A calibration method, apparatus, device, and storage medium for an analyte sensor.

Abstract

This application discloses a calibration method, apparatus, device, and storage medium for an analyte sensor, relating to the field of biomedical technology. The method includes: constructing time-series data using sensitivity data of the analyte sensor collected within a first preset time interval; constructing a rate-of-change model of the sensitivity data at a target time scale based on a mapping relationship between a second preset time interval and the first preset time interval, and using a preset calibration curve model; wherein the second preset time interval is greater than the first preset time interval; determining the actual rate of change of the sensitivity data at the target time scale based on the mapping relationship, and constructing a loss function based on the rate-of-change model and the actual rate of change; obtaining target parameters by minimizing the loss function, and updating the preset calibration curve model based on the target parameters to obtain the calibration curve of the target analyte sensor. This application enables efficient and unified calibration of analyte sensors based on the common characteristics of sensor sensitivity change rates.

Description

Technical Field

[0001] This invention relates to the field of biomedical technology, and in particular to a calibration method, apparatus, device, and storage medium for an analytical analyte sensor. Background Technology

[0002] Existing continuous analyte monitoring systems rely on subcutaneous sensors to perform continuous monitoring for more than 10 days. During the operation of the analyte sensor under the skin, the sensitivity will drift over time. Based on the time drift of the sensor sensitivity, a time drift function of the sensitivity can be constructed, enabling the analyte sensor to operate without calibration during its service life.

[0003] During the manufacturing process of sensors in a continuous analyte monitoring system, differences in raw materials and process parameters can affect their core parameters, specifically manifesting as variations in sensor sensitivity. On the other hand, since the working principles of the sensors are the same, their reaction properties with the analyte are consistent; specifically, this means that the rate of change of the analyte sensor's sensitivity over time is consistent during operation.

[0004] Different types of analyte sensors may have inconsistent continuous output intervals, but they can generally complete measurement outputs within minutes. When using conventional regression methods to fit the sensitivity of analyte sensors to high-frequency raw signals, noise can be amplified, leading to numerical instability and making it difficult to guarantee consistency across different time scales. At the same time, the high frequency of analyte sensors can easily lead to instability in their continuous differences, and the sensitivity differences of analyte sensors are only consistent at lower resolutions.

[0005] Existing technologies fail to effectively balance the need for capturing subtle drifts in high-resolution signals with the need for stable modeling at low resolution. They are unable to isolate the interference of absolute differences between individual sensors on the drift model, nor can they achieve uniform calibration of different batches of sensors, resulting in large batch-to-batch errors and insufficient calibration efficiency and consistency.

[0006] In summary, how to achieve efficient and unified calibration of analyte sensors based on the common characteristics of sensor sensitivity change rates is a technical problem that urgently needs to be solved. Summary of the Invention

[0007] In view of this, the purpose of this invention is to provide a calibration method, apparatus, device, and storage medium for analyte sensors, which can achieve efficient and uniform calibration of analyte sensors based on the common characteristics of the rate of change of sensor sensitivity. The specific solution is as follows: In a first aspect, this application provides a calibration method for an analyte sensor, comprising: Time series data are constructed using the sensitivity data of the analyte sensor collected within the first preset time interval; Based on the mapping relationship between the second preset time interval and the first preset time interval, and using a preset calibration curve model, a rate of change model of the sensitivity data at the target time scale is constructed; the second preset time interval is greater than the first preset time interval. Based on the mapping relationship, the actual rate of change of the sensitivity data at the target time scale is determined, and a loss function is constructed based on the rate of change model and the actual rate of change; The target parameters are obtained by minimizing the loss function, and the preset calibration curve model is updated based on the target parameters to obtain the calibration curve of the target analyte sensor.

