Exercise risk evaluation method based on fuzzy entropy, system, device and medium

Abstract

SOLUTION: To provide an exercise risk evaluation method capable of reliably quantifying a disorder degree and a complexity degree of heart rate variability, analyzing a nonlinear growth law of exercise load in real time, effectively quantifying a relation between exercise risk and exercise intensity and further improving reliability and accuracy of exercise risk evaluation.EFFECT: An exercise risk evaluation method includes the steps of: acquiring historical electrocardiographic signals at exercise intensity at a time of waiting for evaluation and previous exercise intensity different therefrom of an exerciser waiting for evaluation; inputting heartbeat interval sequences of different time periods, obtained according to the historical electrocardiographic signals, into a fuzzy entropy model respectively, to calculate fuzzy entropy and acquire exercise fuzzy entropy sequences; and inputting the time point of waiting for evaluation, the exercise fuzzy entropy sequences and the exercise intensities into a sequence change point identification model to perform risk evaluation based on the change point identification results obtained by performing change point identification, and a risk state transition model, and acquiring the risk evaluation result at the time point of waiting for evaluation.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to the technical field of exercise risk assessment, and in particular to a method, system, computer device and storage medium for exercise risk assessment based on fuzzy entropy. [Background technology]

[0002] With the continuous improvement of people's living standards and quality, exercise and fitness have become indispensable activities in people's lives. However, not all athletes have received professional training or guidance, and in exercise and fitness, the risk of physical injury due to excessive exercise intensity is very likely to occur, for example, myocardial ischemia can be induced if the exercise intervention intensity is too high. Therefore, it is of great significance to provide timely and effective exercise risk assessment for athletes and bodybuilders.

[0003] Conventional exercise risk assessment methods are mostly limited to calculating heart rate variability (HRV), such as the standard deviation of the heart rate interval (RR interval), the frequency domain and / or time domain characteristics of the time series signal, etc., and can meet the demands of risk assessment to a certain extent. However, such assessment methods are only similar to ex post risk assessment, do not have strong real-time characteristics, and cannot provide quantified risk assessment results. They cannot make reliable prior predictions of exercise risks that may exist at a future point in time according to changes in real-time exercise load during the exercise period, and do not involve research on the relationship between exercise intensity and exercise risk. In other words, they cannot truly and effectively avoid the occurrence of potential exercise risks, and cannot provide reliable protection for the life and health of exercisers. Summary of the Invention

[0004] The objective of the present invention is to provide an exercise risk assessment method based on fuzzy entropy, which uses HRV, a quantitative index of physiological load intensity, as an index for judging exercise load, and combines it with the calculation of fuzzy entropy at different exercise intensities, prediction of sudden change points of fuzzy entropy series and prediction of state transition risk to analyze the nonlinear growth law of exercise load in real time, provide real-time exercise risk assessment results, effectively solve the application defects of traditional exercise risk assessment methods, effectively quantify the relationship between exercise risk and exercise intensity, and improve the reliability and accuracy of exercise risk assessment, thereby effectively preventing exercise risks caused by excessive increase in exercise load.

[0005] In order to achieve the above objectives, it is necessary to provide an exercise risk assessment method, system, computer device and storage medium based on fuzzy entropy for the above technical problems.

[0006] In a first embodiment of the present invention, the method includes the steps of: acquiring exercise intensity of an awaiting evaluation time point of an awaiting evaluation exerciser and historical exercise data before the awaiting evaluation time point, the historical exercise data including historical electrocardiogram signals at different exercise intensities; obtaining a sequence of cardiac intervals for different time periods based on the historical electrocardiogram signals; Inputting the heartbeat interval series of different time periods into the pre-constructed fuzzy entropy model respectively to perform fuzzy entropy calculation, and obtain the corresponding motion fuzzy entropy series; inputting the motion fuzzy entropy sequence and the motion intensity into a pre-constructed sequence change point identification model to perform change point identification and obtain a change point identification result, the change point identification result including a sudden change point location and a corresponding change point motion intensity; and A step of performing risk assessment based on the evaluation waiting time point, the change point identification result, and a pre-constructed risk state transition model to obtain a risk assessment result at the evaluation waiting time point, wherein the risk state transition model includes a state level quantile and a state transition risk function. A method for exercise risk assessment based on fuzzy entropy is provided.

[0007] Furthermore, before the step of obtaining the exercise intensity of the athlete at the time of waiting for evaluation and the historical exercise data before the time of waiting for evaluation, obtaining electrocardiogram signal data of different exercisers at different exercise intensities and constructing an exercise electrocardiogram signal dataset; obtaining a corresponding set of heart beat interval sequences based on the exercise electrocardiogram signal data set, the set of heart beat interval sequences including heart beat interval sequences of different exercisers at different exercise intensities; and The method further comprises the step of constructing the fuzzy entropy model, the sequence change point identification model, and the risk state transition model based on the set of cardiac interval sequences.

[0008] Furthermore, the step of constructing the fuzzy entropy model includes: obtaining heartbeat interval sequences at a maximum exercise intensity and a minimum exercise intensity from the set of heartbeat interval sequences; preset several groups of fuzzy entropy calculation parameters, and perform fuzzy entropy calculation for each of the heart beat interval sequences at the highest exercise intensity and the lowest exercise intensity based on each group of fuzzy entropy calculation parameters to obtain corresponding highest intensity fuzzy entropy value and lowest intensity fuzzy entropy value; Calculating the difference between the maximum intensity fuzzy entropy value and the minimum intensity fuzzy entropy value of each fuzzy entropy calculation parameter group, and determining the parameter group of the fuzzy entropy model corresponding to the maximum difference as the parameter group optimal for fuzzy entropy calculation; and The method further comprises the step of obtaining the fuzzy entropy model based on a set of parameters that are optimal for calculating the fuzzy entropy.

[0009] Furthermore, the step of constructing the sequence change point identification model includes: Obtaining a heart rate variability fuzzy entropy sequence set based on the heart rate interval sequence set and the fuzzy entropy model, the heart rate variability fuzzy entropy sequence set including heart rate variability fuzzy entropy sequences of different exercisers at different exercise intensities; and The method includes a step of estimating change-point locations and fitting distribution parameters to the heart rate variability fuzzy entropy series at different exercise intensities based on the likelihood ratio test method and the circular binary segmentation algorithm, and obtaining the series change-point identification model.

