Heart rate monitoring system based on cloud computing

Abstract

The present invention relates to the field of biomedical monitoring technology, specifically to a heart rate monitoring system based on cloud computing, the system comprising a heart rate data analysis module, a heartbeat abnormality recognition module, a heart disease risk assessment module, and a patient status monitoring module. In the present invention, by comprehensive time series analysis and heart rate pattern research, heart rhythm abnormality recognition and heart disease risk assessment are achieved, by analyzing the heart rate time series, combined with the local fluctuation of the heart rate and the frequency domain characteristic analysis, the types of heartbeat abnormality and arrhythmia can be more carefully identified, not only improving the diagnostic accuracy of arrhythmia, but also providing more abundant data support for the risk assessment of heart disease, by combining the arrhythmia level with historical health data, the risk results obtained by the assessment are more forward-looking and personalized, the patient status is continuously monitored, and the heart rate deviating from the normal range is marked in real time, enhancing the continuity of patient monitoring and the emergency response capability of the system.

Description

Technical Field

[0001] The present invention relates to the field of biomedical monitoring technology, and in particular to a heart rate monitoring system based on cloud computing. Background Art

[0002] The field of biomedical monitoring technology involves the use of various sensors, devices and systems to track and analyze individual physiological parameters to provide real-time health status information. The core of the field is to interpret the data collected from biomedical sensors through advanced data analysis technologies such as machine learning and pattern recognition. The data includes heart rate, blood pressure, blood oxygen level, etc., which are used to monitor the health status and physiological responses of individuals. The technology in this field not only supports the application of medical professionals in diagnosis, treatment decision-making and patient monitoring, but also realizes telemedicine and home health management, and provides the public with better health management tools.

[0003] Among them, the cloud computing-based heart rate monitoring system is a system that uses cloud technology to achieve real-time collection, storage and analysis of heart rate data. This system mainly collects data through heart rate sensors installed on individuals and then transmits the data to the cloud platform. On the cloud platform, the data is not only stored securely, but can also be processed and analyzed through advanced data analysis tools, thereby providing real-time health status feedback to users or medical professionals. The main uses of this system include real-time health monitoring, disease prevention, medical intervention, and long-term tracking of health data, and is particularly suitable for continuous monitoring of heart patients or performance evaluation of athletes.

[0004] The main shortcoming of existing heart rate monitoring systems is that they rely on more traditional data processing and analysis technologies, which limits the ability to perform in-depth analysis of complex physiological data. Although existing technologies support basic physiological parameter monitoring and simple pattern recognition, they cannot effectively process large-scale data or identify more complex types of arrhythmias in real time. For example, when processing highly irregular heart rate data, traditional technologies cannot accurately distinguish the specific types of arrhythmias, resulting in delays in early warning and intervention measures. The existing systems are also relatively limited in their functions in personalized health management and long-term disease monitoring, making it difficult to provide patients with continuous and personalized health status feedback, resulting in poor results of patient management plans, increased risks of misdiagnosis and missed diagnosis, and limited application in chronic disease management and preventive medicine. Summary of the invention

[0005] This application provides a heart rate monitoring system based on cloud computing. The main shortcoming of existing heart rate monitoring systems is that they rely on more traditional data processing and analysis technologies, which limits the ability to perform in-depth analysis in complex physiological data. Although existing technologies support basic physiological parameter monitoring and simple pattern recognition, they cannot effectively process large-scale data or identify more complex types of arrhythmias in real time. For example, when processing highly irregular heart rate data, traditional technologies cannot accurately distinguish the specific types of arrhythmias, resulting in delays in early warning and intervention measures. The existing systems are also relatively limited in their functions in personalized health management and long-term disease monitoring, making it difficult to provide patients with continuous and personalized health status feedback, resulting in poor results in patient management plans, increased risks of misdiagnosis and missed diagnosis, and limited application in chronic disease management and preventive medicine.

[0006] In view of the above problems, the present application provides a heart rate monitoring system based on cloud computing.

[0007] The present application provides a heart rate monitoring system based on cloud computing, wherein the system comprises:

[0008] The heart rate data analysis module collects heart rate data based on the photoelectric volumetric pulse wave sensor on the cloud platform, constructs a time series, analyzes the time point of each heartbeat, obtains the heart rate time series, inputs the heart rate time series into the processing link, preliminarily classifies the heartbeat pattern, and obtains the extended classification output result;

[0009] The heartbeat abnormality recognition module receives the extended classification output result, identifies the heartbeat abnormality by analyzing the local fluctuation and frequency domain characteristics of the heart rate, evaluates the irregularity of the heart rate, generates an abnormal diagnosis result, and classifies the arrhythmia based on the type and frequency of the abnormal diagnosis result to obtain an arrhythmia level output result;

[0010] The heart disease risk assessment module uses the arrhythmia level output result, combines historical health data and heart rate patterns for analysis, assesses the predicted risk of heart disease, obtains a risk assessment result, and uses the risk assessment result to optimize the patient's monitoring plan to obtain optimized patient monitoring recommendations;

[0011] Based on the optimized patient monitoring suggestions, the patient status monitoring module updates the patient's heart rate status record, records each heart activity, analyzes the frequency and stability of the heart rate, identifies abnormal heart rates that deviate from the normal range, marks the abnormal heart rates, summarizes the abnormal data, and compiles the heart rate monitoring results.

[0012] Preferably, the step of acquiring the heart rate time series is specifically as follows:

[0013] Based on the cloud platform, photoelectric volume pulse wave sensors are used to collect heart rate data, collect raw heart rate data, and generate raw heart rate data sets.

[0014] Analyze the time point of each heartbeat in the original heart rate data set, and calculate the time interval between two adjacent heartbeats using the formula:

[0015]

[0016] Get the heartbeat time interval array, where Δt i is the heartbeat time interval array, t i represents the time point of the i-th heartbeat, t i+1 represents the time point of the i+1th heartbeat, α is the linear adjustment coefficient, β is the nonlinear adjustment coefficient within the square root, Δt max represents the maximum time interval, θ is the minimum time interval proportional adjustment coefficient, Δt min Represents the minimum time interval;

[0017] The heart rate time series is constructed using the heartbeat time interval array using the formula:

[0018]

[0019] Calculate the number of heart beats per minute and generate a heart rate time series, where HR(t) represents the number of heart beats per minute, Δt represents the heart beat time interval array, γ is the linear correction coefficient, κ is the inverse proportional adjustment factor, and ∈ is a small positive number.

