Intelligent assessment and prediction method for heart rehabilitation effect
By analyzing the incidental and memory indicators of heart rate data using the STL decomposition algorithm, a predictive model was constructed, which solved the problems of nonlinear and sudden changes in heart rate, and achieved accurate assessment and prediction of cardiac rehabilitation effects.
Patent Information
- Application Number
- CN202511275611.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-08
- Publication Date
- 2025-12-19
AI Technical Summary
In existing technologies, conventional prediction models struggle to capture the nonlinear relationships and sudden changes in heart rate, leading to a decrease in the accuracy of cardiac rehabilitation effect assessment.
The STL decomposition algorithm is used to construct a prediction model by acquiring occasional and memory indicators of heart rate data. The model analyzes the sudden changes and regularities in heart rate data, and combines trend, seasonal and residual terms to decompose and predict heart rate data.
It improves the effectiveness and accuracy of cardiac rehabilitation outcome assessment, enabling more accurate prediction of future heart rate changes and helping to optimize treatment plans and manage medical resources.
Smart Images
Figure CN121171584A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of heart rate data prediction, and in particular to an intelligent evaluation and prediction method for cardiac rehabilitation effect. BACKGROUND
[0002] Cardiac rehabilitation, as a class IA recommendation for the prevention and treatment of cardiovascular diseases, can effectively reduce the rehospitalization rate and mortality rate of cardiovascular diseases and patients, and improve the quality of life of patients. Cardiac rehabilitation effect evaluation is a process of systematically evaluating the progress and effect of patients with heart disease in the cardiac rehabilitation plan. This evaluation can help doctors and rehabilitation teams understand the rehabilitation of patients, adjust treatment plans, and determine the next rehabilitation plan for patients. Predicting the rehabilitation effect is also an effective means to improve the rehabilitation effect and reduce potential risks, and is one of the important tools for optimizing treatment plans and managing medical resources.
[0003] The existing technology for evaluating and predicting the effect of cardiac rehabilitation is usually based on real-time heart rate data of patients. By monitoring the heart rate of patients in real time and constructing a corresponding prediction model, it can be determined which risks may occur to patients during the rehabilitation process. The prediction model can better represent the trend and seasonal characteristics of heart rate, and has a significant effect on revealing the long-term change rule of heart rate. However, not all patients' heart rates will follow similar linear change rules, and there may be complex nonlinear relationships. Conventional prediction models are difficult to fully capture such relationships, and some patients may have acute changes in heart rate data due to sudden heart problems or unexpected stress, which can easily lead to a decrease in the accuracy of the prediction results of the model and affect the effectiveness of the evaluation of the effect of cardiac rehabilitation. SUMMARY
[0004] In order to solve the technical problem that the conventional prediction model cannot consider the heart rate change rule of all patients, and some patients may have acute changes in heart rate, leading to a decrease in the accuracy of the prediction results of the model and affecting the effectiveness of the evaluation of the effect of cardiac rehabilitation, the purpose of the present application is to provide an intelligent evaluation and prediction method for cardiac rehabilitation effect, and the technical scheme adopted is as follows:
[0005] An intelligent evaluation and prediction method for cardiac rehabilitation effect, the method comprising:
[0006] obtaining heart rate data of a patient at each preset time in each preset real time period;
[0007] Optionally, the heart rate data at a preset time in a preset real time period is taken as target heart rate data; a data variation degree of the target heart rate data is obtained according to a data variation characteristic of the target heart rate data in a preset time neighborhood; an accident index of the target heart rate data is obtained according to a variation difference between the target heart rate data and other heart rate data in the preset time neighborhood and the data variation degree of the target heart rate data; a time window of the target heart rate data is obtained according to a variation characteristic of the target heart rate data relative to other heart rate data in the same preset real time period; a memory index of the target heart rate data is obtained according to a similarity characteristic between the time window of the target heart rate data and a time window of heart rate data at the same preset time in each preset real time period and a heart rate data variation characteristic in the time window of the target heart rate data;
[0008] A prediction model is constructed for all the heart rate data and is decomposed to obtain a trend item, a seasonal item and a residual item at each future preset time; heart rate data at each future preset time is obtained according to the trend item, the seasonal item and the residual item at each future preset time and the accident index and the memory index of all the heart rate data in the preset time window; a heart recovery index in a preset future time period is obtained according to a difference characteristic of each heart rate data in the preset future time period relative to all the heart rate data in the preset real time period and a data variation degree of each heart rate data in the preset future time period;
[0009] The heart recovery effect is evaluated according to the heart recovery index.
