Cancer survivor job stress monitoring system based on big data

Abstract

The present application relates to the technical field of health monitoring, in particular to a cancer survivor work pressure monitoring system based on big data, which comprises a physiological parameter analysis module, a stress cause analysis module, a work load distribution module and an abnormal stress monitoring module. In the present application, the sensitivity and accuracy of data analysis are improved through the difference analysis of heart rate fluctuation, skin conductance and blood pressure range, the data points with large fluctuations are screened and classified, the efficiency of abnormal data capture and classification is optimized, the stress cause identification accuracy and distribution law insight are improved, the key stress factors are determined through sensitivity sorting, the stress evaluation is more scientific and quantitative, the priority dynamic adjustment is realized in load classification and grouping, the work pressure distribution is optimized, the abnormal stress point monitoring and grading judgment are strengthened, the stress abnormality grade fine monitoring is realized, the data quantification and dynamic monitoring capability are enhanced, the stress management efficiency is improved, and the targeted data support is provided for personalized health intervention.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of health monitoring, and particularly relates to a cancer survivor work stress monitoring system based on big data. BACKGROUND

[0002] The technical field of health monitoring includes technologies for monitoring, recording, and analyzing human health status through computer technology, data analysis technology, and information processing technology. The core content of this technical field is the collaborative application of big data, sensor technology, and information technology for real-time collection and processing of personal health data, thereby achieving comprehensive monitoring and evaluation of individual health-related status. The overall technical field mainly covers health data collection, storage, management, and analysis, including stress monitoring, sleep monitoring, and exercise monitoring. Through accurate data processing and dynamic monitoring, data support and decision-making references are provided for health management.

[0003] Among them, the cancer survivor work stress monitoring system refers to a system for monitoring and analyzing the psychological and physiological stress status of individuals in the rehabilitation stage of cancer in the workplace. This patent subject mainly targets the special needs of cancer survivors, collects multi-dimensional data on stress factors in the workplace through big data collection technology, including work environment data, task load data, and physiological state data. Specifically, based on sensor collection of individual physiological parameters such as heart rate and skin conductance, data analysis technology is used to dynamically quantify work stress, and data mining technology is used to identify the rules and influencing factors of individual stress changes. Ultimately, a panoramic monitoring and analysis of stress status is formed.

[0004] The existing technology lacks accurate capture and in-depth analysis of the range and fluctuation amplitude of physiological parameter changes in stress monitoring. It can only record basic physiological parameters and cannot effectively identify severe fluctuations and abnormal states in a short period of time. Since stress cause analysis mainly relies on a single dimension of data source, it ignores the interactive effects of physiological parameters and task intensity and other stress factors, leading to one-sidedness in identifying the source of stress and difficulty in accurately locating key stress causes. In terms of work load management, the existing technology cannot dynamically adjust the priority of the load, and the matching degree of work tasks and stress is low. It cannot optimize the work allocation mode according to the individual needs of cancer survivors. In terms of abnormal stress monitoring, the existing technology mainly uses static threshold warning, which cannot dynamically evaluate the fluctuation amplitude and distribution characteristics of node stress, leading to lag in monitoring abnormal stress and inability to timely discover and intervene in high-risk stress states. This results in insufficient monitoring accuracy, missing key cause identification, and delayed abnormal stress management, thereby affecting the comprehensiveness and timeliness of health management. SUMMARY

[0005] The object of the present application is to solve the drawbacks existing in the prior art, and the cancer survivor work stress monitoring system based on big data is proposed.

[0006] In order to achieve the above object, the present application adopts the following technical scheme: the cancer survivor work stress monitoring system based on big data comprises:

[0007] The physiological parameter analysis module performs difference analysis on the heart rate variation range and skin conductance fluctuation amplitude in the time period based on the heart rate fluctuation, skin conductance and blood pressure range of the cancer survivor, and screens and classifies the data points with large fluctuations to generate physiological parameter fluctuation feature results;

[0008] The stress inducement analysis module compares the heart rate fluctuation amplitude, skin conductance difference and task intensity in the data points one by one based on the physiological parameter fluctuation feature results, generates a stress inducement difference set, and performs sensitivity sorting on the parameter fluctuation range in the stress inducement difference set to generate a stress inducement key distribution result;

[0009] The work load distribution module classifies and groups the task intensity, work duration and stress peak of the cancer survivor one by one based on the stress inducement key distribution result to obtain work load classification data, adjusts the classification interval load intensity in the work load classification result in priority, and generates a load adjustment path result;

[0010] The abnormal stress monitoring module analyzes the task intensity distribution and stress fluctuation amplitude distribution of the node pressure based on the load adjustment path result, generates an abnormal stress distribution data set, and classifies the distribution and fluctuation amplitude of the abnormal stress points to generate a stress abnormality level classification monitoring result.

