A method and system for calculating the cause contribution degree of glycometabolic imbalance and evaluating joint intervention

Abstract

The application relates to the technical field of health monitoring data analysis, and discloses a cause contribution degree calculation and joint intervention evaluation method and system for sugar homeostasis imbalance, wherein the method comprises the following steps: acquiring a multi-source original monitoring data set; performing a rhythm disturbance basic input set construction task; performing a periodization abnormality representation set construction task; performing a preprocessing task; performing a cause contribution degree time sequence calculation task; performing joint intervention evaluation processing; and outputting a joint intervention evaluation result set. Compared with a scheme for performing coarse-grained evaluation on sugar metabolism abnormalities by relying on a daily average blood glucose index in the prior art, especially under the condition of long-term shift work and circadian rhythm disorder, the technical problem that dynamic migration identification of sugar homeostasis imbalance causes in a continuous time window cannot be realized. Since the application introduces multi-source heterogeneous data alignment modeling under a unified time axis and combines an individual circadian baseline fitting mechanism, the cause explanation capability of sugar homeostasis abnormality evaluation is improved.

Description

Technical Field

[0001] This invention relates to the field of health monitoring data analysis technology, and in particular to a method and system for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions. Background Technology

[0002] Currently, most methods for assessing glucose homeostasis imbalance rely primarily on fasting blood glucose, postprandial blood glucose, glycated hemoglobin, or single continuous glucose monitoring results, combined with limited lifestyle information, to make a static assessment of an individual's glucose metabolism status. While these methods have some application value in general physical examinations, basal metabolic risk assessments, and routine chronic disease management, their assessment granularity typically remains at the level of "whether it is abnormal" or "grading the degree of abnormality." They lack the ability to further analyze the pathogenic factors underlying glucose homeostasis imbalance, and are particularly difficult to identify the dynamic changes in which different causes alternately dominate at different times of the day.

[0003] For example, in individuals with long-term shift work, sleep deprivation, frequent nighttime eating, or circadian rhythm disorders, abnormal glucose homeostasis does not manifest as a single stable pattern. Instead, it may primarily present as morning hyperglycemia driven by cortisol rhythm shifts during the morning, elevated basal blood glucose during the night, and further as a decreased recovery rate and prolonged fluctuation duration after meals. Such abnormalities are usually influenced by multiple factors simultaneously, including sleep phase shifts, autonomic nervous system dysregulation, nighttime eating behavior, misaligned physical activity times, and circadian metabolic rhythm imbalances in the gut microbiota. Current technologies often use daily aggregation or single-test value analysis, making it difficult to quantify the coupled effects of these multiple factors over continuous time windows, and also unable to clearly indicate the changing contribution of each etiology across different time periods.

[0004] Furthermore, existing technologies, in generating intervention recommendations, often rely on overall blood glucose levels or simple genotyping results, directly outputting general suggestions such as dietary control, lifestyle adjustments, exercise management, or probiotic supplementation. They lack the ability to assess interventions at specific timeframes, targeting the dominant etiological processes. Especially under conditions of circadian rhythm disruption, the time windows, intensity, and synergistic effects of different interventions vary significantly. If a uniform static intervention model is still adopted, it can easily lead to mismatched intervention timing, unstable combined intervention effects, and generalized assessment results. Existing technologies cannot fully meet the needs for refined identification of glucose homeostasis imbalances, dynamic interpretation of etiologies, and synergistic assessment of combined interventions.

[0005] Therefore, there is an urgent need for a method that can continuously calculate the contribution of glucose homeostasis imbalance to the etiology and conduct collaborative assessment of joint interventions even under conditions of circadian rhythm disorder, interleaved changes in multi-source heterogeneous monitoring data, and alternating dominance of multiple etiologies. This would improve the accuracy of glucose homeostasis abnormality identification, the clarity of etiological explanation, and the pertinence and reliability of joint intervention decisions. Summary of the Invention

[0006] To address the aforementioned technical shortcomings, the purpose of this invention is to propose a method for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions. This method aims to solve the technical problem that existing technologies rely on daily average blood glucose levels for coarse-grained assessment of abnormal glucose metabolism, especially under conditions of long-term shift work and disrupted circadian rhythms, which cannot achieve dynamic migration identification of causes of glucose homeostasis imbalance within a continuous time window.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a method for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint intervention.

[0008] The methods for calculating the etiological contribution of glucose homeostasis imbalance and evaluating joint interventions include:

[0009] Step S10: Obtain the multi-source raw monitoring dataset of the target object within the preset monitoring period, and perform the rhythm perturbation basic input set construction task based on the multi-source raw monitoring dataset using a unified time axis reconstruction and time sequence alignment mechanism, and output the rhythm perturbation basic input set;

[0010] Step S20: Based on the rhythmic perturbation input set, the task of constructing a time-based anomaly representation set is performed using the individual day-night baseline fitting and rhythmic offset analysis mechanism, and the time-based anomaly representation set is output.

[0011] Step S30: Based on the time-delayed anomaly representation set, a time-delayed response decomposition and etiology candidate state mapping mechanism is used to perform preprocessing tasks and output the etiology candidate state input set;

[0012] Step S40: Based on the etiology candidate state input set, the constrained state space recursive solution mechanism is used to perform the etiology contribution time series calculation task, and output the etiology contribution time series result set;

[0013] Step S50: Based on the time series results of etiological contribution, use the Bliss independent collaborative evaluation method to perform joint intervention assessment and output the joint intervention assessment result set.

[0014] Preferably, step S10, which involves obtaining the multi-source raw monitoring dataset of the target object within a preset monitoring period, performing the rhythmic perturbation basic input set construction task based on the multi-source raw monitoring dataset using a unified time axis reconstruction and time series alignment mechanism, and outputting the rhythmic perturbation basic input set, specifically includes:

[0015] Step S101: Obtain the multi-source raw monitoring dataset of the target object within the preset monitoring period. The multi-source raw monitoring dataset includes continuous blood glucose monitoring data, sleep duration and sleep stage data, heart rate variability data, cortisol rhythm data, food time distribution data, physical activity period data, and gut microbiota diurnal metabolic rhythm data.