[0010] Further, the data variation degree obtaining method comprises:
[0011] The data variation degree of the target heart rate data is obtained by averaging a data difference between each two adjacent heart rate data in a preset time neighborhood of the target heart rate data.
[0012] Further, the accident index obtaining method comprises:
[0013] The accident index is obtained according to an accident index calculation formula, and the accident index calculation formula is as follows:
[0014]
[0015] In the formula, t represents the serial number of the target heart rate data; δ t represents the accident index of the target heart rate data;
[0016] B t represents the data variation degree of the target heart rate data; represents the mean of the data variation degree of all the heart rate data in the preset real time period in which the target heart rate data is located; N represents the number of all the heart rate data in the preset real time period in which the target heart rate data is located; G represents the number of the heart rate data on the right side of the target heart rate data in the preset time neighborhood; ΔR t+i represents the difference between every two adjacent heart rate data in the preset time neighborhood of the target heart rate data; ΔR n+i represents the difference between every two adjacent heart rate data in the preset time neighborhood of the nth other heart rate data in the preset real time period in which the target heart rate data is located, wherein n is not equal to t; max() represents the maximum value function; norm[] represents the normalization function.
[0017] Further, the method for obtaining the time window comprises:
[0018] taking the mode of all the heart rate data in the preset real time period in which the target heart rate data is located as the resting heart rate data in the preset real time period;
[0019] calculating the difference between the target heart rate data and the resting heart rate data as the relative heart rate deviation of the target heart rate data;
[0020] when the relative heart rate deviation of the target heart rate data is less than a preset first threshold, taking the heart rate data as the non-change heart rate data, and when the relative heart rate deviation of the target heart rate data is not less than the preset first threshold, taking the heart rate data as the change heart rate data;
[0021] traversing forward and backward along the preset time point in which the target heart rate data is located until the first non-change heart rate data appears, taking the set composed of all the preset time points corresponding to the change heart rate data as the time window of the target heart rate data; traversing all the heart rate data to obtain the time window of each heart rate data.
[0022] Further, the method for obtaining the memory index comprises:
[0023] obtaining the memory index according to the memory index calculation formula, and the memory index calculation formula is as follows:
[0024]
[0025] In the formula, t represents the serial number of the target heart rate data; μ t represents the memory index of the target heart rate data; C t represents the time window length of the target heart rate data; M represents the number of other preset real time periods different from the preset real time period in which the target heart rate data is located; C m represents the time window length of the heart rate data at the same preset time point as the target heart rate data in the mth preset real time period; a mean of data variation degrees of all heart rate data in the time window of the target heart rate data; DTW(C t ,C m ) represents a matching degree between the time window of the target heart rate data and the time window of the heart rate data at the same preset time point in the mth preset real time period; norm() represents a normalization function.
[0026] Further, the method for acquiring the heart rate data at each future preset time point comprises:
[0027] The heart rate data at each future preset time point is acquired according to a calculation formula of the heart rate data at each future preset time point, and the calculation formula of the heart rate data at each future preset time point is as follows:
[0028]
[0029] In the formula, R p represents the heart rate data at the pth future preset time point; T p represents a trend item at the pth future preset time point; represents a mean of all heart rate data in a preset time window at the pth future preset time point; represents a mean of all heart rate data in a preset time window at the pth future preset time point; S p represents a seasonal item at the pth future preset time point; L p represents a residual item at the pth future preset time point.
[0030] Further, the method for acquiring the heart rate data at each future preset time point comprises:
[0031] The mode of all heart rate data in the preset future time period is taken as the predicted resting heart rate data in the preset future time period.
[0032] The heart rate data at each future preset time point is acquired according to a calculation formula of the heart rate data at each future preset time point, and the calculation formula of the heart rate data at each future preset time point is as follows:
[0033]
[0034] In the formula, γ represents the heart rate data at each future preset time point; represents a mean of all heart rate data in the preset future time period; R fes (K) represents the predicted resting heart rate data in the preset future time period; represents a mean of data variation degrees of all heart rate data in the time window of the target heart rate data; DTW(C max represents a maximum value of heart rate data in the preset future time period; R minrepresents a minimum value of heart rate data in a preset future period; norm[] represents a normalization function.
[0035] An intelligent evaluation and prediction system for cardiac rehabilitation effect, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the above-mentioned method for intelligent evaluation and prediction of cardiac rehabilitation effect when executing the computer program.
