In-vivo calibration parameter analysis method, apparatus, device, and medium based on continuous analyte sensor

Abstract

The application discloses an in-vivo calibration parameter analysis method and device based on a continuous analyte sensor, an equipment and a medium, relates to the field of data processing, and comprises the following steps: calculating a target in-vivo calibration parameter of a target sample based on current data of an analyte sensor and blood parameters; determining a parameter consistency statistical quantity corresponding to the target in-vivo calibration parameter, and judging whether the parameter consistency statistical quantity is greater than a preset threshold; if yes, determining that the in-vivo performance of the target sample is unqualified; if no, determining an optimal calibration parameter based on the target in-vivo calibration parameter, and recalculating the current data through the optimal calibration parameter to determine whether an evaluation index obtained by the recalculating meets a preset evaluation requirement and obtain a corresponding evaluation result; and determining whether the in-vivo performance of the target sample is qualified based on the evaluation result. Therefore, the optimal in-vivo calibration parameter of an analyte can be found more quickly and accurately, and the accuracy of parameter screening and the efficiency of in-vivo calibration parameter analysis are improved.

Description

Technical Field

[0001] This invention relates to the field of data processing applications, and in particular to a method, apparatus, device, and medium for analyzing in vivo calibration parameters based on a continuous analyte sensor. Background Technology

[0002] In recent years, with the rapid development of sensor technology, electrochemical technology, wireless communication technology, and mobile internet technology, continuous analyte monitoring devices based on human physiological indicators (such as blood glucose, lactate, and blood ketones) have been gradually applied in fields such as health management and disease prevention. Continuous analyte sensor monitoring devices utilize factory calibration technology, eliminating the need for fingertip calibration, reducing user discomfort, and improving the user experience.

[0003] Currently, most factory calibration methods use in vivo calibration parameter calculations based on mean values. This method is susceptible to outliers, and the calibration parameters obtained from mean calculations may not be the optimal in vivo calibration parameters to meet sample accuracy requirements. Another method involves standard least squares fitting of the effective current and analyte reference point values ​​to obtain a fitting equation, followed by grid search with a wide range of threshold values ​​for the calibration parameters. However, this method typically uses a large threshold range, and as the number of samples and calibration parameters increases, the time required for grid search increases exponentially, significantly increasing the time spent optimizing calibration parameters, resulting in slow processing and reduced overall data processing efficiency. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method, apparatus, device, and medium for in vivo calibration parameter analysis based on a continuous analyte sensor, which can find the optimal calibration parameters within the analyte more quickly and accurately, improving the accuracy of parameter selection and the efficiency of in vivo calibration parameter analysis. The specific solution is as follows:

[0005] In a first aspect, this application discloses an in vivo calibration parameter analysis method based on a continuous analyte sensor, comprising:

[0006] The target in vivo calibration parameters of the target sample are calculated based on the current data from the analyte sensor and blood parameters.

[0007] Determine the parameter consistency statistic corresponding to the calibration parameters in the target body, and determine whether the parameter consistency statistic is greater than a preset threshold;

[0008] If the parameter consistency statistic is greater than the preset threshold, the in vivo performance of the target sample is determined to be unqualified.

[0009] If the parameter consistency statistic is not greater than the preset threshold, then the optimal calibration parameter is determined based on the calibration parameter in the target body, and the current data is back-calculated using the optimal calibration parameter to determine whether the obtained evaluation index meets the preset evaluation requirements, and the corresponding evaluation result is obtained.

[0010] Based on the evaluation results, it is determined whether the in vivo performance of the target sample is qualified.

[0011] Optionally, the calculation of the target in vivo calibration parameters of the target sample based on the current data of the analyte sensor and blood parameters includes:

[0012] Collect current data and sample data of the analyte sensor during the wearing of the target sample, and determine the raw current data for each day and the filtered current value after offline filtering of the raw current data.

[0013] Calculate the difference between each of the current data and the corresponding filtered current value to obtain the daily current noise data;

[0014] Identify the current noise data to be filtered that is greater than a preset current noise threshold in the current noise data, and determine the current data to be removed for the day corresponding to the current noise data to be filtered.

[0015] The current data to be removed is removed from the original current data to obtain the target current data;

[0016] The target current data and blood parameters are fitted to obtain the target in vivo calibration parameters of the target sample.

[0017] Optionally, fitting the target current data and blood parameters to obtain the target in vivo calibration parameters of the target sample includes:

[0018] The target current data is processed by a preset compensation function to obtain compensated current data.

[0019] Based on the compensated current data, blood parameters, and in vivo calibration parameters, an expression function for the mean absolute relative deviation is constructed, and the minimum value of the expression function is calculated.

[0020] The in vivo calibration parameter corresponding to the minimum value of the function is used as the in vivo calibration parameter to be screened, and the target in vivo calibration parameter that meets the preset threshold range is selected from the in vivo calibration parameters to be screened.

[0021] Optionally, determining the parameter consistency statistic corresponding to the calibration parameters within the target body, and judging whether the parameter consistency statistic is greater than a preset threshold, includes:

[0022] Calculate the standard deviation and mean value of the calibration parameters in the target body, and use the ratio between the standard deviation and the mean value as the parameter consistency statistic of the calibration parameters in the target body.

[0023] Determine whether the parameter consistency statistic is greater than a preset threshold.

[0024] Optionally, if the parameter consistency statistic is not greater than the preset threshold, then determining the optimal calibration parameters based on the in vivo calibration parameters of the target body includes:

[0025] If the parameter consistency statistic is not greater than the preset threshold, the average value of the in-body calibration parameters is taken as the prior value, and the least squares fitting is performed based on the prior value using the target loss function to obtain the calibration parameters to be verified corresponding to the in-body calibration parameters.

[0026] The calibration parameter to be verified is verified. If the calibration parameter to be verified passes the verification, it is taken as the optimal calibration parameter.