[0008] Optionally, after constructing the time series data, the method further includes: The time axis corresponding to the time series data is divided into several time intervals; For each time interval, a corresponding first benchmark sensitivity model is constructed based on the time series data within the time interval to characterize the sensitivity variation of the analyte sensor within the time interval. A second benchmark sensitivity model is constructed based on the first benchmark sensitivity model corresponding to each time interval and the weights corresponding to each first benchmark sensitivity model. In this model, the weights of the first benchmark sensitivity model corresponding to each time interval are summed to 1; the second benchmark sensitivity model is used to represent the sensitivity data drift pattern of sensors of the same type of analyte.

[0009] Optionally, the calibration method for the analyte sensor further includes: Determine the target product between the first preset parameter corresponding to the analyte sensor and the second reference sensitivity model, and construct the preset calibration curve model based on the sum of the target product and the second preset parameter corresponding to the analyte sensor.

[0010] Optionally, the step of constructing a rate-of-change model of the sensitivity data at the target time scale based on the mapping relationship between the second preset time interval and the first preset time interval, and using a preset calibration curve model, includes: Based on the mapping relationship, the time scale transformation relationship between the first preset time interval and the second preset time interval is determined; Based on the time scale transformation relationship and the preset calibration curve model, a rate of change model corresponding to the sensitivity data is constructed; the rate of change model is used to represent the predicted change degree of the sensitivity data under the target time scale, where the target time scale is the time scale corresponding to the second preset time interval.

[0011] Optionally, the time series data includes sensitivity data and the time index corresponding to the sensitivity data; Accordingly, determining the actual rate of change of the sensitivity data at the target time scale based on the mapping relationship includes: For any of the aforementioned time indices, a target time index is determined based on the time index and the mapping relationship; From the time series data, determine the first sensitivity data corresponding to the time index and the second sensitivity data corresponding to the target time index; The first actual rate of change corresponding to the time index is determined based on the first sensitivity data and the second sensitivity data; Based on the first actual rate of change corresponding to each time index, the actual rate of change corresponding to the sensitivity data is determined; the actual rate of change is the actual degree of change of the sensitivity data at the target time scale.

[0012] Optionally, constructing the loss function based on the rate of change model and the actual rate of change includes: The loss function is constructed based on the rate of change model, the actual rate of change, the preset regularization term, and the preset regularization weight coefficient; The preset regularization term is used to constrain the parameters of the second reference sensitivity model in the preset calibration curve model.

[0013] Optionally, the step of obtaining the target parameters by minimizing the loss function and updating the preset calibration curve model based on the target parameters to obtain the calibration curve of the target analyte sensor includes: The target parameters are obtained by minimizing the loss function; The target parameters are input into the preset calibration curve model to obtain the target calibration curve model; The calibration curve of the target analyte sensor is obtained based on the target calibration curve model and the preset individual parameters of the target analyte sensor.

[0014] Secondly, this application provides a calibration apparatus for an analyte sensor, comprising: The time series data construction module is used to construct time series data using the sensitivity data of the analyte sensor collected within a first preset time interval; The rate of change model construction module is used to construct a rate of change model of the sensitivity data at a target time scale based on the mapping relationship between the second preset time interval and the first preset time interval, and using a preset calibration curve model; the second preset time interval is greater than the first preset time interval. The loss function construction module is used to determine the actual rate of change of the sensitivity data at the target time scale based on the mapping relationship, and to construct a loss function based on the rate of change model and the actual rate of change; The calibration curve determination module is used to obtain target parameters by minimizing the loss function, and update the preset calibration curve model based on the target parameters to obtain the calibration curve of the target analyte sensor.

[0015] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the aforementioned calibration method for the analyte sensor.

[0016] Fourthly, this application provides a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned calibration method for an analyte sensor.