[0010] Furthermore, the step of constructing the risk state transition model includes: performing change point identification for each of the heart rate variability fuzzy entropy series of different exercisers at different exercise intensities among the heart rate variability fuzzy entropy series set based on the series change point identification model, to obtain a change point identification result set; performing a statistical analysis on the change-point identification result set to obtain the state-level quantiles; obtaining an exercise state of each exerciser in the heart rate variability fuzzy entropy sequence set based on the state level quantile; constructing the state transition risk function based on the motion states of all the exercisers; and Deriving the risk state-transition model based on the state-level quantiles and the state-transition risk function.

[0011] Further, the step of constructing a state transition risk function based on the motion states of all the above-mentioned exercisers includes: obtaining individual physiological characteristics of each exerciser in the heart rate variability fuzzy entropy sequence set, the individual physiological characteristics including height, weight and lung capacity; estimating characteristic state transition risk relationship values ​​by maximizing a pre-defined partial likelihood function based on the individual physiological characteristics of all exercisers; Obtaining a baseline risk function based on the motion states of all exercisers; and The method further comprises the step of obtaining the state transition risk function based on the characteristic state transition risk relation value and the reference risk function.

[0012] Furthermore, the step of performing risk assessment based on the evaluation waiting time point, the change point identification result, and a pre-constructed risk state transition model to obtain a risk assessment result at the evaluation waiting time point includes: obtaining a fuzzy entropy at the evaluation waiting time based on the evaluation waiting time, the exercise intensity and the fuzzy entropy model, and obtaining an exercise state at the evaluation waiting time based on the evaluation waiting time fuzzy entropy and the state level quantile; obtaining a sudden change time fuzzy entropy according to the change-point motion intensity, the sudden change point location and the fuzzy entropy model, and obtaining a next motion state according to the sudden change time fuzzy entropy and the state level quantile; obtaining physiological characteristics of the awaiting exerciser, and inputting the physiological characteristics, the awaiting exercise state, the next exercise state, and the awaiting exercise time into the state transition risk function to obtain a corresponding state transition risk probability; and Obtaining the risk assessment result based on the state transition risk probability, the next motion state and the corresponding sudden change point location.

[0013] In a second embodiment of the present invention, a data acquisition module is provided for acquiring exercise intensity of an awaiting evaluation time point of an awaiting evaluation athlete and historical exercise data prior to the awaiting evaluation time point, the historical exercise data including historical electrocardiogram signals at different exercise intensities; an index analysis module for obtaining cardiac interval sequences for different time periods based on the historical electrocardiogram signals; a fuzzy entropy calculation module for inputting the heartbeat interval series of different time periods into a pre-constructed fuzzy entropy model respectively to perform fuzzy entropy calculation and obtain a corresponding motion fuzzy entropy series; a change-point identification module for inputting the motion fuzzy entropy sequence and the motion intensity into a pre-constructed sequence change-point identification model to perform change-point identification and obtain a change-point identification result, the change-point identification result including several sudden change points and corresponding sudden change motion intensities; and A risk assessment module for performing risk assessment based on the evaluation waiting time, the change point identification result, and a pre-constructed risk state transition model to obtain a risk assessment result at the evaluation waiting time, the risk state transition model including a state level quantile and a state transition risk function. An exercise risk assessment system based on fuzzy entropy is provided.

[0014] In a third embodiment of the present invention, there is further provided a computing device including a memory, a processor, and a computer program stored in the memory and executable by the processor, the computing device implementing the steps of the method when the computer program is executed by the processor.

[0015] In a fourth embodiment of the present invention, there is further provided a computer-readable storage medium storing a computer program, the computer program implementing the steps of the method when executed by a processor.

[0016] The present application provides a method, system, computer device and storage medium for exercise risk assessment based on fuzzy entropy, and the method realizes a technical solution of obtaining a historical electrocardiogram signal of an athlete waiting for evaluation at the same exercise intensity as the evaluation waiting time before the evaluation waiting time, obtaining heart beat interval series of different time periods based on the historical electrocardiogram signal, and inputting the heart beat interval series of different time periods into a pre-constructed fuzzy entropy model respectively to perform fuzzy entropy calculation to obtain a corresponding exercise fuzzy entropy series, and then inputting the exercise fuzzy entropy series and exercise intensity into a pre-constructed series change point identification model to perform change point identification and obtain a change point identification result, and performing risk assessment based on the evaluation waiting time, the change point identification result and the pre-constructed risk state transition model to obtain a risk assessment result at the evaluation waiting time. Compared with conventional technologies, this fuzzy entropy-based exercise risk assessment method can not only reliably quantify the disorder and complexity of heart rate variability, but also analyze the nonlinear growth law of exercise load in real time, effectively quantify the relationship between exercise risk and exercise intensity, and further improve the reliability and accuracy of exercise risk assessment, effectively prevent exercise risks caused by excessive increase in exercise load, and provide a reliable guarantee for the life and health of exercisers. [Brief description of the drawings]

[0017] [Figure 1] FIG. 2 is a schematic diagram of an application scenario of the exercise risk assessment method based on fuzzy entropy in an embodiment of the present invention. [Diagram 2] FIG. 1 is a schematic diagram of the flow of a fuzzy entropy-based exercise risk assessment method in an embodiment of the present invention. [Diagram 3] FIG. 1 is a schematic diagram showing the results of an analog simulation of a heart rate variability fuzzy entropy sequence in an embodiment of the present invention. [Figure 4] FIG. 2 is a schematic diagram of exercise risk state transitions during the course of an exercise in an embodiment of the present invention. [Diagram 5] FIG. 1 is a schematic diagram showing the configuration of an exercise risk assessment system based on fuzzy entropy in an embodiment of the present invention. [Figure 6] FIG. 2 is a diagram showing the internal configuration of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0018] In order to make the purpose, technical solution and beneficial effects of this application clearer, the present invention will be described in more detail below in conjunction with the accompanying drawings and embodiments, and it is clear that the embodiments described below are only some of the embodiments of the present invention, and are only for the purpose of explaining the present invention, and do not limit the scope of the present invention. Based on the embodiments in the present invention, all other embodiments that can be obtained by those skilled in the art without creative labor fall within the scope of protection of the present invention.