[0011] As a further scheme of the present application, the acquisition step of the physiological parameter fluctuation feature result is specifically:

[0012] Based on the heart rate fluctuation, skin conductance and blood pressure range of the cancer survivor, the range of heart rate in each time period is analyzed, the standard deviation is calculated through the skin conductance data, and the heart rate variation range and skin conductance fluctuation amplitude feature data table is obtained;

[0013] The heart rate variation range and skin conductance fluctuation amplitude feature data table are subjected to difference analysis, and the formula is:

[0014]

[0015] The difference index value is calculated;

[0016] Wherein, D represents the difference index value, HR max represents the maximum value of heart rate in the monitoring time period, HR minσ represents the minimum value of the heart rate in the monitoring time period SC σ represents the standard deviation of the fluctuation amplitude of the skin conductance, and w1 and w2 are weight parameters.

[0017] Based on the difference index value, a threshold parameter is set, a logical judgment is performed on the difference index, fluctuation data points satisfying the condition that the difference index is greater than the threshold are screened, the data points meeting the condition are counted and classified, and a physiological parameter fluctuation feature result is generated.

[0018] As a further scheme of the present application, the acquisition step of the stress inducement difference set is specifically:

[0019] Based on the physiological parameter fluctuation feature result, the heart rate fluctuation amplitude value in the data points is extracted, the values are sorted and classified into a set, the skin conductance difference value and the task intensity value are extracted and sorted, and a physiological parameter arrangement result is obtained;

[0020] Based on the physiological parameter arrangement result, the heart rate fluctuation amplitude is selected, the difference value of adjacent data points is analyzed, the skin conductance difference is processed according to the difference value calculation rule of adjacent data points, the task intensity value change is counted, and a parameter change difference set is obtained.

[0021] Based on the parameter change difference set, the heart rate fluctuation and the skin conductance difference value set are read, the data points in the range are counted, the associated data area is screened, and a stress inducement difference set is obtained.

[0022] As a further scheme of the present application, the acquisition step of the stress inducement key distribution result is specifically:

[0023] According to each parameter in the stress inducement difference set, the standard deviation and the mean of the fluctuation amplitude are identified by counting the fluctuation range of the parameter, the parameter fluctuation feature is analyzed, the standard deviation is used to reflect the discreteness of the fluctuation degree, the corresponding size of the fluctuation level is judged, and a fluctuation feature set is generated.

[0024] Based on the fluctuation feature set, the sensitivity of each parameter is weighted and sorted by combining the weight, and the calculation formula is:

[0025]

[0026] A sorted stress inducement sensitivity list is generated.

[0027] Wherein, S represents the sensitivity sorting value, M i σ represents the minimum value of the heart rate in the monitoring time period i σ represents the standard deviation of the fluctuation amplitude of the skin conductance, and w1 and w2 are weight parameters. i σ represents the standard deviation of the fluctuation amplitude of the skin conductance, and w1 and w2 are weight parameters.

[0028] Based on the sorted pressure inducement sensitivity list, by comparing the influence of each parameter sensitivity on the pressure inducement distribution, the fluctuation characteristics of the weight parameters in the sensitivity ranking are screened, the sensitivity parameters are selected as the key pressure inducement, and the key distribution result of the pressure inducement is established.

[0029] As a further scheme of the present application, the acquisition step of the work load classification data is specifically:

[0030] Based on the key distribution result of the pressure inducement, the task intensity influence range analysis is carried out, the work duration and the pressure peak value are classified and compared respectively, the correlation parameters are calculated, the low correlation task intensity data is eliminated, and the work load distribution parameter is obtained;

[0031] For the work load distribution parameter, the task intensity, work duration and pressure peak value data are grouped and classified, the average value, standard deviation and coefficient of variation of the characteristic value of each group of data are identified, the load distance is calculated, and the formula is:

[0032]

[0033] The work load characteristic distance parameter is obtained;

[0034] Wherein, L represents the load distance, T a represents the single task intensity, T m represents the overall average of the task intensity, W a represents the single task work duration, W m represents the overall average of the work duration, and P represents the pressure peak value of the single task.

[0035] According to the work load characteristic distance parameter, grouping analysis and load distance screening are carried out, the task intensity, work duration and pressure peak value load distance are sorted and labeled with characteristic values, classified according to the load distance, and work load classification data is obtained.

[0036] As a further scheme of the present application, the acquisition step of the load adjustment path result is specifically:

[0037] Based on the work load classification result, the load intensity of the classification interval is extracted, the load intensity values are sorted and checked one by one, the corresponding order and values of the load intensity in the sorted classification interval are recorded, and the load intensity sorting result is generated;

[0038] Based on the load intensity sorting result, the priority is sequentially distributed, the load intensity value is set priority number one by one and corresponding, the priority number distribution order is recorded, the mapping relationship between the priority number and the classification interval load intensity is analyzed, and the load priority distribution result is obtained;

[0039] Based on the load priority allocation result, the load intensity of each classification interval is adjusted item by item, the adjustment path is determined in combination with the priority number, the adjusted load intensity value is compared with the load intensity value before adjustment, the adjustment path parameter change and the adjustment amplitude are recorded, and a load adjustment path result is generated.

[0040] As a further scheme of the present application, the step of obtaining the abnormal pressure distribution dataset specifically comprises:

[0041] Based on the load adjustment path result, the time sequence information in the node pressure data is called, the task intensity distribution is normalized with the numerical value of the pressure fluctuation amplitude, the time sequence data is divided into fixed intervals, and whether the fluctuation amplitude exceeds the threshold value is compared to obtain a pressure fluctuation classification result.

