CN115762782A - Blood glucose analysis method based on CGM dynamic graph and edge calculation device - Google Patents

Abstract

The invention provides a blood sugar analysis method and an edge computing device based on a CGM dynamic map, belonging to the technical field of intelligent medical devices. The present invention includes: using the daily CGM atlas data to predict the risk probability of daily blood sugar fluctuations; using the daytime CGM dynamic atlas data and the previous day's nighttime CGM dynamic atlas data to predict the current day's nighttime hypoglycemia risk probability; for the long-term CGM dynamic atlas data, Predict the risk probability of long-term blood sugar fluctuations, and give the risk level of blood sugar control effect according to the risk probability of long-term blood sugar fluctuations and daily analysis results. The present invention includes an edge computing device, which utilizes the above-mentioned blood sugar analysis method based on a CGM dynamic map. The invention can effectively prevent the risk of hypoglycemia at night, reflect the blood sugar control situation of the patient to a certain extent, and give risk prompt in time.

Description

A blood sugar analysis method and edge computing device based on CGM dynamic map

technical field

The invention relates to the technical field of intelligent medical devices, in particular to a blood glucose analysis method and an edge computing device based on a CGM dynamic map.

Background technique

In recent years, due to the impact of many factors such as living standards, diet structure and sub-healthy lifestyles, the prevalence of diabetes has increased year by year. In order to effectively control blood sugar, blood sugar monitoring related technologies are gradually developing towards intelligence and informatization. Therefore, the continuous blood glucose monitoring (CGM) system has gradually received extensive attention from people in related fields. It can detect hyperglycemia after meals, nighttime Hypoglycemia, dawn phenomenon and Somogyi phenomenon are of great significance in the clinical treatment of diabetes. The CGM system can generate an ambulatory glucose profile (AGP), which can output blood glucose values, blood glucose graphs, and blood glucose graph summaries at fixed daily sampling intervals (blood glucose fluctuations, average values, time in normal range (TIR), time above normal range (TAR) , time below normal range (TBR) and area under the blood glucose curve, etc.); the indicator information provided by the map can better adjust the treatment plan for diabetes and evaluate the control effect of blood glucose.

However, the utilization rate of the dynamic blood glucose map of the existing CGM system is low and the technical threshold is high. It is difficult for patients to personally interpret the report of the map. Poor, CGM alarms are frequent, which makes the user alarm fatigued, and patients will customize the blood sugar alarm threshold settings to ignore the alarm information, resulting in a decrease in the application effect of the CGM system. Usually, users pay attention to their blood sugar changes during the day, and may ignore the hidden risk of hypoglycemia at night, which poses a threat to the life and health of users. If the risk of hypoglycemia at night can be predicted based on the daytime and historical nighttime blood sugar The probability, given to the user as a reference, will more effectively prevent the risk of hypoglycemia at night. In addition, there are differences in blood glucose fluctuations over time, which can reflect the patient's blood glucose control to a certain extent, and these differences cannot be provided on the current dynamic blood glucose map. In addition, under the current conditions, the evaluation of the effect of blood sugar control and the risk warning of the need for hospitalization are not timely enough.

Chinese invention patent CN111110249B discloses a blood sugar fluctuation evaluation method and evaluation device, which is a blood sugar fluctuation calculation method based on gradient variation for CGM data. Although the proposed gradient variation parameter can simultaneously characterize the blood sugar fluctuation amplitude and blood sugar fluctuation frequency, but Part of the information in the CGM data may still be lost; however, the map analysis information of the present invention not only includes the blood sugar fluctuation range and blood sugar fluctuation frequency, but also enables dynamic analysis of the information in the CGM data.

Chinese invention patent CN113948207A discloses a blood sugar data processing method for hypoglycemia early warning, using a pre-trained support vector machine model to obtain a blood sugar prediction value corresponding to real-time blood sugar data, although there is a correlation between real-time blood sugar data and time , so that the output of the early warning model is more accurate than relying only on a single variable, but it is only based on real-time blood sugar data and time for early warning, and there is no distinction between night and day alarms; and the present invention is not only based on each CGM map Data for early warning, and a separate early warning for hypoglycemia at night.

Contents of the invention

In order to solve the technical problems existing in the prior art, the present invention provides a blood sugar analysis method based on CGM dynamic atlas: using daily CGM atlas data to predict the risk probability of daily blood sugar fluctuation; One day and night CGM dynamic map data predicts the risk probability of hypoglycemia at night; for the long-term CGM dynamic map data of the first set time period, predicts the risk probability of long-term blood sugar fluctuations, according to the risk probability of long-term blood sugar fluctuations and the second set time period The results are analyzed daily to give a risk level for glycemic control effectiveness for the second set time period. The present invention provides an edge computing device, which utilizes the above-mentioned blood sugar analysis method based on a CGM dynamic map. The invention can more effectively prevent the risk of hypoglycemia at night, reflect the patient's blood sugar control situation to a certain extent, and can prompt the risk in time.

The invention provides a blood glucose analysis method based on CGM dynamic atlas, comprising the following steps:

Convert user information and CGM dynamic map data into structured data through data preprocessing;

Obtain map information;

Use the daily CGM map data and combine the map information to predict the risk probability of daily blood sugar fluctuations;

Use the daytime CGM dynamic map data and the previous day's nighttime CGM dynamic map data, and combine the map information to predict the risk probability of hypoglycemia at night;

Convert daily CGM dynamic map data and daily analysis results into structured data;

According to the long-term CGM dynamic map data of the first set time period, combined with the map information to predict the risk probability of long-term blood sugar fluctuations;

According to the long-term blood sugar fluctuation risk probability and the daily analysis results of the second set time period, the risk level of blood sugar control effect in the second set time period is given.