[0016] Step S102: Map the multi-source raw monitoring dataset to a unified time axis, and perform segmented sampling with a preset time window length Δt to obtain the unified observation vector for the k-th time window. , ;in, This represents the continuous blood glucose feature vector at the k-th time window. Represents the sleep feature vector. This represents the heart rate variability eigenvector. Represents the cortisol feature vector. This represents the feeding feature vector. Represents the activity feature vector. Represents the metabolic feature vector of the microbial community;

[0017] Step S103: and based on the unified observation vector The rhythmic perturbation base input set is constructed using a missing segment length selection interpolation completion strategy. .

[0018] Preferably, in step S101, the sampling interval for continuous blood glucose monitoring data is no more than 15 minutes; sleep duration and sleep stage data include the time of falling asleep, the time of waking up, the proportion of deep sleep, and the number of sleep interruptions; heart rate variability data includes RMSSD and LF / HF; cortisol rhythm data includes at least the first value after waking up in the morning, the daytime average, and the nighttime end value; food time distribution data includes at least the timestamp of each meal, carbohydrate intake, and calorie value; physical activity time data includes at least the start and end times of activity, steps, or metabolic equivalents; and gut microbiota diurnal metabolic rhythm data includes at least the short-chain fatty acid concentration values ​​and the gut microbiota rhythm index collected at two different intraday time periods.

[0019] Preferably, in step S20, the step of constructing a time-based anomaly representation set based on the rhythmic perturbation input set using an individual diurnal baseline fitting and rhythmic offset analysis mechanism, and outputting the time-based anomaly representation set, specifically includes:

[0020] Step S201: A diurnal baseline modeling unit is pre-constructed using a one-dimensional convolutional neural network and a long short-term memory network. This unit includes a local fluctuation feature extraction layer, a long-period temporal dependency modeling layer, and a baseline parameter regression output layer. The local fluctuation feature extraction layer extracts local fluctuation features from continuous time windows within the circadian perturbation input set. The long-period temporal dependency modeling layer models long-term dependencies across multiple diurnal cycles. The baseline parameter regression output layer outputs individual blood glucose diurnal baseline models, individual cortisol diurnal baseline models, and individual sleep phase reference models. The circadian perturbation input set is then input into the diurnal baseline modeling unit, which outputs the individual blood glucose diurnal baseline model. Individual cortisol diurnal baseline model and individual sleep phase reference model ;

[0021] Step S202: Based on individual blood glucose diurnal baseline model Individual cortisol diurnal baseline model and individual sleep phase reference model Calculate the blood glucose shift, cortisol shift, and sleep phase shift within a continuous time window;

[0022] Step S203: Linearly fuse the blood glucose offset, cortisol offset, and sleep phase offset under continuous time windows to construct and output a time-based abnormal characterization set.

[0023] Preferably, in step S202, the individual blood glucose diurnal baseline model is used. Individual cortisol diurnal baseline model and individual sleep phase reference model The steps for calculating blood glucose shift, cortisol shift, and sleep phase shift over a continuous time window specifically include:

[0024] Step S2021: Obtain the measured blood glucose value for the k-th time window. And based on individual blood glucose diurnal baseline model Obtain individual baseline blood glucose levels throughout the day Based on the measured blood glucose value and individual baseline blood glucose levels Constructing blood glucose offset , ;

[0025] Step S2022: Obtain the measured cortisol value for the k-th time window and base it on the individual cortisol diurnal baseline model. Obtain individual diurnal baseline values ​​of cortisol, and construct cortisol offset based on measured cortisol values;

[0026] Step S2023: Obtain the measured sleep phases for the k-th time window and base them on the individual sleep phase reference model. Obtain the individual sleep phase baseline value, and construct the sleep phase offset based on the measured sleep phase and the individual sleep phase baseline value.

[0027] Preferably, step S30, which involves performing a preprocessing task based on a time-delayed response decomposition and etiology candidate state mapping mechanism using a time-dependent anomaly representation set to output an etiology candidate state input set, specifically includes:

[0028] Step S301: Using the blood glucose shift, cortisol shift, and sleep phase shift of the time-phased abnormal characterization set as core abnormal variables, calculate the time-delay response of the core abnormal variables to continuous blood glucose fluctuations based on the convolutional time-delay response kernel function.

[0029] Step S302: Construct a candidate state vector for etiology based on the time-delay response quantity combined with the basic input set of rhythmic perturbation; wherein, the candidate state vector for etiology includes at least the state quantity of diurnal phase misalignment, the state quantity of enhanced hepatic glucose output driven by cortisol rhythm abnormality, the state quantity of decreased insulin sensitivity driven by autonomic nervous system dysregulation, the state quantity of delayed postprandial recovery driven by nighttime eating, and the state quantity of metabolic amplification driven by diurnal metabolic imbalance of the microbiota.

[0030] Step S303: Vectorize and normalize the candidate state vectors of etiology and output the input set of candidate states of etiology.

[0031] Preferably, step S40, which involves performing the time-series calculation of etiology contribution based on the etiology candidate state input set using a constrained state-space recursive solution mechanism and outputting the time-series result set of etiology contribution, specifically includes:

[0032] Step S401: Construct a state transition model and an observation model with the etiology candidate state input set as the hidden state. The state transition model is used to describe the recursive relationship of each etiology state quantity between adjacent time windows, and the observation model is used to describe the influence relationship of each etiology state quantity on the continuous abnormal blood glucose observation value.

[0033] Step S402: Based on the state transition model and observation model, perform filtering and updating on the etiology candidate state input set to obtain the etiology state estimate for each time window; wherein, the state transition equation used in the filtering and updating process is expressed as follows: The observation equation used is expressed as follows: ;in, Indicates the first State transition equations under a time window Indicates the first The pathogenesis state vector under each time window Represents the external input vector. Indicates process noise. Represents the state transition matrix. Represents the input action matrix; Represent the observation equation, Represents the observation matrix. Observation noise;

[0034] Step S403: Perform non-negative constraint normalization on the estimated values ​​of etiological state under each time window to obtain the etiological contribution vector corresponding to each time window. Based on the etiological contribution vector corresponding to each time window, construct the etiological contribution time series result set by splicing in the order of time windows.