[0017] In this application, firstly, time-series data is constructed using the sensitivity data of the analyte sensor collected within a first preset time interval; then, based on the mapping relationship between the second preset time interval and the first preset time interval, and using a preset calibration curve model, a rate of change model of the sensitivity data at a target time scale is constructed; the second preset time interval is greater than the first preset time interval; subsequently, the actual rate of change of the sensitivity data at the target time scale is determined based on the mapping relationship, and a loss function is constructed based on the rate of change model and the actual rate of change; finally, the target parameter is obtained by minimizing the loss function, and the preset calibration curve model is updated based on the target parameter to obtain the calibration curve of the target analyte sensor. As can be seen from the above, this application first collects sensitivity data of the analyte sensor according to a first preset time interval to construct time series data; then, based on the mapping relationship between the second and first preset time intervals and combined with a preset calibration curve model, a sensitivity change rate model at the target time scale is constructed to avoid differential instability caused by high-frequency data through low-resolution scale conversion, making the change rate characterization more consistent; subsequently, the actual change rate of sensitivity data at the target time scale is determined based on the mapping relationship, and a loss function is constructed in combination with the change rate model; finally, the target parameters are obtained by minimizing the loss function, and the preset calibration curve model is updated based on the target parameters to obtain the calibration curve of the target analyte sensor using the updated model. In this way, this application effectively avoids noise interference and differential instability of high-frequency data by modeling at a low-resolution target time scale, and at the same time, it isolates individual differences of sensors by using a unified change rate model to focus on the common drift patterns of analyte sensors, realizing efficient and unified calibration of batch analyte sensors. This not only improves the consistency and continuity of calibration results at different time scales, but also reduces batch errors, providing reliable support for the accurate and stable operation of continuous analytes. Attached Figure Description

[0018] 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.

[0019] Figure 1 A flowchart of a calibration method for an analyte sensor provided in this application; Figure 2 A schematic diagram showing the change of sensitivity data of a specific sensor for different analytes over time, provided in this application; Figure 3A schematic diagram illustrating the actual rate of change of sensitivity data for different analyte sensors provided in this application; Figure 4 A schematic diagram of a calibration device for an analyte sensor provided in this application; Figure 5 This application provides a structural diagram of an electronic device. Detailed Implementation

[0020] 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.

[0021] Existing continuous analyte monitoring systems rely on subcutaneously implanted sensors to achieve continuous monitoring for more than 10 days. During this subcutaneous operation, the sensitivity of these sensors exhibits time-varying drift. The manufacturing process of these sensors can be affected by variations in raw materials and process parameters, impacting their core parameters and resulting in differences in sensitivity between sensors. On the other hand, since the working principles of the sensors are the same, their reaction properties with the analyte are consistent, specifically, the rate of change of sensitivity over time is uniform during operation. While the continuous output intervals of different types of analyte sensors may differ, they generally complete measurement outputs within minutes. Using conventional regression methods to fit the sensitivity of analyte sensors to high-frequency raw signals can easily amplify noise, leading to numerical instability and difficulty in ensuring consistency across different time scales. Furthermore, the high frequency of analyte sensors can cause instability in their continuous differential values; only at lower resolutions do the sensitivity differences of analyte sensors exhibit consistency. Therefore, this application provides a calibration scheme for analyte sensors that can achieve efficient and uniform calibration based on the common characteristics of the sensor sensitivity change rate.

[0022] See Figure 1 As shown in the figure, an embodiment of the present invention discloses a calibration method for an analyte sensor, which may include: Step S11: Construct time series data using the sensitivity data of the analyte sensor collected within the first preset time interval.

[0023] In this embodiment, sensitivity data of the analyte sensor is first acquired at a high-resolution timescale. It should be noted that the high-resolution timescale corresponds to a first preset time interval for acquiring the sensitivity data. The first preset time interval refers to the physical time interval of continuous measurement by the analyte sensor; generally, the first preset time interval is preferably in the minute range, such as 1 minute, 3 minutes, 5 minutes or other equivalent time intervals.

[0024] Next, time series data are constructed based on the collected sensitivity data and the corresponding time indices. As shown below: ; in, It is a discrete-time index of a time series. The time index is The sensitivity data of the analyte sensor at that time. It can be seen that this embodiment introduces a discrete-time index at a high-resolution time scale. Sort the sensitivity data, i.e. This is used to indicate the sequential position of sensitivity data in time series data.