[0027] Optionally, the step of verifying the calibration parameter to be verified, and if the calibration parameter to be verified passes the verification, then using the calibration parameter to be verified as the optimal calibration parameter, includes:

[0028] Set the grid search step size, search loop range, and error range;

[0029] Several verification calibration parameters within the search cycle range are determined by grid search based on the search step size and the search cycle range;

[0030] Determine the minimum value among the plurality of verification calibration parameters, and determine whether the difference between the calibration parameter to be verified and the minimum value is within the error range;

[0031] If the difference between the calibration parameter to be verified and the minimum value is within the error range, then the calibration parameter to be verified is taken as the optimal calibration parameter.

[0032] Optionally, the step of back-calculating the current data using the optimal calibration parameters to determine whether the obtained evaluation indicators meet the preset evaluation requirements and obtain the corresponding evaluation results includes:

[0033] The current data is back-calculated using the optimal calibration parameters to obtain the analyte concentration data;

[0034] The system statistically analyzes several analyte concentration evaluation indicators corresponding to the preset evaluation requirements, and determines whether the analyte concentration data meets the several analyte concentration evaluation indicators.

[0035] If the analyte concentration data meets the several analyte concentration evaluation indicators, then an evaluation result indicating that the target sample is qualified in vivo is generated.

[0036] If any of the analyte concentration data does not meet the evaluation criteria for the concentration of the analytes, an evaluation result indicating that the target sample is unqualified in vivo will be generated.

[0037] Secondly, this application discloses an in vivo calibration parameter analysis device based on a continuous analyte sensor, comprising:

[0038] The parameter calculation module is used to calculate the target in vivo calibration parameters of the target sample based on the current data of the analyte sensor and blood parameters.

[0039] The data comparison module is used to determine the parameter consistency statistics corresponding to the calibration parameters in the target body, and to determine whether the parameter consistency statistics are greater than a preset threshold.

[0040] The first performance determination module is used to determine that the in vivo performance of the target sample is unqualified if the parameter consistency statistic is greater than the preset threshold.

[0041] The indicator evaluation module is used to determine the optimal calibration parameters based on the calibration parameters in the target body if the parameter consistency statistic is not greater than the preset threshold, and to perform back calculation on the current data using the optimal calibration parameters to determine whether the obtained evaluation indicators meet the preset evaluation requirements and obtain the corresponding evaluation results.

[0042] The second performance determination module is used to determine whether the in vivo performance of the target sample is qualified based on the evaluation results.

[0043] Thirdly, this application discloses an electronic device, comprising:

[0044] Memory, used to store computer programs;

[0045] A processor is used to execute the computer program to implement the in vivo calibration parameter analysis method based on a continuous analyte sensor as described above.

[0046] Fourthly, this application discloses a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the in vivo calibration parameter analysis method based on a continuous analyte sensor as described above.

[0047] In this application, target in vivo calibration parameters of a target sample can be calculated based on current data from an analyte sensor and blood parameters; the parameter consistency statistic corresponding to the target in vivo calibration parameters can be determined, and it can be determined whether the parameter consistency statistic is greater than a preset threshold; if the parameter consistency statistic is greater than the preset threshold, the in vivo performance of the target sample is determined to be unqualified; if the parameter consistency statistic is not greater than the preset threshold, the optimal calibration parameters are determined based on the target in vivo calibration parameters, and the current data is back-calculated using the optimal calibration parameters to determine whether the obtained evaluation index meets the preset evaluation requirements, thereby obtaining the corresponding evaluation result; based on the evaluation result, it is determined whether the in vivo performance of the target sample is qualified.

[0048] Therefore, the method of this application allows for the calculation of in vivo calibration parameters for a target sample based on current data from the analyte sensor and blood parameters. Then, the target in vivo calibration parameters need to be determined. The in vivo performance of the target sample is initially confirmed by comparing the parameter consistency statistic corresponding to the target in vivo calibration parameters with a preset threshold. If the parameter consistency statistic is greater than the preset threshold, the in vivo performance of the target sample can be directly determined to be unqualified. Conversely, if the parameter consistency statistic is greater than the preset threshold, further judgment is required. The optimal calibration parameters need to be determined using the target in vivo calibration parameters, and then the current data is recalculated using the optimal calibration parameters to determine whether the in vivo performance of the target sample is qualified based on the obtained evaluation indicators. This approach allows for faster and more accurate identification of the optimal calibration parameters within the analyte, facilitating factory calibration and in vivo / in vitro model construction. It provides a streamlined and modular approach to in vivo data analysis and parameter optimization, improving the accuracy of parameter selection and increasing the efficiency of in vivo calibration parameter analysis. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0050] Figure 1 This is a flowchart of an in vivo calibration parameter analysis method based on a continuous analyte sensor disclosed in this application;

[0051] Figure 2 This is a schematic diagram of the processing sequence of an in vivo calibration parameter analysis method based on a continuous analyte sensor disclosed in this application;

[0052] Figure 3aThis is a schematic diagram of the noise curve of a glucose sensor A disclosed in this application;

[0053] Figure 3b This is a schematic diagram of the noise curve of a glucose sensor B disclosed in this application;

[0054] Figure 4a This is a schematic diagram of the noise statistics of a glucose sensor A disclosed in this application;

[0055] Figure 4b This is a schematic diagram of the noise statistics of a glucose sensor B disclosed in this application;

[0056] Figure 5a This is a schematic diagram of the noise curve of a lactic acid sensor disclosed in this application;

[0057] Figure 5b This is a schematic diagram of the noise statistics of a lactic acid sensor disclosed in this application;

[0058] Figure 6 This is a schematic diagram of a grid search result disclosed in this application;

[0059] Figure 7 This is a schematic diagram of an in vivo calibration parameter analysis system based on a continuous analyte sensor disclosed in this application;

[0060] Figure 8 This is a schematic diagram of a cloud management platform and algorithm unit structure disclosed in this application;

[0061] Figure 9 This is a schematic diagram of the in vivo calibration parameter analysis device based on a continuous analyte sensor disclosed in this application;

[0062] Figure 10 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] Currently, in vivo calibration parameters are typically calculated using mean values. This method is susceptible to outliers, and the calibration parameters obtained from mean calculations may not be the optimal in vivo calibration parameters to meet sample accuracy requirements. Another approach involves standard least squares fitting of the effective current and analyte reference point values ​​to obtain a fitting equation, followed by grid search with a wide range of threshold values ​​for the calibration parameters. However, this method usually uses a large threshold range, and as the number of samples and calibration parameters increases, the time required for grid search increases exponentially, significantly increasing the optimization time for calibration parameters and slowing down the overall data processing efficiency.