[0036] A computer-readable storage medium storing a computer program, the computer program implementing the steps of the above-mentioned method for intelligent evaluation and prediction of cardiac rehabilitation effect when executed by a processor.
[0037] A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the above-mentioned method for intelligent evaluation and prediction of cardiac rehabilitation effect when executing the computer program.
[0038] The present application has the following advantages:
[0039] The present application obtains heart rate data at each preset time in each preset period; since the heart rate of a cardiac rehabilitation patient is easily affected by an instantaneous psychological reaction, the mutation of the heart rate data, i.e., the degree of data change of the heart rate data, is analyzed to obtain an occasional index of the heart rate data; since in most cases, the change of the heart rate of a cardiac rehabilitation patient has regularity, i.e., the heart rate data is usually in memory for a long period of time, which is manifested as a strong periodicity and stable time window of the heart rate data, a time window of target heart rate data is defined, and a memory index of the target heart rate data is analyzed; the occasional index and the memory index of the heart rate data are used to construct an STL decomposition prediction model, and the heart rate data at each preset time in the future is predicted; since the judgment of the cardiac rehabilitation effect by a staff member is mainly based on the overall stability of the heart rate, the change of the resting heart rate data relative to the preset real period, and the range of the degree of data change of the heart rate data, the cardiac rehabilitation index in the preset future period is obtained according to the difference features of each heart rate data in the preset future period relative to all heart rate data in the preset real period and the degree of data change of each heart rate data in the preset future period; and the cardiac rehabilitation effect is evaluated according to the cardiac rehabilitation index. The present application obtains a relatively accurate prediction model, and finally improves the effectiveness of the evaluation of the cardiac rehabilitation effect. BRIEF DESCRIPTION OF DRAWINGS
[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, below will briefly introduce the drawings required by the embodiments or prior art description. Obviously, the drawings in the following description only represent some embodiments of the present application, and for those skilled in the field, other drawings can be obtained from these drawings without any creative effort.
[0041] Figure 1 A flow chart of a heart rehabilitation effect intelligent evaluation and prediction method provided by an embodiment of the present application. DETAILED DESCRIPTION
[0042] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined purposes, below will combine the drawings and preferred embodiments to specifically describe the specific implementation, structure, features and effects of the heart rehabilitation effect intelligent evaluation and prediction method according to the present application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.
[0043] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.
[0044] Below will specifically describe the specific scheme of the heart rehabilitation effect intelligent evaluation and prediction method provided by the present application in combination with the drawings.
[0045] Please refer to Figure 1 which shows a heart rehabilitation effect intelligent evaluation and prediction method provided by an embodiment of the present application. The method comprises:
[0046] Step S1: acquiring heart rate data of a patient at each preset time point in each preset real time period.
[0047] The embodiment of the present application is mainly applied to the patient heart rehabilitation effect prediction scene. In order to analyze the heart rehabilitation effect of the patient, the heart rate data of the patient in a period of time is acquired to establish a prediction model, and the heart rate data of the patient in a future period of time is predicted, and the heart rehabilitation effect of the patient is predicted and evaluated through the predicted heart rate data. Therefore, the embodiment of the present application acquires the heart rate data of the patient at each preset time point in each preset real time period.
[0048] In an embodiment of the present application, the preset real time period is set to 1 day, and the preset time point is set to 1 second, that is, the heart rate data of the patient every 1 second in a day is acquired. It should be noted that in other embodiments of the present application, the preset real time period and the preset time point can be set by the user, which is not limited here.
[0049] In addition, in one embodiment of the present application, all heart rate data of 10 days is acquired, while in other embodiments of the present application, the number of days of acquisition can be set by the user and is not limited herein.
[0050] In actual situations, a heart recovery patient will have a recovery trend in a certain period of time, that is, the heart rate data will show a trend, and the human heart rate also has a circadian rhythm, so it will show strong seasonality. Therefore, in the embodiment of the present application, the STL algorithm is used, and an STL decomposition prediction model is constructed to further obtain the predicted heart rate data of the heart recovery patient in the future period. And the STL algorithm is used in the following processes.
[0051] Step S2: optionally, heart rate data at a preset time in a preset real period is taken as target heart rate data; according to the data variation characteristics of the target heart rate data in a preset time neighborhood, the data variation degree of the target heart rate data is obtained; according to the variation difference between the target heart rate data and other heart rate data in the preset time neighborhood and the data variation degree of the target heart rate data, the accidental index of the target heart rate data is obtained; according to the variation characteristics of the target heart rate data relative to other heart rate data in the same preset real period, the time window of the target heart rate data is obtained; according to the similarity characteristics between the time window of the target heart rate data and the time window of heart rate data at the same preset time in each preset real period, and the variation characteristics of heart rate data in the time window of the target heart rate data, the memory index of the target heart rate data is obtained.