[0020] Preferably, the step of obtaining the extended classification output result is specifically:

[0021] The heart rate time series is input into the data processing link and classified using the heartbeat interval feature using the formula:

[0022]

[0023] Get the heartbeat pattern feature score, where S i represents the heartbeat pattern feature score, x j is the jth feature in the heart rate time series, w j is the weight of the jth feature, λ is the parameter to adjust the nonlinear effect, μ is the adjustment parameter for feature weighting, and n is the number of features;

[0024] A threshold is determined for the heartbeat pattern feature score of each sample using the following formula:

[0025]

[0026] Perform a preliminary classification of the heartbeat pattern and obtain the preliminary classification results, among which C irepresents the preliminary classification result, S is the heartbeat pattern feature score, θ is the classification threshold, σ is the standard deviation, ρ is the offset parameter, 1 indicates an abnormal heartbeat pattern, and 0 indicates a normal heartbeat pattern;

[0027] According to the preliminary classification results, the classification is refined and the formula is applied:

[0028]

[0029] Generate extended classification output results, where D represents the extended classification output results, C k is the preliminary classification result, v k is the weight of the corresponding classification label, m is the number of classification labels, and δ is the logistic regression intercept adjustment parameter.

[0030] Preferably, the steps of obtaining the abnormal diagnosis result are specifically:

[0031] Based on the extended classification output results, the local fluctuations in the heart rate time series are analyzed using the formula:

[0032]

[0033] Generate volatility results, where V is the volatility result, HR j is the heart rate at the jth time point, is the mean heart rate, σ j is the standard deviation at the jth time point, ∈ is a small positive number that avoids division by zero;

[0034] Based on the fluctuation results, the frequency domain characteristics of the heart rate sequence are analyzed using Fourier transform and frequency domain weighted formula:

[0035]

[0036] Generate frequency domain analysis results, where F is the frequency domain analysis result, FFT (HR k ) is the Fourier transform result, w k is the weight of frequency k, δ k is the frequency domain attenuation factor;

[0037] Combining the volatility results and the frequency domain analysis results, the weighted sum method is adopted and the nonlinear combination formula is used:

[0038]

[0039] Generate anomaly diagnosis results, where A represents the anomaly diagnosis result, V is the volatility result, F is the frequency domain analysis result, and α and β are adjustment coefficients.

[0040] Preferably, the step of obtaining the arrhythmia level output result is specifically:

[0041] Based on the abnormal diagnosis results, the distribution and frequency of abnormal types are analyzed using the formula:

[0042]

[0043] Calculate and generate the abnormal type distribution score, where T represents the abnormal type distribution score, f i is the frequency of abnormal occurrence, t i is the severity index of the anomaly type, k is a parameter used to adjust the impact of each anomaly severity, and g i is the severity level of each anomaly, and h is the decision threshold;

[0044] A weighted score is calculated based on the abnormality type distribution score and the clinical relevance of the abnormality using the formula:

[0045]

[0046] The arrhythmia severity score is calculated, where SA is the arrhythmia severity score, T j is the abnormal type distribution score, c j is the clinical relevance weight, δ is the global adjustment coefficient, r j is the response coefficient for each anomaly type;

[0047] Arrhythmias are graded using a grading scale based on the arrhythmia severity score, using the formula:

[0048]

[0049] The arrhythmia level output result is obtained, where L represents the arrhythmia level output result, SA is the arrhythmia severity score, α is the adjustment increase coefficient, {thresholds} is the classification threshold, and classify is the classification function.

[0050] Preferably, the steps for obtaining the risk assessment results are specifically:

[0051] Integrate the arrhythmia level output results and the patient's historical health data, and assign influence to risk factors in a weighted manner, applying the formula:

[0052]

[0053] The risk assessment value is calculated, where FA represents the risk assessment value, and w i is the weight of the ith risk factor, h i is the risk value corresponding to the historical health data, k i is the gain factor, l iis the risk factor corresponding to the arrhythmia level;

[0054] Based on the risk assessment value, heart rate variability analysis is performed using the heart rate pattern data using the formula:

[0055]

[0056] Integrate the contribution of heart rate data and calculate the heart rate pattern risk score, where R is the heart rate pattern risk score, p j is the parameter of the heart rate mode, v j is the heart rate variability analysis score, λ j is the normalization coefficient;

[0057] The risk assessment value and the heart rate pattern risk score are combined and integrated in a weighted manner, using the formula:

[0058]

[0059] Calculate the predicted risk of heart disease and generate a risk assessment result, where Q represents the risk assessment result, FA is the risk assessment value, R is the heart rate pattern risk score, and γ and δ are adjustment coefficients used to balance FA and R.

[0060] Preferably, the steps of obtaining the optimized patient monitoring suggestion are specifically:

[0061] Analyze the risk assessment results and calculate the risk-weighted sum using the formula:

[0062]

[0063] Where P represents the risk-weighted sum, u i is the importance coefficient of the risk indicator, v i is the value of the risk indicator, s i is the stability adjustment factor, ∈ is a small constant;

[0064] Based on the risk-weighted sum, adjust the treatment plan according to the risk level, synchronize information to the medical team, determine the level of monitoring, and apply the formula:

[0065]

[0066] Among them, Q represents the monitoring level, P is the risk-weighted sum, γ is the response sensitivity parameter, and δ is the response threshold;

[0067] Based on the level of supervision, a personalized supervision plan is designed using the formula:

[0068] N = ∫Q(t)·dt

[0069] The time-weighted method is used to calculate the monitoring needs within the duration and obtain the optimized patient monitoring recommendation, where N represents the personalized monitoring plan, Q(t) is the time function of the monitoring level, and t is the time variable.