[0035] This invention also provides a system for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions, including:

[0036] The rhythm perturbation basic input set construction module is used to obtain the multi-source raw monitoring dataset of the target object within a preset monitoring period. Based on the multi-source raw monitoring dataset, a unified time axis reconstruction and time sequence alignment mechanism is used to perform the rhythm perturbation basic input set construction task and output the rhythm perturbation basic input set.

[0037] The time-based anomaly characterization set construction module is used to perform the task of constructing a time-based anomaly characterization set based on the rhythmic perturbation input set using the individual day-night baseline fitting and rhythmic offset analysis mechanism, and outputs the time-based anomaly characterization set.

[0038] The etiology candidate state input set construction module is used to perform preprocessing tasks based on the time-period-based abnormal characterization set using a time-delay response decomposition and etiology candidate state mapping mechanism, and output the etiology candidate state input set.

[0039] The causal contribution time series calculation module is used to perform the causal contribution time series calculation task based on the causal candidate state input set using a constrained state space recursive solution mechanism, and output the causal contribution time series result set;

[0040] The joint intervention assessment module is used to perform joint intervention assessment processing based on the time series results set of etiological contribution using the Bliss independent collaborative evaluation method, and outputs the joint intervention assessment result set.

[0041] The present invention also provides a device for calculating and evaluating the contribution of causes to glucose homeostasis imbalance and for joint intervention, comprising: a memory, a processor, and a program for calculating and evaluating the contribution of causes to glucose homeostasis imbalance stored in the memory and executable on the processor. When the program for calculating and evaluating the contribution of causes to glucose homeostasis imbalance is executed by the processor, a method for calculating and evaluating the contribution of causes to glucose homeostasis imbalance is implemented.

[0042] The present invention also provides a computer program product, including a program for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions, wherein the program for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions is executed by a processor to implement the method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions.

[0043] The beneficial effects of this invention are as follows: By constructing a unified time axis reconstruction of multi-source heterogeneous monitoring data for continuous time windows, individual day-night baseline fitting, time-delay response decomposition, and constrained state space recursive solution link, this invention can improve the glucose homeostasis imbalance from traditional single detection results or static anomaly judgment to dynamic quantitative calculation of the contribution of etiology in the context of day-night rhythm disturbance. This effectively improves the identification accuracy and temporal interpretation ability of different etiological dominant relationships such as day-night phase misalignment, cortisol rhythm abnormality, autonomic nervous system disorder, nighttime eating influence, and day-night metabolic imbalance of gut microbiota.

[0044] Based on the time-series results of etiological contribution, this invention further employs a joint intervention evaluation mechanism to quantitatively analyze the synergistic effects among multiple intervention measures. This enables the output of more targeted joint intervention evaluation results for the dominant etiological factors within different time windows, thereby effectively avoiding the problems of static, generalized, and time-appropriate intervention recommendations in existing technologies. This improves the precision of glucose homeostasis imbalance intervention decisions, the ability of collaborative evaluation, and the reliability of practical applications. Attached Figure Description

[0045] 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a flowchart illustrating the first embodiment of a method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions according to the present invention.

[0047] Figure 2 This is a schematic diagram of continuous blood glucose curves and missing information completion effects, representing the first embodiment of a method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions according to the present invention.

[0048] Figure 3 This diagram illustrates the effects of sleep staging and sleep interruption on the first embodiment of a method for calculating the contribution of glucose homeostasis imbalance to the cause and evaluating joint interventions according to the present invention.

[0049] Figure 4This is a schematic diagram comparing the measured blood glucose curve with the individual's diurnal blood glucose baseline, representing the first embodiment of a method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions according to the present invention.

[0050] Figure 5 This is a schematic diagram of the offset and comprehensive abnormal characterization intensity of the first embodiment of the method for calculating the causal contribution of glucose homeostasis imbalance and evaluating joint intervention according to the present invention.

[0051] Figure 6 This is a schematic diagram of the device for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions according to the present invention. Detailed Implementation

[0052] 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.

[0053] Example 1: As Figure 1 The diagram shown is a flowchart of the first embodiment of the method for calculating the causal contribution of glucose homeostasis imbalance and evaluating joint intervention according to the present invention. The first embodiment of the method for calculating the causal contribution of glucose homeostasis imbalance and evaluating joint intervention according to the present invention is presented.

[0054] In the first embodiment, the method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint intervention includes:

[0055] Step S10: Obtain the multi-source raw monitoring dataset of the target object within the preset monitoring period, and perform the rhythm perturbation basic input set construction task based on the multi-source raw monitoring dataset using a unified time axis reconstruction and time sequence alignment mechanism, and output the rhythm perturbation basic input set;

[0056] It should be noted that the purpose of this step is not merely to summarize various monitoring data, but rather to pre-construct a basic input structure with a unified time scale, unified sampling granularity, and unified feature organization method for subsequent time-series calculation of etiological contribution. Since glucose homeostasis imbalance in circadian rhythm disorders typically manifests as coupled fluctuations across time periods, modalities, and physiological channels, if the traditional processing mode of "separate collection and storage by each monitoring device, followed by rough splicing" is still adopted, it is easy to cause problems such as temporal misalignment between data from different sources, inconsistent sampling density, and unclear correspondences of key events, thereby affecting the accuracy of subsequent abnormal characterization extraction and etiological state recursive solution. Therefore, in this step, the multi-source raw monitoring dataset preferably includes continuous blood glucose monitoring data, sleep duration and sleep stage data, heart rate variability data, cortisol rhythm data, food intake time distribution data, physical activity period data, and gut microbiota circadian metabolic rhythm data. The aforementioned data correspond to different dimensions of influence in the glucose homeostasis regulation process. Continuous blood glucose monitoring data characterizes glucose homeostasis fluctuations; sleep duration and sleep stage data characterize sleep structure disturbances; autonomic nervous system-related heart rate variability data characterizes sympathetic-parasympathetic regulation; cortisol rhythm data characterizes diurnal endocrine rhythm changes; food intake time distribution data characterizes the phase of exogenous energy intake; physical activity time data characterizes energy expenditure and peripheral metabolic activation periods; and gut microbiota diurnal metabolic rhythm data characterizes the potential amplification or buffering effect of gut microbiota rhythms on glucose metabolism. In practical implementation, the sampling timestamps corresponding to each type of raw monitoring data can be extracted first, and a unified reference clock can be established. Then, the data from different modalities can be uniformly resampled according to a preset time window length Δt to form aligned observation segments within the same time window. The preset time window length Δt can be set according to the basic sampling interval of continuous blood glucose monitoring data, preferably any one of 5 minutes, 10 minutes, 15 minutes or 30 minutes, to ensure that the continuous blood glucose fluctuation characteristics are not overly smoothed, while taking into account the mapping stability of low-frequency data such as sleep, activity and cortisol.