[0019] The exercise risk assessment method based on fuzzy entropy provided by the present invention is a real-time exercise risk assessment method that uses HRV, an index of heart rate variability at different exercise intensities, as basic data, employs a fuzzy entropy algorithm to quantify the disorder and complexity of heart rate variability, performs sudden change point identification analysis on the obtained exercise fuzzy entropy series at different exercise intensities (output, unit w), and uses the corresponding change point identification result as the basis for state transition risk assessment to estimate the probability of exercise state transition (appearance of a sudden change point in exercise fuzzy entropy) at a certain exercise intensity based on exercise history information, and can be applied to a terminal or server as shown in Figure 1. Here, the terminal may be, but is not limited to, various personal computers, notebook computers, smartphones, tablet computers, and portable wearable devices, and the server may be realized by an independent server or a server cluster consisting of multiple servers. The server can adopt the exercise risk assessment method based on fuzzy entropy to perform real-time exercise risk assessment for the exerciser, and send the obtained risk assessment result to the terminal for analysis by the terminal, or store it in the server for later research. The following embodiment details the exercise risk assessment method based on fuzzy entropy provided by the present invention.

[0020] In one embodiment, as shown in FIG. 2, a fuzzy entropy-based exercise risk assessment method is provided, which includes the following steps:

[0021] S11, obtain the exercise intensity of the awaiting evaluation athlete at the time of awaiting evaluation and the historical exercise data before the time of awaiting evaluation, and the historical exercise data includes the historical ECG signals at different exercise intensities. Here, the exercise intensity can be understood as the physical load that the athlete receives during the actual exercise process, and in the entire exercise process, as the length of exercise time increases, the athlete can adjust the exercise intensity according to the actual exercise demand at different time periods, and accordingly, the ECG signals at different exercise intensities are different, and the HRV index reflected in different ECG signal data is also different, and the potential exercise risks included are also different. In actual risk assessment, the higher the HRV index, the stronger the cardiovascular function and stress resistance, and the lower the exercise risk, and conversely, the higher the exercise risk. Considering that exercise intensity is directly related to exercise risk, and the length of actual exercise time and switching between different exercise intensities also affect the exercise risk level, risk assessment at a certain exercise intensity at a certain exercise time is reliable only by analysis based on all historical data. In this embodiment, when risk assessment is performed for a certain exerciser at a certain exercise time (a time awaiting evaluation), it is preferable to obtain the exercise intensity at this exercise time and all historical electrocardiogram signal data of this exerciser prior to the time awaiting evaluation as basic data for subsequent evaluation and analysis.

[0022] In actual risk assessment applications, in order to realize real-time intelligent assessment of exercise risk and ensure the real-time and efficiency of exercise risk assessment, after the awaiting evaluation data of the awaiting evaluation athletes is obtained, it needs to be automatically analyzed, therefore, it is necessary to collect electrocardiogram signal data of different athletes at different exercise intensities to build the required automated analysis model, that is, before the step of obtaining the exercise intensity at the awaiting evaluation time of the awaiting evaluation athletes and the historical exercise data before the awaiting evaluation time, the following is further included:

[0023] The electrocardiogram signal data of different exercisers at different exercise intensities is obtained, and an exercise electrocardiogram signal data set is constructed. Here, the electrocardiogram signal data of different exercisers at different exercise intensities can adopt the related electrocardiogram signal data collected in conventional exercise risk research, and the specific acquisition method is not limited here. Note that, the exercise electrocardiogram signal data set constructed in this embodiment is the modeling basis of the fuzzy entropy model, the sequence change point identification model, and the risk state transition model used in the risk assessment of the present invention.

[0024] According to the exercise electrocardiogram signal data set, a corresponding heart beat interval sequence set is obtained, and the heart beat interval sequence set includes heart beat interval sequences of different exercisers at different exercise intensities. Here, the heart beat interval sequence can be understood as an interval time series based on peak R points in the electrocardiogram signal data at different exercise intensities, and can also be called an electrocardiogram RR interval sequence. In actual risk assessment, a larger RR interval difference in the heart beat interval sequence indicates a higher HRV and a lower exercise risk, and conversely, a smaller RR interval difference in the heart beat interval sequence indicates a lower HRV and a higher exercise risk.

[0025] The fuzzy entropy model, the sequence change point identification model, and the risk state transition model are constructed based on the heartbeat interval sequence set.

[0026] The fuzzy entropy model can be understood as a data model for performing fuzzy entropy calculation for a heartbeat interval sequence, which is a fuzzy entropy calculation formula that determines optimal fuzzy entropy calculation parameters based on exercise electrocardiogram signal data sets. Assuming that an athlete continues to exercise on a bicycle ergometer with a certain exercise intensity, and measures heartbeat interval sequence u={u(1),u(2),...,u(N)} in a certain time period [th,t+h], where m, r and N respectively represent the number of dimensions of the phase space, the similarity tolerance and the length of the heartbeat interval sequence, the specific process of using the fuzzy entropy calculation formula to perform fuzzy entropy calculation is as follows:

[0027] 1) Introduce a non-negative integer m (m≦N-2) and reconstruct the topological space for u. JPEG2025016433000002.jpg185162(8)

[0028] Since the value of fuzzy entropy FuzzyEn(m,r,N) in formula (8) depends on the selection of calculation parameters m, r and N, in order to ensure that fuzzy entropy is adopted to accurately quantify heart rate variability, in this embodiment, a method for selecting the optimal values ​​of calculation parameters m, r and N according to the purpose is designed, and a more ideal fuzzy entropy model is obtained. Specifically, the construction step of the fuzzy entropy model comprises the following:

[0029] Obtain the heart beat interval sequences at the highest and lowest exercise intensities from the heart beat interval sequence set, where the heart beat interval sequence set includes the heart beat interval sequences calculated from the electrocardiogram data at different exercise intensities as described above, order the exercise intensities from the sequence set, find the highest exercise intensity value and the lowest exercise intensity value, and obtain the corresponding heart beat interval sequences to serve as a basis for selecting the optimal values ​​of parameters m, r and N.

