An artificial intelligence-based patient blood glucose monitoring and risk intervention system

Abstract

The present invention discloses a patient blood glucose monitoring and risk intervention system based on artificial intelligence, including a blood glucose data monitoring module, which records the current patient's blood glucose data and digitizes it according to the individual differences of the patient, for matching other patients with a high enough similarity to the current patient's blood glucose data; and further includes a blood glucose data prediction module, which further forms a predicted value of the current patient's blood glucose data based on the blood glucose data characteristics of other patients, and makes corresponding medication recommendations for the current patient according to the predicted value. In the present invention, since the average blood glucose index, the maximum blood glucose index, and the minimum blood glucose index of the current patient are predicted, the cognitive level of the blood glucose patient regarding their own blood glucose data is enhanced, enabling the blood glucose patient to timely predict their own situation of abnormal blood glucose, thereby improving the safety of the current blood glucose patient during the illness period.

Description

Technical Field

[0001] The present invention belongs to the technical field of blood glucose prediction, and specifically relates to a patient blood glucose monitoring and risk intervention system based on artificial intelligence. Background Art

[0002] Traditional blood glucose monitoring methods mainly rely on discrete fingertip blood glucose detection, making it difficult for patients to obtain continuous and comprehensive blood glucose data. At the same time, this method cannot effectively combine individual differences for accurate risk assessment, nor can it respond to blood glucose abnormalities in real time and provide targeted intervention suggestions.

[0003] In the prior art, the Chinese invention patent with the publication number CN118366595A [An integrated comprehensive management system for diabetic patients] records a technical solution for obtaining the blood glucose parameters of each diabetic patient corresponding to the current monitoring period, and visually monitoring and specifically analyzing the blood glucose stability status in the fasting state, postprandial state, and bedtime state among the blood glucose parameters of each diabetic patient corresponding to the current monitoring period. However, this technology mainly conducts predictions by tracking the periodic data of a single patient. Therefore, this solution requires a large amount of data of the patient to be measured to be recorded for a long time in order to accurately predict the patient himself. At the same time, when the health status of this patient changes irregularly, it is very difficult to accurately predict based on the patient's previous health status.

[0004] Therefore, a technical solution for predicting blood glucose values based on big data of patients with the same blood glucose is proposed. The characteristics of the short-term health data of the patient to be measured are matched with a large number of patients with the same symptoms, and the matching results are applied to the blood glucose value prediction of the patient to be measured to achieve a fast and accurate prediction effect. Summary of the Invention

[0005] The purpose of the present invention is: In order to address the above problems, to meet the need for accurate recording and processing of individual patient blood glucose data, attempt to generate accurate medication suggestions and intervention strategies, avoid the difficulty for patients to detect blood glucose risks, and thus improve the quality of life of blood glucose patients, a patient blood glucose monitoring and risk intervention system based on artificial intelligence is proposed.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A patient blood glucose monitoring and risk intervention system based on artificial intelligence at least includes:

[0008] A blood glucose data monitoring module that records the blood glucose data of the current patient and constructs a blood glucose curve on a coordinate system. The blood glucose curve includes a normal blood glucose curve and an abnormal blood glucose curve, and an abnormal curve set is constructed based on the abnormal blood glucose curve to determine an abnormal normalization circle;

[0009] Construct a set of normal curves based on the normal blood glucose curve to determine the normal normalized circle. Establish an abnormal normalized circle and a normal normalized circle in the coordinate system, obtain the normalized circle diagram of the current patient, obtain the normalized circle diagrams of other patients, and calculate the similarity α;

[0010] Create an abnormal circle set and a normal circle set based on the abnormal curve set and the normal curve set respectively. Obtain the normal circle diagram based on the normal normalized circle and the normal circle set, and obtain the abnormal circle diagram based on the abnormal normalized circle and the abnormal circle set;

[0011] Obtain the normal circle diagrams and abnormal circle diagrams of multiple other patients, and thereby calculate the similarity β between other patients and the current patient;

[0012] Combine the similarity α and the similarity β to match other patients with blood glucose data similar to that of the current patient;

[0013] It also includes a blood glucose data prediction module, which forms a predicted value of the current patient's blood glucose data based on the blood glucose data of other patients, and performs risk intervention on the current patient according to the predicted value.