[0052] Since the heart rate of a heart recovery patient is easily affected by immediate psychological reactions, such as stress reaction, short-term intense exercise, environmental change, etc., the influence on the heart rate data is usually short-term fluctuation, and the STL algorithm cannot process the heart rate data with mutation, so in the subsequent process, inaccurate decomposition items will be obtained, so it is necessary to analyze the mutation of the heart rate data, and then obtain the accidental index of the heart rate data. In the embodiment of the present application, first, the mutation of the heart rate data is analyzed, that is, according to the data variation characteristics of the target heart rate data in a preset time neighborhood, the data variation degree of the target heart rate data is obtained.
[0053] Preferably, in one embodiment of the present application, the method for obtaining the data variation degree comprises:
[0054] The data difference between each adjacent two heart rate data in the preset time neighborhood of the target heart rate data is averaged to obtain the data variation degree of the target heart rate data. In one embodiment of the present application, the data variation degree calculation formula is as follows:
[0055]
[0056] In the formula, t represents the serial number of the target heart rate data; B tdenotes the data variation degree of the target heart rate data; G denotes the number of heart rate data on the right side of the target heart rate data in the preset time neighborhood; and AR t+i denotes the difference between each adjacent two heart rate data in the preset time neighborhood of the target heart rate data; and || denotes an absolute value function.
[0057] In the data variation degree calculation formula, denotes the data difference mean value between each adjacent two heart rate data in the preset time neighborhood of the target heart rate data from t-G to t+G, and the greater the data difference mean value, the greater the data variation degree of the target heart rate data.
[0058] In an embodiment of the present application, the preset time neighborhood is set as the target heart rate data and each two heart rate data before and after the target heart rate data, that is, G=2. It should be noted that in other embodiments of the present application, the preset time neighborhood can be set by itself, which is not limited herein.
[0059] In the embodiment of the present application, the sporadic index of the target heart rate data is obtained according to the variation difference between the target heart rate data and other heart rate data in the preset time neighborhood and the data variation degree of the target heart rate data.
[0060] Preferably, in an embodiment of the present application, the method for obtaining the sporadic index comprises:
[0061] The sporadic index is obtained according to the sporadic index calculation formula, and the sporadic index calculation formula is as follows:
[0062]
[0063] In the formula, t denotes the serial number of the target heart rate data; δ t denotes the sporadic index of the target heart rate data; B t denotes the data variation degree of the target heart rate data; denotes the data variation degree mean value of all heart rate data in the preset real time period in which the target heart rate data is located; N denotes the number of all heart rate data in the preset real time period in which the target heart rate data is located; I denotes the number of heart rate data on the right side of the target heart rate data in the preset time neighborhood; and AR t+i denotes the difference between each adjacent two heart rate data in the preset time neighborhood of the target heart rate data; and ΔR n+i denotes the difference between each adjacent two heart rate data in the preset time neighborhood of the nth other heart rate data in the preset real time period in which the target heart rate data is located, wherein n is not the same as t; max( ) denotes a maximum value function; and norm[ ] denotes a normalization function.
[0064] In the sporadic index calculation formula, a change abnormality feature of the target heart rate data, wherein when the change abnormality feature of the target heart rate data exceeds the mean value of the data change degree The more the change abnormality feature of the target heart rate data, the greater the sporadic index of the target heart rate data, and when the change abnormality feature of the target heart rate data does not exceed the mean value of the data change degree The change abnormality feature of the target heart rate data is 0, and the sporadic index of the target heart rate data is 0; the data difference between each two adjacent heart rate data in the preset time neighborhood of the target heart rate data and the difference ΔR between each two adjacent heart rate data in the preset time neighborhood of each other heart rate data t+i -ΔR n+i The greater the difference between the target heart rate data and the other heart rate data in the preset real time period, the more unusual the change mode of the target heart rate data, and the greater the sporadic index of the target heart rate data.
[0065] In most cases, the heart rate of a cardiac rehabilitation patient changes regularly, for example, the heart rate change after behaviors such as diet, regular exercise or drug use is usually similar, and the long-term prediction of the heart rate data is greatly helpful by identifying these regular changes, that is, the heart rate data is usually in memory for a long period of time, which is manifested as a strong periodicity and stable time window of the heart rate data, so in the embodiment of the present application, the time window of the target heart rate data is defined, and the memory index of the target heart rate data is analyzed.