[0025] It should be noted that high-resolution time-series data is used to characterize the subtle drift of the analyte sensor sensitivity over time and serves as the basis for subsequent model inputs. Furthermore, the sensitivity data of different analyte sensors over time can be found in [reference needed]. Figure 2 As shown.

[0026] Step S12: Based on the mapping relationship between the second preset time interval and the first preset time interval, and using a preset calibration curve model, construct a rate of change model of the sensitivity data at the target time scale; the second preset time interval is greater than the first preset time interval.

[0027] It is understandable that sensitivity data will drift over time. Based on this time drift, a time drift function, i.e., a calibration curve model, can be constructed to enable the analyte sensor to operate without calibration during its service life. To construct the preset calibration curve model, the specific process may include: first, dividing the time axis corresponding to the time series data into several time intervals; then, for each time interval, constructing a corresponding first benchmark sensitivity model based on the time series data within that time interval to characterize the sensitivity change pattern of the analyte sensor within that time interval; subsequently, constructing a second benchmark sensitivity model based on the first benchmark sensitivity model corresponding to each time interval and the weights corresponding to each first benchmark sensitivity model; wherein the sum of the weights corresponding to the first benchmark sensitivity model corresponding to each time interval is 1; the second benchmark sensitivity model is used to represent the sensitivity data drift pattern of the same type of analyte sensor; finally, determining the target product between the first preset parameter corresponding to the analyte sensor and the second benchmark sensitivity model, and constructing the preset calibration curve model based on the sum of the target product and the second preset parameter corresponding to the analyte sensor.

[0028] Specifically, based on high-resolution time series data, the time axis is divided into m stage intervals, and a corresponding first baseline sensitivity model is constructed based on the time series data within each time interval. Then, based on the first baseline sensitivity model corresponding to each time interval and the weights corresponding to each first baseline sensitivity model, a second baseline sensitivity model is constructed, as shown below: ; in, This is the weighting function used to control the first reference sensitivity model in the j-th segment, and the weighting function must satisfy the condition that the sum of the weights corresponding to the first reference sensitivity model in each time interval is 1, i.e. .

[0029] In one specific implementation, The specific form is as follows: ; in, It is the index of the first sensitivity data point on the left side of the j-th time interval in the time series data. It is the index of the first sensitivity data point on the right side of the j-th time interval in the time series data, and h is the preset window width. It is the sigmoid function or characteristic function.

[0030] It should be noted that, It is a piecewise function used to control the rate and magnitude of change of time series data within the j-th time interval. Its form can be one or more combinations of linear, logarithmic, and exponential functions. It is a piecewise function The parameters can be adjusted according to the characteristics of the data, such as using linear functions in the steady phase and exponential functions or higher-order polynomials in the rapidly changing phase.

[0031] In the first specific implementation, when the piecewise function is a linear function, it can be constructed .

[0032] In the second specific implementation, when the piecewise function is in the form of a higher-order polynomial, it can be constructed .

[0033] In the third specific implementation, when the piecewise function is in the form of an exponential function, it can be constructed .

[0034] In the fourth specific implementation, when the piecewise function is in the form of a logarithmic function, it can be constructed .

[0035] Next, a preset calibration curve model for the sensitivity data is constructed, as shown below: ; in, It is the discrete-time index of the time series, where j is the j-th time interval. These are the preset parameters for different analyte sensors. It is the preset sensor gain. These are preset bias parameters.

[0036] As can be seen, the preset calibration curve model constructed in this embodiment to describe the time drift of the analyte sensor sensitivity can be transformed into a mathematical model with estimable parameters using a lower resolution, which is difficult to model directly.

[0037] It should be noted that, in this embodiment, a rate of change model for high-resolution time series data can be constructed at a low-resolution time scale based on a preset time scale transformation rule and a preset calibration curve model. Based on the mapping relationship between a second preset time interval and a first preset time interval, and using the preset calibration curve model, a rate of change model for sensitivity data at a target time scale is constructed. The specific process may include: first, determining the time scale transformation relationship between the first preset time interval and the second preset time interval based on the mapping relationship; then, constructing the rate of change model corresponding to the sensitivity data based on the time scale transformation relationship and the preset calibration curve model; the rate of change model is used to represent the predicted degree of change of the sensitivity data at the target time scale, where the target time scale is the time scale corresponding to the second preset time interval.