Preferably, the daily analysis results include the risk probability of daily blood sugar fluctuation and the risk probability of nightly hypoglycemia fluctuation.

Preferably, the sampling time of the daily CGM dynamic atlas data is 24 hours, and the sampling interval is 5 minutes.

Preferably, the first set time period is not less than 14 days, the second set time period is not less than 7 days, during the day from 6:00 to 24:00 on the same day, and from 0:00 to 6:00 on the same day at night on the previous day, The period of three hours before the countdown of the day is from 21:00 to 24:00 of the day.

Preferably, the map information includes map analysis information and map categories; the map analysis information includes: time period map glucose maximum value Max, time period map glucose minimum value Min, time period map slope Slope, time period map glucose variation coefficient GV, time period map standard deviation SD, Period map variance Var, time period map glucose mean Mean, time period map average blood sugar fluctuation range MAGE, and time period map average blood glucose absolute difference MODD; the map categories include daytime map, nighttime map, full-time map and optional time-period map.

Preferably, for the full-time graph, the graph analysis information includes the time percentages corresponding to the times in which all glucose concentration segments appear in a day.

Preferably, the glucose concentration segments are divided into segments according to the number of mg of glucose concentration contained in each deciliter.

Preferably, predicting the risk probability of daily blood sugar fluctuation is completed by a daily blood sugar fluctuation evaluation model; the daily blood sugar fluctuation evaluation model obtains the daily risk probability of blood sugar fluctuation according to the input data x, x is as follows:

x = [SD, Mean, GV, MAGE, MODD].

Preferably, the daily blood glucose volatility evaluation model is shown in the following formula:

Among them, p represents the risk probability of daily blood sugar fluctuation, y represents the risk result of manual calibration, and the values of 1 and 0 in y represent high risk and low risk of blood sugar fluctuation respectively. F(x) includes random forest classification model and Logistic For regression classification models, ∈ represents the error term.

Preferably, the prediction of the nighttime hypoglycemia risk probability is completed by the daytime nighttime hypoglycemia risk warning model; the daytime nighttime hypoglycemia risk warning model obtains the daytime nighttime hypoglycemia risk probability according to the input data X, and X is as follows:

X=[Gpremin _{at night} , Gmin _{during the day} , G_max_re, Gmin _-3h , Gmean _-3h , Gslope _-3h ];

Among them, G_max_re represents the ratio of the maximum value G_max of the daytime dynamic map glucose value _{of the day} to the maximum value G_Pre_max of the daytime dynamic map glucose value _{of the previous day, and Gpremin night represents} the minimum value of the nighttime dynamic map glucose value of the previous day , Gmin _{during the day} indicates the minimum value of the glucose value of the daytime dynamic map of the day, and Gmin _-3h , Gmean _-3h and Gslope _-3h respectively represent the minimum value and average value of the glucose value of the three hours before the reciprocal of the daytime dynamic map of the day and slope.

Preferably, the day-night hypoglycemia risk warning model is as follows:

Among them, _plow indicates the probability of hypoglycemia at night, S indicates the nighttime hypoglycemia result of manual calibration data, and the values of S 1 and 0 respectively indicate hypoglycemia at night and no hypoglycemia at night, F(X) includes random forest Classification model, Logistic regression classification model, XGboost classification model and LightGBM classification model, ∈ represents the error term.

Preferably, predicting the risk probability of long-term blood sugar fluctuations is completed by a long-term blood sugar fluctuation evaluation model; the long-term blood sugar fluctuation evaluation model obtains the risk probability of long-term blood sugar fluctuations according to the X _length , and the X _length is as follows:

X _length = [Clu_index _mean , SD _mean , MEAN _mean , GV _mean , MAGE _mean , MODD _mean ];

Among them, Clu_index _mean , SD _mean , MEAN _mean , GV _mean , MAGE _mean and MODD _mean are respectively the average values of the clustering index Clu_index, SD, MEAN, GV, MAGE and MODD of the daily CGM dynamic map data within the set period .

Preferably, the cluster index Clu_index of the daily CGM dynamic atlas data is obtained using a hierarchical clustering model.

Preferably, the long-term blood glucose volatility evaluation model is shown in the following formula:

Among them, p _length is the risk probability of long-term blood glucose fluctuation, F(X _length ) includes random forest classification model and Logistic regression classification model, and ∈ represents the error term.

Preferably, the risk level of blood glucose control effect in the second set time period is obtained by using a decision tree model.

The present invention provides an edge computing device based on CGM dynamic map analysis, which uses the blood sugar analysis method based on CGM dynamic map as described in any one of the preceding items to perform blood sugar analysis. The edge computing device includes an MCU controller and a data communication module, and the MCU controls The device is connected to the data communication module; the data communication module transmits the data between the edge computing device and the terminal and the cloud data processing center; the MCU controller performs data preprocessing on the CGM dynamic map data received from the terminal and converts it into structured data and performs Analysis; wherein, the MCU controller includes a blood sugar fluctuation evaluation unit, a nighttime hypoglycemia risk warning unit and a blood sugar control periodic evaluation unit;

The blood sugar fluctuation evaluation unit predicts the risk probability of daily blood sugar fluctuation and the risk probability of long-term blood sugar fluctuation;

The nocturnal hypoglycemia risk warning unit predicts the probability of hypoglycemia risk at night;

The blood sugar control periodic evaluation unit evaluates the risk level of blood sugar control effect.