[0014] Among them, the blood glucose data monitoring module records the blood glucose data of the current patient in a cycle and constructs a blood glucose curve on the coordinate system;

[0015] Extract the abnormal curves corresponding to the abnormal states from the abnormal blood glucose curves, connect the endpoints between multiple adjacent abnormal curves to form an abnormal curve set, connect the head and tail endpoints of the abnormal curve set to form a line segment set that penetrates the abnormal curve set, and combine the head and tail endpoints of the line segment penetrating the abnormal curve set to form a circular structure and bend it into an abnormal normalized circle towards a perfect circle;

[0016] Extract the normal curves corresponding to the normal states from the normal blood glucose curves, connect the endpoints between multiple adjacent normal curves to form a normal curve set, connect the head and tail endpoints of the normal curve set to form a line segment set that penetrates the normal curve set, and combine the head and tail endpoints of the line segment penetrating the normal curve set to form a circular structure and bend it into a normal normalized circle towards a perfect circle.

[0017] Among them, the similarity α between the normalized circle diagrams of the current patient and multiple other patients is calculated according to the following relational formula:

[0018]

[0019] Among them, A 0 represents the area of the abnormal normalized circle of the current patient:

[0020]

[0021] is the outer radius of the normalized circular graph for the current patient, is the inner radius of the normalized circular graph for the current patient,

[0022] represents the abnormal normalized circular area of the i-th other patient:

[0023]

[0024] where, is the outer radius of the normalized circular graph of the blood glucose data of the i-th other patient, is the inner radius of the normalized circular graph of the blood glucose data of the i-th other patient;

[0025] The other patients corresponding to the normalized circular graphs that meet the similarity range α are determined as proportion-matched patients.

[0026] Among them, the normal curve set is bent according to the curvature of the normal normalized circle to form a normal circle set, and the normal normalized circle and the normal circle set are drawn with the same point as the center in the coordinate system, which is defined as the normal circular graph of the current patient's blood glucose data;

[0027] In the normal circular graph, the coordinates of multiple parallel points of the normal circle set relative to the normal normalized circle are recorded as M nor0 、M nor1 、M nor2 、……、M norn , n represents the number of parallel points of the normal circular graph, and the distances R nor0 、R nor1 、R nor2 、……、R norn .

[0028] Among them, the abnormal curve set is bent according to the curvature of the abnormal normalized circle to form an abnormal circle set that fluctuates around the abnormal normalized circle;

[0029] In the coordinate system, the abnormal normalized circle and the abnormal circle set are drawn at the center point F, which is defined as the abnormal circular graph of the current patient's blood glucose data. In the abnormal circular graph, the coordinates of multiple parallel points of the abnormal circle set relative to the abnormal normalized circle are recorded as M abn0 、M abn1 、M abn2 、……、M abnm , m represents the number of parallel points of the abnormal circular graph, and the distances R abn0 、R abn1 、R abn2 、……、R abnm .