[0030] A number of fuzzy entropy calculation parameter groups are preset, and fuzzy entropy calculation is performed for the heartbeat interval series at the highest exercise intensity and the lowest exercise intensity according to each fuzzy entropy calculation parameter group, respectively, to obtain the corresponding highest intensity fuzzy entropy value and the lowest intensity fuzzy entropy value. Here, the fuzzy entropy calculation parameter groups can be understood as different candidate groups set according to experience, or as a candidate group obtained by sequentially listing and combining the range of values ​​that each parameter can take, and the specific manner of presetting the fuzzy entropy calculation parameter groups can be determined according to actual application needs, and no specific limitation is made here. Accordingly, in each preset fuzzy entropy calculation parameter group, it is necessary to calculate the fuzzy entropy value corresponding to the heartbeat interval series at the highest exercise intensity and the fuzzy entropy value corresponding to the heartbeat interval series at the lowest exercise intensity.

[0031] The difference between the maximum intensity fuzzy entropy value and the minimum intensity fuzzy entropy value of each fuzzy entropy calculation parameter group is calculated, and the parameter group of the fuzzy entropy model corresponding to the maximum difference is determined as the parameter group optimal for fuzzy entropy calculation. Here, the principle of selecting the parameter group optimal for fuzzy entropy calculation is a selection principle that is preferably based on the consideration that the fuzzy entropy calculation reflects the difference in heart rate variability at different exercise intensities as much as possible and reliably quantifies the heart rate variability, and ensures that the difference in fuzzy entropy value is also the largest when the difference in exercise intensity is the largest. In other words, the parameter group optimal for fuzzy entropy calculation can be understood as the fuzzy entropy calculation parameter group that has the largest difference between the maximum intensity fuzzy entropy value and the minimum intensity fuzzy entropy value among all the preset fuzzy entropy calculation parameter groups.

[0032] The fuzzy entropy model is obtained based on the optimal parameter group for the fuzzy entropy calculation, where the fuzzy entropy model can be understood as an optimal fuzzy entropy calculation formula that can reliably quantify heart rate variability at different exercise intensities, determined after replacing each of the calculation parameters m, r, and N in formula (8) with the corresponding parameter value from the optimal parameter group for the fuzzy entropy calculation obtained by the steps of the method.

[0033] In addition, the optimal parameter set for calculating the fuzzy entropy may be determined by using a concept called cross validation, in which all cardiac interval sequences in the cardiac interval sequence set are divided into a training set and a test set at a fixed ratio, the training set is used to find optimal possible values ​​of m, r, and N, and the test set is used to verify the validity of the optimal possible values ​​of the parameters.

[0034] The sequence change point identification model can be understood as a normal distribution model that can be used to perform effective change point identification for the fuzzy entropy series at different exercise intensities obtained by fitting the fuzzy entropy series at different exercise intensities calculated based on the heartbeat interval sequence set and the fuzzy entropy model. Specifically, the step of constructing the sequence change point identification model includes the following steps.

[0035] A heart rate variability fuzzy entropy sequence set is obtained based on the heart rate interval sequence set and the fuzzy entropy model, and the heart rate variability fuzzy entropy sequence set includes heart rate variability fuzzy entropy sequences of different athletes at different exercise intensities. Here, the heart rate variability fuzzy entropy sequence can be understood as a fuzzy entropy sequence consisting of optimal parameter fuzzy entropy FuzzyEnOptim(x,t) corresponding to each observation time point t at a certain exercise intensity (the magnitude of load output is x). Note that the optimal parameter fuzzy entropy FuzzyEnOptim(x,t) corresponding to the observation time point t can be understood as an optimal parameter fuzzy entropy calculated based on a subsequence in a time period [th,t+h] at an exercise intensity (output is x), and h is the adjustment interval for the time period. The observation point time series t={t 1 ,...,t k} for K observation times t k Then, the entire heartbeat interval sequence is divided into K observations t k For example, a certain athlete was observed for 6 hours. Now, the entire observed time series is divided into 3 subsequences and the fuzzy entropy is calculated for each subsequence. K=3, t 1 = 1, t 2 = 3, t 3 = 5, h = 1, and the three subsequences obtained are the first subsequence in the [0,2] time slot, the second subsequence in the [2,4] time slot, and the third subsequence in the [4,6] time slot. Each of the divided subsequences is k -h,t k +h], and by adopting a fuzzy entropy model, the fuzzy entropy of each subsequence can be calculated, and the entire heart rate variability fuzzy entropy sequence can be obtained.

[0036] Based on the likelihood ratio test method and the circular binary segmentation algorithm, the change point location is estimated and the distribution parameters are fitted for the heart rate variability fuzzy entropy series at different exercise intensities; The series change point identification model is obtained, where a change point location can be understood as a point location (time point) where a significant change occurs in a time series, and can indicate a sudden change in the distribution, mean, trend, or other statistical characteristics of the data; For example, the length of a time series is K=500, and there are 500 observations. As can be seen from the analog simulation results shown in Figure 3, there are three change points in this series, and the positions of the change points are JPEG2025016433000003.jpg68164

[0037] The series change point identification model can be understood as a distribution model that can be applied to the identification and prediction of the change points of the fuzzy entropy obtained by calculating all subsequences at all exercise intensities, which is obtained by consolidating the fuzzy entropies corresponding to all exercise intensities and all subsequences into one data set and estimating based on the data set. In practical applications, specifically, a likelihood ratio test method and a generic binary segmentation algorithm are adopted to generate a time series FuzzyEnOptim(x,t),t=t with a length of K. 1 ,...,t k For , the detailed process of identifying the number of change points, predicting the change point positions, and estimating the parameters of the normal distribution model corresponding to the subsequences divided by each change point is as follows:

[0038] Step 1: FuzzyEnOptim(x,t),t=t, a time series with total length K 1 ,...,t k First, find the first change point that divides the entire time series into two segments, and assume that the change point is located at one of the time t_k. However, it is unknown at what time the change point is located, that is, what possible values ​​it can take. First, we define the change point position as k (i.e. JPEG2025016433000004.jpg29164

[0039] JPEG2025016433000005.jpg45165

[0040] JPEG2025016433000006.jpg21163

[0041] JPEG2025016433000007.jpg43164

[0042] Step 2: For the two subsequences divided in the previous step, the operations of parameter estimation in the previous step and likelihood ratio test are performed respectively to check whether there are new change points in each of the two subsequences. For example, if new change points are found in both of the two subsequences, each will continue to be divided into two, forming a total of four subsequences. Of course, it is also possible that there are no new change points in either of the two subsequences obtained in step 1, or that there is a new change point in only one of the subsequences.