[0065] To overcome the aforementioned technical problems, this application discloses a method, apparatus, device, and medium for analyzing in vivo calibration parameters based on a continuous analyte sensor. This method can find the optimal calibration parameters within the analyte more quickly and accurately, thereby improving the accuracy of parameter selection and the efficiency of in vivo calibration parameter analysis.

[0066] See Figure 1 As shown, this embodiment of the invention discloses an in vivo calibration parameter analysis method based on a continuous analyte sensor, comprising:

[0067] Step S11: Calculate the target in vivo calibration parameters of the target sample based on the current data of the analyte sensor and blood parameters.

[0068] In this embodiment, it is necessary to calculate the in vivo calibration parameters based on the data from the analyte sensor. Specifically, it is first necessary to collect the current data of the analyte sensor during the wearing of the target sample and the sample data, and determine the raw current data for each day and the filtered current value after offline filtering of the raw current data, such as... Figure 2 As shown, it is necessary to calculate the current noise data of a single analyte sensor within the same batch, i.e., the sensor during the target sample's wearing period. This requires determining the raw current data for each day using the analyte sensor's current data. And determine the filtered current value after offline filtering of the original current data. Then, it is necessary to calculate the difference between the daily current data and the corresponding filtered current values ​​to obtain the daily current noise data. It should be noted that the offline filtering method for the raw current data is SG (Savitzky-Golay) filtering. Further, it is necessary to identify the current noise data exceeding a preset current noise threshold to be filtered, and to determine the corresponding current data to be removed for the current day. Then, the current data to be removed is discarded from the raw current data to obtain the target current data.

[0069] Taking a glucose sensor as an example, with a signal sampling time of 3 minutes, a designed window length of 21, and a polynomial order of 2, the extracted noise curve is as follows: Figure 3a as well as Figure 3b As shown, the signal noise of sensor B is significantly higher than that of sensor A in the mid-term. In order to quantify the daily noise, statistical indicators of the daily noise can be calculated, such as the daily noise mean, noise 3 / 4 digit, noise 4 / 5 digit, and noise maximum value. The statistical indicators are shown in Table 1.

[0070] Table 1 Noise Statistics Table

[0071]

[0072] The preset current noise threshold needs to be determined based on noise statistics. For example, according to the data in Table 1, the 4 / 5 digit of the noise level can be used as the noise statistic using an empirical method. The optimal noise threshold is 0.2 nA, and the control threshold is 0.4 nA. The noise statistics of the noise curves of glucose sensor A and glucose sensor B are shown below. Figure 4a as well as Figure 4b As shown, the noise statistics of sensor A are all below the control threshold, while the noise statistics of sensor B exceed the control threshold on days 6, 7, 8, 9, and 10. Therefore, these data need to be discarded to avoid affecting the accuracy of the analysis. It should be noted that in this embodiment, only a glucose sensor is used as an example of the analyte sensor. The sensors used in this method include, but are not limited to, different types of analyte sensors such as lactate sensors, blood ketone sensors, uric acid sensors, blood urea nitrogen sensors, creatinine sensors, and ion sensors.

[0073] Taking lactic acid as the analyte sensor as an example, the sensor's signal sampling time is 1 minute, the designed window length is 61, the polynomial order is 2, and the noise level is empirically determined by taking 9 / 10 digits of the noise as the noise statistic. The optimal noise threshold is 0.1 nA, and the control threshold is 0.25 nA. The extracted noise curve is as follows. Figure 5a As shown, the noise statistics are as follows Figure 5b As shown, the sensor noise statistics reached the control threshold on the second day, and the noise statistics on the other days were all less than the control threshold.

[0074] Furthermore, such as Figure 2 As shown, it is necessary to fit the target current data and blood parameters to obtain the target in vivo calibration parameters of the target sample. Specifically, the target current data needs to be processed by a preset compensation function to obtain compensated current data. The preset compensation function is a temperature compensation function, and the expression of the compensated current data is as follows:

[0075] ;

[0076] in, The target current data is the response current of the analyte at time t. To compensate for the current data, the function 𝑓 is a temperature compensation function, and it is generally given by an empirical function from an in vitro temperature response experiment.

[0077] Then, based on the compensated current data, blood parameters, and in vivo calibration parameters, an expression function for the average absolute relative deviation needs to be constructed. Specifically, the compensated current data and blood parameters have a linear relationship, which can be expressed as:

[0078] ;

[0079] in, To predict the analyte values, The background current is typically provided by in vitro testing, and k is the in vivo calibration parameter. Furthermore, if an effective current value is set... Therefore, the expression showing a linear relationship between the compensated current data and blood parameters can be expressed as:

[0080] ;

[0081] Furthermore, the expression function for Mean Absolute Relative Difference (MARD) can be represented as:

[0082] ;

[0083] in, Here, k represents the blood reference point value, and N is the number of data pairs. The calibration parameter k mentioned above can be a candidate calibration parameter for a single sample, representing the blood reference point value. The median of the comparison values, k_median, is used as an approximation. Upper and lower thresholds, ths = ±0.2, can be set. Using a grid search method, the k value is calculated back to find the k that minimizes the MARD value as the optimal in vivo calibration parameter. These parameters are then summarized to obtain the in vivo calibration parameters to be screened. It should be noted that the compensated current data and blood parameters can also satisfy nonlinear relationships, such as polynomial, power function, and exponential function. The calibration parameters for the sample can be fitted using the least squares method based on the paired values ​​of the response current and blood reference points. Upper and lower thresholds can also be set, and the optimal calibration parameters for a single sample can be found through grid search.