[0030] Among them, calculate the preliminary coordinate matching degree Sum distance matching degree , as follows;

[0031]

[0032]

[0033] Among them, and are respectively set as the weights of coordinates and distances in the normal circular diagram, and are respectively set as the weights of coordinates and distances in the abnormal circular diagram;

[0034] Set the normal circular diagrams and abnormal circular diagrams of the blood glucose data of other patients as the control group C k ;

[0035] Candidate control group C k The matching error of the normal circular diagram of the candidate control group C and the normal circular diagram of the current patient's blood glucose data in terms of coordinate features :

[0036]

[0037] is the position of the 𝑗th parallel point in the normal circular diagram of the candidate control group C k , and is the position of the 𝑑th parallel point of the normal circular set relative to the normal normalized circular in the normal circular diagram of the current patient;

[0038] Candidate control group C k The matching error of the normal circular diagram of the candidate control group C and the normal circular diagram of the current patient's blood glucose data in terms of distance features :

[0039]

[0040] Among them, is the distance between the 𝑗th parallel point and point E in the normal circular diagram of the candidate control group C k , and is the distance between the 𝑑th parallel point of the normal circular set relative to the normal normalized circular and the corresponding point E in the normal circular diagram of the current patient;

[0041] Candidate control group C k The matching error of the abnormal circular diagram of the candidate control group C and the abnormal circular diagram of the current patient's blood glucose data in terms of coordinate features :

[0042]

[0043] is the candidate control group Ck The position of the q-th parallel point in the abnormal circular diagram, is the position of the r-th parallel point of the abnormal circular set relative to the abnormal normalized circle in the abnormal circular diagram of the current patient;

[0044] Candidate control group C k The matching error in distance characteristics between the abnormal circular diagram of :

[0045]

[0046] Among them, is the distance between the q-th parallel point in the abnormal circular diagram of candidate control group C and point F, k and is the distance between the r-th parallel point of the abnormal circular set relative to the abnormal normalized circle in the abnormal circular diagram of the current patient and the corresponding point F.

[0047] Among them, the similarity β between the abnormal circular diagram of the control group C k and the abnormal circular diagram of the current patient is specifically calculated as follows:

[0048]

[0049] Among them, and are weighting coefficients, and ;

[0050] Obtain multiple amplitude-matching patients corresponding to the similarity β for the control group C k .

[0051] Among them, based on the calculation of the similarity α and , obtain the set of proportion-matching patients and the set of amplitude-matching patients , judge the fluctuation range of the blood glucose data of the current patient, and judge the degree of blood glucose abnormality and the level of medication required for the current patient;

[0052] Divide the and sets into three sets again, that is, the set that is simultaneously in and is , the set that is in and not in is , and the set that is in and not in is :

[0053] In the set , and calculate the average value of the average blood glucose index respectively:

[0054] , , ;

[0055] Average value of the maximum blood glucose index:

[0056] , , ;

[0057] And the average value of the minimum blood glucose index:

[0058] , , ;

[0059] At this time, from the screened set of matching patients, combined with the weighting coefficient, calculate the predicted average blood glucose index , maximum blood glucose index and minimum blood glucose index of the current patient:

[0060]

[0061] Among them, , , are the weighting indices of sets , and respectively;

[0062] Predict the fluctuation range of the blood glucose data of the current patient and perform risk intervention according to the predicted average blood glucose index, maximum blood glucose index and minimum blood glucose index of the current patient.

[0063] To sum up, due to the adoption of the above technical solutions, the beneficial effects of the present invention are:

[0064] 1. In the present invention, due to the adoption of the blood glucose data monitoring module, the blood glucose data of the current patient is recorded, and the similarities α and β between other patients and the current patient are calculated therefrom. Further, other patients with blood glucose data similar to that of the current patient are matched in combination with the similarity α and the similarity β, and a predicted value of the blood glucose data of the current patient is formed according to the blood glucose data of other patients, and risk intervention is performed on the current patient according to the predicted value, so as to achieve the purpose of rapid and accurate prediction.

[0065] 2. In the present invention, since a large number of similar patient groups to the current patient are matched, the further development trend of the current patient's current condition can be predicted based on the development status of the blood glucose conditions of this similar patient group, assisting the current patient to have a more accurate expectation of the treatment requirements and treatment effects of the blood glucose condition.