[0043] Step 3: Repeat the above operations until no new change points are found after running the cyclic binary segmentation algorithm on all current subsequences, and then determine the number of change points in the entire sequence, the location of each change point, and obtain the corresponding distribution model.

[0044] In this embodiment, the likelihood ratio test method and the circular binary segmentation algorithm are adopted to estimate the change point location and fit the distribution parameters to the heart rate variability fuzzy entropy series at different exercise intensities. The obtained series change point identification model can effectively verify that the possible values ​​of the heart rate variability fuzzy entropy series are related to both the exercise intensity x and the observation time series t, and each fuzzy entropy divided by the change point can be identified. At the same time, the obtained sequence change point identification model can be applied to accurately identify the change points of fuzzy entropy sequence under different exercise intensities, and has good universality and versatility, providing a solid basis for subsequent risk status assessment.

[0045] In practical application, the risk state transition model can be understood as an intelligent state analysis model that directly provides a risk assessment result including the next motion state, the sudden change time (the time of entering the next state), and the state transition risk probability of transitioning from the motion state corresponding to the evaluation waiting time to the next motion state based on the change point identification result of the heart rate variability fuzzy entropy sequence corresponding to the evaluation waiting time and the historical electrocardiogram data of the awaiting evaluation athlete. This model also needs to be pre-trained based on a certain data set, and in order to ensure the consistency of application with the fuzzy entropy model and the sequence change point identification model, and further ensure the accuracy and reliability of the risk assessment, in this embodiment, it is preferable to perform analysis and construction based on the data obtained by analyzing the motion electrocardiogram signal data set using the already trained fuzzy entropy model and the sequence change point identification model. Specifically, the construction step of the risk state transition model includes the following:

[0046] Based on the series change-point identification model, change-point identification is performed on each of the heart-rate variability fuzzy entropy series of different athletes at different exercise intensities in the heart-rate variability fuzzy entropy series set to obtain a change-point identification result set, where the change-point identification result set can be understood as a group of each series change-point identification result obtained by performing change-point identification on each heart-rate variability fuzzy entropy series in the heart-rate variability fuzzy entropy series set based on the above-obtained series change-point identification model.

[0047] The change-point identification result set is subjected to statistical analysis to obtain the condition level quantiles. Here, the number of condition level quantiles and the specific possible values ​​of each quantile can be set according to the actual application needs. For example, quintiles can be selected and determined to be 95%, 75%, 50%, 25%, and 5% quantile values, respectively. Then, the change-point positions and exercise intensities in the actually obtained change-point identification result set are statistically collected to obtain the change-point range Q where the exercise condition risk increases as shown in FIG. 4. 1 ,...,Q 5In practical applications, when the fuzzy entropy sequence of the exerciser's heart rate variability reaches a quantile of one change point, it indicates that one exercise state transition has occurred. As shown in Figure 4, the path "Q 1 -Q 2 ” indicates that an exerciser goes from a very low exercise risk state to a low risk state, and “Q 4 -Q 5 " indicates that the athlete enters into an extremely high exercise risk state from a high risk state. Note that the state level quantile obtained here can only be used to determine the corresponding exercise state based on the fuzzy entropy value at a certain time point, and to determine the next exercise state based on the next nearby change point position, but cannot quantify the risk probability that the state transition actually occurs. In order to quantify and accurately evaluate the exercise risk, in this embodiment, it is preferable to construct a mathematical model by the following method, which can accurately predict the time and probability that the athlete will enter the next exercise state from the current exercise state based on the physiological characteristics of the athlete and the evaluation time point.

[0048] A motion state of each exerciser in the heart rate variability fuzzy entropy sequence set is obtained based on the state level quantile.

[0049] Based on the motion states of all the exercisers, the state transition risk function is constructed, where the state transition risk function is expressed as follows: JPEG2025016433000009.jpg8100

[0050] JPEG2025016433000010.jpg38164

[0051] JPEG2025016433000011.jpg66164

[0052] Obtain individual physiological characteristics of each exerciser in the heart rate variability fuzzy entropy sequence set, and the individual physiological characteristics include height, weight and lung capacity.

[0053] Based on the individual physiological characteristics of all exercisers, the characteristic state transition risk relationship values ​​are estimated by maximizing a pre-defined partial likelihood function, where JPEG2025016433000012.jpg25121

[0054] JPEG2025016433000013.jpg36164

[0055] Based on the exercise states of all exercisers, a reference risk function is obtained, which is preferably calculated by the Breslow estimator and is expressed as follows: JPEG2025016433000014.jpg15111

[0056] JPEG2025016433000015.jpg6121

[0057] The state transition risk function is obtained based on the characteristic state transition risk relation value and the reference risk function.

[0058] The risk state transition model is obtained based on the state level quantiles and the state transition risk function.

[0059] In this embodiment, the quantiles of the change points of the fuzzy entropy time series, which can define various motion states, are determined by statistical analysis, and a state transition risk function, which can accurately give the transition risk probability between adjacent motion states, is estimated by a method of maximizing the partial likelihood function. In this way, a reliable real-time assessment can be directly and effectively performed based on the motion state transitions and corresponding transition risks in each time period of each exerciser's historical ECG signal data, thereby providing a reliable guarantee for real-time intelligent risk assessment.