[0042] The task intensity distribution and the pressure fluctuation amplitude data in the pressure fluctuation classification result are analyzed, and a formula is used:

[0043]

[0044] The pressure abnormal distribution coefficient is calculated.

[0045] Wherein, Q represents the pressure abnormal distribution coefficient, P max represents the maximum value of the pressure, P avg represents the average value of the pressure, sigma P represents the standard deviation of the pressure, T max represents the maximum value of the task intensity, T min represents the minimum value of the task intensity.

[0046] According to the pressure abnormal distribution coefficient, the screened abnormal node pressure data is reclassified, the pressure fluctuation amplitude of the abnormal node is analyzed in linkage with the corresponding task intensity, the abnormal data is marked according to the amplitude of the task intensity change, and an abnormal pressure distribution dataset is generated.

[0047] As a further scheme of the present application, the step of obtaining the pressure abnormal level classification monitoring result specifically comprises:

[0048] Based on the abnormal pressure distribution dataset, the pressure value and the fluctuation amplitude of the pressure point data are extracted, the data is classified and processed, and the pressure mean value and the standard deviation are calculated, the data with the standard deviation lower than the set threshold value is removed, and a pressure value preliminary screening result is obtained.

[0049] Based on the pressure value preliminary screening result, the pressure points are classified and judged, the pressure level of each pressure point is analyzed, and a formula is used:

[0050]

[0051] Calculate the stress level score, generate the stress point level judgment result;

[0052] Wherein, Y represents the stress level score, H represents the stress value, mu P is the average value of the stress, N is the standard deviation of the stress, Delta P represents the stress change rate, and Lambda is a weight coefficient adjusted according to the data distribution self-matching;

[0053] The stress point level judgment result is summarized, the number and distribution characteristics of the differentiated level stress points are obtained, the data set is classified and statistically analyzed according to the abnormal level, the stress range and typical fluctuation characteristics are marked, and the stress abnormal level classification monitoring result is generated.

[0054] Compared with the prior art, the advantages and positive effects of the present application are:

[0055] In the present application, by analyzing the differences in heart rate fluctuation, skin conductance and blood pressure range, the change range and fluctuation amplitude of physiological parameters in a time period are accurately quantified, the sensitivity and accuracy of data analysis are significantly improved, the data points with large fluctuations are screened and classified, the capture efficiency and classification accuracy of abnormal data are optimized, the physiological parameter fluctuation results are compared item by item and analyzed in multiple dimensions, the correlation analysis of heart rate fluctuation amplitude, skin conductance difference and task intensity and other indicators is combined, the recognition accuracy of stress inducement and the ability to understand the distribution law are improved, based on the sensitivity sorting of stress inducement difference set, the key influencing factors in stress change are determined, the stress evaluation is more scientific and quantitative, in the load classification and grouping of task intensity, working time and stress peak value, through the dynamic adjustment of load intensity priority, the working stress distribution optimization is realized, more reasonable working load suggestions are provided for cancer survivors, through the comprehensive analysis of node stress distribution and stress fluctuation amplitude, the monitoring ability of abnormal stress points is strengthened, and through the grading judgment of abnormal stress points, the fine monitoring of stress abnormal level is realized. The whole processing logic optimizes the whole link processing process of physiological data to stress state, enhances the ability of accurate quantization of data and dynamic monitoring of stress state, significantly improves the stress management efficiency in the working environment of cancer survivors, and provides more targeted data support for personalized health intervention. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 The system flowchart of the present application;

[0057] Figure 2 The flowchart of physiological parameter fluctuation characteristic result in the present application;

[0058] Figure 3 The flowchart of stress inducement difference set in the present application;

[0059] Figure 4Flow chart for pressure cause key distribution result in the present application;

[0060] Figure 5 Flow chart for workload classification data in the present application;

[0061] Figure 6 Flow chart for load adjustment path result in the present application;

[0062] Figure 7 Flow chart for abnormal pressure distribution data set in the present application;

[0063] Figure 8 Flow chart for pressure anomaly level classification monitoring result in the present application. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0065] In the description of the present application, it should be understood that the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, in the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise specifically limited.

[0066] Please refer to Figure 1 The cancer survivor work pressure monitoring system based on big data comprises:

[0067] The physiological parameter analysis module performs difference analysis on the heart rate variation range and skin conductance fluctuation amplitude within a time period based on the heart rate fluctuation, skin conductance and blood pressure range of the cancer survivor, and screens and classifies the data points with large fluctuations to generate physiological parameter fluctuation characteristic results;

[0068] The pressure cause analysis module compares the heart rate fluctuation amplitude, skin conductance difference and task intensity in the data points one by one based on the physiological parameter fluctuation characteristic results, generates a pressure cause difference set, sorts the parameter fluctuation range in the pressure cause difference set for sensitivity, and generates a pressure cause key distribution result;

[0069] The work load distribution module classifies and groups the task intensity, work duration and stress peak of the cancer survivor based on the stress inducement key distribution result, obtains work load classification data, adjusts the classification interval load intensity in the work load classification result, and generates a load adjustment path result;

[0070] The abnormal stress monitoring module analyzes the task intensity distribution and stress fluctuation amplitude distribution of the node pressure based on the load adjustment path result, generates an abnormal stress distribution data set, classifies the distribution and fluctuation amplitude of the abnormal stress point, and generates a stress abnormality level classification monitoring result.