[0066] 3. In the present invention, since the average blood glucose index, maximum blood glucose index, and minimum blood glucose index of the current patient are predicted, it is possible to determine the stage to which the blood glucose data of the current patient belongs and the possibility that the blood glucose data changes to exceed the normal range, realizing targeted prediction and planning for the blood glucose data of the current patient, enhancing the cognitive level of the blood glucose patient regarding their own blood glucose data, enabling the blood glucose patient to promptly predict their own situation of abnormal blood glucose, and thereby improving the safety of the current blood glucose patient during the illness period. Detailed implementation manners

[0067] In order to make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with embodiments. 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.

[0068] Embodiment 1, an artificial intelligence-based patient blood glucose monitoring and risk intervention system, at least including:

[0069] A blood glucose data monitoring module, which records the blood glucose data of the current patient and digitalizes it according to the individual differences of the patient, and is used to match other patients with a high enough similarity to the blood glucose data of the current patient;

[0070] It further includes a blood glucose data prediction module, which further forms a predicted value for the blood glucose data of the current patient based on the blood glucose data characteristics of other patients, and makes corresponding medication suggestions for the current patient according to the predicted value.

[0071] Among them, during the daily exercise and diet of the current patient, the blood glucose data of the current patient is recorded in a cycle (24 hours) and a blood glucose curve is constructed on a coordinate system. The blood glucose curve includes a normal blood glucose curve and an abnormal blood glucose curve;

[0072] The abnormal blood glucose curve and normal blood glucose curve data are output, used to compare with the big data of the corresponding control group of existing blood glucose patients, respectively obtain groups with a relatively high similarity to the abnormal blood glucose curve and normal blood glucose curve of the current patient, and predict the blood glucose data of the current patient in the next cycle according to the blood glucose change characteristics of this group with a relatively high similarity, and output the predicted blood glucose result of the current patient;

[0073] Prompt the predicted blood glucose results of the current patient, as well as the corresponding blood glucose medications and medical advice, which can ensure that the patient pays more attention to taking medications on time and in the right amount. When the current patient's condition is severe, it can also prompt the timely intervention of calling the nursing staff to maintain safety.

[0074] Further, after detecting abnormal fluctuations in the blood glucose data of the current patient and obtaining the abnormal fluctuation curve of the current patient, including two characteristics: blood glucose value and blood glucose change trend, determine whether the real-time blood glucose value and the blood glucose change trend are in an abnormal state, where:

[0075] The abnormal state of the blood glucose value means that the blood glucose value exceeds the safe range, and the safe threshold of the blood glucose value is (3.9 mmol / L - 6.1 mmol / L);

[0076] The abnormal state of the blood glucose change trend means that the blood glucose value suddenly rises or suddenly drops, resulting in a sharp fluctuation, and the fluctuation amplitude exceeds the preset safe value of the change curvature threshold (the speed change of the blood glucose value within 1 hour is less than 1.1 mmol / L);

[0077] If there are multiple abnormal states (the real-time blood glucose value and the blood glucose change trend are mixed with each other), and each abnormal state corresponds to an abnormal curve. Among them, the abnormal curve of the real-time blood glucose value refers to the time period from when the real-time blood glucose value exceeds the safe threshold of the blood glucose value to when it returns to the safe threshold of the blood glucose value again, and the abnormal curve of the blood glucose change trend refers to the time period when the curvature of the increase / decrease of the blood glucose value is greater than the change curvature threshold;

[0078] Extract the abnormal curves corresponding to the abnormal states from the abnormal blood glucose curves, connect the endpoints between multiple adjacent abnormal curves to form an abnormal curve set, connect the head and tail endpoints of the abnormal curve set to form a line segment set that runs through the abnormal curve set, and unify the head and tail endpoints of the line segment running through the abnormal curve set to form a circular structure. Further, bend the circular structure into a regular circular shape to form an abnormal normalized ring;

[0079] Similarly, extract the normal curves corresponding to the normal states from the normal blood glucose curves,