[0070] Preferably, the steps of obtaining the heart rate monitoring results are as follows:

[0071] Based on the optimized patient monitoring recommendations, the patient's heart rate status record is updated using the formula:

[0072]

[0073] Calculate the updated heart rate state score, where U represents the updated heart rate state score, r i is the heart rate value recorded for the i-th cardiac activity, s i is the stability score of the corresponding heart rate state, δ is the coefficient for adjusting the influence of heart rate deviation, and μ is the average heart rate;

[0074] Based on the updated heart rate status score, abnormal heart rates that deviate from the normal range are marked using the formula:

[0075]

[0076] Generate an abnormal heart rate score, where VI is the abnormal heart rate score, f j is the heart rate value of the jth cardiac activity, ν is the average value of the normal heart rate, σ is the standard deviation of the normal heart rate, ∈ is the adjustment parameter, d j is the duration of the i-th record;

[0077] Based on the abnormal heart rate score, the abnormal data is summarized and the compiled heart rate monitoring result is generated using the formula:

[0078]

[0079] Among them, SI represents the compiled heart rate monitoring result, κ j is the severity weight of the jth abnormal heart rate, and VI is the abnormal heart rate score.

[0080] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0081] Through comprehensive time series analysis and in-depth research on heart rate patterns, more accurate abnormal heart rhythm identification and heart disease risk assessment can be achieved. Through detailed analysis of the heart rate time series, combined with local fluctuations in heart rate and frequency domain characteristic analysis, the types of abnormal heartbeats and arrhythmias can be identified more finely. This detailed identification process not only improves the diagnostic accuracy of arrhythmias, but also provides richer data support for heart disease risk assessment. By combining the arrhythmia level with historical health data, the risk results obtained are more forward-looking and personalized, making medical intervention more targeted and timely, continuously monitoring the patient's status, and marking the heart rate that deviates from the normal range in real time, further enhancing the continuity of patient monitoring and the system's emergency response capabilities.

[0082] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 The present invention proposes an overall flow chart of a heart rate monitoring system based on cloud computing;

[0084] Figure 2 A flowchart of the heart rate time series of a heart rate monitoring system based on cloud computing is proposed for the present invention;

[0085] Figure 3 A flow chart of the extended classification output results of the cloud computing-based heart rate monitoring system is proposed for the present invention;

[0086] Figure 4 A flow chart of abnormal diagnosis results of a heart rate monitoring system based on cloud computing is proposed for the present invention;

[0087] Figure 5 A flow chart of the arrhythmia level output results of the cloud computing-based heart rate monitoring system is proposed for the present invention;

[0088] Figure 6 A flow chart of the risk assessment results of the heart rate monitoring system based on cloud computing is proposed for the present invention;

[0089] Figure 7 A flowchart of the present invention for providing optimized patient monitoring recommendations for a cloud computing-based heart rate monitoring system;

[0090] Figure 8 The present invention proposes a flow chart of compiling heart rate monitoring results of a heart rate monitoring system based on cloud computing. DETAILED DESCRIPTION

[0091] The present application provides a heart rate monitoring system based on cloud computing.

[0092] Application Overview

[0093] The main shortcoming of the heart rate monitoring system in the existing technology is that it relies on more traditional data processing and analysis technology, which limits the ability to perform in-depth analysis in complex physiological data. Although the existing technology supports basic physiological parameter monitoring and simple pattern recognition, it cannot effectively process large-scale data or identify more complex types of arrhythmias in real time. For example, when processing highly irregular heart rate data, traditional technology cannot accurately distinguish the specific type of arrhythmia, resulting in delays in early warning and intervention measures. The existing system is also relatively limited in its functions in personalized health management and long-term disease monitoring, making it difficult to provide patients with continuous and personalized health status feedback, resulting in poor results of patient management plans, increased risks of misdiagnosis and missed diagnosis, and limited application in chronic disease management and preventive medicine.

[0094] In response to the above technical problems, the overall idea of ​​the technical solution provided by this application is as follows:

[0095] Example

[0096] like Figure 1 As shown, the present application provides a heart rate monitoring system based on cloud computing, wherein the system includes:

[0097] The heart rate data analysis module collects heart rate data based on the photoelectric volumetric pulse wave sensor on the cloud platform, constructs a time series, analyzes the time point of each heartbeat, obtains the heart rate time series, inputs the heart rate time series into the processing link, preliminarily classifies the heartbeat pattern, and obtains the extended classification output result;

[0098] The heartbeat abnormality recognition module receives the extended classification output result, identifies the heartbeat abnormality by analyzing the local fluctuation and frequency domain characteristics of the heart rate, evaluates the irregularity of the heart rate, generates an abnormal diagnosis result, and classifies the arrhythmia based on the type and frequency of the abnormal diagnosis result to obtain an arrhythmia level output result;

[0099] The heart disease risk assessment module uses the arrhythmia level output results, combines historical health data and heart rate patterns for analysis, evaluates the predicted risk of heart disease, obtains risk assessment results, and uses the risk assessment results to optimize the patient's monitoring plan and obtain optimized patient monitoring recommendations;

[0100] The patient status monitoring module updates the patient's heart rate status record based on optimized patient monitoring recommendations, records each heart movement, analyzes the frequency and stability of the heart rate, identifies abnormal heart rates that deviate from the normal range, marks abnormal heart rates, summarizes abnormal data, and compiles heart rate monitoring results.

[0101] The extended classification output results include frequency data, time intervals, and waveform characteristics; the arrhythmia level output results include arrhythmia type, frequency deviation, and critical value judgment; optimized patient monitoring recommendations include monitoring frequency, monitoring intensity, and preventive measures; and the compilation of heart rate monitoring results includes data abnormality records, heart rate fluctuation monitoring, and cumulative abnormal events.

[0102] Specifically, if Figure 2 As shown in the figure, the steps for obtaining the heart rate time series are as follows:

[0103] Based on the cloud platform, photoelectric volume pulse wave sensors are used to collect heart rate data, collect raw heart rate data, and generate raw heart rate data sets.

[0104] Analyze the time point of each heartbeat in the original heart rate data set and calculate the time interval between two adjacent heartbeats using the formula:

[0105]

[0106] Get the heartbeat time interval array, where Δt i is the heartbeat time interval array, t i represents the time point of the i-th heartbeat, t i+1 represents the time point of the i+1th heartbeat, α is the linear adjustment coefficient, β is the nonlinear adjustment coefficient within the square root, Δt max represents the maximum time interval, θ is the minimum time interval proportional adjustment coefficient, Δt min Represents the minimum time interval;

[0107] Use the heartbeat time interval array to construct the heart rate time series using the formula:

[0108]

[0109] Calculate the number of heart beats per minute and generate a heart rate time series, where HR(t) represents the number of heart beats per minute, Δt represents the heart beat time interval array, γ is the linear correction coefficient, κ is the inverse proportional adjustment factor, and ∈ is a small positive number.