[0057] It is understandable that a unified timeline reconstruction and timing alignment mechanism is not limited to simple timestamp sorting and synchronization mapping, but may also include the following processing logic:

[0058] Firstly, for continuous blood glucose monitoring data, the original sampling interval can be directly used as a high-frequency reference axis, and threshold screening or local smoothing can be performed on abnormal jump values.

[0059] Secondly, for sleep stage data, the entire night's sleep segments can be divided into continuous time windows, and then the sleep stage labels can be mapped to quantifiable stage feature values, such as the encoding values ​​or proportion values ​​of the wakefulness, light sleep, deep sleep and REM sleep stages.

[0060] Third, for heart rate variability data, RMSSD, LF / HF and the rate of change of adjacent time windows can be calculated according to time windows to reflect the strength of autonomic nervous regulation and its fluctuation trend.

[0061] Fourth, for cortisol rhythm data, since its original detection frequency is usually lower than the continuous blood glucose monitoring frequency, the characteristic values ​​under the continuous time window can be formed by segmented interpolation of adjacent sampling points or by constraint completion based on the diurnal rhythm reference curve.

[0062] Fifth, for data on the distribution of eating time and the period of physical activity, an event window mapping method can be used to extend and map the start time, duration, intensity level, and corresponding calories or metabolic equivalent of each eating or activity period to each time window that overlaps with its time.

[0063] Sixth, for the diurnal metabolic rhythm data of gut microbiota, since it is often not continuous high-frequency sampling data, the rhythm index, short-chain fatty acid concentration and peak-valley difference characteristics of multiple sampling points within the day can be used to map to the corresponding time period, and then window-level representation can be completed through trend preservation.

[0064] For example, such as Figure 2 As shown in the figure, the horizontal axis represents a continuous time window under a unified time axis, and the vertical axis represents continuous blood glucose values. The dashed curve represents the original blood glucose monitoring results containing missing segments, the solid curve represents the continuous blood glucose curve after time axis alignment and completion processing, and the shaded area represents the time period when the original data was missing and processed by the completion strategy. As can be seen from the figure, after the nighttime eating event, the target subject's blood glucose level rose significantly and remained at a high level for a period of time afterward, indicating that abnormal glucose homeostasis has a significant temporal manifestation under circadian rhythm disruption. At the same time, by processing the missing segments, the blood glucose curve remains continuous overall, thus preserving the dynamic process of nighttime hyperglycemia formation, persistence, and decline more completely. Figure 3 As shown in the figure, the horizontal axis represents a continuous time window under a unified time axis, and the vertical axis represents the discrete state of different sleep stages. The awake, light sleep, deep sleep, and REM sleep stages are represented by different stage codes. The figure shows that the target subject exhibits significant sleep interruptions during nighttime sleep, and the overall sleep start and end times are delayed compared to a normal sleep schedule, indicating that their sleep rhythm has deviated from the conventional circadian rhythm pattern.

[0065] Step S20: Based on the rhythmic perturbation input set, the task of constructing a time-based anomaly representation set is performed using the individual day-night baseline fitting and rhythmic offset analysis mechanism, and the time-based anomaly representation set is output.

[0066] It should be noted that the "individual circadian baseline fitting" in this step refers to fitting the target subject's own circadian rhythm changes, cortisol rhythm changes, and sleep phase distribution patterns to the pre-established circadian rhythm perturbation input set within a preset monitoring period, constructing an individualized reference baseline that matches the target subject's actual physiological rhythm. The individualized reference baseline includes the individual's circadian blood glucose baseline, individual's circadian cortisol baseline, and individual's sleep phase reference baseline. If necessary, it may also include reference trends for autonomic nervous system fluctuations, eating phases, or activity phases closely related to the circadian rhythm. "Rhythm offset analysis" refers to comparing each observed value with its corresponding individual baseline value window by window for the actual observation results within a continuous time window, obtaining offset information used to characterize the degree of abnormality. The "time-specific abnormal characterization set" in this step refers to a set of temporal abnormal features formed by structuring and organizing abnormal characterization quantities such as the degree of blood glucose deviation, endocrine rhythm deviation, and sleep rhythm deviation within each time window, using continuous time windows as the basic organizational unit.

[0067] Understandably, this invention can transform the raw, multi-source rhythmic fluctuation information of the target object into more interpretable "relative offset information." For complex problems such as glucose homeostasis imbalance, which are influenced by diurnal rhythms, behavioral timing, and endocrine fluctuations, direct analysis of raw values ​​often makes it difficult to accurately determine whether a fluctuation belongs to normal rhythmic fluctuations or is part of an abnormal imbalance process. However, after establishing an individualized diurnal baseline and analyzing the offset through this step, the observation results under each time window have clear reference significance, making it easier to identify whether abnormalities accumulate at night, whether they are synchronized with morning hormone changes, or whether they occur in conjunction with sleep interruptions. The technical effect of this step is also reflected in its ability to significantly enhance the temporal localization capability of abnormality identification. Although traditional methods such as overall mean, daily mean, or single detection can draw some overall abnormality conclusions, they often cannot specify whether the abnormality mainly occurs at night, in the morning, in the early postprandial period, or in the postprandial recovery phase. This step, by constructing offsets and time-specific abnormality representations through time windows, makes the abnormality no longer an abstract overall conclusion, but a dynamic result that can be precisely pinpointed to a specific time period. This ability to pinpoint time periods is directly valuable for identifying whether the condition is dominated by "day-night phase misalignment," "cortisol rhythm abnormalities," or "sleep phase disorder."