[0066] Preferably, in an embodiment of the present application, the method for obtaining the time window comprises:
[0067] Since the human body is in a resting state in most cases, the mode of all the heart rate data in the preset real time period of the target heart rate data is taken as the resting heart rate data in the preset real time period.
[0068] The difference between the target heart rate data and the resting heart rate data is calculated as the relative heart rate deviation of the target heart rate data.
[0069] When the relative heart rate deviation of the target heart rate data is less than a preset first threshold, the heart rate data is taken as non-change heart rate data, and when the relative heart rate deviation of the target heart rate data is not less than the preset first threshold, the heart rate data is taken as change heart rate data. In an embodiment of the present application, the preset first threshold is set to be It should be noted that in other embodiments of the present application, the preset first threshold can be set by itself, which is not limited herein.
[0070] Traverse forward and backward along the preset time of the target heart rate data until the first non-changing heart rate data appears, and a set composed of the preset time of all the changing heart rate data is taken as the time window of the target heart rate data; traverse all the heart rate data to obtain the time window of each heart rate data.
[0071] Preferably, in one embodiment of the present application, the memory index acquisition method comprises:
[0072] The memory index is calculated according to the memory index calculation formula as follows:
[0073]
[0074] In the formula, t represents the serial number of the target heart rate data; μ t represents the memory index of the target heart rate data; C t represents the length of the time window of the target heart rate data; M represents the number of other preset real time periods different from the preset real time period where the target heart rate data is located; C m represents the length of the time window of the heart rate data at the same preset time as the target heart rate data in the mth preset real time period; represents the average data variation degree of all the heart rate data in the time window of the target heart rate data; DTW(C t , C m ) represents the matching degree between the time window of the target heart rate data and the time window of the heart rate data at the same preset time as the target heart rate data in the mth preset real time period, which can be directly obtained by the prior art; norm() represents a normalization function.
[0075] In the memory index calculation formula, the average matching degree between the time window of the target heart rate data and the time window of the heart rate data at the same preset time in each preset real time period is smaller, the greater the similarity between the two time windows is, the stronger the periodicity of the target heart rate data is, the stronger the memory of the target heart rate data in different preset real time periods is, i.e., the greater the memory index of the target heart rate data is; the greater the length of the time window C t of the target heart rate data is, the stronger the memory of the target heart rate data is, the greater the memory index of the target heart rate data is; the smaller the average data variation degree of all the heart rate data in the time window of the target heart rate data is , the stronger the stability of the heart rate data in the time window of the target heart rate data is, the greater the memory index of the target heart rate data is.
[0076] Step S3: constructing a prediction model for all heart rate data and decomposing to obtain a trend item, a seasonal item and a residual item for each future preset time; obtaining heart rate data for each future preset time according to the trend item, the seasonal item and the residual item for each future preset time, and the occasional index and the memory index of all heart rate data in the preset time window; and obtaining a heart recovery index in the preset future period according to the difference feature of each heart rate data in the preset future period relative to all heart rate data in the preset real period and the data change degree of each heart rate data in the preset future period.
[0077] The occasional index and the memory index of all preset time heart rate data can be obtained through the foregoing process, and then the STL decomposition prediction model is constructed. It is known that, in the process of predicting heart rate data at a future preset time by using a conventional STL algorithm, the predicted heart rate data is obtained by respectively calculating the size of the trend item, the seasonal item and the residual item of the heart rate data at the future preset time and then adding them together. However, the heart rate data of a heart recovery patient has complex and unstable changes, and the heart rate data obtained by using the fixed combination method often has a large deviation from the true value. Therefore, in the embodiment of the present application, the three basic prediction decomposition items for each future preset time are first calculated, and then the prediction decomposition items are adaptively combined according to the real-time heart rate data features to obtain accurate heart rate prediction data. Therefore, in the embodiment of the present application, the heart rate data for each future preset time is obtained according to the trend item, the seasonal item and the residual item for each future preset time, and the occasional index and the memory index of all heart rate data in the preset time window.
[0078] In an embodiment of the present application, a method for obtaining a prediction decomposition item at a future pth preset time is provided, which specifically comprises:
[0079] ① Trend item T p : STL decomposition is performed on all heart rate data, and for a sequence composed of trend items, a least square method is used to fit the sequence to capture the long-term trend of the data, thereby obtaining a prediction trend item T p for the future pth preset time.