[0110] The formula for the heartbeat time interval array is derived as follows:

[0111]

[0112] t i and t i+1 : The time point between two consecutive heartbeats;

[0113] α, β, θ: preset adjustment coefficients;

[0114] Δt max : Maximum time interval between heartbeats;

[0115] Δt min : The minimum time interval between heartbeats.

[0116] First calculate the time interval between two heartbeats: Δt raw =t i+1 -t i ;

[0117] Compute the absolute value of the difference from the maximum interval and take the square root:

[0118] If the heartbeat interval is less than the preset minimum value, adjust the ratio:

[0119] Combine the above calculation results with the adjustment factor: Δt i =α·Δt raw +β·adjustment max +θ·adjustment min Assume that t i =1.00s, t i+1 =1.80s, Δt max =2.00s, Δt min =0.50s, α=1.0, β=0.5, θ=0.2,

[0120] Calculate the original interval: Δt raw =1.80-1.00=0.80s,

[0121] Maximum interval adjustment:

[0122] Minimum interval adjustment:

[0123] Comprehensive adjustment results:

[0124] Δt i =1.0·0.80+0.5·1.10+0.2·0.625=0.80+0.55+0.125=1.475s.

[0125] The formula for the number of heart beats per minute is:

[0126]

[0127] Δt: heartbeat time interval array;

[0128] γ and κ: preset adjustment coefficients;

[0129] ∈: A small positive number to prevent division by zero errors.

[0130] Based on the intervals obtained, calculate the basal heart rate:

[0131] Adjust using factor: HR adjusted =HR base γ o ;

[0132] Calculate the correction term:

[0133] Final calculation result: HR(t) = HR adjusted -HR correction .

[0134] Continue to use the above Δt i =1.475s, γ=2.0, κ=1.5, ∈=0.01,

[0135] Basic heart rate:

[0136] Adjust heart rate: HR a adjusted=40.68·2.0=81.36BPM,

[0137] Abnormal heart rate correction:

[0138] Final heart rate: HR(t)=81.36-1.01=80.35 BPM.

[0139] Specifically, if Figure 3 As shown, the steps for obtaining the extended classification output results are as follows:

[0140] The heart rate time series is input into the data processing link, and the heartbeat interval feature is used for classification, using the formula:

[0141]

[0142] Get the heartbeat pattern feature score, where S i represents the heartbeat pattern feature score, x j is the jth feature in the heart rate time series, w j is the weight of the jth feature, λ is the parameter to adjust the nonlinear effect, μ is the adjustment parameter for feature weighting, and n is the number of features;

[0143] The threshold value of each sample heartbeat pattern feature score is judged using the judgment formula:

[0144]

[0145] Perform a preliminary classification of the heartbeat pattern and obtain the preliminary classification results, among which C irepresents the preliminary classification result, S is the heartbeat pattern feature score, θ is the classification threshold, σ is the standard deviation, ρ is the offset parameter, 1 indicates an abnormal heartbeat pattern, and 0 indicates a normal heartbeat pattern;

[0146] According to the preliminary classification results, the classification is refined and the formula is applied:

[0147]

[0148] Generate extended classification output results, where D represents the extended classification output results, C k is the preliminary classification result, v k is the weight of the corresponding classification label, m is the number of classification labels, and δ is the logistic regression intercept adjustment parameter.

[0149] Formula for calculating heartbeat pattern feature score:

[0150]

[0151] Feature Square: Calculate the square of each feature This reflects the amplitude enhancement of the feature;

[0152] Eigenroot and Logarithmic Weighting: Calculation Extract the fundamental characteristics of each feature and calculate log(x j ), to capture the sensitivity of features to changes at lower amplitudes;

[0153] Weighted sum: Add the above calculated result to the corresponding weight w j , λ, and μ are multiplied and then summed, where λ and μ adjust the influence of the square root and logarithmic terms.

[0154] Assume that for a heartbeat time series, there are two features x1=4 and x2=9, weights w1=0.5, w2=0.3, adjustment coefficients λ=0.2, μ=0.1,

[0155]

[0156] The result is S i ≈20.69.

[0157] Preliminary classification results C:

[0158] Set dynamic threshold: Calculate the dynamic threshold θ·σ+ρ, where θ is the threshold multiple, σ is the standard deviation of the heartbeat pattern feature scores, and ρ is the offset used to adjust the threshold to adapt to the data distribution;

[0159] Classification decision: According to S i Comparison results with dynamic thresholds are classified:

[0160]

[0161] S: score from the heartbeat pattern features;

[0162] θ, σ, ρ: threshold multiple, standard deviation and offset respectively;

[0163] Assume θ = 0.8, σ = 1.5, ρ = 0.05,

[0164] S = 1.3 (derived from step 1),

[0165] Calculate the threshold θ·σ+ρ=0.8·1.5+0.05=1.25,

[0166] Compare S to see if it is greater than or equal to 1.25 to allocate C

[0167] The formula for expanding the classification output results is:

[0168]

[0169] Logistic regression model application: Use the logistic regression model to refine the classification results;

[0170] C k : The kth classification label in the preliminary classification result;

[0171] v k : The weight of the corresponding classification label;

[0172] m: the number of classification labels;

[0173] Assume m = 2, v1 = 0.4, v2 = 0.6, δ = -0.1, C1 = 1, C2 = 0,

[0174] calculate

[0175] Applying the logistic regression model, we get

[0176] Specifically, if Figure 4 As shown, the specific steps for obtaining abnormal diagnosis results are:

[0177] Based on the extended classification output, the local fluctuations in the heart rate time series are analyzed using the formula:

[0178]

[0179] Generate volatility results, where V is the volatility result, HR j is the heart rate at the jth time point, is the mean heart rate, σ j is the standard deviation at the jth time point, ∈ is a small positive number that avoids division by zero;

[0180] Based on the fluctuation results, the frequency domain characteristics of the heart rate sequence are analyzed using Fourier transform and frequency domain weighted formula:

[0181]

[0182] Generate frequency domain analysis results, where F is the frequency domain analysis result, FFT (HR k ) is the Fourier transform result, w k is the weight of frequency k, δ k is the frequency domain attenuation factor;

[0183] Combining the volatility results with the frequency domain analysis results, the weighted sum method is adopted, and the nonlinear combination formula is used:

[0184]

[0185] Generate anomaly diagnosis results, where A represents the anomaly diagnosis result, V is the volatility result, F is the frequency domain analysis result, and α and β are adjustment coefficients.