[0084] Finally, target in vivo calibration parameters that meet the preset threshold range need to be selected from the selected in vivo calibration parameters. It should be noted that the preset threshold range is the threshold range of the MARD value. Since the MARD value corresponds to the in vivo calibration parameter, the in vivo calibration parameter corresponding to the selected MARD value is the target in vivo calibration parameter. For example, if the accuracy index is the MARD value and the threshold is set to 15%, samples with an accuracy index exceeding 15% are considered abnormal and are excluded. Table 2 shows the calculated MARD values ​​and in vivo calibration parameters for a batch of sensors implanted in subjects, as shown below:

[0085] Table 2 Sensor Parameter Data Table

[0086]

[0087] According to Table 2, the MARD values ​​of groups 7, 8, 9, and 10 exceed the threshold, so their corresponding in vivo calibration parameters need to be removed. The remaining in vivo calibration parameters are the target in vivo calibration parameters.

[0088] Step S12: Determine the parameter consistency statistics corresponding to the calibration parameters in the target body, and determine whether the parameter consistency statistics are greater than a preset threshold.

[0089] In this embodiment, it is necessary to calculate the parameter consistency statistic based on the in vivo calibration parameters of the target body. Specifically, it is necessary to calculate the standard deviation and mean value of the in vivo calibration parameters of the target body, and use the ratio between the standard deviation and the mean value as the parameter consistency statistic of the in vivo calibration parameters of the target body. It should be noted that in this embodiment, the parameter consistency statistic is represented by the coefficient of variation (CV), and its expression is as follows:

[0090] ;

[0091] in, For in vivo calibration parameters of the target, express standard deviation express The average value, taking the data in Table 2 as an example, after removing the data, the parameter consistency statistic corresponding to the calibration parameters in the target body is 4.99%. Then it is necessary to determine whether the obtained parameter consistency statistic is greater than the preset threshold.

[0092] Step S13: If the parameter consistency statistic is greater than the preset threshold, then the in vivo performance of the target sample is determined to be unqualified.

[0093] In this embodiment, as Figure 2As shown, if the parameter consistency statistic is greater than the preset threshold, the in vivo performance of the target sample can be directly determined to be unqualified.

[0094] Step S14: If the parameter consistency statistic is not greater than the preset threshold, then the optimal calibration parameter is determined based on the calibration parameter in the target body, and the current data is back-calculated using the optimal calibration parameter to determine whether the obtained evaluation index meets the preset evaluation requirements, and the corresponding evaluation result is obtained.

[0095] In this embodiment, if the parameter consistency statistic is not greater than a preset threshold, the optimal calibration parameters need to be determined first based on the in vivo transfer parameters. Specifically, the average value of the target in vivo calibration parameters needs to be calculated. As prior values, the target loss function is used to perform least-squares fitting based on the prior values ​​to obtain the calibration parameters to be verified corresponding to the calibration parameters in the target body. It should be noted that the residual function of the standard least-squares method is as follows:

[0096] ;

[0097] in, For the residual function, Blood reference point values, To predict analyte values, and based on the residuals optimized by the accuracy metric, if the accuracy metric is MARD, the optimized expression for the residuals as a function of the relative deviation is as follows:

[0098] ;

[0099] Furthermore, depending on the analyte being detected, a concentration threshold θ can be set to determine different accuracy indicators. Below the concentration threshold, absolute deviation can be used, while above the concentration threshold, relative deviation can be used. The optimized residual function is as follows:

[0100] ;

[0101] Where S1 is the scaling factor for absolute bias and S2 is the scaling factor for relative bias, used to standardize the mixed residuals. The formula for the scaling factor S is as follows:

[0102] ;

[0103] ;

[0104] in, This is the residual vector.

[0105] In a specific application scenario, if the analyte is glucose, the concentration threshold θ can be 3.9 mmol / L; if the analyte is lactic acid, the concentration threshold θ can be 4.5 mmol / L; and if the analyte is blood ketone, the concentration threshold θ can be 5.5 mmol / L.

[0106] In another specific application scenario, the analyte may have different emphases in different population segments. More granular segmentation intervals can be set, and weights can be assigned to the residuals for each interval, thus optimizing the residual function. as follows:

[0107] ;

[0108] in, For weight parameters, the square root value is used for ease of calculation. Let be the residual function, and if the analyte is glucose, then The range of values ​​for is as follows:

[0109] ;

[0110] If the analyte is lactic acid, then The range of values ​​for is as follows:

[0111] ;

[0112] Depending on the specific application scenario, an appropriate residual function can be selected to calculate the calibration parameters to be verified corresponding to the calibration parameters within the target body. However, to increase the robustness of the fit, the L2 loss of the standard least squares method can be replaced with Cauchy loss to construct an optimized target loss function. Let n be the number of data points. To obtain the residual vector from the residual function, the expression for the constructed target loss function L is as follows:

[0113] .

[0114] Furthermore, the calibration parameters to be verified, obtained by least-squares fitting through the target loss function, need to be validated to ensure that the optimal calibration parameters are found. Specifically, for example... Figure 2As shown, the grid search step size, search loop range, and error range can be set. Based on the search step size and search loop range, several verification calibration parameters within the search loop range are determined through grid search. It should be noted that calibration parameters are usually retained to two decimal places. To prevent errors caused by precision loss, the in vivo calibration parameter k fitted in Table 2, which is 1.07, can be used as a candidate value. The grid search step size is set to 0.01, the search loop range to 7 steps, and the error range ths = ±0.03 for grid search to ensure that the optimal calibration parameter is found. The minimum value among several verification calibration parameters is determined, and it is determined whether the difference between the calibration parameter to be verified and the minimum value is within the error range. If the difference between the calibration parameter to be verified and the minimum value is within the error range, then the calibration parameter to be verified is taken as the optimal calibration parameter. Figure 6 The results of the grid search are shown below, combined with... Figure 6 The search results were compared with the results of the standard least squares-large-scale grid search, and the comparison results are shown in Table 3 below:

[0115] Table 3 Comparison of Optimization Results

[0116]

[0117] Analysis of the data in Table 3 shows that the calibration parameters obtained using standard least squares differ from the optimal calibration parameters by 4 to 12 steps, while the calibration parameters obtained using the optimized objective loss function differ from the optimal calibration parameters by 0 to 1 step.