[0060] S12, obtain heart beat interval sequences of different time periods according to the historical electrocardiogram signals, where the different time periods of the heart beat interval sequences can be obtained according to a division manner when constructing the fuzzy entropy model, and the heart beat interval sequences can also be obtained by referring to the processing method for the exercise electrocardiogram signal data set, so as to ensure the consistency between the processing of the model training data and the actual application data, and provide reliable data support for the subsequent risk assessment using the heart beat interval sequences.

[0061] S13, the heartbeat interval series of different time periods are input into the pre-constructed fuzzy entropy model respectively to perform fuzzy entropy calculation, and obtain the corresponding motion fuzzy entropy series. Here, the process of obtaining the motion fuzzy entropy series can refer to the relevant content in the construction step of the fuzzy entropy model, and thus will not be described in detail here.

[0062] S14, input the motion fuzzy entropy sequence and the motion intensity into a pre-constructed sequence change point identification model to perform change point identification, and obtain a change point identification result, the change point identification result including a sudden change point position and a corresponding change point motion intensity, where the sudden change point position can be understood as a position of entering a next motion state, and the corresponding sudden change point position can be understood as a time corresponding to a sudden change point in the sequence, i.e., a time point of entering a next motion state.

[0063] S15, performing risk assessment based on the evaluation waiting time, the change point identification result, and a pre-constructed risk state transition model, and obtaining a risk assessment result at the evaluation waiting time, the risk state transition model including a state level quantile and a state transition risk function. Here, the state level quantile can be understood as a risk state level division range that determines the motion state at the evaluation waiting time and the motion state corresponding to the sudden change point position based on the evaluation waiting time and the sudden change point position, and the state transition risk function can be understood as a model that directly obtains the potential risk magnitude of the corresponding state transition based on the motion state at the evaluation waiting time and the motion state corresponding to the sudden change point, and the physiological characteristics of the exerciser. Specifically, the step of performing risk assessment based on the above-mentioned evaluation waiting time, the change point identification result, and a pre-constructed risk state transition model, and obtaining a risk assessment result at the evaluation waiting time includes the following:

[0064] According to the evaluation waiting time, the exercise intensity and the fuzzy entropy model, obtain the evaluation waiting time fuzzy entropy, and according to the evaluation waiting time fuzzy entropy and the state level quantile, obtain the evaluation waiting time exercise state. Here, the process of obtaining the evaluation waiting time exercise state can be understood as judging the state level quantile range in which the evaluation waiting time fuzzy entropy exists, and when the evaluation waiting time fuzzy entropy is greater than the value of one quantile and less than the value of the second quantile, the corresponding exercise state is the first exercise state, and other situations can be inferred with reference to this example, and will not be described in detail here.

[0065] According to the change point motion intensity, the change point position and the fuzzy entropy model, obtain a sudden change time fuzzy entropy, and according to the sudden change time fuzzy entropy and the state level quantile, obtain a next motion state, where the next motion state can be understood as the motion state corresponding to the sudden change point position, and the specific obtaining process can refer to the description of obtaining the evaluation waiting time motion state, so it will not be described in detail here.

[0066] The physiological characteristics of the athlete awaiting evaluation are obtained, and the physiological characteristics, the motion state at the time of awaiting evaluation, the next motion state, and the time point at which evaluation is awaited are input into the state transition risk function to obtain a corresponding state transition risk probability. Here, the state transition risk probability can be understood as a predicted value of the probability that the athlete awaiting evaluation will enter the next motion state from the motion state at the time of awaiting evaluation at the time point at which evaluation is awaited, obtained by inputting the physiological characteristics of the athlete awaiting evaluation, the motion state at the time of awaiting evaluation, the next motion state, and the time point at which evaluation is awaited into the state transition risk function shown in formula (10).

[0067] According to the state transition risk probability, the next motion state and the corresponding sudden change point position, the risk assessment result is obtained. Here, the risk assessment result can be used to give real-time hints to the exerciser to avoid the occurrence of potential motion risks, and can be used to inversely estimate the time interval during which motion is safe and low risk according to the risk of different motion times determined in the analysis process, for example, when the motion time is 5 minutes, the calculated probability of transition to the dangerous state is 0.2, when the motion time is 10 minutes, the transition probability is 0.4, and when the motion time is 20 minutes, the transition probability is 0.9, so that it can be roughly determined that the motion time needs to be less than 10 minutes in order to reduce the risk to less than 50%, and according to the analysis result, the appropriate exercise time length at a certain motion intensity can be given to the exerciser, which is convenient for the exerciser to know the motion risk in real time, and can provide advice on the reasonable exercise time length, which can effectively improve the user experience.

[0068] In an embodiment of the present application, a historical electrocardiogram signal of an athlete waiting for evaluation is obtained at the same exercise intensity as the awaiting evaluation time point before the awaiting evaluation time point, and a heart beat interval series of different time periods is obtained based on the historical electrocardiogram signal. The heart beat interval series of different time periods are respectively input into a pre-constructed fuzzy entropy model to perform fuzzy entropy calculation to obtain a corresponding exercise fuzzy entropy series. Then, the exercise fuzzy entropy series and the exercise intensity are input into a pre-constructed series change point identification model to perform change point identification to obtain a change point identification result. Then, a risk assessment is performed based on the awaiting evaluation time point, the change point identification result and the pre-constructed risk state transition model to obtain a risk assessment result at the awaiting evaluation time point. This technical solution can not only reliably quantify the disorder and complexity of heart rate variability, but also analyze the nonlinear growth law of exercise load in real time, effectively quantify the relationship between exercise risk and exercise intensity, and further improve the reliability and accuracy of exercise risk assessment, effectively prevent exercise risks caused by excessive increase in exercise load, and provide a reliable guarantee for the life and health of the athlete.

[0069] Although the steps in the flowchart are shown sequentially as indicated by the arrows, the steps are not necessarily performed in the order indicated by the arrows. Unless otherwise expressly stated in this specification, there is no strict order restriction on the execution of the steps, and the steps may be performed in other orders.