[0080] Connect and bend the endpoints between multiple normal curves to form a normal normalized ring,

[0081] Draw the abnormal normalized ring and the normal normalized ring through the same center in the coordinate system;

[0082] Thus, obtain the normalized ring diagram of the blood glucose data of the current patient;

[0083] Separate the historical blood glucose data of the known multiple other patients' blood glucose data in the same way with 24 hours as a cycle to form multiple cycle data, and thus obtain the normalized ring diagrams of the multiple other patients' blood glucose data;

[0084] According to the normalized circular diagrams of multiple other patients obtained, compare the normalized circular diagram of the current patient's blood glucose data one by one, calculate the similarity α, select other patients whose similarity α to the current patient's data is close as the subsequent reference objects, and predict the blood glucose change trend of the current patient by comparing the reference objects with the current patient's blood glucose data, and adjust the intervention plan;

[0085] Calculate the similarity α between the normalized circular diagram of the current patient and the normalized circular diagrams of multiple other patients. The specific formula is as follows:

[0086]

[0087] Among them, the abnormal normalized circular area A of the current patient 0 :

[0088]

[0089] is the outer radius of the normalized circular diagram of the current patient, is the inner radius of the normalized circular diagram of the current patient,

[0090] The abnormal normalized circular area of the i-th other patient :

[0091]

[0092] Among them, is the outer radius of the normalized circular diagram of the blood glucose data of the i-th other patient, is the inner radius of the normalized circular diagram of the blood glucose data of the i-th other patient;

[0093] Thus, the similarity α between the normalized circular diagram of the current patient and the normalized circular diagram of each other patient is obtained. Set the similarity range to determine the other patients corresponding to the normalized circular diagrams that meet the similarity range α as proportion-matched patients (similarity 1 > α ≥ 0.9);

[0094] Further, bend the abnormal curve set according to the curvature of the abnormal normalized circle, and combine the abnormal normalized circles to form an abnormal circle set that fluctuates around the abnormal normalized circle;

[0095] Similarly, bend the normal curve set according to the normal normalized circle to form a normal circle set,

[0096] Set points E and F in the coordinate system, and draw a normal normalized circle and a normal circle set with point E as the center;

[0097] Similarly, draw an abnormal normalized circle and an abnormal circle set at point F in the same proportion;

[0098] Thus, the normal circular diagram () and abnormal circular diagram () of the current patient's blood glucose data are obtained;

[0099] In the normal circular diagram, the coordinates of multiple parallel points of the normal circular set relative to the normal normalized circle are recorded as M nor0 、M nor1 、M nor2 、……、M norn , where n represents the number of parallel points in the normal circular diagram, and the distances R nor0 、R nor1 、R nor2 、……、R norn ;

[0100] Correspondingly, in the abnormal circular diagram, the coordinates of multiple parallel points of the abnormal circular set relative to the abnormal normalized circle are recorded as M abn0 、M abn1 、M abn2 、……、M abnm , where m represents the number of parallel points in the abnormal circular diagram, and the distances R abn0 、R abn1 、R abn2 、……、R abnm ;

[0101] Among them, the parallel points are set as the inflection points on the curves of the normal circular set and the abnormal circular set that are parallel to the arcs of the corresponding normal normalized circle and abnormal normalized circle, that is, the connection line between the parallel point and the center of the normal circular diagram / abnormal circular diagram, and the intersection of the connection line between the parallel point and the center and the normal normalized circle / abnormal normalized circle is perpendicular;

[0102] Similarly, according to the normalized circular diagrams of multiple other known patients, the normal circular diagrams and abnormal circular diagrams of the corresponding other patients' blood glucose data are obtained. Thus, the normal circular diagrams and abnormal circular diagrams of the other patients' blood glucose data are set as the control group C k , and the coordinate features and distance features of the control group and the current patient's blood glucose data are obtained respectively, and the preliminary coordinate matching degree S k and distance matching degree T k ;