[0118] The calibration parameters obtained by the objective loss function are more accurate in locating the optimal calibration parameters compared to the standard least squares method, and the time consumption for small-scale grid search is also less than that for large-scale grid search. This verifies the effectiveness of combining the objective loss function with small-scale grid search for optimization.

[0119] Furthermore, the current data needs to be back-calculated using optimal calibration parameters to determine whether the obtained evaluation indicators meet the preset evaluation requirements and obtain the corresponding evaluation results. Specifically, the current data is back-calculated using optimal calibration parameters to obtain analyte concentration data. Several analyte concentration evaluation indicators corresponding to the preset evaluation requirements are statistically analyzed, and it is determined whether the analyte concentration data meets several analyte concentration evaluation indicators. For example, if the analyte is glucose, the accuracy indicators can be evaluated using grid error analysis, deviation analysis, and the pass rate within different deviation ranges; if the analyte is lactic acid, the accuracy indicators can be evaluated using deviation analysis and the deviation of the aerobic and anaerobic threshold points found during incremental load exercise; if the analyte is blood ketones, the accuracy indicators can be evaluated using segmented deviation analysis, such as segmented MARD analysis and Bland-Altman analysis. If the analyte concentration data meets several analyte concentration evaluation indicators, an evaluation result indicating that the target sample's in vivo performance is qualified is generated. It should be noted that qualified in vivo performance means that each indicator meets the corresponding indicator requirements. If the analyte concentration data contains concentration data that does not meet several analyte concentration evaluation indicators, an evaluation result indicating that the target sample's in vivo performance is unqualified will be generated. It should be noted that if any one of the indicators fails to meet the requirements, it indicates that the in vivo performance is unqualified.

[0120] Step S15: Determine whether the in vivo performance of the target sample is qualified based on the evaluation results.

[0121] In this embodiment, it is necessary to determine whether the in vivo performance of the target sample is qualified based on the evaluation results. Qualified in vivo performance means that each indicator meets the corresponding indicator requirements, and if any one indicator fails to meet the requirements, the in vivo performance is considered unqualified. Furthermore, as... Figure 2 As shown, if the performance of all samples in this batch is qualified, the in vivo calibration parameters corresponding to this batch of sensors can be generated for reference during subsequent data analysis.

[0122] In this application, target in vivo calibration parameters of a target sample can be calculated based on current data from an analyte sensor and blood parameters; the parameter consistency statistic corresponding to the target in vivo calibration parameters can be determined, and it can be determined whether the parameter consistency statistic is greater than a preset threshold; if the parameter consistency statistic is greater than the preset threshold, the in vivo performance of the target sample is determined to be unqualified; if the parameter consistency statistic is not greater than the preset threshold, the optimal calibration parameters are determined based on the target in vivo calibration parameters, and the current data is back-calculated using the optimal calibration parameters to determine whether the obtained evaluation index meets the preset evaluation requirements, thereby obtaining the corresponding evaluation result; based on the evaluation result, it is determined whether the in vivo performance of the target sample is qualified. Therefore, the method of this application can calculate the in vivo calibration parameters of the target sample based on the current data of the analyte sensor and blood parameters. Then, it is necessary to determine the target in vivo calibration parameters. The in vivo performance of the target sample is initially confirmed by comparing the parameter consistency statistic corresponding to the target in vivo calibration parameters with a preset threshold. If the parameter consistency statistic is greater than the preset threshold, the in vivo performance of the target sample can be directly determined as unqualified. On the other hand, if the parameter consistency statistic is greater than the preset threshold, further judgment is required. The optimal calibration parameters need to be determined using the target in vivo calibration parameters, and then the current data is recalculated using the optimal calibration parameters to determine whether the in vivo performance of the target sample is qualified based on the obtained evaluation index. In this way, the method of this application can effectively remove noisy and abnormal data from the sample, perform single-sample accuracy evaluation, effectively remove abnormal samples, and ensure the quality of the modeled samples. This method optimizes residual function design based on analyte accuracy metrics, selects loss functions with strong outlier resistance and robustness, and obtains in vivo calibration parameters that are closer to the optimal calibration coefficients. Small-scale grid search and batch sensor performance verification ensure the reliability of the selected calibration parameters. This method can find the optimal calibration parameters for the analyte more quickly and accurately, facilitating factory calibration and in vivo / in vitro model construction. It provides a streamlined and modular approach to in vivo data analysis and parameter optimization, improving the accuracy of parameter selection and enhancing the efficiency of in vivo calibration parameter analysis.

[0123] As a preferred embodiment, such as Figure 7 As shown, this application discloses an in vivo calibration parameter analysis system based on a continuous analyte sensor, including a continuous analyte monitoring device and continuous analyte monitoring application software connected thereto, as well as a cloud management platform connected to the continuous analyte monitoring application software and application software for access control of the platform.

[0124] The continuous analyte monitoring device monitors the current in the subcutaneous interstitial fluid of the implantee, which characterizes the analyte concentration, as well as the implantee's surface temperature. It receives calibration parameters from the continuous analyte monitoring application software, converts the current value into an estimated analyte concentration value, and sends it to the application software. The implantee can be a human or an animal, and the analytes include, but are not limited to, glucose, lactic acid, blood ketones, uric acid, urea nitrogen, creatinine, and ions. The continuous analyte monitoring device and its connected application software can connect via Bluetooth Low Energy. The application software receives calibration parameters from the cloud management platform and writes them to the device; it receives raw current values, temperature values, and estimated analyte concentration values ​​from the device and provides real-time display of analyte concentration, trends, and fluctuation characteristics. It can also record analyte reference values ​​and send this information to the cloud management platform. The application software supports platforms including, but not limited to, Android, iOS, and HarmonyOS.