Preferably, the cloud data processing center stores map information.

Preferably, the risk probability of hypoglycemia during the day and night is transmitted to the terminal by the data communication module; the risk level of blood sugar control effect is transmitted to the terminal and the cloud data processing center by the data communication module, and the cloud data processing center judges whether the user of the terminal needs to be hospitalized.

Preferably, the MCU controller also includes a slice priority evaluation unit, which classifies users into 5G network slices with different network speeds according to the risk level of the blood sugar control effect, wherein users with worse blood sugar control effects are allocated to the corresponding 5G network slices A network slice with a higher network speed priority.

Preferably, the MCU controller further includes a CGM atlas data structured processing unit for converting the CGM dynamic atlas data into structured data.

Preferably, the MCU controller also includes an expandable unit for adding new units.

Preferably, the data communication module includes a 5G module and a communication interface unit, the 5G module transmits data between the edge computing device and the cloud data processing center and a terminal supporting 5G communication, and the communication interface unit transmits data between the edge computing device and the cloud data processing center. Data transmission between the processing center and terminals supporting non-5G communication.

Preferably, the edge computing device also includes a memory connected to the MCU controller, a power supply, an indicator light, a buzzer and a touch screen; the memory stores local data; the power supply is used to supply power to the edge computing device; The display of the running status; the buzzer is used for equipment abnormal alarm reminder; the touch screen is used to display and input the parameter settings of the edge computing device, and check the power status, space capacity, memory consumption and process status information of the edge computing device.

Compared with the prior art, the beneficial effects of the present invention are as follows:

1. A blood glucose analysis method and edge computing device based on CGM dynamic atlas of the present invention uses the daytime CGM dynamic atlas data and the previous day's nighttime CGM dynamic atlas data to predict the risk probability of hypoglycemia at night, and calculates the nighttime hypoglycemia risk probability of the current day. The risk probability is sent to the terminal to avoid the risk of threatening the user's life and health because the user may ignore the hidden hypoglycemia risk at night. By predicting the risk probability of hypoglycemia at night for the user as a reference, it is more effective to prevent nighttime hypoglycemia. Blood sugar risk.

2. A blood glucose analysis method and edge computing device based on a CGM dynamic atlas of the present invention uses daily CGM atlas data to predict the risk probability of daily blood glucose fluctuations, and can predict the risk probability of daily blood glucose fluctuations, while blood glucose fluctuations that change over time Differences in gender can reflect the patient's blood sugar control to a certain extent.

3. A blood glucose analysis method and edge computing device based on CGM dynamic atlas of the present invention, for long-term CGM dynamic atlas data longer than a set time period, predicts the risk probability of long-term blood sugar fluctuations, and gives the risk level of blood sugar control effect, so that Users can obtain the evaluation of the effect of blood sugar control, and timely prompt the risk of hospitalization.

4. A blood glucose analysis method and edge computing device based on CGM dynamic atlas of the present invention supports 5G network communication and provides more favorable transmission conditions for the collection of CGM dynamic atlas information.

5. A blood glucose analysis method and edge computing device based on CGM dynamic map of the present invention, the data is calculated on the edge computing device, and then transmitted to the cloud data processing center, reducing cloud pressure, speeding up data processing speed and risk warning Efficiency: It improves the utilization rate of CGM dynamic map data, effectively taps the value of potential map data, effectively identifies blood sugar risks, improves user experience, and then effectively controls blood sugar, improving the actual application effect.

Description of drawings

Fig. 1 is a flow chart of a blood glucose analysis method based on a CGM dynamic atlas according to an embodiment of the present invention;

FIG. 2 is a flow chart of a blood sugar analysis system integrating a blood sugar analysis method based on CGM dynamic map analysis and an edge computing device according to an embodiment of the present invention;

3 is a schematic diagram of the connection structure between an edge computing device based on CGM dynamic map analysis and a terminal supporting 5G communication, a continuous glucose sensor module, and a cloud data processing center according to an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of an edge computing device based on CGM dynamic map analysis according to an embodiment of the present invention;

FIG. 5 is a schematic structural diagram of various modules in an MCU controller of an edge computing device based on CGM dynamic graph analysis according to an embodiment of the present invention.

In the figure, 1. MCU controller; 2. memory; 3. data communication module; 31. communication interface unit; 32. 5G module; 4. power supply; 5. indicator light; 6. buzzer; 7. touch screen; 11. CGM atlas data structured processing unit; 12. Blood glucose volatility evaluation unit; 13. Night hypoglycemia risk early warning unit; 14. Blood glucose control periodic evaluation unit; 15. Slice priority evaluation unit; 16. Expandable unit.

Detailed ways

In order to make the purpose, 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 in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments It is some embodiments of the present invention, but not all of them. Based on the implementation manners in the present invention, all other implementation manners obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

Example 1

Referring to FIG. 1 , according to a specific embodiment of the present invention, a blood glucose analysis method based on a CGM dynamic map of the present invention will be described in detail below.