[0068] For example, such as Figure 4As shown in the figure, the horizontal axis represents a continuous time window under a unified time axis, and the vertical axis represents blood glucose values. The solid line represents the measured blood glucose curve of the target subject within a continuous monitoring period, and the dashed line represents the individual's diurnal blood glucose baseline obtained by fitting the input set based on rhythmic perturbations. The shaded area represents a representative abnormal clustering period formed under the combined effects of nighttime eating, sleep rhythm disturbances, and related physiological perturbations. As can be seen from the figure, the measured blood glucose curve of the target subject during this period deviates significantly from the individual's diurnal blood glucose baseline. Furthermore, this deviation is not a momentary fluctuation at a single discrete point, but rather a continuous abnormal rise along multiple consecutive time windows. It is evident that this invention emphasizes "deviation relative to the individual diurnal baseline" rather than "crossing the boundary relative to a fixed threshold of the population." Even if the target subject itself exhibits a stable rhythmic shift, it can still identify whether an abnormal rise exceeding its normal rhythmic contour has occurred within a specific time period, thus providing a more targeted basis for rhythmic offset analysis and etiological contribution calculation in subsequent steps. Figure 5 As shown in the figure, the horizontal axis represents a continuous time window under a unified time axis, and the vertical axis represents various offsets and the overall anomaly intensity. Specifically, the first curve represents the blood glucose offset of the measured blood glucose relative to the individual's diurnal blood glucose baseline; the second curve represents the cortisol offset of the measured cortisol relative to the individual's diurnal cortisol baseline; the third curve represents the sleep phase offset intensity of the measured sleep phase relative to the individual's sleep phase reference baseline; and the fourth curve represents the overall anomaly representation intensity based on the fusion of multiple offset components. Shaded areas indicate periods of anomaly aggregation. As can be seen from the figure, within a specific period at night, the blood glucose offset, cortisol offset, and sleep phase offset intensity all continuously increase, forming a more significant peak aggregation area in the overall anomaly representation layer. Therefore, this invention does not simply output isolated judgments such as "elevated blood glucose," "sleep abnormality," or "high cortisol," but rather maps these offset information onto the same time scale, thereby identifying whether anomalies occur together within the same time window and whether they form a synergistic amplification effect.

[0069] Step S30: Based on the time-delayed anomaly representation set, a time-delayed response decomposition and etiology candidate state mapping mechanism is used to perform preprocessing tasks and output the etiology candidate state input set;

[0070] It should be noted that the "time-delay response decomposition" in this step refers to: analyzing the delayed effects and continuous transmission relationships between various abnormal representations and their preceding perturbation factors in the time-segmented abnormal representation set output in step S20, in order to identify that the abnormalities within the current time window are not solely determined by factors at the current moment, but may be formed by the cumulative effects of eating, sleep interruption, activity misalignment, cortisol rhythm shift, and gut microbiota rhythm imbalance in previous time windows. Preceding perturbation factors include nighttime eating events, sleep phase misalignment events, sleep interruption events, nighttime activity residue events, cortisol abnormal shift events, and gut microbiota diurnal rhythm amplitude decrease events. The "causal candidate state mapping mechanism" in this step refers to: jointly associating the time-delay response decomposition results with the basic input set of rhythmic perturbations in step S10 and the offset results in step S20, and mapping them into candidate state quantities with clear physiological significance according to the preset causal explanation framework. Candidate state quantities include state quantities with diurnal phase misalignment, state quantities with enhanced hepatic glucose output driven by cortisol rhythm abnormalities, state quantities with decreased insulin sensitivity driven by autonomic nervous system dysregulation, state quantities with delayed postprandial recovery driven by nighttime eating, and state quantities with metabolic amplification driven by diurnal metabolic imbalance of the gut microbiota. The "etiological candidate state input set" in this step refers to the set of state inputs formed by vectorizing and organizing the candidate state quantities under each time window, using continuous time windows as the organizational unit. Its purpose is to provide initial state inputs with etiological explanatory boundaries for the temporal calculation of etiological contribution in subsequent step S40, rather than directly outputting the final etiological conclusion.

[0071] Understandably, the anomaly representation results obtained in step S20 are further transformed into candidate causal state results that can participate in subsequent state recursive solutions, thus elevating the anomaly information from the "offset phenomenon layer" to the "causal state layer." By introducing time-delay response decomposition, this invention can identify the temporal transmission relationship between anomalies and preceding perturbations, thereby avoiding the simplistic interpretation of the current anomaly as a result directly triggered by the current cause. By introducing causal candidate state mapping, this invention can also consolidate dispersed multimodal anomaly features into state quantities with clear structures and explicit meanings, improving the stability, continuity, and interpretability of subsequent causal contribution calculations.

[0072] It should be understood that, compared to traditional techniques that directly infer the cause of an illness based on a single abnormal indicator at the current moment, this step no longer adopts the static attribution logic of "current phenomenon corresponds to current cause." Instead, it first analyzes the time-lag process of abnormal formation and then constructs candidate cause state inputs. Traditional methods often ignore the delayed effects of factors such as eating, sleep, hormones, and gut microbiota on abnormal glucose homeostasis, which can easily lead to one-sided or temporally misaligned cause judgments. This step, by explicitly incorporating preceding perturbation factors into the computational chain, makes the formation of candidate cause states more consistent with the actual transmission law of glucose homeostasis imbalance in the context of circadian rhythm disorders. Therefore, it has stronger causal coherence and cause explanation ability compared to traditional techniques. For example, if a target subject exhibits elevated blood glucose and cortisol shifts and persistent sleep phase shifts within the current time window corresponding to 03:00 AM, relying solely on the results of the current time window would typically only indicate the presence of obvious abnormalities, but would be difficult to determine the main cause of the abnormality. In this step, the preceding time window is further traced back to identify previous disturbances such as nighttime eating, continuous sleep interruption, and elevated cortisol levels at the end of the night. Based on these disturbances, the time-delayed response relationships of the current anomaly are calculated. Subsequently, these time-delayed response results are mapped to candidate etiological states such as postprandial recovery delay, diurnal phase misalignment, and enhanced liver glucose output. Therefore, the output of this step is no longer simply "the current anomaly is strong," but rather a candidate etiological state input that further characterizes "the current anomaly is mainly driven by the delayed nighttime eating effect, sleep phase disorder, and hormonal rhythm abnormalities," thus providing a more targeted basis for subsequent time-series calculations of etiological contribution.