[0125] Furthermore, such as Figure 8 As shown, the cloud management platform includes a data transmission unit, a data storage unit, a data preprocessing unit, an algorithm unit, and a log monitoring unit; it coordinates and controls the normal operation of the calibration parameter analysis system. The data transmission unit supports gRPC (Google Remote Procedure Call) and ZeroMQ (ZeroMessage Queue), responsible for receiving raw current, temperature, and concentration values ​​from the analyte sensor, as well as the actual reference value of the analyte, transmitted from the application software; it then transmits the obtained calibration parameters to the continuous analyte monitoring application software. The data storage unit contains various databases, which, depending on their function, can be used to record account and device information (e.g., PostgreSQL), a general-purpose database for storing logs (e.g., MongoDB), and a time-series database specifically for storing analyte time-series data (e.g., QuestDB). The data preprocessing unit, based on a Python environment, includes data cleaning, calculation, and time processing components such as pandas, numpy, and datetime, which can convert raw current into effective current and pair the actual reference value of the analyte with the effective current value at the time closest to the reference value.

[0126] Furthermore, such as Figure 8As shown, the algorithm unit includes interfaces for noise statistics and evaluation, single-sample optimal calibration parameter calculation, outlier exclusion, consistency evaluation, calibration parameter optimization and confirmation, and batch sensor performance evaluation. To implement these interface functions, the algorithm unit includes various Python environments and components such as scikit-learn, pytorch, scipy, pandas, numpy, and matplotlib visualization components required for model training and testing. Specifically, the noise statistics and evaluation interface can be used to extract current noise during analyte sensor wear and calculate statistics; the single-sample optimal calibration parameter calculation interface is used to calculate and find the optimal calibration parameters for a single sample; the outlier exclusion interface is used to exclude data with abnormal noise statistics and samples with accuracy indicators exceeding thresholds; the consistency evaluation interface is used to calculate calibration parameter consistency statistics and determine whether they exceed thresholds; the calibration parameter optimization and confirmation interface is used to calculate and confirm the optimal calibration parameters; and the batch sensor performance evaluation interface is used to statistically evaluate the performance indicators of batch analyte sensors. The log monitoring unit uses a Graylog cluster, indexed by Elasticsearch, and utilizes Kibana for real-time visualization and alarm configuration, monitoring system health and recording key events.

[0127] It should be noted that the access control application software includes front-end components such as the Angular framework to provide user interaction and to push and subscribe to cloud messages through Server-Sent Events, facilitating access to and control of the cloud management platform.

[0128] Taking a specific scenario as an example, the processing flow of the in vivo calibration parameter analysis system based on a continuous analyte sensor is as follows: The user activates the continuous analyte device, establishing a connection between the continuous analyte application software and the device. The application software requests calibration parameters from the cloud management platform based on the device's serial number and writes them into the analyte device. Then, the continuous analyte device sends the raw current value, temperature value, and estimated analyte concentration value to the application software via Bluetooth. The application software sends the account, sensor information, and analyte-related numerical information to the cloud management platform. The cloud management platform receives the data and stores it on the cloud server. Cloud engineers can access and control the application software to search for in vivo sample data from the analyte sensor. After data preprocessing and algorithm unit implementation, valid data is screened through noise statistics, interference is filtered through an abnormal sample exclusion mechanism, the optimal calibration parameters are confirmed using the calibration parameter optimization interface, and finally, the batch sensor performance is verified by the performance evaluation interface. The selected in vivo calibration parameters can be paired with the in vitro calibration parameters of the same batch to construct an in vivo and in vitro parameter model.

[0129] See Figure 9As shown, this embodiment of the invention discloses an in vivo calibration parameter analysis device based on a continuous analyte sensor, comprising:

[0130] The parameter calculation module 11 is used to calculate the target in vivo calibration parameters of the target sample based on the current data of the analyte sensor and blood parameters.

[0131] The data comparison module 12 is used to determine the parameter consistency statistics corresponding to the calibration parameters in the target body, and to determine whether the parameter consistency statistics are greater than a preset threshold.

[0132] The first performance determination module 13 is used to determine that the in vivo performance of the target sample is unqualified if the parameter consistency statistic is greater than the preset threshold.

[0133] The indicator evaluation module 14 is used to determine the optimal calibration parameter based on the calibration parameter in the target body if the parameter consistency statistic is not greater than the preset threshold, and to perform back calculation on the current data through the optimal calibration parameter to determine whether the obtained evaluation indicator meets the preset evaluation requirements and obtain the corresponding evaluation result.

[0134] The second performance determination module 15 is used to determine whether the in vivo performance of the target sample is qualified based on the evaluation results.

[0135] In this embodiment, the target in vivo calibration parameters of the target sample can be calculated based on the current data of the analyte sensor and blood parameters; the parameter consistency statistic corresponding to the target in vivo calibration parameters is determined, and it is determined whether the parameter consistency statistic is greater than a preset threshold; if the parameter consistency statistic is greater than the preset threshold, the in vivo performance of the target sample is determined to be unqualified; if the parameter consistency statistic is not greater than the preset threshold, the optimal calibration parameters are determined based on the target in vivo calibration parameters, and the current data is back-calculated using the optimal calibration parameters to determine whether the obtained evaluation index meets the preset evaluation requirements, and the corresponding evaluation result is obtained; based on the evaluation result, it is determined whether the in vivo performance of the target sample is qualified. Therefore, the method of this application allows for the calculation of in vivo calibration parameters for a target sample based on current data from the analyte sensor and blood parameters. Then, the target in vivo calibration parameters need to be determined. The in vivo performance of the target sample is initially confirmed by comparing the parameter consistency statistic corresponding to the target in vivo calibration parameters with a preset threshold. If the parameter consistency statistic is greater than the preset threshold, the in vivo performance of the target sample can be directly determined to be unqualified. Conversely, if the parameter consistency statistic is greater than the preset threshold, further judgment is required. The optimal calibration parameters need to be determined using the target in vivo calibration parameters, and then the current data is recalculated using the optimal calibration parameters to determine whether the in vivo performance of the target sample is qualified based on the obtained evaluation indicators. This approach allows for faster and more accurate identification of the optimal calibration parameters within the analyte, facilitating factory calibration and in vivo / in vitro model construction. It provides a streamlined and modular approach to in vivo data analysis and parameter optimization, improving the accuracy of parameter selection and increasing the efficiency of in vivo calibration parameter analysis.