[0038] Specifically, determine the second preset time interval. With the first preset time interval Mapping relationship between As shown below: ; in, It is a positive integer.

[0039] In this way, the time scale transformation relationship between the first preset time interval and the second preset time interval can be obtained. That is, the time index interval at a low-resolution time scale is as follows: ; Furthermore, a model of the rate of change of sensitivity data at a low-resolution time scale is constructed, as shown below: ; in, For the rate of change model, For the preset calibration curve model, This represents a time-scale transformation relationship. This is the time index corresponding to the sensitivity data. The parameters are those in the preset calibration curve model. This represents the total number of sensitivity data points in the time series data.

[0040] Understandably, the rate of change model Based on time index interval It was constructed, but its index step size is different from the preset calibration curve model. They are the same, both being 1, which can be achieved using a sliding window.

[0041] In one specific implementation, when the first preset time interval The second preset time interval is 5 minutes. When the interval is 1 day, the time index interval is... .

[0042] Therefore, the rate of change model of sensitivity data with time index t at a low-resolution time scale can be obtained, as shown below: .

[0043] Step S13: Determine the actual rate of change of the sensitivity data at the target time scale based on the mapping relationship, and construct a loss function based on the rate of change model and the actual rate of change.

[0044] In this embodiment, the actual rate of change of sensitivity data for different analyte sensors is shown in [reference needed]. Figure 3 As shown. The process of determining the actual rate of change of sensitivity data at a target time scale based on the mapping relationship can include: for any given time index, firstly, determining a target time index based on the time index and the mapping relationship; then, determining the first sensitivity data corresponding to the time index and the second sensitivity data corresponding to the target time index from the time series data; subsequently, determining the first actual rate of change corresponding to the time index based on the first sensitivity data and the second sensitivity data; finally, determining the actual rate of change corresponding to the sensitivity data based on the first actual rate of change corresponding to each time index; the actual rate of change is the actual degree of change of the sensitivity data at the target time scale. Specifically, when the time index is... At that time, the target time index is determined based on the time index and mapping relationship. Then, based on the time index... Corresponding first sensitivity data With target time index Corresponding second sensitivity data To determine the actual rate of change of the sensitivity data at a low-resolution time scale, as shown below: .

[0045] It should be noted that constructing the loss function based on the rate of change model and the actual rate of change can specifically include: constructing the loss function based on the rate of change model, the actual rate of change, a preset regularization term, and a preset regularization weight coefficient; wherein, the preset regularization term is used to constrain the parameters of the second benchmark sensitivity model in the preset calibration curve model. Specifically, the loss function is as follows: ; in, The parameters are those in the preset calibration curve model. This is a predefined regularization term used to constrain the smoothness and physical plausibility of parameter and time series data estimates. The preset regularization weight coefficients are used. In one specific implementation, regularization... .

[0046] Step S14: Obtain the target parameters by minimizing the loss function, and update the preset calibration curve model based on the target parameters to obtain the calibration curve of the target analyte sensor.

[0047] In this embodiment, the target parameters are obtained by minimizing the loss function, and the preset calibration curve model is updated based on the target parameters to obtain the calibration curve of the target analyte sensor. The specific process may include: first, obtaining the target parameters by minimizing the loss function; then, inputting the target parameters into the preset calibration curve model to obtain the target calibration curve model; and finally, obtaining the calibration curve of the target analyte sensor according to the target calibration curve model and the preset individual parameters of the target analyte sensor.