[0080] ② Seasonal item S p : The seasonal item reflects the periodic characteristics of the heart rate data. According to the living rhythm of the human body, the period of the heart rate data is one day, and the size of the seasonal item at the same position in each period is basically the same. Therefore, the seasonal items of the heart rate data at the same preset time in other preset periods are added and averaged to obtain a prediction seasonal item S p for the future pth preset time.
[0081] ③ Residual item L pThe residual term reflects the random fluctuation part of the heart rate data, and thus the prediction residual term L at the future pth preset time point is obtained by using a moving average method on the residual terms at the same time point as the future pth preset time point in a preset number of preset time periods in the future p . In the present embodiment, the preset number is set to 5, and in other embodiments of the present application, the preset number can be set by the user and is not limited herein.
[0082] The STL algorithm is a well-known technical means to those skilled in the art, and thus is not described herein.
[0083] Preferably, in an embodiment of the present application, the method for obtaining the heart rate data at each preset time point comprises:
[0084] The heart rate data at each future preset time point is obtained according to the calculation formula of the heart rate data at each future preset time point, and the calculation formula of the heart rate data at each future preset time point is as follows:
[0085]
[0086] In the formula, R p represents the heart rate data at the future pth preset time point; T p represents the trend term at the future pth preset time point; represents the mean of the occurrence index of all heart rate data in the preset time window at the future pth preset time point; represents the mean of the memory index of all heart rate data in the preset time window at the future pth preset time point; S p represents the seasonal term at the future pth preset time point; and L p represents the residual term at the future pth preset time point.
[0087] In the calculation formula of the heart rate data at each future preset time point, the smaller the mean of the occurrence index and the mean of the memory index of all heart rate data in the preset time window at the future pth preset time point, the greater the weight of the prediction trend term T p at the future pth preset time point; the greater the mean of the occurrence index of all heart rate data in the preset time window at the future pth preset time point, the greater the weight of the prediction residual term L p at the future pth preset time point; and the greater the mean of the memory index of all heart rate data in the preset time window at the future pth preset time point, the greater the weight of the prediction seasonal term S p at the future pth preset time point.
[0088] The heart recovery effect of the patient is evaluated by using the heart rate data of each preset time in the preset future period. Since in the real situation, the judgment of the heart recovery effect is mainly based on the overall stability of the heart rate, the change of the resting heart rate data relative to the preset real period, and the range of the data change degree of the heart rate data, in the embodiment of the present application, the heart recovery index in the preset future period is obtained according to the difference characteristics of each heart rate data in the preset future period relative to all heart rate data in the preset real period and the data change degree of each heart rate data in the preset future period.
[0089] In an embodiment of the present application, the preset future period is set to 1 day, that is, the heart rate data of every 1 second in the future 1 day is predicted. It should be noted that in other embodiments of the present application, the preset future period can be set by itself, which is not limited here.
[0090] Preferably, in an embodiment of the present application, the method for obtaining the heart recovery index comprises:
[0091] The mode of all heart rate data in the preset future period is taken as the predicted resting heart rate data in the preset future period;
[0092] The heart recovery index is obtained according to the heart recovery index calculation formula, and the heart recovery index calculation formula is as follows:
[0093]
[0094] In the formula, γ represents the heart recovery index in the preset future period; R represents the average value of all resting heart rate data in the preset real period; fes R represents the predicted resting heart rate data in the preset future period; R represents the data change degree average value of all heart rate data in the preset future period; max R represents the maximum value of the heart rate data in the preset future period; min R represents the minimum value of the heart rate data in the preset future period; and norm[] represents a normalization function.
[0095] In the heart recovery index calculation formula, the difference between the predicted resting heart rate data in the preset future period and the average value of all resting heart rate data in the preset real period is greater, which means that the heart recovery effect in the preset future period is better, that is, the heart recovery index in the preset future period is greater; the data change degree average value of all heart rate data in the preset future period is smaller, which means that the overall stability of the heart rate in the preset future period is stronger, and at this time, the heart recovery effect in the preset future period is better, that is, the heart recovery index in the preset future period is greater; the maximum difference (R max -Rmin ) the smaller, the smaller the heart rate fluctuation range, and the better the heart recovery effect of the preset future period, that is, the greater the heart recovery index in the preset future period.
[0096] At this point, the heart recovery index in the preset future period is obtained.
[0097] Step S4: evaluating the heart recovery effect according to the heart recovery index.