[0186] The calculation formula of volatility result is:

[0187]

[0188] Collect heart rate data: Assume there is a heart rate sequence HR = [80, 82, 78, 85, 80], where N = 5;

[0189] Calculate heart rate average

[0190] Determine the standard deviation σ j and a small positive number ∈: Assume that for simplicity of calculation, the standard deviation σ at each time point j All are set to 1.5, and ∈ = 0.01;

[0191] Calculate the contribution at each time point: For each j, calculate

[0192] j=1:

[0193] j=2:

[0194] j=3:

[0195] j=4:

[0196] j=5:

[0197] Calculate volatility V:

[0198] The formula for calculating the frequency domain analysis results is:

[0199]

[0200] Perform Fourier transform: Assume that the result of FFT(HR) is [8.6, -1.5+2.3i, 2.5-1.1i, 2.5+1.1i, -1.5-2.3i], where K = 5;

[0201] Set the weight w k and the attenuation factor δ k : w=[1.0, 0.8, 0.5, 0.5, 0.8], δ=

[0202] 0.1, 0.2, 0.3, 0.3, 0.2];

[0203] Compute the weighted contribution of each frequency component:

[0204] k=1:

[0205] k=2:

[0206] k=3:

[0207] k=4:

[0208] k=5:

[0209] Calculate the frequency domain analysis result F: F=7.82+2.59+1.22+1.22+2.59=15.44.

[0210] Abnormal diagnosis result formula:

[0211]

[0212] Use the V and F calculated above and select appropriate α and β (e.g., α = 0.6, β = 0.4);

[0213] Calculate abnormal diagnosis result A:

[0214]

[0215] Specifically, if Figure 5 As shown, the steps for obtaining the arrhythmia level output result are as follows:

[0216] Based on the abnormal diagnosis results, the distribution and frequency of abnormal types are analyzed using the formula:

[0217]

[0218] Calculate and generate the abnormal type distribution score, where T represents the abnormal type distribution score, f i is the frequency of abnormal occurrence, t i is the severity index of the anomaly type, k is a parameter used to adjust the impact of each anomaly severity, and g i is the severity level of each anomaly, and h is the decision threshold;

[0219] A weighted score was calculated based on the abnormality type distribution score and the clinical relevance of the abnormality using the formula:

[0220]

[0221] The arrhythmia severity score is calculated, where SA is the arrhythmia severity score, T j is the abnormal type distribution score, c j is the clinical relevance weight, δ is the global adjustment coefficient, r j is the response coefficient for each anomaly type;

[0222] Arrhythmias are graded using a grading scale based on the arrhythmia severity score, using the formula:

[0223]

[0224] The arrhythmia level output result is obtained, where L represents the arrhythmia level output result, SA is the arrhythmia severity score, α is the adjustment increase coefficient, {thresholds} is the classification threshold, and classify is the classification function.

[0225] The formula for the anomaly type distribution score is:

[0226]

[0227] f i : The frequency of occurrence of the i-th anomaly type, assuming that this is obtained by counting the number of occurrences of each anomaly in the past period of time and dividing it by the total number of observations;

[0228] t i : The severity index of the i-th abnormality type, which can be obtained through expert scoring or historical data analysis;

[0229] k: induction intensity coefficient, assumed to be a constant, used to adjust the shape of the sigmoid function, which can be adjusted according to past data to match the actual situation;

[0230] g i: The severity level of a given abnormality, which is determined based on assessments from clinical data or historical records;

[0231] h: Reference threshold, which can be determined based on the average severity in historical data.

[0232] Assume there are three exception types with the following parameters:

[0233] f = [0.2, 0.3, 0.5],

[0234] t = [2, 5, 3],

[0235] k=1,g=[1,2,1.5],

[0236] h = 1.5,

[0237] Calculate T:

[0238]

[0239] T≈0.4+1.5+1.5=3.4.

[0240] The calculation formula for the arrhythmia severity score is:

[0241]

[0242] T j : The distribution score of the jth abnormal type;

[0243] c j : The weight coefficient of clinical relevance is determined by historical case analysis and expert evaluation;

[0244] δ: adjustment coefficient, used to adjust the response intensity of different anomalies, determined based on historical data analysis;

[0245] r j : Abnormal response coefficient, which reflects the influence of different abnormal types and can be obtained through data analysis.

[0246] Assumptions: c = [1, 2, 1.5], δ = 0.5, r = [1, 0.8, 1.2], T = [0.4, 1.5, 1.5],

[0247] Calculate SA:

[0248]

[0249] SA≈0.27+2.4+1.46=4.13.

[0250] The calculation formula for the arrhythmia level output result is:

[0251]

[0252] SA: Arrhythmia Severity Score;

[0253] α: coefficient increase, used to adjust the sensitivity of the classification, determined based on clinical needs;

[0254] {thresholds}: grading thresholds, which are set based on medical standards and historical data;

[0255] Assumptions: α = 2, thresholds = [1, 2, 3, 4], SA = 4.13,

[0256] Calculate L:

[0257]

[0258] L=classify(4.06, [1, 2, 3, 4]), given the threshold, L should correspond to the 4th level.

[0259] Specifically, if Figure 6 As shown in the figure, the steps for obtaining risk assessment results are as follows:

[0260] Integrate the arrhythmia level output results and the patient's historical health data to assign influence to risk factors in a weighted manner, applying the formula:

[0261]

[0262] The risk assessment value is calculated, where FA represents the risk assessment value, and w i is the weight of the ith risk factor, h i is the risk value corresponding to the historical health data, k i is the gain factor, l i is the risk factor corresponding to the arrhythmia level;

[0263] Based on the risk assessment value, heart rate variability analysis is performed using the heart rate pattern data using the formula:

[0264]

[0265] Integrate the contribution of heart rate data and calculate the heart rate pattern risk score, where R is the heart rate pattern risk score, p j is the parameter of the heart rate mode, v j is the heart rate variability analysis score, λ j is the normalization coefficient;

[0266] The risk assessment value and the heart rate pattern risk score are combined and integrated in a weighted manner using the formula:

[0267]

[0268] Calculate the predicted risk of heart disease and generate a risk assessment result, where Q represents the risk assessment result, FA is the risk assessment value, R is the heart rate pattern risk score, and γ and δ are adjustment coefficients used to balance FA and R.