The invention provides a blood glucose analysis method based on CGM dynamic atlas, comprising the following steps:

Convert user information and CGM dynamic map data into structured data through data preprocessing;

Obtain map information, map information includes map analysis information and map category; said map analysis information includes: time period map glucose maximum value, time period map glucose minimum value Min, time period map slope Slope, time period map glucose variation coefficient GV, time period map standard Deviation SD, time period map variance Var, time period map glucose mean Mean, time period map average blood sugar fluctuation range MAGE and time period map average blood glucose absolute difference MODD; the map categories include daytime map, nighttime map, full time map and optional time period Atlas;

Use the daily CGM map data and combine the map information to predict the risk probability of daily blood sugar fluctuations;

Use the daytime CGM dynamic map data and the previous day's nighttime CGM dynamic map data, and combine the map information to predict the risk probability of hypoglycemia at night;

Convert daily CGM dynamic map data and daily analysis results into structured data;

According to the long-term CGM dynamic map data of the first set time period, combined with the map information to predict the risk probability of long-term blood sugar fluctuations;

According to the long-term blood sugar fluctuation risk probability and the daily analysis results of the second set time period, the risk level of blood sugar control effect in the second set time period is given.

Example 2

2-3, according to a specific embodiment of the present invention, a system of the present invention integrating a blood glucose analysis method based on CGM dynamic map analysis and an edge computing device will be described in detail below.

The present invention provides a system that integrates a blood sugar analysis method based on CGM dynamic map analysis and an edge computing device. Before blood sugar analysis, the following steps are required:

S1: The user opens the blood glucose monitoring APP on the terminal to register, enters basic information and sets whether to allow access to the edge computing device;

Among them, basic information includes user information and blood glucose sensor information; basic information includes name, age, height, weight, diabetes type, and recent fasting blood glucose value; sensor information includes sensor factory date, id number, daytime blood glucose map collection time and nighttime Blood glucose map collection time, blood glucose sensor information can be obtained by scanning the QR code through the APP.

S2: If access is allowed, transmit user information, blood glucose sensor information and daily CGM dynamic map data to the edge computing device;

Among them, the daily map data includes 24 hours, the sampling interval is 5 minutes, and the sensor output blood glucose value of a total of 288 data points;

The edge computing device requests the transmission of CGM dynamic map data twice, the first time is for daytime blood sugar collection, and the second time is nighttime blood sugar collection.

The default collection time range of daytime blood glucose is 6:00-24:00 every day, users can set the time range of daytime blood glucose collection according to their daily work and rest habits, such as 5:30-23:30;

The default collection time of nighttime blood glucose is from 24:00 of the current day to 6:00 of the next day. The collection time of nighttime blood glucose is calculated according to the set daytime blood glucose collection time range. After the nighttime blood glucose collection is completed, a custom 24h CGM is allowed. Spectrum data transfer is complete.

Example 3

2-3, according to a specific embodiment of the present invention, a system of the present invention integrating a blood glucose analysis method based on CGM dynamic map analysis and an edge computing device will be described in detail below.

The present invention provides a system that integrates a blood sugar analysis method based on CGM dynamic map analysis and an edge computing device. When blood sugar analysis is performed, the following steps are included:

S3: MCU controller 1 converts user information and CGM map data into structured data and stores it in the cloud data processing center.

Among them, the cloud data processing center establishes user information table, sensor information table, blood sugar value record table, map analysis record table and user blood sugar control risk evaluation table, as shown in Table 1, Table 2, Table 3, Table 4 and Table 5 below .

As shown in Table 1, the fields of the user information table include user ID, user information, network access category, network priority, creation time, modification time and expandable fields, and the primary key is the user ID;

The network access category includes non-5G communication protocols and 5G communication protocols, represented by integers 1 and 2 respectively;

The network priority includes 4 slices with different network speed priorities, represented by integers 0, 1, 2, and 3 respectively, and 0 represents the highest priority network slice.

Table 1 User Information Table

As shown in Table 2, the sensor information table fields include sensor ID, sensor attribute, user ID, sensor status and expandable fields. The primary key is the sensor ID, and the foreign key is the user ID. The user information table can be associated with the user ID.

Table 2 Sensor information table

As shown in Table 3, the fields of the blood glucose record table include record serial number, blood glucose value, time, sensor ID and expandable fields. The primary key is the record serial number, and the foreign key is the sensor ID. The sensor information table can be associated with the sensor ID.

Table 3 blood glucose value recording table

As shown in Table 4, the fields of the map analysis record table include: record number, map analysis information, time, map category, sensor ID and expandable fields. The primary key is the record number, and the foreign key is the sensor ID. Sensor information can be associated with the sensor ID surface.

Spectrum analysis information includes: maximum value of glucose at time period map, minimum value of glucose at time period map Min, slope of time period map Slope, coefficient of variation GV of glucose at time period map, standard deviation SD of time period map, variance Var of time period map, average value of glucose at time period map, time period The average blood glucose fluctuation range MAGE of the map and the absolute difference MODD of the average blood glucose absolute difference of the time map.

For the full segment map, the map analysis information also includes the time percentage corresponding to the time of all glucose concentration segments in a day: TAR1 (181-250 mg/dL), TAR2 (>250 mg/dL), TIR (70-180 mg/dL), TIR (70-180 mg/dL), dL), TBR1 (54-69 mg/dL), and TBR2 (<54 mg/dL).