[0073] Step S40: Based on the etiology candidate state input set, the constrained state space recursive solution mechanism is used to perform the etiology contribution time series calculation task, and output the etiology contribution time series result set;

[0074] It should be noted that the "constrained state-space recursive solution mechanism" in this step refers to: using the candidate state input set as the initial input of the hidden state, establishing the dynamic transition relationship of the state of the cause between adjacent time windows and the mapping relationship of the state of the cause to the abnormal observation results, and recursively updating it on continuous time windows to obtain the dynamic contribution results of various causes throughout the entire monitoring period. "Constraint" means that in the recursive solution process, not only the temporal continuity and observation consistency of the state of the cause are considered, but also the non-negativity, contribution normalization, and smoothness of abnormal fluctuations of the state of the cause are restricted to avoid results with unclear physiological meaning or numerical instability. The "time-series calculation task of cause contribution" in this step calculates the contribution ratio of various candidate causes to the current abnormal result for each of the multiple consecutive time windows within the preset monitoring period, and organizes the contribution results under each time window into a time-series sequence. The "Etiological Contribution Time Series Results Set" refers to a continuous time window set of etiological contributions, consisting of the contribution of diurnal phase misalignment, the contribution of increased hepatic glucose output driven by cortisol rhythm abnormalities, the contribution of decreased insulin sensitivity driven by autonomic nervous system dysfunction, the contribution of delayed postprandial recovery driven by nighttime eating, and the contribution of metabolic amplification driven by diurnal metabolic imbalance of the gut microbiota.

[0075] It should be understood that, compared to traditional etiology analysis methods based on single-time-point rule judgments, single-time-point classifications, or single-time-point scoring, the significant improvement in this step lies in the fact that it no longer treats each time window as an isolated, independent sample, but rather as a dynamic system with inherent connections within continuous time windows. Traditional methods often determine the etiology type or risk level separately at each moment. While this is computationally simple, it easily leads to frequent jumps in etiology conclusions between adjacent time windows. That is, the previous time period might be judged as dominated by sleep factors, the next time period immediately judges it as dominated by hormone factors, and the time after that period shifts to dominated by eating factors. This results in a lack of continuous logic in etiology explanation, which is not conducive to clinical or management personnel understanding the abnormality formation process. In this step, the present invention introduces a constrained state space recursive solution mechanism, so that the update of etiology status is simultaneously constrained by both the state of the previous time window and the current observation information, thereby improving the calculation of etiology contribution from "discrete static judgment" to "continuous dynamic estimation". For example, a target subject exhibits a persistently elevated blood glucose level between 01:00 and 04:00 during a continuous monitoring period, accompanied by sleep phase disturbances, elevated nocturnal cortisol levels, and a prior nighttime eating event. After processing in step S30, candidate etiology states are generated for each of these adjacent time windows. The states driven by nighttime eating (delayed postprandial recovery), diurnal phase misalignment, and enhanced hepatic glucose output) are all at high levels, while the state driven by diurnal metabolic imbalance in gut microbiota (metabolic amplification) remains at a moderate level. Using a traditional single-time-point judgment method, the anomaly might be interpreted as nighttime eating at 01:00, sleep disturbance at 02:00, and hormonal factors at 03:00, resulting in a discrete and abrupt interpretation. In this step, the etiology state estimation results from the previous time window are used to recursively update the current time window, and the results are recalibrated based on the current anomaly observations, resulting in a smoother and more continuous evolution of etiology contribution. For example, between 1:00 AM and 2:00 AM, the contribution of delayed postprandial recovery driven by nighttime eating may be relatively high; between 2:00 AM and 3:00 AM, the contributions of diurnal phase misalignment and increased hepatic glucose output driven by cortisol rhythm abnormalities gradually increase; after 3:00 AM, if sleep interruption persists, the contribution of decreased insulin sensitivity driven by autonomic nervous system dysfunction will also gradually increase. This time-series result set of etiological contributions not only explains which etiologies are driving the abnormality at the current moment, but also further illustrates the strength and weakness of each etiology across consecutive time windows. In other words, the output is no longer just a static conclusion at a single point in time, but rather provides a dynamic etiological evolution process where "the nighttime eating effect initially plays a dominant role, followed by a gradual increase in rhythm misalignment and hormonal abnormalities, further compounded by autonomic nervous system factors."

[0076] Step S50: Based on the time series results of etiological contribution, use the Bliss independent collaborative evaluation method to perform joint intervention assessment and output the joint intervention assessment result set.

[0077] It should be noted that the "Bliss Independent-Synergistic Evaluation Method" in this step refers to: based on the time-series results set of etiological contribution obtained in step S40, independently calculating the probability of effect of individual interventions targeting different etiologies within each time window, and estimating the expected synergistic effect when multiple interventions are implemented in combination according to the Bliss formula. The so-called "combined intervention evaluation treatment" includes calculating the independent effect probability and combined effect probability of different intervention combinations (e.g., time-restricted eating intervention, sleep correction intervention, cortisol regulation intervention, gut microbiota regulation intervention, and stem cell-assisted intervention) in each time window, forming a comprehensive result set that can be directly used for intervention strategy decision-making. The "combined intervention evaluation result set" includes the individual suitability score, combined synergistic effect score, and dominant etiological match degree for each intervention in each time window.

[0078] Understandably, by combining the contribution of etiology with the probability of intervention effect, it is possible to dynamically assess the suitability and potential synergistic effects of various interventions within a specific time window, thus providing a quantitative reference for individualized and time-phased joint interventions. Even if the target subject has multiple etiological abnormalities simultaneously, this step can calculate the expected improvement probability of single and multiple interventions based on the Bliss independent synergistic formula, making the intervention strategy more scientifically grounded and targeted.

[0079] It should be understood that, compared to traditional methods that typically evaluate only the effects of a single intervention or a static intervention strategy, the innovation of this step lies in introducing the time-series results of etiological contribution as input and employing the Bliss Independent Collaborative Evaluation method for joint probability calculation. This gives the joint intervention assessment temporal continuity, etiological matching, and the ability to quantify the synergistic effects of multiple interventions. Traditional static assessments often fail to reveal the synergistic effects of interventions under multi-factor coupling and struggle to distinguish which interventions are most effective against the dominant etiology within a specific time window. This step can simultaneously consider the interactions of multiple interventions, different time windows, and different etiologies, thereby optimizing the intervention combination and implementation timing.