[0048] Specifically, the target parameters are obtained by minimizing the loss function. , target parameters Substituting these values ​​into the second reference sensitivity model of the preset calibration curve model yields the target reference sensitivity model under the low-resolution rate of change constraint, as shown below: ; Next, the target reference sensitivity model is substituted into the preset calibration curve model to obtain the target calibration curve model. In application, only the preset individual parameters of the target analyte sensor need to be used. By substituting the values ​​into the target calibration curve model, the calibration curve of the target analyte sensor can be obtained, thereby achieving the calibration of the continuous analyte sensor.

[0049] In this way, by introducing a modeling method based on actual rate of change, this embodiment decouples the absolute differences between different analyte sensors caused by manufacturing variations and initial sensitivity differences. This avoids the impact of absolute shifts in the sensitivity of different analyte sensors on the preset calibration curve model, and ensures that different analyte sensors exhibit consistent relative rate of change characteristics at low-resolution time scales, thereby achieving unified modeling of analyte sensors. More specifically, this embodiment is based on the sensitivity of the analyte sensor at time intervals of... The rate of change model within the window width is used to invert the target reference sensitivity model of the analyte sensor. For analyte sensors from different batches or dates, a consistent target reference sensitivity model helps reduce batch-to-batch errors. When analyte sensors from different batches or production dates have different numerical parameters, this embodiment can, based on a consistent target reference sensitivity model, configure only the parameters related to the corresponding batch or production date in the algorithm program. and This allows for the integration of the target reference sensitivity model of the analyte sensor. Automatically generate calibration curves for the corresponding analyte sensors.

[0050] As can be seen from the above, in this embodiment, time series data is first constructed using the sensitivity data of the analyte sensor collected within a first preset time interval; then, based on the mapping relationship between the second preset time interval and the first preset time interval, and using a preset calibration curve model, a rate of change model of the sensitivity data at a target time scale is constructed; the second preset time interval is greater than the first preset time interval; subsequently, the actual rate of change of the sensitivity data at the target time scale is determined based on the mapping relationship, and a loss function is constructed based on the rate of change model and the actual rate of change; finally, the target parameter is obtained by minimizing the loss function, and the preset calibration curve model is updated based on the target parameter to obtain the calibration curve of the target analyte sensor. As can be seen from the above, in this embodiment, sensitivity data of the analyte sensor is first collected according to a first preset time interval to construct time series data. Then, based on the mapping relationship between the second and first preset time intervals and combined with a preset calibration curve model, a sensitivity change rate model at the target time scale is constructed to avoid differential instability caused by high-frequency data through low-resolution scale conversion, making the change rate characterization more consistent. Subsequently, the actual change rate of sensitivity data at the target time scale is determined based on the mapping relationship, and a loss function is constructed in combination with the change rate model. Finally, the target parameters are obtained by minimizing the loss function, and the preset calibration curve model is updated based on the target parameters to obtain the calibration curve of the target analyte sensor using the updated model. In this way, this embodiment effectively avoids noise interference and differential instability of high-frequency data by modeling at a low-resolution target time scale. At the same time, by using a unified change rate model to isolate individual differences of sensors and focus on the common drift patterns of analyte sensors, efficient and unified calibration of batch analyte sensors is achieved. This not only improves the consistency and continuity of calibration results at different time scales but also reduces batch-to-batch errors, providing reliable support for the accurate and stable operation of continuous analytes.

[0051] Accordingly, see Figure 4 As shown in the embodiments of this application, a calibration device for an analyte sensor is also provided, which may include: The time series data construction module 11 is used to construct time series data using the sensitivity data of the analyte sensor collected within a first preset time interval; The rate of change model construction module 12 is used to construct a rate of change model of the sensitivity data at a target time scale based on the mapping relationship between the second preset time interval and the first preset time interval, and using a preset calibration curve model; the second preset time interval is greater than the first preset time interval. The loss function construction module 13 is used to determine the actual rate of change of the sensitivity data at the target time scale based on the mapping relationship, and to construct a loss function based on the rate of change model and the actual rate of change; The calibration curve determination module 14 is used to obtain the target parameters by minimizing the loss function, and update the preset calibration curve model based on the target parameters to obtain the calibration curve of the target analyte sensor.