[0136] In some embodiments, the parameter calculation module 11 may specifically include:

[0137] The data acquisition submodule is used to collect current data and sample data of the analyte sensor during the wearing of the target sample, and to determine the raw current data for each day and the filtered current value after offline filtering of the raw current data.

[0138] The difference calculation submodule is used to calculate the difference between each of the current data and the corresponding filtered current value to obtain the current noise data for each day.

[0139] The data determination submodule is used to determine the current noise data to be filtered that is greater than a preset current noise threshold in the current noise data, and to determine the current data to be removed for the current noise data to be filtered on the same day.

[0140] The data removal submodule is used to remove the current data to be removed from the original current data to obtain the target current data;

[0141] The first parameter fitting submodule is used to fit the target current data and blood parameters to obtain the target in vivo calibration parameters of the target sample.

[0142] In some embodiments, the first parameter fitting submodule may specifically include:

[0143] The data compensation unit is used to process the target current data through a preset compensation function to obtain the compensated current data.

[0144] The data calculation unit is used to construct an expression function of the average absolute relative deviation based on the compensated current data, blood parameters and in vivo calibration parameters, and to calculate the minimum value of the expression function.

[0145] The first parameter determination unit is used to take the in vivo calibration parameter corresponding to the minimum value of the function as the in vivo calibration parameter to be screened, and to screen out the target in vivo calibration parameter that meets the preset threshold range from the in vivo calibration parameter to be screened.

[0146] In some embodiments, the data comparison module 12 may specifically include:

[0147] The data determination unit is used to calculate the standard deviation and average value of the calibration parameters in the target body, and to use the ratio between the standard deviation and the average value as the parameter consistency statistic corresponding to the calibration parameters in the target body.

[0148] The data judgment unit is used to determine whether the parameter consistency statistic is greater than a preset threshold.

[0149] In some embodiments, the first performance determination module 13 may specifically include:

[0150] The second parameter fitting submodule is used to take the average value of the calibration parameters in the target body as the prior value if the parameter consistency statistic is not greater than the preset threshold, and perform least squares fitting based on the prior value through the target loss function to obtain the calibration parameters to be verified corresponding to the calibration parameters in the target body.

[0151] The parameter verification submodule is used to verify the calibration parameter to be verified. If the calibration parameter to be verified passes the verification, the calibration parameter to be verified is taken as the optimal calibration parameter.

[0152] In some embodiments, the parameter verification submodule may specifically include:

[0153] The parameter setting unit is used to set the grid search step size, search cycle range, and error range;

[0154] The verification parameter determination unit is used to determine a number of verification calibration parameters within the search cycle range based on the search step size and the search cycle range through grid search;

[0155] The parameter determination unit is used to determine the minimum value among the plurality of verification calibration parameters, and to determine whether the difference between the calibration parameter to be verified and the minimum value is within the error range;

[0156] The second parameter determination unit is used to determine the calibration parameter to be verified as the optimal calibration parameter if the difference between the calibration parameter to be verified and the minimum value is within the error range.

[0157] In some embodiments, the indicator evaluation module 14 may specifically include:

[0158] A data back-calculation unit is used to back-calculate the current data using the optimal calibration parameters to obtain analyte concentration data;

[0159] The indicator judgment unit is used to statistically analyze several analyte concentration evaluation indicators corresponding to preset evaluation requirements, and to determine whether the analyte concentration data meets the several analyte concentration evaluation indicators.

[0160] The first evaluation result generation unit is used to generate an evaluation result indicating that the target sample is qualified in vivo if the analyte concentration data meets the several analyte concentration evaluation indicators.

[0161] The second evaluation result generation unit is used to generate an evaluation result indicating that the in vivo performance of the target sample is unqualified if there are concentration data in the analyte concentration data that do not meet the evaluation indicators of the plurality of analyte concentrations.

[0162] Furthermore, embodiments of this application also disclose an electronic device, Figure 10 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.

[0163] Figure 10 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the in vivo calibration parameter analysis method based on a continuous analyte sensor disclosed in any of the foregoing embodiments. Alternatively, the electronic device 20 in this embodiment may specifically be a computer.

[0164] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0165] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0166] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the in vivo calibration parameter analysis method based on a continuous analyte sensor executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0167] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned disclosed in vivo calibration parameter analysis method based on a continuous analyte sensor. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0168] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0169] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0170] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0171] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0172] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for analyzing in vivo calibration parameters based on a continuous analyte sensor, characterized in that, include: The target in vivo calibration parameters of the target sample are calculated based on the current data from the analyte sensor and blood parameters. Determine the parameter consistency statistic corresponding to the calibration parameters in the target body, and determine whether the parameter consistency statistic is greater than a preset threshold; If the parameter consistency statistic is greater than the preset threshold, the in vivo performance of the target sample is determined to be unqualified. If the parameter consistency statistic is not greater than the preset threshold, then the optimal calibration parameter is determined based on the calibration parameter in the target body, and the current data is back-calculated using the optimal calibration parameter to determine whether the obtained evaluation index meets the preset evaluation requirements, and the corresponding evaluation result is obtained. Based on the evaluation results, determine whether the in vivo performance of the target sample is qualified; Wherein, if the parameter consistency statistic is not greater than the preset threshold, then determining the optimal calibration parameters based on the in vivo calibration parameters of the target body includes: If the parameter consistency statistic is not greater than the preset threshold, the average value of the in-body calibration parameters is taken as the prior value, and the least squares fitting is performed based on the prior value using the target loss function to obtain the calibration parameters to be verified corresponding to the in-body calibration parameters. The calibration parameter to be verified is verified. If the calibration parameter to be verified passes the verification, the calibration parameter to be verified is taken as the optimal calibration parameter. The step of verifying the calibration parameter to be verified, and if the calibration parameter to be verified passes the verification, then the calibration parameter to be verified is taken as the optimal calibration parameter, includes: Set the grid search step size, search loop range, and error range; Several verification calibration parameters within the search cycle range are determined by grid search based on the search step size and the search cycle range; Determine the minimum value among the plurality of verification calibration parameters, and determine whether the difference between the calibration parameter to be verified and the minimum value is within the error range; If the difference between the calibration parameter to be verified and the minimum value is within the error range, then the calibration parameter to be verified is taken as the optimal calibration parameter.

