference between two adjacent times of high blood glucose and the time difference between two adjacent times of low blood glucose are obtained as features. Among them, the coefficient of variation helps the subsequent recognition model to perceive the stability of blood glucose control. Unstable blood glucose may increase the risk of deterioration of diabetes and its complications, such as kidney disease. Excessive numbers of high blood glucose and low blood glucose may both have a negative impact on the kidneys. The time difference between two adjacent times of high blood glucose and the time difference between two adjacent times of low blood glucose can reflect the persistence of this situation. A too short time interval also means a greater negative impact on kidney disease. Therefore, in the physical sign feature vector of this embodiment, the larger the coefficient of variation of the blood glucose, the more times of high blood glucose, the more times of low blood glucose, and the smaller the time difference between two adjacent times of high blood glucose and the smaller the time difference between two adjacent times of low blood glucose will all make the values of the feature parameters therein larger. Based on this setting, the ability of the subsequent kidney disease prediction module to perceive the dynamic changes and risk rhythm of blood glucose is improved.

[0116] Finally, the kidney disease prediction module includes:

[0117] An input layer for inputting the feature vector;

[0118] A feature fusion layer for performing multi-modal fusion on the feature vector to obtain a fusion feature matrix;

[0119] An output layer for outputting a prediction result on whether nephropathy lesions occur based on the fused feature matrix.

[0120] As a further preferred solution, the method for performing multimodal fusion on the feature vector is as follows:

[0121] For the predicted corrected feature vector and the physical sign feature vector perform a linear transformation:

[0122] ;

[0123] ;

[0124] , ;

[0125] , ;

[0126] ;

[0127] where and are respectively the and after linear transformation, and are the linear transformation weights, and are the linear transformation biases, is the target dimension of linearization, is 's dimension, is 's dimension, max is the maximum value function, and respectively represent real number matrices with dimensions of and , represents a one-dimensional vector with dimension d;

[0128] Obtain the concatenated feature vector S:

[0129] ;

[0130] Obtain the attention weight :

[0131] ;

[0132] where represents the softmax function, is the attention weight matrix, is the attention bias;

[0133] Obtain the fused feature matrix :

[0134] ;

[0135] wherein, is the dot product operator of the matrix.

[0136] The predicted corrected feature vector and the physical sign feature vector obtained in this embodiment have different dimensions and different dimensions, and it is easy to be unbalanced when directly splicing or weighting. The fusion method adopted in this embodiment enables features of different dimensions and different sources to be in a relatively fair expression space during fusion. Therefore, the feature fusion in this embodiment can dynamically allocate the weight values of various features according to the individual differences and feature expression conditions of different patients, thereby highlighting the key features that have a greater impact on the risk of kidney disease lesions and weakening the influence of interfering features on the prediction results.

[0137] In addition, the method for outputting the prediction result of whether kidney disease lesions occur according to the fused feature vector is as follows:

[0138] Perform a linear transformation on the fused feature matrix to obtain an intermediate parameter :

[0139] ;

[0140] wherein, is the transformation weight and the dimension is , is the transformation bias and is a constant;

[0141] Obtain the probability of kidney disease lesions through sigmoid :

[0142] ;

[0143] wherein, e is the natural constant.

[0144] In this embodiment, the judgment of right and wrong is realized by introducing the sigmoid function into the model. For example, if the output result of the probability of kidney disease lesions is greater than 0.5, it can be judged that kidney disease lesions are predicted to occur, otherwise it is judged that kidney disease lesions are not predicted to occur. The kidney disease lesions here can include two situations: the deterioration of the kidney disease situation within a relatively short time period, such as within one month, or a person without kidney disease developing kidney disease.

[0145] Specifically, when predicting using the nephropathy prediction module, the labels used are determined by the glomerular filtration rate. When the glomerular filtration rate is lower than the preset threshold, the label indicates the occurrence of nephropathy lesions; otherwise, the label indicates the non-occurrence of nephropathy lesions.

[0146] Embodiment 2

[0147] A diabetes data supervision system based on real-time collection is applied to the above-mentioned diabetes data supervision method based on real-time collection. Refer to Figure 2 , and includes:

[0148] A static collection module for collecting static characteristic parameters of a patient, where the static characteristic parameters include the patient's personal information and treatment information;

[0149] A dynamic collection module for periodically collecting dynamic characteristic parameters of a patient, where the dynamic characteristic parameters include the patient's blood glucose data;

[0150] A feature construction module for constructing a feature vector based on the static and dynamic characteristic parameters. The feature vector includes a prediction correction feature vector and a physical sign feature vector. The larger the value in the prediction correction feature vector and the physical sign feature vector, the greater the negative impact on diabetes;

[0151] An evaluation module for evaluating the risk of nephropathy lesions through the nephropathy prediction module according to the feature vector.