[0071] The physiological parameter fluctuation feature result includes heart rate variation range difference, skin conductance fluctuation amplitude difference, and fluctuation data point classification. The stress inducement difference set includes heart rate fluctuation amplitude difference, skin conductance difference, and task intensity difference. The stress inducement key distribution result includes sensitive parameter sorting, stress inducement distribution position, and fluctuation range feature. The work load classification data includes task intensity classification, work duration classification, and stress peak classification. The load adjustment path result includes classification interval load priority, load adjustment optimization path, and load distribution optimization scheme. The abnormal stress distribution data set includes task intensity distribution, stress fluctuation amplitude distribution, and abnormal stress point distribution. The stress abnormality level classification monitoring result includes abnormal stress classification, stress fluctuation amplitude classification, and stress level division.

[0072] Please refer to Figure 2 The acquisition steps of the physiological parameter fluctuation feature result are as follows:

[0073] Based on the heart rate fluctuation, skin conductance and blood pressure range of the cancer survivor, the range of heart rate in each time period is analyzed, the standard deviation is calculated through the skin conductance data, and a heart rate variation range and skin conductance fluctuation amplitude feature data table is obtained.

[0074] The range of heart rate variation and the amplitude of skin conductance fluctuation data of cancer survivors in the monitoring period are extracted, and the monitoring data is recorded with time as the axis. Through preprocessing, abnormal data points are removed. For heart rate data, first, the effective interval of the data is set according to the physiological range, and all data points outside the effective interval range are marked and deleted. After deletion, the interpolation method is used to supplement the missing area in the data to ensure the integrity of the time series. For skin conductance data, it is segmented into fixed time periods, and the skin conductance value in each time period is statistically analyzed to extract the average value and maximum value in that time period, and the standard deviation is calculated to quantify the fluctuation degree. After segmenting the time series data, the fluctuation trend between different time periods can be further analyzed according to the time period length. The results of heart rate data and skin conductance fluctuation amplitude data are summarized to construct a heart rate variation range and skin conductance fluctuation amplitude feature data table. The feature data table lists the maximum and minimum values of the heart rate data and the fluctuation amplitude of the skin conductance in each time period, providing complete feature data support for subsequent difference analysis and screening.

[0075] The heart rate variation range and skin conductance fluctuation amplitude feature data table are subjected to difference analysis, using the formula:

[0076]

[0077] The difference index value is calculated;

[0078] Where D represents the difference index value, HR max represents the maximum value of the heart rate in the monitoring period, HR min represents the minimum value of the heart rate in the monitoring period, σ SC represents the standard deviation of the skin conductance fluctuation amplitude, and w1 and w2 are weight parameters.

[0079] The advantage of the formula is that by combining the two different types of data of heart rate variation range and skin conductance fluctuation amplitude into a single index and adjusting the relative influence of the two through weight parameters, the fluctuation difference in individual physiological state can be more intuitively reflected, and the influence of extreme data points on the result is effectively reduced through the denominator smoothing function, increasing the robustness of the index.

[0080] Let the maximum heart rate in the monitoring period be 140, the minimum heart rate be 70, the skin conductance fluctuation standard deviation be 2.5, and the weight parameters be 1.5 and 0.8 respectively. The formula is calculated as follows:

[0081] Calculate the heart rate difference: |HR max -HR min | = |140-70| = 70.

[0082] Calculate the numerator: |HRmax - HR min | + w1 · s SC = 70 + 1.5 · 2.5 = 70 + 3.75 = 73.75;

[0083] Calculate the denominator part:

[0084] Calculate the difference index:

[0085] The result shows that the difference index value is 27.61, indicating that the individual's heart rate change and skin conductance fluctuation difference is relatively high in the current time period, and the value will be used as a screening basis in subsequent steps to judge the significance of the individual's fluctuation state.

[0086] Based on the difference index value, set the threshold parameter, make a logical judgment on the difference index, screen the fluctuation data points that meet the difference index greater than the threshold, count the data points that meet the conditions and classify them to generate the physiological parameter fluctuation feature result;

[0087] First, set the screening logic. The logical judgment process needs to be based on the predefined difference threshold to group process the difference index. Take the threshold as the distinguishing standard to determine whether the difference index of each time period is greater than or less than the threshold. If the difference index is greater than the threshold, mark the time period as a significant fluctuation interval. For the difference index of the significant fluctuation interval, further analyze the specific characteristic values of the heart rate change and skin conductance fluctuation in each significant fluctuation interval, and divide them into different categories by induction of similar characteristics. The specific classification rules are based on the range of heart rate change and the fluctuation amplitude level of skin conductance. At the same time, in order to ensure the classification accuracy, compare the characteristic values of adjacent time periods. If the characteristic values of consecutive time periods are consistent, merge the time periods into the same classification. Through point-by-point analysis and classification operation, generate the physiological parameter fluctuation feature result, which records in detail the classification attribution, characteristic value range and difference index state of each time period.