2. The in vivo calibration parameter analysis method based on a continuous analyte sensor according to claim 1, characterized in that, The calculation of the target in vivo calibration parameters for the target sample based on the current data from the analyte sensor and blood parameters includes: Collect current data and sample data of the analyte sensor during the wearing of the target sample, and determine the raw current data for each day and the filtered current value after offline filtering of the raw current data. Calculate the difference between each of the current data and the corresponding filtered current value to obtain the daily current noise data; Identify the current noise data to be filtered that is greater than a preset current noise threshold in the current noise data, and determine the current data to be removed for the day corresponding to the current noise data to be filtered. The current data to be removed is removed from the original current data to obtain the target current data; The target current data and blood parameters are fitted to obtain the target in vivo calibration parameters of the target sample.

3. The in vivo calibration parameter analysis method based on a continuous analyte sensor according to claim 2, characterized in that, The process of fitting the target current data and blood parameters to obtain the target in vivo calibration parameters of the target sample includes: The target current data is processed by a preset compensation function to obtain compensated current data. Based on the compensated current data, blood parameters, and in vivo calibration parameters, an expression function for the mean absolute relative deviation is constructed, and the minimum value of the expression function is calculated. The in vivo calibration parameter corresponding to the minimum value of the function is used as the in vivo calibration parameter to be screened, and the target in vivo calibration parameter that meets the preset threshold range is selected from the in vivo calibration parameters to be screened.

4. The in vivo calibration parameter analysis method based on a continuous analyte sensor according to claim 1, characterized in that, The step of determining the parameter consistency statistic corresponding to the calibration parameters in the target body, and judging whether the parameter consistency statistic is greater than a preset threshold, includes: Calculate the standard deviation and mean value of the calibration parameters in the target body, and use the ratio between the standard deviation and the mean value as the parameter consistency statistic of the calibration parameters in the target body. Determine whether the parameter consistency statistic is greater than a preset threshold.

5. The in vivo calibration parameter analysis method based on a continuous analyte sensor according to any one of claims 1 to 4, characterized in that, The step of back-calculating the current data using the optimal calibration parameters to determine whether the obtained evaluation indicators meet the preset evaluation requirements and obtain the corresponding evaluation results includes: The current data is back-calculated using the optimal calibration parameters to obtain the analyte concentration data; The system statistically analyzes several analyte concentration evaluation indicators corresponding to the preset evaluation requirements, and determines whether the analyte concentration data meets the several analyte concentration evaluation indicators. If the analyte concentration data meets the several analyte concentration evaluation indicators, then an evaluation result indicating that the target sample is qualified in vivo is generated. If any of the analyte concentration data does not meet the evaluation criteria for the concentration of the analytes, an evaluation result indicating that the target sample is unqualified in vivo will be generated.

6. An in vivo calibration parameter analysis device based on a continuous analyte sensor, characterized in that, include: The parameter calculation module is used to calculate the target in vivo calibration parameters of the target sample based on the current data of the analyte sensor and blood parameters. The data comparison module is used to determine the parameter consistency statistics corresponding to the calibration parameters in the target body, and to determine whether the parameter consistency statistics are greater than a preset threshold. The first performance determination module is used to determine that the in vivo performance of the target sample is unqualified if the parameter consistency statistic is greater than the preset threshold. The indicator evaluation module is used to determine the optimal calibration parameters based on the calibration parameters in the target body if the parameter consistency statistic is not greater than the preset threshold, and to perform back calculation on the current data using the optimal calibration parameters to determine whether the obtained evaluation indicators meet the preset evaluation requirements and obtain the corresponding evaluation results. The second performance determination module is used to determine whether the in vivo performance of the target sample is qualified based on the evaluation results. The first performance determination module includes: The second parameter fitting submodule is used to take the average value of the calibration parameters in the target body as the prior value if the parameter consistency statistic is not greater than the preset threshold, and perform least squares fitting based on the prior value through the target loss function to obtain the calibration parameters to be verified corresponding to the calibration parameters in the target body. The parameter verification submodule is used to verify the calibration parameter to be verified. If the calibration parameter to be verified passes the verification, the calibration parameter to be verified is taken as the optimal calibration parameter. The parameter verification submodule includes: The parameter setting unit is used to set the grid search step size, search cycle range, and error range; The verification parameter determination unit is used to determine a number of verification calibration parameters within the search cycle range based on the search step size and the search cycle range through grid search; The parameter determination unit is used to determine the minimum value among the plurality of verification calibration parameters, and to determine whether the difference between the calibration parameter to be verified and the minimum value is within the error range; The second parameter determination unit is used to determine the calibration parameter to be verified as the optimal calibration parameter if the difference between the calibration parameter to be verified and the minimum value is within the error range.

7. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the in vivo calibration parameter analysis method based on a continuous analyte sensor as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, Used to store a computer program, wherein the computer program, when executed by a processor, implements the in vivo calibration parameter analysis method based on a continuous analyte sensor as described in any one of claims 1 to 5.

Images