[0103]

[0104]

[0105] Among them, 、 are respectively set as the weights of coordinates and distances in the normal circular diagram, satisfying ;

[0106] , are respectively set as the weights of coordinates and distances in the abnormal circular diagram, satisfying ;

[0107] Candidate control group C k The matching error of the normal circular diagram of the candidate control group C :

[0108]

[0109] is the position of the 𝑗-th parallel point in the normal circular diagram of the candidate control group C k , and is the position of the d-th parallel point of the normal circular set relative to the normal normalized circle in the normal circular diagram of the current patient;

[0110] Candidate control group C k The matching error of the normal circular diagram of the candidate control group C :

[0111]

[0112] Among them, is the distance between the 𝑗-th parallel point and point E in the normal circular diagram of the candidate control group C k , and is the distance between the d-th parallel point of the normal circular set relative to the normal normalized circle and the corresponding point E in the normal circular diagram of the current patient;

[0113] Candidate control group C k The matching error of the abnormal circular diagram of the candidate control group C :

[0114]

[0115] is the position of the q-th parallel point in the abnormal circular diagram of the candidate control group C k , and is the position of the r-th parallel point of the abnormal circular set relative to the abnormal normalized circle in the abnormal circular diagram of the current patient;

[0116] Candidate control group C k The matching error of the abnormal circular diagram of the candidate control group C :

[0117]

[0118] Among them, is the candidate control group C k The distance between the q-th parallel point and point F in the abnormal circular graph is the distance between the r-th parallel point of the abnormal circular ring set relative to the abnormal normalized circular ring and the corresponding point F in the abnormal circular graph of the current patient;

[0119] Multiply the coordinate matching degree S k and the distance matching degree T k to obtain the similarity β between the abnormal circular graph of this control group C k and the abnormal circular graph of the current patient's blood glucose data. The specific formula is as follows:

[0120]

[0121] Among them, and are weighting coefficients, and ;

[0122] In the above, it should be noted that by comparing the normal circular graph and the abnormal circular graph of the current patient's blood glucose data with the control group C k to obtain multiple amplitude-matching patients (similarity 1 > k ≥ 0.9) of the control group C .

[0123] In this way, by calculating the similarity α and , the set of proportion-matching patients and the set of amplitude-matching patients are obtained. Thus, through and the blood glucose characteristics of the set, the fluctuation range of the current patient's blood glucose data is judged, and further the degree of blood glucose abnormality and the level of medication required for the current patient are judged;

[0124] Among these, and the set is re-divided into three sets, that is, the set that is simultaneously in and is , the set that is in and not in is , and the set that is in and not in is :

[0125] In the sets , and , calculate the average value of the blood glucose average index respectively:

[0126] , , ;

[0127] Mean of the maximum blood glucose index:

[0128] , , ;

[0129] And the mean of the minimum blood glucose index:

[0130] , , ;

[0131] At this time, from the set of screened matching patients, combined with the weighting coefficient, calculate the predicted average blood glucose index , maximum blood glucose index and minimum blood glucose index :

[0132]

[0133] Among them, , , are the weighting indices of the sets , and respectively, and > + , + + = 1.

[0134] According to the predicted average blood glucose index, maximum blood glucose index, and minimum blood glucose index of the current patient, it is possible to judge the stage to which the blood glucose data of the current patient belongs and the possibility that the blood glucose data changes to exceed the normal range, so as to realize targeted prediction and planning for the blood glucose data of the current patient, enhance the cognitive level of blood glucose patients about their own blood glucose data, enable blood glucose patients to predict in time their own abnormal blood glucose situation, and thus improve the safety of current blood glucose patients during the illness.

[0135] Finally, it is worth noting that if the prediction of the current patient on the current date is successful through the samples provided by big data, the blood glucose data of the current patient on the current date can be trained as the closest sample and recorded, thereby further improving the prediction accuracy of subsequent predictions.