[0080] For example, in the target subjects during the 2:00 AM to 3:00 AM time window, blood glucose shifts were primarily driven by nighttime eating, accompanied by sleep phase disturbances and cortisol rhythm abnormalities. Step S40 calculated the contribution of each etiology as follows: nighttime eating 0.6, sleep phase disturbance 0.25, and cortisol abnormality 0.15. The independent improvement probability for nighttime eating intervention was 0.65; the improvement probability for sleep correction intervention was 0.50; and the improvement probability for gut microbiota regulation intervention was 0.45. According to the Bliss independent synergistic evaluation method, the expected improvement probability of combining nighttime eating intervention and sleep correction intervention was 0.825; if gut microbiota regulation intervention was added, the combined improvement probability was 0.911. This quantifies the expected improvement effect of single interventions and multiple intervention combinations within this time window and matches it with the contribution of each etiology, forming a dynamic combined intervention evaluation result set. Decision-makers can use this information to determine which nighttime eating interventions and sleep correction interventions should be prioritized during this period, and adjust the timing of gut microbiota regulation interventions based on the weight of causal contribution, thereby achieving precise, individualized, and multi-factor synergistic interventions.

[0081] Example 2: Furthermore, the present invention provides a system for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions, which employs a method for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions as described in the above embodiments. This system can solve the technical problem of calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions. The beneficial effects of the system for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions provided by the present invention are the same as those of the method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions as described in the above embodiments. Other technical features of the system for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0082] Example 3: This invention provides a device for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions. Please refer to... Figure 6A device for calculating the contribution of causes to glucose homeostasis imbalance and for joint intervention assessment includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to execute the method for calculating the contribution of causes to ... A device for calculating the contribution of causes to glucose homeostasis imbalance and for joint intervention assessment may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An I / O interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows a device for calculating the contribution of causes to glucose homeostasis imbalance and for joint intervention assessment to exchange data with other devices wirelessly or via wired communication. Although the figure shows a device for calculating the contribution of causes to glucose homeostasis imbalance and for joint intervention assessment with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems may be implemented alternatively.

[0083] Example 4: This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for calculating the contribution of causes to glucose homeostasis imbalance and for joint intervention assessment. The computer program product provided by this invention can solve the technical problem of calculating the contribution of causes to ...

[0084] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this invention.

[0085] It should be understood that the various parts disclosed in this invention can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0086] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions, characterized in that, The methods include: Step S10: Obtain the multi-source raw monitoring dataset of the target object within the preset monitoring period, and perform the rhythm perturbation basic input set construction task based on the multi-source raw monitoring dataset using a unified time axis reconstruction and time sequence alignment mechanism, and output the rhythm perturbation basic input set; Step S20: Based on the rhythmic perturbation input set, the task of constructing a time-based anomaly representation set is performed using the individual day-night baseline fitting and rhythmic offset analysis mechanism, and the time-based anomaly representation set is output. Step S30: Based on the time-delayed anomaly representation set, a time-delayed response decomposition and etiology candidate state mapping mechanism is used to perform preprocessing tasks and output the etiology candidate state input set; Step S40: Based on the etiology candidate state input set, the constrained state space recursive solution mechanism is used to perform the etiology contribution time series calculation task, and output the etiology contribution time series result set; Step S50: Based on the time series results of etiological contribution, use the Bliss independent collaborative evaluation method to perform joint intervention assessment and output the joint intervention assessment result set.

2. The method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions as described in claim 1, characterized in that, Step S10 involves obtaining the multi-source raw monitoring dataset of the target object within a preset monitoring period, and then using a unified time axis reconstruction and time series alignment mechanism to construct the basic input set of rhythmic perturbation based on the multi-source raw monitoring dataset, and outputting the basic input set of rhythmic perturbation. Specifically, this includes: Step S101: Obtain the multi-source raw monitoring dataset of the target object within the preset monitoring period. The multi-source raw monitoring dataset includes continuous blood glucose monitoring data, sleep duration and sleep stage data, heart rate variability data, cortisol rhythm data, food time distribution data, physical activity period data, and gut microbiota diurnal metabolic rhythm data. Step S102: Map the multi-source raw monitoring dataset to a unified time axis, and perform segmented sampling with a preset time window length Δt to obtain the unified observation vector for the k-th time window. , ;in, This represents the continuous blood glucose feature vector at the k-th time window. Represents the sleep feature vector. This represents the heart rate variability eigenvector. Represents the cortisol feature vector. This represents the feeding feature vector. Represents the activity feature vector. Represents the metabolic feature vector of the microbial community; Step S103: and based on the unified observation vector The rhythmic perturbation base input set is constructed using a missing segment length selection interpolation completion strategy. .

3. The method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions as described in claim 2, characterized in that, In step S101, the sampling interval for continuous blood glucose monitoring data is no more than 15 minutes; sleep duration and sleep stage data include the time of sleep onset, the time of wakefulness, the proportion of deep sleep, and the number of sleep interruptions; heart rate variability data includes RMSSD and LF / HF; cortisol rhythm data includes at least the first time value after waking up, the daytime mean, and the nighttime end value; food time distribution data includes at least the timestamp of each meal, carbohydrate intake, and calorie value; physical activity time data includes at least the start and end times of activity, steps, or metabolic equivalents; gut microbiota diurnal metabolic rhythm data includes at least the short-chain fatty acid concentration values ​​and the gut microbiota rhythm index collected at two different intraday time periods.