[0088] Please refer to Figure 3 , the acquisition steps of the stress inducement difference set are as follows:

[0089] Based on the physiological parameter fluctuation feature result, extract the heart rate fluctuation amplitude value in the data points, sort the values and classify them into a set. Extract the skin conductance difference value and task intensity value and sort them to obtain the physiological parameter arrangement result.

[0090] First, the upper and lower limit values of the heart rate fluctuation amplitude are determined, and the fluctuation amplitude interval is divided according to 5bpm intervals, for example, for the heart rate fluctuation amplitude in the range of 50 to 150bpm, 20 intervals can be divided, after each data point is classified into the corresponding interval, the number of data points in each interval is counted, the average heart rate fluctuation amplitude of each interval is calculated, and the dispersion degree of data points in each interval is evaluated through variance analysis, the skin conductance difference value and the task intensity value are extracted and sorted, for the skin conductance data, the difference value between adjacent time points is calculated based on the continuous data sampled per second, the mean value and fluctuation range of the difference value are recorded, and the task intensity value is segmented and classified, for example, the task intensity is divided into low (0-3), medium (4-7) and high (8-10) three levels, the matching degree of the average skin conductance difference value corresponding to each level of task intensity and the heart rate fluctuation amplitude is calculated respectively, and finally the physiological parameter sorting result is formed.

[0091] Based on the physiological parameter sorting result, the heart rate fluctuation amplitude is selected, the difference value of adjacent data points is analyzed, the skin conductance difference is processed according to the difference value calculation rule of adjacent data points, the task intensity value change is counted, and the parameter change difference set is obtained.

[0092] First, the difference value set between the two continuous data points in the time sequence is extracted according to the heart rate fluctuation amplitude, for example, for the heart rate fluctuation amplitude data points [60, 65, 70, 68], the difference value of adjacent points is calculated as [5, 5, -2], the average value and change direction of the difference value are recorded, for the skin conductance difference value, according to the difference value calculation rule of adjacent data points, the absolute value threshold of the difference value is set as 0.05μS, the proportion of data points exceeding this threshold is counted, at the same time, the difference value set exceeding the threshold is classified into the high change group, and the rest is classified into the low change group, the average difference value and the maximum difference value of the two groups of data are calculated, the task intensity value change is counted, according to the time sequence of the task intensity, the standard deviation of the task intensity is calculated in a sliding window manner (for example, every 10 minutes), and the corresponding analysis is carried out with the change trend of the heart rate fluctuation and the skin conductance difference, and finally the parameter change difference set containing the linkage relationship of heart rate, skin conductance and task intensity is obtained.

[0093] Based on the parameter change difference set, the heart rate fluctuation and skin conductance difference value set are read, the region is divided according to the numerical range, the number of data points in the region is counted, the associated data region is screened, and the stress inducing difference set is obtained.

[0094] First, the difference set of heart rate fluctuation is divided into low amplitude interval (0-5 bpm), medium amplitude interval (6-10 bpm) and high amplitude interval (>10 bpm) according to the fluctuation amplitude, and the difference set of skin conductance is divided into low difference interval (0-0.05 μS), medium difference interval (0.06-0.1 μS) and high difference interval (>0.1 μS) according to the absolute value, the number of data points in each interval is counted, for example, there are 50 data points in the low amplitude interval, the proportion of data points in each interval is calculated, and the data points with significant task intensity change are further screened, the corresponding heart rate fluctuation and skin conductance difference interval range are extracted, when screening the associated data area, the threshold condition is set as the simultaneous occurrence of heart rate fluctuation difference value exceeding 6 bpm and skin conductance difference value higher than 0.05 μS, which is marked as the stress trigger area, the number of data points in the area is counted, the average heart rate fluctuation difference value and skin conductance difference value in the area are recorded, and the stress trigger difference set is formed for further analysis of subsequent work stress monitoring.

[0095] Please refer to Figure 4 The acquisition steps of the stress trigger key distribution result are as follows:

[0096] According to each parameter in the stress trigger difference set, the standard deviation and mean of the fluctuation amplitude are identified by counting the fluctuation range, the parameter fluctuation characteristics are analyzed, the standard deviation is used to reflect the dispersion of the fluctuation degree, the corresponding size of the fluctuation level is judged, and the fluctuation characteristic set is generated.

[0097] First, the historical data set of each parameter is collected, the data is arranged in time sequence, and the change trend is fitted by linear regression model, the fluctuation range of each parameter in the whole time period is obtained, then the standard deviation and mean of each parameter are calculated, the standard deviation reflects the dispersion of the fluctuation value of each parameter, and the mean reflects the concentration level of the parameter, by comparing the ratio of the standard deviation and the mean, the parameter set with larger fluctuation range is further identified, the parameter fluctuation range is associated with the corresponding time point, and the parameters with significant fluctuation characteristics are further extracted by time series analysis to generate the fluctuation characteristic set, which represents the fluctuation range and distribution trend of each parameter.

[0098] Based on the fluctuation characteristic set, the sensitivity of each parameter is weighted and sorted by combining the weight, and the calculation formula is:

[0099]

[0100] A sorted stress trigger sensitivity list is generated.