[0136] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A patient blood glucose monitoring and risk intervention system based on artificial intelligence, characterized in that: At least: The blood glucose data monitoring module records the current patient's blood glucose data and constructs a blood glucose curve on the coordinate system. The blood glucose curve includes a normal blood glucose curve and an abnormal blood glucose curve. An abnormal curve set is constructed based on the abnormal blood glucose curve to determine the abnormal normalization circle. According to the normal blood glucose curve, a normal curve set is constructed to determine the normal normalized circle, and an abnormal normalized circle and a normal normalized circle are established in the coordinate system to obtain the normalized circle diagram of the current patient, obtain the normalized circle diagrams of other patients, and calculate the similarity α; An abnormal ring set and a normal ring set are created according to the abnormal curve set and the normal curve set respectively, a normal ring diagram is obtained based on the normal naturalized ring and the normal ring set, and an abnormal ring diagram is obtained based on the abnormal naturalized ring and the abnormal ring set; Obtain normal donut diagrams and abnormal donut diagrams of multiple other patients, and thereby calculate the similarity β between the other patients and the current patient; Combine similarity α and similarity β to match other patients with similar blood glucose data to the current patient; It also includes a blood sugar data prediction module, which forms a predicted value of the blood sugar data of the current patient based on the blood sugar data of other patients, and performs risk intervention on the current patient based on the predicted value; The blood sugar data monitoring module records the blood sugar data of the current patient in one cycle and constructs a blood sugar curve on the coordinate system; Extract the abnormal curve corresponding to the abnormal state from the abnormal blood glucose curve, connect the endpoints of multiple adjacent abnormal curves to form an abnormal curve set, connect the head and tail endpoints of the abnormal curve set to form a line segment set that runs through the abnormal curve set, and combine the head and tail endpoints of the line segment that runs through the abnormal curve set to form a ring structure that bends into a perfect circle to become an abnormal normalized ring; Extracting a normal curve corresponding to a normal state from the normal blood glucose curve, connecting the endpoints of multiple adjacent normal curves to form a normal curve set, connecting the head and tail endpoints of the normal curve set to form a line segment set that penetrates the normal curve set, and merging the head and tail endpoints of the line segment that penetrates the normal curve set to form a ring structure that bends into a perfect circle to become a normalized ring; The similarity α between the normalized donut diagrams of the current patient and multiple other patients is calculated according to the following relationship: Among them, A0 represents the abnormal normalized annulus area of ​​the current patient: is the outer radius of the normalized donut diagram of the current patient, is the inner radius of the normalized donut diagram of the current patient, Represents the abnormal normalized annulus area of ​​the i-th other patient: in, is the outer radius of the normalized donut chart of the i-th other patient’s blood glucose data, The inner radius of the normalized donut chart for the blood glucose data of the i-th other patient; Other patients corresponding to the normalized donut diagram that meet the similarity range α are determined as proportion-matched patients.

2. The artificial intelligence-based patient blood glucose monitoring and risk intervention system according to claim 1, characterized in that: The normal curve set is bent according to the curvature of the normal normalized circle to form a normal circle set, and a normal normalized circle and a normal circle set are drawn with the same point as the center in the coordinate system, which is defined as a normal circle diagram of the current patient's blood sugar data; In the normal ring diagram, the coordinates of multiple parallel points of the normal ring set relative to the normal normalized ring are recorded as M nor0 、M nor1 、M nor2 ,……,M norn , n represents the number of parallel points in the normal donut diagram, and records the distance R between each parallel point and the corresponding center point E nor0 , R nor1 , R nor2 ,……,R norn .