[0070] In one embodiment, as shown in FIG. 5 , a data acquisition module 1 is provided for acquiring exercise intensity of an awaiting evaluation time point and historical exercise data before the awaiting evaluation time point, the historical exercise data including historical electrocardiogram signals at different exercise intensities; an index analysis module 2 for obtaining cardiac interval series for different time periods based on the historical electrocardiogram signal; a fuzzy entropy calculation module 3 for inputting the heartbeat interval series of different time periods into a pre-constructed fuzzy entropy model respectively to perform fuzzy entropy calculation and obtain a corresponding motion fuzzy entropy series; a change-point identification module 4 for inputting the motion fuzzy entropy sequence and the motion intensity into a pre-constructed sequence change-point identification model to perform change-point identification and obtain a change-point identification result, the change-point identification result including several sudden change points and corresponding sudden change motion intensities; and A risk assessment module 5 for performing risk assessment based on the evaluation waiting time, the change point identification result, and a pre-constructed risk state transition model to obtain a risk assessment result at the evaluation waiting time, the risk state transition model including a state level quantile and a state transition risk function. An exercise risk assessment system based on fuzzy entropy is provided.

[0071] It should be noted that the specific limitations of the fuzzy entropy-based exercise risk assessment system can be referred to the limitations of the above-mentioned fuzzy entropy-based exercise risk assessment method, and the corresponding technical effects can be obtained equally, so they will not be described in detail here. Each module in the above-mentioned fuzzy entropy-based exercise risk assessment system can be realized in whole or in part by software, hardware, and a combination thereof. Each module may be embedded in a processor in a computer device as hardware, or may be independent from a processor in a computer device, or may be stored in a memory in a computer device as software, and facilitate the processor to call and execute the operations corresponding to each of the above modules.

[0072] FIG. 6 shows an internal block diagram of a computer device in one embodiment, and the computer device may specifically be a terminal or a server. As shown in FIG. 6, the computer device includes a processor, a memory, a network interface, a display, and an input device connected by a system bus. Here, the processor of the computer device is for providing calculation and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an execution environment for the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is for communicating with an external terminal via a network connection. When the computer program is executed by the processor, the exercise risk assessment method based on fuzzy entropy is realized. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, a key, a trackball, or a touch pad provided on the housing of the computer device, or an external keyboard, touch pad, mouse, etc.

[0073] Those skilled in the art will understand that the structure shown in FIG. 6 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied, and a specific computer device may include more or fewer components than those shown in the figure, or may combine some components, or may have the same component arrangement.

[0074] In one embodiment, a computing device is provided that includes a memory, a processor, and a computer program stored in the memory and executable by the processor, the computer device implementing the steps of the method when the computer program is executed by the processor.

[0075] In one embodiment, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method.

[0076] To sum up, the embodiments of the present invention provide a fuzzy entropy-based exercise risk assessment method, system, computer device and storage medium. The fuzzy entropy-based exercise risk assessment method obtains historical exercise data including the exercise intensity at the time of waiting for evaluation of an athlete and historical electrocardiogram signals at different exercise intensities prior to the time of waiting for evaluation, obtains heart beat interval series of different time periods based on the historical electrocardiogram signals, inputs the heart beat interval series of different time periods into pre-constructed fuzzy entropy models respectively to perform fuzzy entropy calculation to obtain corresponding exercise fuzzy entropy series, and then inputs the exercise fuzzy entropy series and exercise intensity into a pre-constructed series change point identification model to perform change point identification and obtain a change point identification result, and then performs risk assessment based on the change point identification result and the pre-constructed risk state transition model to obtain a risk assessment result at the time of waiting for evaluation. This exercise risk assessment method based on fuzzy entropy can not only reliably quantify the disorder and complexity of heart rate variability, but also analyze the nonlinear growth law of exercise load in real time, effectively quantify the relationship between exercise risk and exercise intensity, further improve the reliability and accuracy of exercise risk assessment, effectively prevent exercise risks caused by excessive increase in exercise load, and provide a reliable guarantee for the life and health of exercisers.

[0077] Each embodiment in this specification is described in a stepwise manner, and the same or similar parts of each embodiment may be referred to each other, and each embodiment will be described focusing on the differences from other embodiments. In particular, the system-related embodiment is basically similar to the method-related embodiment, so the description is relatively simple, and the relevant points may be referred to the partial description of the method-related embodiment. Note that each technical feature of the above embodiments can be arbitrarily combined, and in order to simplify the description, all possible combinations of each technical feature in the above embodiments are not described, but as long as the combination of these technical features is not contradictory, it should be considered as being within the scope described in this specification.

[0078] The above-mentioned embodiments merely show some preferred embodiments of the present application, and although the description is more specific and detailed, it cannot be understood as a limitation of the scope of the claims of the invention. It should be noted that, for a person skilled in the art, some improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered as the scope of protection of the present application. Therefore, the scope of protection of the patent of the present application shall be equivalent to the scope of protection of the claims.

Claims

1. acquiring exercise intensity of the awaiting evaluation time point and historical exercise data before the awaiting evaluation time point, the historical exercise data including historical electrocardiogram signals at different exercise intensities; obtaining a sequence of cardiac intervals for different time periods based on the historical electrocardiogram signals; inputting the heartbeat interval series of different time periods into a pre-constructed fuzzy entropy model to perform fuzzy entropy calculation, and obtain a corresponding motion fuzzy entropy series; inputting the motion fuzzy entropy sequence and the motion intensity into a pre-constructed sequence change point identification model to perform change point identification and obtain a change point identification result, the change point identification result including a sudden change point location and a corresponding change point motion intensity; A step of performing risk assessment based on the evaluation waiting time, the change point identification result, and a pre-constructed risk state transition model to obtain a risk assessment result at the evaluation waiting time, the risk state transition model including a state level quantile and a state transition risk function; The step of performing risk assessment based on the evaluation waiting time point, the change point identification result, and a pre-constructed risk state transition model and obtaining a risk assessment result at the evaluation waiting time point includes: Obtaining a fuzzy entropy at the evaluation waiting time based on the evaluation waiting time, the exercise intensity and the fuzzy entropy model, and obtaining an exercise state at the evaluation waiting time based on the evaluation waiting time fuzzy entropy and the state level quantile; Obtaining a sudden change time fuzzy entropy according to the change point motion intensity, the sudden change point location and the fuzzy entropy model, and obtaining a next motion state according to the sudden change time fuzzy entropy and the state level quantile; Obtaining physiological characteristics of the awaiting exerciser, and inputting the physiological characteristics, the awaiting exercise state, the next exercise state and the awaiting exercise time into the state transition risk function to obtain a corresponding state transition risk probability; and obtaining the risk assessment result according to the state transition risk probability, the next motion state and the corresponding sudden change point location.