Map categories include: day map, night map, full-time map and optional time map (map selected according to a certain time period).

Table 4 Spectrum Analysis Record Form

As shown in Table 5, the fields of user blood glucose control risk evaluation table include record serial number, evaluation type, evaluation time, evaluation result, remark information, user ID and expandable fields. The primary key is the record serial number, and the foreign key is user ID. Can be associated with the user information table; the types and results of various blood sugar evaluations of the user will be written into the user's blood sugar control risk evaluation table.

Table 5 User Blood Sugar Risk Evaluation Form

Example 4

2-3, according to a specific embodiment of the present invention, a system of the present invention integrating a blood glucose analysis method based on CGM dynamic map analysis and an edge computing device will be described in detail below.

The present invention provides a system that integrates a blood sugar analysis method based on CGM dynamic map analysis and an edge computing device. When blood sugar analysis is performed, the following steps are included:

S4: Evaluate the daily blood glucose volatility for the daily CGM dynamic atlas data;

For the daily CGM dynamic atlas data, the standard deviation SD, segment glucose mean Mean, segment glucose variation coefficient GV, average blood glucose fluctuation range MAGE, and average blood glucose absolute difference MODD are used as data feature input, and the high risk of manual calibration results is 1. The low risk is 0 as the target output, and the daily blood sugar fluctuation evaluation model is established as follows:

Among them, p represents the risk probability of daily blood glucose fluctuation, F(x) includes random forest classification model and Logistic regression classification model, ∈ represents the error term, x represents the characteristic data of daily CGM dynamic map input, x=[SD,MEAN, GV, MAGE, MODD].

Furthermore, the random forest classification model and the Logistic regression classification model are implemented through the RandomForestClassifier module and the LogisticRegression module in the sklearn library in python, and model training is performed using manually calibrated data.

Furthermore, the daily blood sugar fluctuation evaluation model has been trained in the cloud data processing center, and is transmitted to the blood sugar fluctuation evaluation unit 14 of the edge computing device for calling, and outputs the daily blood sugar fluctuation risk probability p.

Furthermore, the formula of segmental glucose variation coefficient GV is as follows:

Furthermore, the average blood glucose fluctuation range MAGE formula is as follows:

Among them, C _max represents the peak glucose value at a fluctuation of the blood glucose curve, C _min represents the minimum value of a fluctuation corresponding to the trough, an effective fluctuation refers to (C _max -C _min )>1SD, and n represents the number of effective fluctuations within 24 hours.

Furthermore, the MODD formula of the mean absolute difference in blood glucose is as follows:

Among them, C _i represents the observed value of glucose at time i of the current day, C _i-24h represents the observed value of glucose at time i of the previous day, and k is the number of observed values.

Further, when the user replaces the blood glucose sensor, the data closest to the date of replacement is used to calculate the average blood glucose absolute difference MODD.

Example 5

2-3, according to a specific embodiment of the present invention, a system of the present invention integrating a blood glucose analysis method based on CGM dynamic map analysis and an edge computing device will be described in detail below.

The present invention provides a system that integrates a blood sugar analysis method based on CGM dynamic map analysis and an edge computing device. When blood sugar analysis is performed, the following steps are included:

S5: Using the nighttime map data of the previous day and the daytime map data of the current day to predict the risk of hypoglycemia at night, if the risk probability of hypoglycemia exceeds the set threshold, the risk warning is transmitted to the terminal.

Among them, the input data feature extraction is the minimum value of the glucose value G_Pre_min of the nighttime dynamic map (default 00:00-6:00) of the previous _day , and the maximum value of the glucose value of the daytime dynamic map (6:00-24:00) at night The ratio G_max_re between G_max _daytime and the maximum value G_Pre_max of _{the daytime} dynamic map glucose value of the previous day, G_max_re, the minimum value Gmin of the map in the last three hours of the daytime map (default 21:00-24:00) _{- 3h} , mean value Gmean _-3h and section slope Gslope _-3h .

Further, the formula of G_max_re is as follows:

Furthermore, if the user customizes the time range of the night segment and the day segment according to living habits, the above data characteristics can be summarized as the following input characteristic data X:

X=[Gpremin _nighttime , _Gmindaytime , G_max_re, Gmin _-3h , Gmean _-3h , Gslope _-3h ].

Further, the characteristic data X is input as the input, and the result of manual calibration data is 1 for hypoglycemia at night, and 0 for no hypoglycemia as the target output, and the early warning model of hypoglycemia risk at night is established, which is expressed as follows:

Among them, _plow indicates the probability of hypoglycemia at night, F(X) includes random forest classification model, Logistic regression classification model, XGboost classification model and LightGBM classification model, and ∈ represents the error term.

Furthermore, the XGboost classification model and the LightGBM regression classification model are implemented through the xgboost module and lightgbm module in python, and the model training is performed using manually calibrated data.

Furthermore, the day-night hypoglycemia warning model has been trained in the cloud data processing center, and is transmitted to the night-time hypoglycemia risk warning unit 13 of the edge computing device for calling, and the day-night hypoglycemia risk probability p _{is low} .

Example 6

2-3, according to a specific embodiment of the present invention, a system of the present invention integrating a blood glucose analysis method based on CGM dynamic map analysis and an edge computing device will be described in detail below.