[0052] In some specific embodiments, the calibration device for the analyte sensor may further include: The time axis division module is used to divide the time axis corresponding to the time series data into several time intervals; The first benchmark sensitivity model construction module is used to construct a corresponding first benchmark sensitivity model for each time interval based on the time series data within the time interval, so as to characterize the sensitivity change law of the analyte sensor within the time interval. The second benchmark sensitivity model construction module is used to construct a second benchmark sensitivity model based on the first benchmark sensitivity model corresponding to each time interval and the weights corresponding to each of the first benchmark sensitivity models; wherein, the sum of the weights corresponding to the first benchmark sensitivity models corresponding to each time interval is 1; the second benchmark sensitivity model is used to represent the sensitivity data drift pattern of the same type of analyte sensor.

[0053] In some specific embodiments, the calibration device for the analyte sensor may further include: The preset calibration curve model construction module is used to determine the target product between the first preset parameter corresponding to the analyte sensor and the second reference sensitivity model, and to construct the preset calibration curve model based on the sum of the target product and the second preset parameter corresponding to the analyte sensor.

[0054] In some specific embodiments, the rate of change model construction module 12 may include: A time scale transformation relationship determination unit is used to determine the time scale transformation relationship between the first preset time interval and the second preset time interval based on the mapping relationship; The rate of change model construction unit is used to construct a rate of change model corresponding to the sensitivity data based on the time scale transformation relationship and the preset calibration curve model; the rate of change model is used to represent the predicted change degree of the sensitivity data under the target time scale, and the target time scale is the time scale corresponding to the second preset time interval.

[0055] In some specific embodiments, the time series data includes sensitivity data and the time index corresponding to the sensitivity data; Accordingly, the loss function construction module 13 may include: A target time index determination unit is configured to determine a target time index for any of the time indices based on the mapping relationship between the time index and the time index. The first sensitivity data determination unit is used to determine, from the time series data, the first sensitivity data corresponding to the time index and the second sensitivity data corresponding to the target time index; The first actual rate of change determination unit is used to determine the first actual rate of change corresponding to the time index based on the first sensitivity data and the second sensitivity data. The actual rate of change determination unit is used to determine the actual rate of change corresponding to the sensitivity data based on the first actual rate of change corresponding to each time index; the actual rate of change is the actual degree of change of the sensitivity data at the target time scale.

[0056] In some specific embodiments, the loss function construction module 13 may include: The loss function construction unit is used to construct the loss function based on the rate of change model, the actual rate of change, the preset regularization term, and the preset regularization weight coefficient; wherein, the preset regularization term is used to constrain the parameters of the second benchmark sensitivity model in the preset calibration curve model.

[0057] In some specific embodiments, the calibration curve determination module 14 may include: The target parameter acquisition unit is used to obtain the target parameters by minimizing the loss function; The target calibration curve model determination unit is used to input the target parameters into the preset calibration curve model to obtain the target calibration curve model; The calibration curve determination unit is used to obtain the calibration curve of the target analyte sensor based on the target calibration curve model and the preset individual parameters of the target analyte sensor.

[0058] Furthermore, embodiments of this application also disclose an electronic device, Figure 5This 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. 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 calibration method for the analyte sensor disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0059] 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.

[0060] 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.

[0061] 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 calibration method of the 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.

[0062] 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 calibration method for the 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.

[0063] 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.

[0064] 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.

[0065] 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.

[0066] 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.

[0067] 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 of calibrating an analyte sensor, the method comprising: The method comprises: constructing time series data by using sensitivity data of the analyte sensor collected in a first preset time interval; constructing a rate model of the sensitivity data in a target time scale according to a mapping relationship between a second preset time interval and the first preset time interval and by using a preset calibration curve model; the second preset time interval is greater than the first preset time interval; determining an actual rate of the sensitivity data in the target time scale based on the mapping relationship, and constructing a loss function according to the rate model and the actual rate; obtaining a target parameter by minimizing the loss function, and updating the preset calibration curve model based on the target parameter to obtain a calibration curve of a target analyte sensor.