[0101] Wherein, S represents the sensitivity sorting value, M i represents the weight coefficient of parameter i, σ i represents the standard deviation of parameter i, μ iThe mean value of the parameter i, n represents the total number of parameters, and ∑ represents the summation symbol;

[0102] The formula has the advantages that by introducing the weight coefficient w of the parameter i and the ratio of the volatility characteristic standard deviation σ i and the mean value μ i , the volatility sensitivity of each parameter can be accurately weighted and calculated to highlight the role of high volatility and important parameters in the ranking;

[0103] First, collect the parameters i in the pressure cause difference set, extract the historical data, calculate the standard deviation σ and the mean value μ of each parameter, where x ij represents the observation value of the parameter i at the time point j, represents the average value of the parameter i, and m represents the number of observation values, then the ratio of σ is calculated according to the above formula, and each parameter is given a weight w i (weight is set according to the importance of the parameter, and the importance is determined by expert scoring or historical data analysis), the sensitivity ranking after weight adjustment is compared, and the final sensitivity ranking value S is calculated;

[0104] For example, for three parameters A, B, and C, it is assumed that σ A = 3, μ A = 15, M A = 0.4, σ B = 2, μ B = 10, M B = 0.3, σ C = 5, μ C = 20, M C = 0.3;

[0105] Substitute the values into the formula:

[0106] The results show that the sensitivity ranking value obtained by comprehensively considering the volatility and importance of the parameters can reflect the comprehensive influence degree of each parameter in the pressure cause, and finally generate a ranked pressure cause sensitivity list.

[0107] Based on the ranked pressure cause sensitivity list, by comparing the influence of the sensitivity of each parameter on the distribution of the pressure cause, the volatility characteristics of the weight parameters in the sensitivity ranking are screened, the sensitivity parameters are selected as the key pressure causes, and the key distribution results of the pressure causes are established;

[0108] First, the parameter set with high sensitivity value is extracted, the sensitivity ranking value is mapped with the corresponding parameter, the screening threshold is set based on the distribution range of the sensitivity ranking value, the parameters with sensitivity value exceeding the threshold are screened out, the screened parameter set is further cross-analyzed with the fluctuation feature set, the time series fluctuation trend of the parameters in the fluctuation feature set is combined, the parameters with high sensitivity value and significant fluctuation are focused on, the parameter set with high sensitivity ranking is selected as the key stress inducement by comparing the fluctuation influence of different parameters one by one, and finally the stress inducement key distribution result is established.

[0109] Please refer to Figure 5 The acquisition step of the workload classification data is specifically:

[0110] Based on the stress inducement key distribution result, the task intensity influence range analysis is performed, the work duration and the stress peak value are compared and classified respectively, and the correlation parameter is calculated, the low correlation task intensity data is removed, and the workload distribution parameter is obtained;

[0111] The task intensity is extracted by the execution times and time proportion of a single task, a data sequence is established with time as the horizontal axis, the average execution times and the deviation amplitude of each task in the sequence are calculated, the work duration is intervalized by collecting the start time and end time of work in a unit period, the interval length is defined as the work duration, and the distribution range and fluctuation of each work interval in the period are evaluated, the stress peak value is calculated by monitoring the physiological parameter fluctuation rate in a unit time, the maximum change rate of the physiological parameter is calculated and the rate is taken as the peak value, the correlation between the task intensity, the work duration and the stress peak value is quantitatively evaluated, the cooperative change relationship between the parameters is constructed, and the low correlation projects that do not meet the requirements are removed through the correlation threshold, and finally the parameter value range of the task intensity, the work duration and the stress peak value that meet the correlation requirements are screened out, and the workload distribution parameter is obtained.

[0112] The workload distribution parameter is grouped and classified according to the task intensity, the work duration and the stress peak value data, the average value, the standard deviation and the coefficient of variation of each group of data characteristic values are identified, the load distance is calculated, and the formula is:

[0113]

[0114] The workload characteristic distance parameter is obtained;

[0115] Wherein, L represents the load distance, T a represents the single task intensity, T m represents the overall average of the task intensity, W a represents the single task work duration, W m represents the overall average of the work duration, and P represents the stress peak value of a single task.

[0116] The formula is beneficial in that, by jointly considering the mean deviation of task intensity and working time length, combined with the normalization processing of stress peak value, the combined load characteristics of work task and stress can be evaluated at the same time, thereby improving the accuracy and adaptability of classification;

[0117] The task intensity data is calculated by the average value of time series distribution, assuming that 5 groups of task intensity data are collected, which are 2.4, 3.1, 2.8, 3.6, and 3.0, and the mean value of task intensity is:

[0118]

[0119] The single task intensity deviation is calculated according to the formula:

[0120] (T a -T m ) 2 =(2.4-3.0) 2 =0.36;

[0121] The stress peak value is assumed to be 120, 135, 125, 140, and 130, and the stress peak value normalization calculation parameter P is calculated:

[0122]

[0123] The working time length data is calculated, and the mean value and single task deviation value are calculated by the same method, and substituted into the formula:

[0124]

[0125] Assuming that the mean value of working time length is 8 hours and the single task working time length is 7 hours, then:

[0126]

[0127] The result shows that the smaller the load distance quantization result L is, the lower the deviation degree of task intensity and working time length relative to the stress peak value is, which meets the load balance requirement.