2. Before the step of acquiring the exercise intensity of the athlete at the time of waiting for evaluation and the historical exercise data before the time of waiting for evaluation, acquiring electrocardiogram signal data of different exercisers at different exercise intensities and constructing an exercise electrocardiogram signal data set; obtaining a corresponding set of heart beat interval sequences based on the exercise electrocardiogram signal data set, the set of heart beat interval sequences including heart beat interval sequences of different exercisers at different exercise intensities; The fuzzy entropy-based exercise risk assessment method according to claim 1, further comprising the steps of: constructing the fuzzy entropy model, the sequence change point identification model, and the risk state transition model based on the heartbeat interval sequence set.

3. The step of constructing the fuzzy entropy model includes: obtaining heartbeat interval sequences at a maximum exercise intensity and a minimum exercise intensity from the set of heartbeat interval sequences; A step of presetting several fuzzy entropy calculation parameter groups, and performing fuzzy entropy calculation for each of the heartbeat interval series at the highest exercise intensity and the lowest exercise intensity based on each fuzzy entropy calculation parameter group to obtain a corresponding highest intensity fuzzy entropy value and a corresponding lowest intensity fuzzy entropy value; calculating a difference between the maximum intensity fuzzy entropy value and the minimum intensity fuzzy entropy value of each fuzzy entropy calculation parameter group, and determining a parameter group of the fuzzy entropy model corresponding to the maximum difference as a parameter group optimal for fuzzy entropy calculation; The method for exercise risk assessment based on fuzzy entropy according to claim 2, further comprising the step of: obtaining the fuzzy entropy model based on a set of parameters that are optimal for calculating the fuzzy entropy.

4. The step of constructing a sequence change point identification model includes: Obtaining a heart rate variability fuzzy entropy sequence set based on the heart rate interval sequence set and the fuzzy entropy model, the heart rate variability fuzzy entropy sequence set including heart rate variability fuzzy entropy sequences of different exercisers at different exercise intensities; The fuzzy entropy-based exercise risk assessment method according to claim 2, further comprising a step of estimating change point locations and fitting distribution parameters to the heart rate variability fuzzy entropy series at different exercise intensities based on a likelihood ratio test method and a circular binary segmentation algorithm, thereby obtaining the series change point identification model.

5. The step of constructing a risk state transition model includes: performing change point identification for each of the heart rate variability fuzzy entropy series of different exercisers at different exercise intensities among the heart rate variability fuzzy entropy series set based on the series change point identification model, to obtain a change point identification result set; performing a statistical analysis on the change-point identification result set to obtain the state level quantiles; obtaining a motion state of each exerciser in the heart rate variability fuzzy entropy sequence set based on the state level quantile; constructing the state transition risk function based on the motion states of all exercisers; and obtaining the risk state transition model based on the state level quantiles and the state transition risk function.

6. The step of constructing a state transition risk function based on the above-mentioned motion states of all the exercisers is as follows: obtaining individual physiological characteristics of each exerciser in the heart rate variability fuzzy entropy sequence set, the individual physiological characteristics including height, weight and lung capacity; estimating characteristic state transition risk relationship values ​​by maximizing a pre-defined partial likelihood function based on the individual physiological characteristics of all exercisers; Obtaining a baseline risk function based on the motion states of all exercisers; and obtaining the state transition risk function based on the characteristic state transition risk relation value and the reference risk function.

7. A data acquisition module for acquiring exercise intensity of the awaiting evaluation time point and historical exercise data before the awaiting evaluation time point, the historical exercise data including historical electrocardiogram signals at different exercise intensities; an index analysis module for obtaining cardiac interval sequences for different time periods based on the historical electrocardiogram signals; a fuzzy entropy calculation module for inputting the heartbeat interval series of different time periods into a pre-constructed fuzzy entropy model to perform fuzzy entropy calculation and obtain a corresponding motion fuzzy entropy series; A change point identification module for inputting the motion fuzzy entropy sequence and the motion intensity into a pre-constructed sequence change point identification model to perform change point identification and obtain a change point identification result, the change point identification result including a sudden change point location and a corresponding change point motion intensity; A risk assessment module for performing risk assessment based on the evaluation waiting time, the change point identification result, and a pre-constructed risk state transition model to obtain a risk assessment result at the evaluation waiting time, the risk state transition model including a state level quantile and a state transition risk function; The step of performing risk assessment based on the evaluation waiting time point, the change point identification result, and a pre-constructed risk state transition model and obtaining a risk assessment result at the evaluation waiting time point includes: Obtaining a fuzzy entropy at the evaluation waiting time based on the evaluation waiting time, the exercise intensity and the fuzzy entropy model, and obtaining an exercise state at the evaluation waiting time based on the evaluation waiting time fuzzy entropy and the state level quantile; Obtaining a sudden change time fuzzy entropy according to the change point motion intensity, the sudden change point location and the fuzzy entropy model, and obtaining a next motion state according to the sudden change time fuzzy entropy and the state level quantile; Obtaining physiological characteristics of the awaiting exerciser, and inputting the physiological characteristics, the awaiting exercise state, the next exercise state and the awaiting exercise time into the state transition risk function to obtain a corresponding state transition risk probability; and obtaining the risk assessment result according to the state transition risk probability, the next motion state and the corresponding sudden change point location.

8. A computing device comprising a memory, a processor and a computer program stored in the memory and executable by the processor, characterized in that said computer program, when executed by the processor, implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that the computer program implements the steps of the method according to any one of claims 1 to 6 when executed by a processor.

Images