2. The method of calibrating an analyte sensor of claim 1, wherein, After the time series data is constructed, the method further comprises: dividing a time axis corresponding to the time series data into a plurality of time intervals; for each time interval, constructing a corresponding first reference sensitivity model based on the time series data in the time interval to represent a corresponding sensitivity change rule of the analyte sensor in the time interval; constructing a second reference sensitivity model based on the first reference sensitivity model corresponding to each time interval and the weight corresponding to each first reference sensitivity model; wherein the sum of the weights corresponding to each first reference sensitivity model is 1; and the second reference sensitivity model is used to represent a sensitivity data drift rule of the same type of analyte sensor. The method further comprises:

3. The method of calibrating an analyte sensor of claim 2, wherein, determining a target product between a first preset parameter corresponding to the analyte sensor and the second reference sensitivity model, and constructing the preset calibration curve model according to the sum of the target product and a second preset parameter corresponding to the analyte sensor. The method further comprises:

4. The method of calibrating an analyte sensor of claim 1, wherein, determining a time scale transformation relationship between the first preset time interval and the second preset time interval based on the mapping relationship; constructing a rate model corresponding to the sensitivity data based on the time scale transformation relationship and the preset calibration curve model; the rate model is used to represent a predicted change degree of the sensitivity data in the target time scale, and the target time scale is a time scale corresponding to the second preset time interval. The time series data comprises sensitivity data and a time index corresponding to the sensitivity data; 5. The method of calibrating an analyte sensor of claim 1, wherein, Correspondingly, the method further comprises: for any time index, determining a target time index based on the time index and the mapping relationship; from the time series data, determining first sensitivity data corresponding to the time index and second sensitivity data corresponding to the target time index; determining a first actual rate corresponding to the time index according to the first sensitivity data and the second sensitivity data; ​ Determine an actual change rate corresponding to the sensitivity data based on the first actual change rate corresponding to each time index; the actual change rate is an actual change degree of the sensitivity data under a target time scale.

6. The method of calibrating an analyte sensor of claim 1, wherein, The loss function is constructed according to the change rate model and the actual change rate, including: The loss function is constructed according to the change rate model, the actual change rate, a preset regularization term, and a preset regularization weight coefficient. The preset regularization term is used to constrain parameters of a second reference sensitivity model in the preset calibration curve model.

7. The method of calibrating an analyte sensor according to any one of claims 1 to 6, wherein, The target parameter is obtained by minimizing the loss function, and the preset calibration curve model is updated based on the target parameter to obtain a calibration curve of a target analyte sensor, including: The target parameter is obtained by minimizing the loss function; The target parameter is input into the preset calibration curve model to obtain a target calibration curve model; The calibration curve of the target analyte sensor is obtained according to the target calibration curve model and preset individual parameters of the target analyte sensor.

8. A calibration device for an analyte sensor, characterized by Including: The time series data construction module is configured to construct time series data by using sensitivity data of the analyte sensor collected in a first preset time interval; The change rate model construction module is configured to construct a change rate model of the sensitivity data under a target time scale according to a mapping relationship between a second preset time interval and the first preset time interval and by using a preset calibration curve model; the second preset time interval is greater than the first preset time interval; The loss function construction module is configured to determine an actual change rate of the sensitivity data under the target time scale based on the mapping relationship, and to construct a loss function according to the change rate model and the actual change rate; The calibration curve determination module is configured to obtain a target parameter by minimizing the loss function, and to update the preset calibration curve model based on the target parameter to obtain a calibration curve of a target analyte sensor.

9. An electronic device, comprising: The electronic device includes a processor and a memory; wherein the memory is used to store a computer program, the computer program is loaded and executed by the processor to realize the calibration method of the analyte sensor as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored, and the computer program is executed by a processor to realize the calibration method of the analyte sensor as claimed in any one of claims 1 to 7.

Images