[0152] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A diabetes data supervision method based on real-time collection, characterized in that: The following steps are involved: Collect the patient's static characteristic parameters, which include the patient's personal information and treatment information; Periodically collecting dynamic characteristic parameters of the patient, the dynamic characteristic parameters including the patient's blood sugar data; Constructing a feature vector based on static feature parameters and dynamic feature parameters, the feature vector includes a prediction correction feature vector and a physical sign feature vector, the larger the value in the prediction correction feature vector and the physical sign feature vector, the greater the negative impact on diabetes; The renal lesion risk is assessed through the renal disease prediction module according to the feature vector; The kidney disease prediction module comprises: An input layer, used for inputting the feature vector; A feature fusion layer, used for performing multi-modal fusion on the feature vectors to obtain a fused feature matrix; The output layer is used to output the prediction result of whether kidney disease occurs according to the fusion feature matrix; The method for performing multimodal fusion on the feature vector is: Correct the feature vector for prediction and the sign feature vector Perform a linear transformation: ； ； ， ； ， ； ； in, and After linear transformation, and , and is the linearly changing weight, and is a linearly varying bias, is the linearized target dimension, for The dimension of for The dimension of , max is the maximum value function, and Representing dimensions and The real matrix of represents a one-dimensional vector of dimension d; Get the concatenated feature vector S: ; Get attention weights : ； in, represents the softmax function, is the attention weight matrix, For attention bias; Get the fusion feature matrix : ； in, is the matrix dot multiplication operator; The method of outputting the prediction result of whether renal disease occurs based on the fused feature vector is: The fusion feature matrix Perform linear transformation to obtain intermediate parameters : ； in, is the transformation weight and has dimension , is the transformation bias and is a constant; Obtain the probability of kidney disease through sigmoid : ； Among them, e is a natural constant.

2. A diabetes data monitoring method based on real-time collection according to claim 1, characterized in that: The personal information includes age, height, weight and usual exercise time; the treatment information includes the duration of illness.

3. A diabetes data monitoring method based on real-time collection according to claim 2, characterized in that: The prediction correction feature vector is constructed based on the static feature parameters, and the method includes: Obtaining personal correction parameters based on the personal information : ； Among them, age represents age, w represents weight in kilograms, h represents height in meters, and t represents the average daily exercise time in minutes. and are adjustment coefficients, b and q are adjustment biases, and the adjustment coefficients and the adjustment biases are empirical parameters; Acquire treatment correction parameters according to the treatment information : ； in, represents the duration of illness and is expressed in years; Construct the predicted corrected feature vector : 。 4. A diabetes data monitoring method based on real-time collection according to claim 3, characterized in that: The method for obtaining the adjustment coefficient and the adjustment bias is: Based on big data, the age, height, weight and habitual exercise time of multiple diabetic patients were obtained; Get the maximum age and minimum value , the maximum average daily exercise duration and minimum value , the maximum value of body mass index and minimum value ; The equation is established and solved to obtain the adjustment coefficient and the adjustment bias: 。 5. A diabetes data monitoring method based on real-time collection according to claim 1, characterized in that: The method for obtaining the physical sign feature vector by using the dynamic feature parameter includes: Collect patients in time period N blood sugar values ​​within and form a blood sugar time series; Setting the Hyperglycemia Threshold and hypoglycemia threshold ; Obtain the value of the blood sugar time series not less than The number of parameters , and record the corresponding blood sugar acquisition time and form a first blood sugar recording time series in chronological order; Obtain the value in the blood sugar time series not higher than The number of parameters , and record the corresponding blood sugar acquisition time and form a second blood sugar recording time series in chronological order; Obtaining the coefficient of variation CV of the blood glucose time series, where the coefficient of variation is the ratio of the standard deviation to the mean; Obtain the difference between every two adjacent values ​​in the first blood glucose record time series, and take the minimum value ; Obtain the difference between every two adjacent values ​​in the second blood glucose record time series, and take the minimum value ; Get the physical sign feature vector : 。 6. A diabetes data monitoring method based on real-time collection according to claim 1, characterized in that: When the kidney disease prediction module makes predictions, the label used is determined by the glomerular filtration rate. When the glomerular filtration rate is lower than a preset threshold, the label is that kidney disease lesions have occurred, otherwise the label is that kidney disease lesions have not occurred.

7. A diabetes data monitoring system based on real-time collection, applied to a diabetes data monitoring method based on real-time collection as claimed in any one of claims 1 to 6, characterized in that: include: Static acquisition module, used to collect static characteristic parameters of patients, Static characteristic parameters include patients’ personal information and treatment information; A dynamic acquisition module, used for periodically acquiring dynamic characteristic parameters of a patient, wherein the dynamic characteristic parameters include blood sugar data of the patient; A feature construction module, used to construct a feature vector based on static feature parameters and dynamic feature parameters, wherein the feature vector includes a prediction correction feature vector and a physical sign feature vector, wherein a larger value in the prediction correction feature vector and the physical sign feature vector represents a greater negative impact on diabetes; The evaluation module is used to evaluate the risk of renal lesions through a renal disease prediction module according to the feature vector.

Images