[0128] According to the working load characteristic distance parameter, grouping analysis and load distance screening are performed, the task intensity, working time length, and stress peak value load distance are sorted and labeled with screening characteristic values, classified according to the load distance, and working load classification data is obtained;

[0129] First, the calculated task intensity, work duration and stress peak load distance are sorted, and each group of load distance range is associated with the task and work duration. The annotation process divides the group range by the upper and lower limit values of the load distance. The group with a load distance less than 0.2 is defined as "low load", the group with a load distance greater than 0.2 and less than 0.5 is defined as "medium load", and the group with a load distance greater than 0.5 is defined as "high load". When annotating the task intensity by grouping the load distance, the task feature data in the corresponding group is called. The stress peak value is re-divided based on the median of the load distance. Finally, combined with the data characteristics after grouping, the annotation result of the load grouping is generated, and the characteristic values of each group are reclassified, and finally the work load classification data is obtained.

[0130] Please refer to Figure 6 The load adjustment path result acquisition step is specifically:

[0131] Based on the work load classification result, the load intensity of the classification interval is extracted. The load intensity values are sorted and checked one by one, the order and values corresponding to the load intensity of the classification interval after sorting are recorded, and the load intensity sorting result is generated;

[0132] First, the work load data is divided into three levels of hours, days and weeks according to different time intervals, and the load intensity values of each level are extracted, for example, the load intensity is calculated from the number of tasks completed per hour. By normalizing the load intensity to the range of 0-100, the comparability of different data sources is ensured. The normalized load intensity values are sorted from high to low, and the values are checked one by one to ensure the accuracy of the sorting result. The order and values corresponding to the load intensity of the classification interval after sorting are recorded, for example, the daily load intensity is recorded as [85, 78, 72, 65, 50, 45, 30] in order of value within a week, and the division basis of high, medium and low load intervals is labeled, for example, the high load interval is 70-100, the medium load interval is 40-69, and the low load interval is 0-39. The load intensity sorting result is generated, and the mean, maximum and minimum values of the load intensity of each classification interval are counted and arranged to provide data support for subsequent priority allocation analysis.

[0133] Based on the load intensity sorting result, the priority is sequentially allocated. The load intensity value is set with priority number one by one and corresponding, the priority number distribution order is recorded, the mapping relationship between the priority number and the classification interval load intensity is analyzed, and the load priority allocation result is obtained;

[0134] The load intensity values are sequentially prioritized and corresponding, and the priority numbers are sequentially assigned from high to low according to the sorting results of the load intensity, for example, the highest load intensity classification interval is marked as priority 1, the second highest is marked as priority 2, and so on, until all the priority numbers of the classification intervals are assigned, and the priority number distribution order is recorded to determine the correspondence between the priority and the load intensity, for example, in a week of load data, if the sorted load intensity is [85, 78, 72, 65, 50, 45, 30], then the corresponding priority numbers are 1 to 7, analyze the mapping relationship between the priority number and the load intensity of the classification interval, calculate the proportion of the load intensity corresponding to each priority number, for example, the load intensity of priority 1 accounts for 85 / 425≈20% of the total load intensity, evaluate the concentration of high-priority tasks, further combined with the work characteristics of cancer survivors, identify key data points in the high-load interval that cause stress, and mark the positions of these data points in the order to form the load priority allocation result, which is used to guide the development of subsequent load adjustment strategies.

[0135] Based on the load priority allocation result, the load intensity of the classification interval is adjusted item by item, the adjustment path is determined combined with the priority number, the adjusted load intensity value is compared with the load intensity value before adjustment, the adjustment path parameter change and the adjustment amplitude are recorded, and the load adjustment path result is generated;

[0136] First, extract the load intensity interval corresponding to the high priority number, for example, in the interval of priority 1 to 3, record the time distribution trend and maximum fluctuation range of the load intensity, analyze whether there is obvious overload in the interval, and adjust the path by introducing a buffer zone in the high-load interval to move part of the load in the high-load interval to the medium-load interval, for example, in daily work, control the task amount in the high-load interval to be less than 30% of the total load intensity, and at the same time, fill the medium-load interval, compare the adjusted load intensity value with the load intensity value before adjustment, calculate the load mean value change, standard deviation change and high-load interval proportion change before and after adjustment, evaluate the adjustment effect, record the adjustment path parameter change and the adjustment amplitude, for example, the buffer amount of the adjustment path, the adjustment step and the target value difference, and finally generate the load adjustment path result, which provides specific data basis for load optimization, supports personalized stress intervention for cancer survivors, and provides adjustment strategy reference for further work stress monitoring.