4. The method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions as described in claim 1, characterized in that, In step S20, the step of constructing a time-based anomaly representation set based on the rhythmic perturbation input set using individual diurnal baseline fitting and rhythmic offset analysis, and outputting the time-based anomaly representation set, specifically includes: Step S201: A diurnal baseline modeling unit is pre-constructed using a one-dimensional convolutional neural network and a long short-term memory network. This unit includes a local fluctuation feature extraction layer, a long-period temporal dependency modeling layer, and a baseline parameter regression output layer. The local fluctuation feature extraction layer extracts local fluctuation features from continuous time windows within the circadian perturbation input set. The long-period temporal dependency modeling layer models long-term dependencies across multiple diurnal cycles. The baseline parameter regression output layer outputs individual blood glucose diurnal baseline models, individual cortisol diurnal baseline models, and individual sleep phase reference models. The circadian perturbation input set is then input into the diurnal baseline modeling unit, which outputs the individual blood glucose diurnal baseline model. Individual cortisol diurnal baseline model and individual sleep phase reference model ; Step S202: Based on individual blood glucose diurnal baseline model Individual cortisol diurnal baseline model and individual sleep phase reference model Calculate the blood glucose shift, cortisol shift, and sleep phase shift within a continuous time window; Step S203: Linearly fuse the blood glucose offset, cortisol offset, and sleep phase offset under continuous time windows to construct and output a time-based abnormal characterization set.

5. The method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions as described in claim 1, characterized in that, In step S202, based on the individual blood glucose diurnal baseline model Individual cortisol diurnal baseline model and individual sleep phase reference model The steps for calculating blood glucose shift, cortisol shift, and sleep phase shift over a continuous time window specifically include: Step S2021: Obtain the measured blood glucose value for the k-th time window. And based on individual blood glucose diurnal baseline model Obtain individual baseline blood glucose levels throughout the day Based on the measured blood glucose value and individual baseline blood glucose levels Constructing blood glucose offset , ; Step S2022: Obtain the measured cortisol value for the k-th time window and base it on the individual cortisol diurnal baseline model. Obtain individual diurnal baseline values ​​of cortisol, and construct cortisol offset based on measured cortisol values; Step S2023: Obtain the measured sleep phases for the k-th time window and base them on the individual sleep phase reference model. Obtain the individual sleep phase baseline value, and construct the sleep phase offset based on the measured sleep phase and the individual sleep phase baseline value.

6. The method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions as described in claim 1, characterized in that, Step S30, which involves performing a preprocessing task based on a time-delayed response decomposition and etiology candidate state mapping mechanism using a time-based anomaly representation set, and outputting an etiology candidate state input set, specifically includes: Step S301: Using the blood glucose shift, cortisol shift, and sleep phase shift of the time-phased abnormal characterization set as core abnormal variables, calculate the time-delay response of the core abnormal variables to continuous blood glucose fluctuations based on the convolutional time-delay response kernel function. Step S302: Construct a candidate state vector for etiology based on the time-delay response quantity combined with the basic input set of rhythmic perturbation; wherein, the candidate state vector for etiology includes at least the state quantity of diurnal phase misalignment, the state quantity of enhanced hepatic glucose output driven by cortisol rhythm abnormality, the state quantity of decreased insulin sensitivity driven by autonomic nervous system dysregulation, the state quantity of delayed postprandial recovery driven by nighttime eating, and the state quantity of metabolic amplification driven by diurnal metabolic imbalance of the microbiota. Step S303: Vectorize and normalize the candidate state vectors of etiology and output the input set of candidate states of etiology.

7. The method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions as described in claim 1, characterized in that, Step S40, which involves performing time-series calculation of etiological contribution based on the etiological candidate state input set using a constrained state-space recursive solution mechanism and outputting a time-series result set of etiological contribution, specifically includes: Step S401: Construct a state transition model and an observation model with the etiology candidate state input set as the hidden state. The state transition model is used to describe the recursive relationship of each etiology state quantity between adjacent time windows, and the observation model is used to describe the influence relationship of each etiology state quantity on the continuous abnormal blood glucose observation value. Step S402: Based on the state transition model and observation model, perform filtering and updating on the etiology candidate state input set to obtain the etiology state estimate for each time window; wherein, the state transition equation used in the filtering and updating process is expressed as follows: The observation equation used is expressed as follows: ;in, Indicates the first State transition equations under a time window Indicates the first The pathogenesis state vector under each time window Represents the external input vector. Indicates process noise. Represents the state transition matrix. Represents the input action matrix; Represent the observation equation, Represents the observation matrix. Observation noise; Step S403: Perform non-negative constraint normalization on the estimated values ​​of etiological state under each time window to obtain the etiological contribution vector corresponding to each time window. Based on the etiological contribution vector corresponding to each time window, construct the etiological contribution time series result set by splicing in the order of time windows.

8. A system for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions, applied to the method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions as described in any one of claims 1 to 7, characterized in that, The system for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint interventions includes: The rhythm perturbation basic input set construction module is used to obtain the multi-source raw monitoring dataset of the target object within a preset monitoring period. Based on the multi-source raw monitoring dataset, a unified time axis reconstruction and time sequence alignment mechanism is used to perform the rhythm perturbation basic input set construction task and output the rhythm perturbation basic input set. The time-based anomaly characterization set construction module is used to perform the task of constructing a time-based anomaly characterization set based on the rhythmic perturbation input set using the individual day-night baseline fitting and rhythmic offset analysis mechanism, and outputs the time-based anomaly characterization set. The etiology candidate state input set construction module is used to perform preprocessing tasks based on the time-period-based abnormal characterization set using a time-delay response decomposition and etiology candidate state mapping mechanism, and output the etiology candidate state input set. The causal contribution time series calculation module is used to perform the causal contribution time series calculation task based on the causal candidate state input set using a constrained state space recursive solution mechanism, and output the causal contribution time series result set; The joint intervention assessment module is used to perform joint intervention assessment processing based on the time series results set of etiological contribution using the Bliss independent collaborative evaluation method, and outputs the joint intervention assessment result set.

9. A device for calculating the contribution of causes to glucose homeostasis imbalance and for evaluating joint intervention, characterized in that, The device for calculating and evaluating the contribution of causes to glucose homeostasis imbalance and for joint intervention includes: a memory, a processor, and a program for calculating and evaluating the contribution of causes to glucose homeostasis imbalance stored in the memory and executable on the processor. When the program for calculating and evaluating the contribution of causes to glucose homeostasis imbalance is executed by the processor, it implements a method for calculating and evaluating the contribution of causes to glucose homeostasis imbalance according to any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a program for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions. When the program is executed by a processor, it implements a method for calculating the contribution of causes to glucose homeostasis imbalance and evaluating joint interventions as described in any one of claims 1 to 7.

Images