The present invention provides a system that integrates a blood sugar analysis method based on CGM dynamic map analysis and an edge computing device. When blood sugar analysis is performed, the following steps are included:

S6: Organize the daily CGM map data and daily analysis results into structured data, store them and send them to the cloud data processing center. The daily analysis results include the daily blood sugar fluctuation risk probability p and the nighttime hypoglycemia probability p _low .

Further, each user _'s daily full-segment map data and map analysis information, daily blood sugar fluctuation risk probability p, and day low blood sugar risk probability p are organized into a json string format and then transmitted to the cloud data processing center.

S7: For the daily CGM dynamic atlas data of the first set time period (not less than 14 days), an evaluation of the long-term fluctuation of blood sugar is given;

Further, extract the TAR1, TAR2, TIR, TBR1 and TBR2 of the daily CGM dynamic map data of the first set time period (not less than 14 days) as feature input data, and use the hierarchical clustering model to obtain the daily CGM dynamic map data The cluster category index Clu_index.

Furthermore, the hierarchical clustering model is in the cloud data processing center. The characteristic data of TAR1, TAR2, TIR, TBR1 and TBR2 of many historical patients are used as the input of the hierarchical clustering model. According to the ratio of TIR from high to low, it is divided into The 35 categories are used as Clu_index, and the larger the Clu_index, the more obvious the fluctuation of blood sugar.

Furthermore, the hierarchical clustering model is trained by the cloud data processing center and saved to the blood glucose fluctuation evaluation unit of the edge computing device for calling.

Further, according to the hierarchical clustering model, the Clu_index of each day in the first set time period (not less than 14 days) is obtained, and the Clu_index in the first set time period (not less than 14 days) is averaged to obtain Clu_index _mean .

Further, SD, MEAN, GV, MAGE and MODD obtained every day within the first set time period (not less than 14 days) are averaged to obtain SD _mean , MEAN _mean , GV _mean , MAGE _mean and MODD _mean .

Further, organize the data into the input feature data X as _follows :

X _length = [Clu_index _mean , SD _mean , MEAN _mean , GV _mean , MAGE _mean , MODD _mean ];

Further, using the input characteristic data X _length , the artificially calibrated long-term blood sugar fluctuation is 1 for high risk and 0 for low risk as the target output, and the following long-term blood sugar fluctuation evaluation model is established as follows:

Among them, p _length is the long-term blood glucose fluctuation probability, and F(X _length ) includes random forest classification model and Logistic regression classification model.

Example 7

2-3, according to a specific embodiment of the present invention, a system of the present invention integrating a blood glucose analysis method based on CGM dynamic map analysis and an edge computing device will be described in detail below.

The present invention provides a system that integrates a blood sugar analysis method based on CGM dynamic map analysis and an edge computing device. When blood sugar analysis is performed, the following steps are included:

S8: According to the long-term blood sugar fluctuation probability and the daily analysis results of the second set time period (not less than 7 days), give the blood sugar control risk level of the second set time period; the daily analysis results include the daily blood sugar fluctuation risk probability and the risk probability of hypoglycemia fluctuations during the night.

Further, the glycosylated hemoglobin value (G HBA1C ) converted from _{the average} value _{plow_ean} of plow in the second set time period, _the average glucose value GV _mean in the second set time period (G _HBA1C ) and calculated The GMI index, as the feature data column X _input is as follows:

X _input = [p _long , p _{low_ean} , G _HBA1C , GMI];

Furthermore, the calculation formula of G _HBA1C is as follows:

G _HBA1C (%)=(GV _mean (/L)+2.59)/1.59;

Furthermore, the calculation formula of GMI is as follows:

GMI(%)＝3.31+0.02392*GV _mean (/dL);

Further, using the decision tree model, the risk level of blood sugar control is given. The risk level of blood sugar control is divided into four levels: low risk, medium risk, medium high risk and high risk, corresponding to 0, 1, 2 and 3 of the decision tree model. category output.

Further, the decision tree model is implemented by DecisionTreeClassifier in the sklearn library in python.

Furthermore, the blood sugar control risk assessment model is trained by the cloud data processing center and saved to the blood sugar control periodic assessment unit 14 of the edge computing device for calling.

S9: For users with medium-high risk and high-risk risk, the edge computing device sends the risk information to the terminal, and transmits the risk information to the cloud data processing center.

S10: The cloud data processing center receives the risk warning, further judges whether admission is necessary, and if so, notifies the user to be admitted to the hospital in time.

S11: If the terminal supports the 5G communication protocol, classify the user into the corresponding 5G slice network according to the blood sugar control risk level of the second set time period (not less than 7 days).

Further, the slice priority evaluation unit 15 will establish the relationship between the blood sugar control risk level of the second set time period (not less than 7 days) and the corresponding 5G network slice, extract the blood sugar control risk level from the user blood sugar risk evaluation table, and Modify the value of the network priority field in the user information table as an identifier for network slicing.

Example 8

3-5, according to a specific embodiment of the present invention, an edge computing device based on CGM dynamic map analysis of the present invention will be described in detail below.

The present invention provides an edge computing device based on CGM dynamic map analysis. The edge computing device uses the blood sugar analysis method based on CGM dynamic map as described in any of the previous embodiments to perform blood sugar analysis; including: MCU controller 1, memory 2, Data communication module 3, power supply 4, indicator light 5, buzzer 6 and touch screen 7;

The MCU controller controls the operation between the memory, power supply, indicator light, buzzer, touch screen and data communication module, and preprocesses the CGM dynamic map data transmitted from the terminal into structured data and analysis.