[0098] In an embodiment of the present application, the preset second threshold is set to 0.7, and the preset third threshold is set to 0.4. When the predicted heart recovery index of the future 7 consecutive days is not less than the preset second threshold, it is considered that the patient's heart recovery state is good, and the more consecutive days that are not less than the preset second threshold, the better the heart recovery effect. If there are two or more days of heart recovery index not greater than the preset third threshold, it is considered that the patient's heart recovery state may be abnormal, and the patient should be notified for review as soon as possible to achieve early prevention of the disease and reduce potential risks. It should be noted that in other embodiments of the present application, the preset second threshold, the preset third threshold, and the consecutive days can be set by the user and are not limited herein.
[0099] At this point, the evaluation and prediction of the heart recovery effect are completed.
[0100] In summary, the heart rate data of the patient at each preset time in each preset real period is obtained; the heart rate data at a preset time in a preset real period is selected as the target heart rate data; the data variation degree of the target heart rate data is obtained according to the data variation characteristics of the target heart rate data in the preset time neighborhood; the sporadic index of the target heart rate data is obtained according to the variation difference between the target heart rate data and other heart rate data in the preset time neighborhood and the data variation degree of the target heart rate data; the time window of the target heart rate data is obtained according to the variation characteristics of the target heart rate data relative to other heart rate data in the same preset real period; the memory index of the target heart rate data is obtained according to the similarity characteristics between the time window of the target heart rate data and the time window of the heart rate data at the same preset time in each preset real period, and the variation characteristics of the heart rate data in the time window of the target heart rate data; a prediction model is constructed for all heart rate data and is decomposed to obtain the trend item, the seasonal item, and the residual item of each future preset time; the heart rate data of each future preset time is obtained according to the trend item, the seasonal item, and the residual item of each future preset time, and the sporadic index and the memory index of all heart rate data of the preset time window; the heart recovery index in the preset future period is obtained according to the difference characteristics of each heart rate data in the preset future period relative to all heart rate data in the preset real period and the data variation degree of each heart rate data in the preset future period; and the heart recovery effect is evaluated according to the heart recovery index.
[0101] An embodiment of the present application provides a heart rehabilitation effect intelligent evaluation and prediction system, which comprises a memory, a processor and a computer program, wherein the memory is used for storing the corresponding computer program, the processor is used for running the corresponding computer program, and the computer program can realize the method described in steps S1-S4 when running in the processor.
[0102] A third object of the embodiment of the present application is to provide a computer device, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, and the processor realizes the method described in steps S1-S4 when executing the computer program.
[0103] A fourth object of the embodiment of the present application is to provide a computer readable storage medium, which stores a computer program, and the computer program realizes the method described in steps S1-S4 when executed by a processor.
[0104] It should be noted that the above-mentioned embodiment sequence of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.
[0105] Each embodiment in the specification is described in a progressive manner, and the same and similar parts between each embodiment can be referred to each other. Each embodiment mainly describes the difference from other embodiments.
Claims
1. A method for intelligent evaluation and prediction of cardiac rehabilitation effect, characterized in that, The method comprises: acquiring heart rate data of each preset time point in each preset real time period of a patient; optionally, acquiring heart rate data of a preset time point in a preset real time period as target heart rate data; obtaining a data variation degree of the target heart rate data according to data variation characteristics of the target heart rate data in a preset time neighborhood; obtaining an accident index of the target heart rate data according to variation differences between the target heart rate data and other heart rate data in the preset time neighborhood and the data variation degree of the target heart rate data; obtaining a time window of the target heart rate data according to variation characteristics of the target heart rate data relative to other heart rate data in the same preset real time period; obtaining a memory index of the target heart rate data according to similarity characteristics between the time window of the target heart rate data and time windows of heart rate data of the same preset time point in each preset real time period and variation characteristics of heart rate data in the time window of the target heart rate data; constructing a prediction model for all heart rate data and decomposing the prediction model to obtain a trend item, a seasonal item and a residual error item of each future preset time point; obtaining heart rate data of each future preset time point according to the trend item, the seasonal item and the residual error item of each future preset time point and the accident index and the memory index of all heart rate data in a preset time window; obtaining a heart recovery index in a preset future time period according to difference characteristics of each heart rate data in the preset future time period relative to all heart rate data in a preset real time period and a data variation degree of each heart rate data in the preset future time period; evaluating a heart recovery effect according to the heart recovery index.