[0269] Formula for calculating risk assessment value:

[0270]

[0271] w i : The weight of the i-th risk factor;

[0272] h i : Risk values ​​in historical health data;

[0273] k i : Adjustment coefficient, used to adjust h i The impact of

[0274] l i : Risk factor of arrhythmia level;

[0275] n: number of total risk factors;

[0276] Determine parameters: Set the weight w of each risk factor i , the risk value h of historical health data i , the adjustment coefficient k for each risk factor i , and the risk factor for arrhythmia i ;

[0277] Calculate the score for each item: For each risk factor, first calculate the

[0278] Applying the Adjustment Factor: Calculating the Adjustment Part

[0279] Calculate the adjusted risk value: Divide by the adjustment part to get the adjusted risk value;

[0280] Apply the arrhythmia factor: multiply the adjusted risk value by

[0281] Weighted sum: multiply the above calculation result by the weight w i And sum up to get the total risk assessment value FA.

[0282] Assume there are two risk factors:

[0283] w1=0.3, h1=2, k1=0.5, l1=1.2,

[0284] w2=0.7, h2=3, k2=0.3, l2=1.5,

[0285]

[0286] Formula for calculating heart rate pattern risk score:

[0287]

[0288] p j : Parameters of the heart rate pattern, representing specific heart rate variability indicators;

[0289] v j : Heart rate variability analysis score;

[0290] λ j : Normalization coefficient, used to adjust the impact of variability score;

[0291] m: total number of heart rate patterns;

[0292] Determine parameters: Set parameters p for each heart rate mode j , variability score v j , and the normalization coefficient λ j ;

[0293] Calculate the variability score: For each heart rate variability score v j Apply logarithmic transformation to improve the nonlinear response of the score, i.e. log91+v j );

[0294] Apply normalization coefficient: Divide the log-transformed scores by the normalization coefficient λ j , to balance the contribution of each heart rate mode;

[0295] Weighted sum: Multiply the above result by the heart rate mode parameter p j And sum up all heart rate patterns to get the total heart rate pattern risk score R.

[0296] Assume there are two heart rate mode parameters:

[0297] p1=0.4, v1=15, λ1=5,

[0298] p2=0.6, v2=10, λ2=3,

[0299]

[0300] Risk assessment result formula:

[0301]

[0302] FA: risk assessment value;

[0303] R: heart rate pattern risk score;

[0304] γ, δ: adjustment coefficients used to balance the effects of F and R;

[0305] Apply powers and roots: Apply 2.5 powers to FA and square roots to R to strengthen the impact of high risk scores and reduce the impact of low risk scores;

[0306] Weighted sum: FA 2.5 and Weighted according to their corresponding adjustment coefficients γ and δ;

[0307] Averaging: Divide the weighted sum by the sum of the adjustment coefficients γ+δ to obtain the final balanced risk assessment result.

[0308] Assume γ = 2, δ = 1, and use the F and R results from the previous step:

[0309] FA=10.71,R=0.702,

[0310]

[0311] Specifically, if Figure 7 As shown in Figure 1, the steps to obtain optimized patient monitoring recommendations are as follows:

[0312] Analyze the risk assessment results and calculate the risk-weighted sum using the formula:

[0313]

[0314] Where P represents the risk-weighted sum, u i is the importance coefficient of the risk indicator, v i is the value of the risk indicator, s i is the stability adjustment factor, ∈ is a small constant;

[0315] Based on the risk-weighted sum, adjust the treatment plan according to the risk level, synchronize information to the medical team, determine the level of monitoring, and apply the formula:

[0316]

[0317] Among them, Q represents the monitoring level, P is the risk-weighted sum, γ is the response sensitivity parameter, and δ is the response threshold;

[0318] Based on the level of guardianship, a personalized guardianship plan is designed using the formula:

[0319] N = ∫Q(t)·dt

[0320] The time-weighted method is used to calculate the monitoring needs within the duration and obtain the optimized patient monitoring recommendation, where N represents the personalized monitoring plan, Q(t) is the time function of the monitoring level, and t is the time variable.

[0321] Formula for calculating risk-weighted sum:

[0322]

[0323] u i : The importance coefficient of the i-th risk indicator is assigned based on historical data and expert knowledge;

[0324] v i : The current value of the i-th risk indicator, obtained directly from the patient's medical record;

[0325] s i : The stability adjustment factor of the risk indicator can be obtained from the standard deviation or coefficient of variation through statistical analysis of long-term monitoring data;

[0326] ∈: A small constant used to prevent division by zero errors, usually set to a very small value such as 0.0001.

[0327] Assume there are three risk indicators:

[0328] u=[0.3, 0.5, 0.2], v=[10, 15, 5], s=[1, 2, 1.5], ∈=0.0001,

[0329] Calculate the score for each metric:

[0330] For indicator 1:

[0331] For indicator 2:

[0332] For indicator 3:

[0333] Multiply the result of each metric by its importance coefficient and sum them:

[0334] P=0.3×2.302+0.5×1.245+0.2×0.721=1.445,

[0335] Formula for determining the level of guardianship:

[0336]

[0337] γ, δ: coefficients are set by fitting historical data or expert opinions to adjust the response intensity and set thresholds;

[0338] Assume γ = 0.1, δ = 2, P = 1.445,

[0339] Calculate the exponential part: γ·P-δ=0.1×1.445-2=-1.855,

[0340] Calculating the level of custody:

[0341] Design a personalized guardianship plan formula:

[0342] N = ∫Q(t)·dt

[0343] Q(t): monitoring level function, which can be calculated in real time or based on a time series prediction model;

[0344] Assuming Q(t) = 0.924 is constant,

[0345] Calculate the cumulative monitoring demand within a certain time period T, for example, T = 24 hours:

[0346]

[0347] Specifically, if Figure 8 As shown, the steps for obtaining the heart rate monitoring results are as follows:

[0348] Based on the optimized patient monitoring recommendations, update the patient's heart rate status record using the formula:

[0349]

[0350] Calculate the updated heart rate state score, where U represents the updated heart rate state score, r i is the heart rate value recorded for the i-th cardiac activity, s i is the stability score of the corresponding heart rate state, δ is the coefficient for adjusting the influence of heart rate deviation, and μ is the average heart rate;

[0351] Based on the updated heart rate status score, mark abnormal heart rates that deviate from the normal range, using the formula:

[0352]

[0353] Generate an abnormal heart rate score, where VI is the abnormal heart rate score, f j is the heart rate value of the jth cardiac activity, ν is the average value of the normal heart rate, σ is the standard deviation of the normal heart rate, ∈ is the adjustment parameter, d j is the duration of the i-th record;

[0354] Based on the abnormal heart rate score, the abnormal data is summarized and the compiled heart rate monitoring results are generated using the formula:

[0355]

[0356] Among them, SI represents the compiled heart rate monitoring result, κ j is the severity weight of the jth abnormal heart rate, and VI is the abnormal heart rate score.