The memory 2 stores local data, and the local data includes CGM dynamic map data, algorithm models and parameter configuration files sent from the terminal.

The data communication module 3 is used for data transmission between the edge computing device and the terminal and the cloud data processing center.

The power supply 4 is used to supply power to the edge computing device.

The indicator light 5 is used to display the operating status of the equipment.

The buzzer 6 is used for equipment abnormal alarm reminder.

The touch screen 7 is used to display and input the parameter settings of the edge computing device, and check the power status, space capacity, memory consumption and process status information of the edge computing device.

Example 9

3-5, according to a specific embodiment of the present invention, an edge computing device based on CGM dynamic map analysis of the present invention will be described in detail below.

The present invention provides an edge computing device based on CGM dynamic map analysis. The edge computing device uses the blood sugar analysis method based on CGM dynamic map as described in any one of the previous embodiments to perform blood sugar analysis; the edge computing device includes an MCU controller 1; MCU The controller 1 includes: a CGM map data structured processing unit 11, a blood sugar fluctuation evaluation unit 12, a nighttime hypoglycemia risk warning unit 13, a blood sugar control periodic evaluation unit 14, a slice priority evaluation unit 15, an expandable unit 16 and a database .

The CGM atlas data structured processing unit 11 converts the CGM dynamic atlas data into structured data.

The blood sugar fluctuation evaluation unit 12 predicts the risk probability of daily blood sugar fluctuation and the risk probability of long-term blood sugar fluctuation.

The nighttime hypoglycemia risk warning unit 13 predicts the probability of the nighttime hypoglycemia risk.

The blood sugar control periodic evaluation unit 14 evaluates the risk level of the blood sugar control effect in the second set time period (not less than 7 days).

The slice priority evaluation unit 15 classifies users into 5G network slices with different network speeds according to the risk level of the blood sugar control effect, wherein users with poorer blood sugar control effects are allocated to users with higher network speed priorities in the corresponding 5G network slices. Network slicing.

The expandable unit 16 is for adding new units.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention are included within the protection scope of the present invention.

Claims (22)

1. A blood sugar analysis method based on CGM dynamic graph, is characterized in that, comprises the following steps: Convert user information and CGM dynamic map data into structured data through data preprocessing; Obtain map information; Use the daily CGM map data and combine the map information to predict the risk probability of daily blood sugar fluctuations; Use the daytime CGM dynamic map data and the previous day's nighttime CGM dynamic map data, and combine the map information to predict the risk probability of hypoglycemia at night; Convert daily CGM dynamic map data and daily analysis results into structured data; According to the long-term CGM dynamic map data of the first set time period, combined with the map information to predict the risk probability of long-term blood sugar fluctuations; According to the long-term blood sugar fluctuation risk probability and the daily analysis results of the second set time period, the risk level of blood sugar control effect in the second set time period is given. 2. A blood glucose analysis method based on CGM dynamic atlas according to claim 1, wherein the daily analysis results include the risk probability of daily blood glucose fluctuation and the risk probability of hypoglycemia fluctuation at night. 3. A blood glucose analysis method based on CGM dynamic atlas according to claim 1, wherein the sampling time of the daily CGM dynamic atlas data is 24 hours, and the sampling interval is 5 minutes. 4. A blood glucose analysis method based on CGM dynamic atlas according to claim 1, wherein the first set time period is not less than 14 days, and the second set time period is not less than 7 days, The daytime of the day is from 6:00 to 24:00 of the day, the nighttime of the previous day is from 0:00 to 6:00 of the day, and the last three hours of the day are from 21:00 to 24:00 of the day. 5. A blood glucose analysis method based on CGM dynamic atlas according to claim 4, wherein said atlas information includes atlas analysis information and atlas category; said atlas analysis information includes: time period atlas glucose maximum value Max, Period map glucose minimum value Min, period map slope Slope, period map glucose variation coefficient GV, period map standard deviation SD, time period map variance Var, time period map glucose mean Mean, time period map average blood glucose fluctuation range MAGE and time period map average blood glucose absolute Poor MODD; the map category includes day map, night map, full time map and optional time map. 6. A blood glucose analysis method based on a CGM dynamic map according to claim 5, characterized in that, for the full-time map, the map analysis information also includes the corresponding time of all glucose concentration segments appearing in a day percentage of time. 7. A blood glucose analysis method based on a CGM dynamic map according to claim 6, wherein the glucose concentration segments are divided into segments according to the number of mg of glucose concentration contained in each deciliter. 8. A blood glucose analysis method based on CGM dynamic atlas according to claim 5, wherein said prediction of daily blood glucose fluctuation risk probability is completed by a daily blood glucose fluctuation evaluation model; said daily blood glucose fluctuation The evaluation model obtains the risk probability of daily blood sugar fluctuations based on the input data x, x is as follows: x = [SD, Mean, GV, MAGE, MODD]. 9. a kind of blood sugar analysis method based on CGM dynamic atlas according to claim 8, is characterized in that, described daily blood sugar fluctuation evaluation model is as shown in the following formula:

Among them, p represents the risk probability of daily blood sugar fluctuation, y represents the risk result of manual calibration, and the values of 1 and 0 in y represent high risk and low risk of blood sugar fluctuation respectively. F(x) includes random forest classification model and Logistic For regression classification models, ∈ represents the error term. 10. A blood glucose analysis method based on CGM dynamic atlas according to claim 6, wherein the prediction of the nighttime hypoglycemia risk probability is completed by the same day night hypoglycemia risk early warning model; the same day night hypoglycemia risk The early warning model obtains the risk probability of hypoglycemia at night based on the input data X, X is as follows: X=[Gpremin _{at night} , Gmin _{during the day} , G_max_re, Gmin _-3h , Gmean _-3h , Gslope _-3h ]; Among them, G_max_re represents the ratio of the maximum value G_max of the daytime dynamic map glucose value _{of the day} to the maximum value G_Pre_max of the daytime dynamic map glucose value _{of the previous day, and Gpremin night represents} the minimum value of the nighttime dynamic map glucose value of the previous day , Gmin _{during the day} indicates the minimum value of the glucose value of the daytime dynamic map of the day, and Gmin _-3h , Gmean _-3h and Gslope _-3h respectively represent the minimum value and average value of the glucose value of the three hours before the reciprocal of the daytime dynamic map of the day and slope. 11. A kind of blood sugar analysis method based on CGM dynamic atlas according to claim 10, it is characterized in that, described day and night hypoglycemia risk warning model is as follows:

Among them, _plow indicates the probability of hypoglycemia at night, S indicates the nighttime hypoglycemia result of manual calibration data, and the values of S 1 and 0 respectively indicate hypoglycemia at night and no hypoglycemia at night, F(X) includes random forest Classification model, Logistic regression classification model, XGboost classification model and LightGBM classification model, ∈ represents the error term. 12. A blood glucose analysis method based on CGM dynamic atlas according to claim 6, wherein said prediction of long-term blood glucose fluctuation risk probability is completed by a long-term blood glucose volatility evaluation model; said long-term blood glucose fluctuation evaluation model is based on X _length , to obtain the risk probability of long-term blood sugar fluctuations, X _length is as follows: X _length = [Clu_index _mean , SD _mean , MEAN _mean , GV _mean , MAGE _mean , MODD _mean l: Among them, Clu_index _mean , SD _mean , MEAN _mean , GV _mean , MAGE _mean and MODD _mean are respectively the average values of the clustering index Clu_index, SD, MEAN, GV, MAGE and MODD of the daily CGM dynamic map data within the set period . 13 . The blood glucose analysis method based on CGM dynamic atlas according to claim 12 , wherein the cluster index Clu_index of the daily CGM dynamic atlas data is obtained using a hierarchical clustering model. 14 . 14. A kind of blood sugar analysis method based on CGM dynamic atlas according to claim 12, is characterized in that, described long-term blood sugar fluctuation evaluation model is shown in the following formula:

Among them, p _length is the risk probability of long-term blood glucose fluctuation, F(X _length ) includes random forest classification model and Logistic regression classification model, and ∈ represents the error term. 15. A blood glucose analysis method based on CGM dynamic atlas according to claim 1, characterized in that, the risk level of blood glucose control effect in the second set time period is obtained by using a decision tree model. 16. An edge computing device based on CGM dynamic map analysis, characterized in that the blood sugar analysis method based on CGM dynamic map according to any one of claims 1-15 is used for blood sugar analysis, and the edge computing device includes MCU controller and data communication module, the MCU controller is connected to the data communication module; the data communication module transmits the data between the edge computing device and the terminal and the cloud data processing center; the MCU controller performs data processing on the CGM dynamic map data received from the terminal After preprocessing, it is converted into structured data and analyzed; among them, the MCU controller includes a blood sugar fluctuation evaluation unit, a nighttime hypoglycemia risk warning unit, and a blood sugar control periodical evaluation unit; The blood sugar fluctuation evaluation unit predicts the risk probability of daily blood sugar fluctuation and the risk probability of long-term blood sugar fluctuation; The nocturnal hypoglycemia risk warning unit predicts the probability of hypoglycemia risk at night; The blood sugar control periodic evaluation unit evaluates the risk level of the blood sugar control effect. 17. An edge computing device based on CGM dynamic map analysis according to claim 16, wherein the cloud data processing center stores map information. 18. An edge computing device based on CGM dynamic map analysis according to claim 16, wherein the risk probability of hypoglycemia at night is transmitted to the terminal by the data communication module; the risk level of the blood sugar control effect is determined by The data communication module transmits to the terminal and the cloud data processing center, and the cloud data processing center judges whether the user of the terminal needs to be hospitalized for treatment. 19. An edge computing device based on CGM dynamic atlas analysis according to claim 16, wherein the MCU controller further includes a slice priority evaluation unit, which assigns the user 5G network slices classified into different network speeds, among which users with poorer blood sugar control effects are assigned to network slices with higher network speed priority in the corresponding 5G network slices. 20. The edge computing device based on CGM dynamic map analysis according to claim 16, wherein the MCU controller also includes a CGM map data structured processing unit, which converts the CGM dynamic map data into a structure data. 21. An edge computing device based on CGM dynamic graph analysis according to claim 16, wherein the MCU controller further includes an expandable unit for adding new units. 22. An edge computing device based on CGM dynamic graph analysis according to claim 16, wherein the data communication module includes a 5G module and a communication interface unit, and the 5G module processes data between the edge computing device and the cloud The data between the center and the terminal supporting 5G communication is transmitted, and the communication interface unit transmits the data between the edge computing device and the cloud data processing center and the terminal supporting non-5G communication.

Images