2. The intelligent evaluation and prediction method of cardiac rehabilitation effect according to claim 1, characterized in that, The method for obtaining the data variation degree comprises: averaging data differences between each adjacent two heart rate data in a preset time neighborhood of the target heart rate data to obtain the data variation degree of the target heart rate data. 3.The intelligent evaluation and prediction method of cardiac rehabilitation effect according to claim 1, characterized in that, The method for obtaining the accident index comprises: obtaining the accident index according to an accident index calculation formula, the accident index calculation formula being as follows: In the formula, t represents the serial number of the target heart rate data; δ t represents the sporadic index of the target heart rate data; B t a data variation degree of the target heart rate data; a data variation degree mean value of all heart rate data in a preset real time period in which the target heart rate data is located; N represents a quantity of all heart rate data in a preset real time period in which the target heart rate data is located; and G represents a quantity of heart rate data on the right side of the target heart rate data in a preset time neighborhood. △R t+i △R represents the difference between any two adjacent heart rate data points within a preset time neighborhood of the target heart rate data. n+i This indicates the number of times the target heart rate data falls within a preset real-time period. n The difference between any two adjacent heart rate data points within a preset time neighborhood, where n and t are not the same; max() represents the maximum value function; norm[] represents a normalization function.
4. The intelligent evaluation and prediction method of cardiac rehabilitation effect according to claim 1, characterized in that, The method for obtaining the time window comprises: taking a mode of all heart rate data in the preset real time period in which the target heart rate data is located as resting heart rate data in the preset real time period; calculating a difference between the target heart rate data and the resting heart rate data as a relative heart rate deviation of the target heart rate data; when the relative heart rate deviation of the target heart rate data is less than a preset first threshold, taking the heart rate data as non-variation heart rate data, and when the relative heart rate deviation of the target heart rate data is not less than the preset first threshold, taking the heart rate data as variation heart rate data; traversing forward and backward along the preset time point in which the target heart rate data is located until the first non-variation heart rate data is found, taking a set composed of preset time points corresponding to all variation heart rate data as a time window of the target heart rate data, and traversing all heart rate data to obtain a time window of each heart rate data.
5. The intelligent evaluation and prediction method of cardiac rehabilitation effect according to claim 1, characterized in that, The method for obtaining the memory index comprises: obtaining the memory index according to a memory index calculation formula, the memory index calculation formula being as follows: In the formula, t represents the serial number of the target heart rate data; μ t represents the memory index of the target heart rate data; C t denotes the length of the time window of the target heart rate data; M denotes the number of other preset real time periods different from the preset real time period where the target heart rate data is located; C m denotes the length of the time window of the heart rate data in the mth preset real time period that is at the same preset time as the target heart rate data; denotes the average of the data variation degrees of all the heart rate data in the time window of the target heart rate data; DTW(C t , C m ) denotes the matching degree between the time window of the target heart rate data and the time window of the heart rate data in the mth preset real time period that is at the same preset time as the target heart rate data; norm() denotes a normalization function.
6. The intelligent evaluation and prediction method of cardiac rehabilitation effect according to claim 1, characterized in that, The method for obtaining heart rate data of each future preset time point comprises: According to a future preset time point heart rate data calculation formula, future preset time point heart rate data is obtained, and the future preset time point heart rate data calculation formula is as follows: wherein R p represents the heart rate data at the pth preset time in the future; T p represents the trend term at the pth preset time in the future; represents the mean of the sporadic index of all heart rate data within the preset time window at the pth preset time in the future; represents the mean of the memory index of all heart rate data within the preset time window at the pth preset time in the future; S p represents the seasonal term at the pth preset time in the future; L p represents the residual term at the pth preset time in the future.
7. The intelligent evaluation and prediction method of cardiac rehabilitation effect according to claim 1, characterized in that, The heart recovery index acquisition method comprises: The mode of all heart rate data in a preset future period is taken as predicted resting heart rate data in the preset future period. According to a heart recovery index calculation formula, the heart recovery index is obtained, and the heart recovery index calculation formula is as follows: In the formula, γ represents a heart recovery index in a preset future period; represents an average value of all resting heart rate data in a preset real period; R fes (K) represents predicted resting heart rate data in a preset future period; a data variability average value representing all heart rate data of a predetermined future period; R max a maximum heart rate data value of a predetermined future period; R min a minimum heart rate data value of a predetermined future period; Norm[] represents a normalization function.
8. An intelligent evaluation and prediction system for cardiac rehabilitation effect, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the heart recovery effect intelligent evaluation and prediction method according to any one of claims 1-7.
9. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 8. The computer program is executed by the processor to implement the steps of the heart recovery effect intelligent evaluation and prediction method according to any one of claims 1-7.
10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the heart recovery effect intelligent evaluation and prediction method according to any one of claims 1-7.