[0357] Updated heart rate status score formula:

[0358]

[0359] r i : The heart rate value of the i-th heart activity record;

[0360] s i : Stability score corresponding to the heart rate state;

[0361] δ: attenuation coefficient, used to adjust the impact of deviation from the average heart rate;

[0362] μ: average heart rate;

[0363] U: Updated heart rate status score.

[0364] First, determine the heart rate value r for each heart activity i and the corresponding stability score s i , calculate the absolute value of the difference between each heart rate value and the average heart rate μ|r i -μ|, multiply the absolute value of the above difference by the attenuation coefficient δ and take the exponential To reduce the impact of heart rate deviation, for each The sum is performed to calculate the total score U;

[0365] Assume that there are three heart events, heart rate value r = [70, 75, 80], stability score s = [0.9, 0.8, 0.7], average heart rate μ = 75, attenuation coefficient δ = 0.1,

[0366] U=0.9×70×e -0.1×|70-75| +0.8×75×e -0.1×|75-75| +0.7×80×e -0.1×|80-75| ,

[0367] U≈62.16+60+54.02=176.18.

[0368] Abnormal heart rate score formula:

[0369]

[0370] f j : The heart rate value of the jth cardiac activity;

[0371] ν: average value of normal heart rate;

[0372] σ: standard deviation of normal heart rate;

[0373] ∈: adjust parameters to reduce the impact of small fluctuations;

[0374] d j : Duration of the jth record;

[0375] VI: Abnormal heart rate score.

[0376] For each heart movement, calculate the heart rate f j The square of the deviation from the normal average heart rate ν (f j -ν) 2 , the standard deviation σ and the adjustment parameter ∈ are used to adjust the square of the deviation and calculate the corrected denominator σ 2 +∈, each heart rate deviation squared divided by the corrected denominator, and then multiplied by the duration d j To obtain the weighted heart rate deviation for each activity, all weighted heart rate deviations are summed to obtain the total abnormal heart rate score VI;

[0377] Assume that there are three heart events, heart rate value f = [120, 80, 100], normal average heart rate ν = 90, standard deviation σ = 10, adjustment parameter ∈ = 0.5, duration d = [30, 45, 15],

[0378]

[0379] VI≈8.42×30+0.89×45+0.89×15=361.35.

[0380] Formula for preparing heart rate monitoring results:

[0381]

[0382] κ j : The severity weight of the jth abnormal heart rate;

[0383] VI: abnormal heart rate score;

[0384] SI: Compile heart rate monitoring results.

[0385] The abnormal score VI of each heart rate activity is multiplied by the corresponding severity weight κ j , all weighted scores were summed and divided by the total number of heart rate activities n to calculate the average abnormal score SI;

[0386] Assume that there are three abnormal scores of cardiac activity VI = [361.35, 250.25, 180.10], and the severity weight κ =

[0387] 0.8, 0.5, 0.3], total number of activities n = 3,

[0388]

[0389] SI≈149.02.

[0390] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the drawings are merely exemplary illustrations of the present application as defined by the appended claims, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include these modifications and variations.

Claims

1. A heart rate monitoring system based on cloud computing, characterized in that: The system comprises: The heart rate data analysis module collects heart rate data based on the photoelectric volumetric pulse wave sensor on the cloud platform, constructs a time series, analyzes the time point of each heartbeat, obtains the heart rate time series, inputs the heart rate time series into the processing link, preliminarily classifies the heartbeat pattern, and obtains the extended classification output result; The heartbeat abnormality recognition module receives the extended classification output result, identifies the heartbeat abnormality by analyzing the local fluctuation and frequency domain characteristics of the heart rate, evaluates the irregularity of the heart rate, generates an abnormal diagnosis result, and classifies the arrhythmia based on the type and frequency of the abnormal diagnosis result to obtain an arrhythmia level output result; The heart disease risk assessment module uses the arrhythmia level output result, combines historical health data and heart rate patterns for analysis, assesses the predicted risk of heart disease, obtains a risk assessment result, and uses the risk assessment result to optimize the patient's monitoring plan to obtain optimized patient monitoring recommendations; The steps for obtaining the risk assessment results are specifically as follows: Integrate the arrhythmia level output results and the patient's historical health data, and assign influence to risk factors in a weighted manner, applying the formula: The risk assessment value is calculated, where FA represents the risk assessment value, and w i is the weight of the ith risk factor, h i is the risk value corresponding to the historical health data, k i is the gain factor, l i is the risk factor corresponding to the arrhythmia level; Based on the risk assessment value, heart rate variability analysis is performed using the heart rate pattern data using the formula: Integrate the contribution of heart rate data and calculate the heart rate pattern risk score, where R is the heart rate pattern risk score, p j is the parameter of the heart rate mode, v j is the heart rate variability analysis score, λ j is the normalization coefficient; The risk assessment value and the heart rate pattern risk score are combined and integrated in a weighted manner, using the formula: Calculate the predicted risk of heart disease and generate a risk assessment result, where Q represents the risk assessment result, FA is the risk assessment value, R is the heart rate pattern risk score, and γ and δ are adjustment coefficients used to balance FA and R; Based on the optimized patient monitoring suggestions, the patient status monitoring module updates the patient's heart rate status record, records each heart activity, analyzes the frequency and stability of the heart rate, identifies abnormal heart rates that deviate from the normal range, marks the abnormal heart rates, summarizes the abnormal data, and compiles the heart rate monitoring results.