3. The patient blood glucose monitoring and risk intervention system based on artificial intelligence as claimed in claim 2, characterized in that: The abnormal curve set is bent according to the curvature of the abnormal normalized ring to form an abnormal ring set that fluctuates around the abnormal normalized ring; Draw an abnormal normalized circle and an abnormal circle set at the center point F in the coordinate system, which is defined as the abnormal circle diagram of the current patient's blood glucose data. The coordinates of multiple parallel points of the abnormal circle set relative to the abnormal normalized circle in the abnormal circle diagram are recorded as M abn0 、M abn1 、M abn2 ,……,M abnm , m represents the number of parallel points in the abnormal donut diagram, and records the distance R between each parallel point and the corresponding point F abn0 , R abn1 , R abn2 ,……,R abnm .

4. The artificial intelligence-based patient blood glucose monitoring and risk intervention system according to claim 3, characterized in that: Calculate the initial coordinate matching and distance matching , as follows; in, , Set as the weights of coordinates and distances in a normal donut diagram, respectively. , They are set as the weights of coordinates and distances in the anomaly donut diagram, respectively; Set the normal donut chart and abnormal donut chart of other patients' blood glucose data as the control group C k ; Candidate control group C k The matching error between the normal donut diagram of the patient's blood glucose data and the normal donut diagram of the patient's blood glucose data in terms of coordinate features : Candidate control group C k The position of the 𝑗th parallel point in the normal donut diagram, is the position of the dth parallel point of the normal ring set relative to the normal normalized ring in the normal ring diagram of the current patient; Candidate control group C k The matching error between the normal donut diagram of the patient's blood glucose data and the normal donut diagram of the current patient's blood glucose data in terms of distance features : in, Candidate control group C k The distance between the 𝑗th parallel point and point E in the normal donut diagram, is the distance between the dth parallel point of the normal ring set relative to the normal normalized ring and the corresponding point E in the normal ring diagram of the current patient; Candidate control group C k The matching error between the abnormal donut diagram of the patient's blood glucose data and the abnormal donut diagram of the current patient's blood glucose data in terms of coordinate features : Candidate control group C k The position of the qth parallel point in the abnormal donut diagram, is the position of the rth parallel point of the abnormal ring set relative to the abnormal normalized ring in the abnormal ring diagram of the current patient; Candidate control group C k The matching error between the abnormal donut diagram of the patient's blood glucose data and the abnormal donut diagram of the current patient's blood glucose data in terms of distance features : in, Candidate control group C k The distance between the qth parallel point and point F in the abnormal donut diagram, is the distance between the rth parallel point of the abnormal ring set relative to the abnormal normalized ring and the corresponding point F in the abnormal ring diagram of the current patient.

5. The patient blood glucose monitoring and risk intervention system based on artificial intelligence as claimed in claim 4, characterized in that: The control group C k The specific calculation formula for the similarity β between the abnormal donut diagram of the patient and the abnormal donut diagram of the current patient is as follows: in, and is the weighting coefficient, and ; Get the similarity β to the control group C k The corresponding multiple amplitudes match the patient.

6. The patient blood glucose monitoring and risk intervention system based on artificial intelligence as claimed in claim 5, characterized in that: Based on the similarity α and Calculation of the proportion matching patient set and amplitude matched patient set , determine the fluctuation range of the current patient's blood sugar data, and determine the current patient's blood sugar abnormality and the level of medication required; Will and The set is re-divided into three sets, namely and The collection is , in and not in The collection is , and in and not in The collection is : In the collection , and , calculate the average value of the average blood sugar index respectively: 、 、 ； Average values ​​of maximum glycemic index: 、 、 ； And the average of the minimum glycemic index: 、 、 ； At this time, the average blood sugar index of the current patient is calculated and predicted from the screened matching patient set and combined with the weighted coefficient. , maximum glycemic index and minimum glycemic index : in, , , Separately , and The weighted index of Based on the predicted average blood sugar index, maximum blood sugar index and minimum blood sugar index of the current patient, the fluctuation range of the current patient's blood sugar data is predicted and risk intervention is performed.

Images