2. The cloud computing-based heart rate monitoring system according to claim 1, characterized in that: The steps for obtaining the heart rate time series are specifically as follows: Based on the cloud platform, photoelectric volume pulse wave sensors are used to collect heart rate data, collect raw heart rate data, and generate raw heart rate data sets. Analyze the time point of each heartbeat in the original heart rate data set, and calculate the time interval between two adjacent heartbeats using the formula: Get the heartbeat time interval array, where Δt i is the heartbeat time interval array, t i represents the time point of the i-th heartbeat, t i+1 represents the time point of the i+1th heartbeat, α is the linear adjustment coefficient, β is the nonlinear adjustment coefficient within the square root, Δt max represents the maximum time interval, θ is the minimum time interval proportional adjustment coefficient, Δt min Represents the minimum time interval; The heart rate time series is constructed using the heartbeat time interval array using the formula: Calculate the number of heart beats per minute and generate a heart rate time series, where HR(t) represents the number of heart beats per minute, Δt represents the heart beat time interval array, γ is the linear correction coefficient, κ is the inverse proportional adjustment factor, and ∈ is a small positive number.

3. The cloud computing-based heart rate monitoring system according to claim 2, characterized in that: The steps for obtaining the extended classification output result are specifically as follows: The heart rate time series is input into the data processing link and classified using the heartbeat interval feature using the formula: Get the heartbeat pattern feature score, where S represents the heartbeat pattern feature score, x j is the jth feature in the heart rate time series, w j is the weight of the jth feature, λ is the parameter to adjust the nonlinear effect, μ is the adjustment parameter for feature weighting, and n is the number of features; A threshold is determined for the heartbeat pattern feature score of each sample using the following formula: Perform preliminary classification on the heartbeat pattern to obtain preliminary classification results, where C represents the preliminary classification result, S is the heartbeat pattern feature score, θ is the classification threshold, σ is the standard deviation, ρ is the offset parameter, 1 represents an abnormal heartbeat pattern, and 0 represents a normal heartbeat pattern; According to the preliminary classification results, the classification is refined and the formula is applied: Generate extended classification output results, where D represents the extended classification output results, C k is the preliminary classification result, v k is the weight of the corresponding classification label, m is the number of classification labels, and δ is the logistic regression intercept adjustment parameter.

4. The cloud computing-based heart rate monitoring system according to claim 3, characterized in that: The steps for obtaining the abnormal diagnosis result are specifically as follows: Based on the extended classification output results, the local fluctuations in the heart rate time series are analyzed using the formula: Generate volatility results, where V is the volatility result, HR j is the heart rate at the jth time point, is the mean heart rate, σ j is the standard deviation at the jth time point, ∈ is a small positive number that avoids division by zero; Based on the fluctuation results, the frequency domain characteristics of the heart rate sequence are analyzed using Fourier transform and frequency domain weighted formula: Generate frequency domain analysis results, where F is the frequency domain analysis result, FFT (HR k ) is the Fourier transform result, w k is the weight of frequency k, δ k is the frequency domain attenuation factor; Combining the volatility results and the frequency domain analysis results, the weighted sum method is adopted and the nonlinear combination formula is used: Generate anomaly diagnosis results, where A represents the anomaly diagnosis result, V is the volatility result, F is the frequency domain analysis result, and α and β are adjustment coefficients.

5. The cloud computing-based heart rate monitoring system according to claim 4, characterized in that: The steps for obtaining the arrhythmia level output result are specifically as follows: Based on the abnormal diagnosis results, the distribution and frequency of abnormal types are analyzed using the formula: Calculate and generate the abnormal type distribution score, where T represents the abnormal type distribution score, f i is the frequency of abnormal occurrence, t i is the severity index of the anomaly type, k is a parameter used to adjust the impact of each anomaly severity, and g i is the severity level of each anomaly, and h is the decision threshold; A weighted score is calculated based on the abnormality type distribution score and the clinical relevance of the abnormality using the formula: The arrhythmia severity score is calculated, where SA is the arrhythmia severity score, T j is the abnormal type distribution score, c j is the clinical relevance weight, δ is the global adjustment coefficient, r j is the response coefficient for each anomaly type; Arrhythmias are graded using a grading scale based on the arrhythmia severity score, using the formula: The arrhythmia level output result is obtained, where L represents the arrhythmia level output result, SA is the arrhythmia severity score, α is the adjustment increase coefficient, {thresholds} is the classification threshold, and classify is the classification function.

6. The cloud computing-based heart rate monitoring system according to claim 1, characterized in that: The steps for obtaining the optimized patient monitoring suggestion are specifically as follows: Analyze the risk assessment results and calculate the risk-weighted sum using the formula: Where P represents the risk-weighted sum, u i is the importance coefficient of the risk indicator, v i is the value of the risk indicator, s i is the stability adjustment factor, ∈ is a small constant; Based on the risk-weighted sum, adjust the treatment plan according to the risk level, synchronize information to the medical team, determine the level of monitoring, and apply the formula: Among them, Q represents the monitoring level, P is the risk-weighted sum, γ is the response sensitivity parameter, and δ is the response threshold; Based on the level of supervision, a personalized supervision plan is designed using the formula: N=∫Q(t)·dt The time-weighted method is used to calculate the monitoring needs within the duration and obtain the optimized patient monitoring recommendation, where N represents the personalized monitoring plan, Q(t) is the time function of the monitoring level, and t is the time variable.

7. The cloud computing-based heart rate monitoring system according to claim 6, characterized in that: The steps for obtaining the heart rate monitoring result are as follows: Based on the optimized patient monitoring recommendations, the patient's heart rate status record is updated using the formula: Calculate the updated heart rate state score, where U represents the updated heart rate state score, r i is the heart rate value recorded for the i-th cardiac activity, s i is the stability score of the corresponding heart rate state, δ is the coefficient for adjusting the influence of heart rate deviation, and μ is the average heart rate; Based on the updated heart rate status score, abnormal heart rates that deviate from the normal range are marked using the formula: Generate an abnormal heart rate score, where VI is the abnormal heart rate score, f j is the heart rate value of the jth cardiac activity, ν is the average value of the normal heart rate, σ is the standard deviation of the normal heart rate, ∈ is the adjustment parameter, d j is the duration of the jth record; Based on the abnormal heart rate score, the abnormal data is summarized and the compiled heart rate monitoring result is generated using the formula: Among them, SI represents the compiled heart rate monitoring result, κ j is the severity weight of the jth abnormal heart rate, and VI is the abnormal heart rate score.

Images