[0137] Please refer to Figure 7 , the steps of obtaining the abnormal stress distribution data set are as follows:

[0138] Based on the load adjustment path result, the time series information in the node stress data is called, the task intensity distribution and the numerical value of the stress fluctuation amplitude are normalized, the time series data is divided into fixed intervals, and it is compared whether the fluctuation amplitude exceeds the threshold value to obtain the stress fluctuation classification result;

[0139] The time series information of node pressure data needs to be first split into multiple intervals according to the monitoring time period, and a distribution model is established inside the interval according to the differentiated characteristics of the change of node pressure. The maximum value and the minimum value in each interval are extracted to calculate the fluctuation range of node pressure. The mean value and the standard deviation in the time series are further calculated from the fluctuation range. The mean value is used to reflect the overall trend of node pressure, and the standard deviation is used to represent the dispersion degree of node pressure fluctuation. The pressure fluctuation amplitude of each interval is normalized to unify the dimension, and the normalized result is classified according to the preset threshold. The interval with higher fluctuation amplitude is selected, and finally the pressure fluctuation classification result is generated by comparing the fluctuation amplitude trend.

[0140] The task intensity distribution and pressure fluctuation amplitude data in the pressure fluctuation classification result are analyzed, and the formula is:

[0141]

[0142] The pressure anomaly distribution coefficient Q is calculated as:

[0143] Wherein, Q represents the pressure anomaly distribution coefficient, P max represents the maximum value of pressure, P avg represents the average value of pressure, σ P represents the standard deviation of pressure, T max represents the maximum value of task intensity, and T min represents the minimum value of task intensity.

[0144] The beneficial effect of the formula is that by comprehensively considering the amplitude deviation of pressure fluctuation and the extreme value ratio of task intensity, the pressure standard deviation is normalized to the pressure deviation, and the nonlinear proportion of task intensity is combined, which can accurately reflect the distribution characteristics of node pressure anomaly and improve the accuracy of data screening.

[0145] The maximum value of node pressure P max is 120, and the time series data of pressure obtained by monitoring are 100, 110, 120, 90 and 105. The average value P avg is calculated as:

[0146]

[0147] The standard deviation of pressure σ P is calculated by the standard deviation formula:

[0148]

[0149] The maximum value T max and the minimum value T min of task intensity are set to 20 and 10, and the formula is substituted.

[0150]

[0151] The result shows that the pressure anomaly distribution coefficient Q is 2.121, indicating that the current node pressure fluctuation is large and the task intensity extremum difference is significant, which belongs to an abnormal node.

[0152] According to the pressure anomaly distribution coefficient, the pressure data of the screened abnormal node is reclassified, the pressure fluctuation amplitude of the abnormal node is linked with the corresponding task intensity, the abnormal data is marked according to the amplitude of the task intensity change, and an abnormal pressure distribution data set is generated;

[0153] The screened abnormal node pressure data needs to be reclassified according to the interval division method. By calculating the distribution characteristics of the abnormal node in the time sequence, the mean and extreme value of the pressure fluctuation amplitude of the abnormal node are first calculated, and then the correlation coefficient of the abnormal node and the task intensity is calculated through the interval division of the pressure fluctuation amplitude. The correlation coefficient is calculated by the ratio of the median of the pressure fluctuation amplitude and the change trend of the task intensity. Finally, the abnormal node is marked according to the amplitude of the task intensity change, and the abnormal class is generated. The results of the correlation analysis of abnormal pressure fluctuation amplitude and task intensity are integrated to generate an abnormal pressure distribution data set.

[0154] Please refer to Figure 8 , the acquisition steps of the pressure abnormality level classification monitoring result are:

[0155] Based on the abnormal pressure distribution data set, the pressure value and fluctuation amplitude of the pressure point data are extracted, the data is classified and processed, and the pressure mean and standard deviation are calculated. The data with a standard deviation lower than the set threshold is removed to obtain a preliminary screening result of the pressure value;

[0156] The time sequence of the pressure point data is disassembled, the pressure data of each node is divided into intervals according to a fixed time interval, the mean and standard deviation of the pressure point in each time interval are calculated, the standard deviation is used as a quantitative indicator of the fluctuation amplitude, the data distribution of all nodes is normalized, which facilitates the unified analysis of nodes with different pressure ranges, and nodes with mean and fluctuation amplitude exceeding the preset threshold range are screened out as abnormal pressure points. Through further refinement of the time distribution characteristics of the abnormal pressure points, the pressure fluctuation amplitude in the time interval is grouped and classified, and finally the preliminary screening result of the pressure value is obtained.

[0157] Based on the preliminary screening result of the pressure value, the grading of the pressure point is judged, the pressure grade of each pressure point is analyzed, and the formula is used:

[0158]

[0159] The pressure level score is calculated to generate the level judgment result of the pressure point;

[0160] Wherein, Y represents the pressure level score, H represents the pressure value, μ P is the average value of the pressure, N is the standard deviation of the pressure, ΔP represents the pressure change rate, and λ is a weight coefficient adjusted according to the data distribution;

[0161] The formula has the advantages that by comprehensively considering the relative deviation degree of the pressure point value and the dynamic influence of the fluctuation rate, higher precision and flexibility are provided for the grading judgment of the pressure anomaly, so that the classification result can better reflect the pressure point distribution characteristics;

[0162] The initial value of the pressure value H is 120, and the pressure point values obtained from the data set are 110, 120, 125, 115 and 130. The average value of the pressure μ P is calculated as follows:

[0163]

[0164] The pressure standard deviation N is calculated as follows:

[0165]

[0166] The pressure change rate ΔP is the average value of the pressure difference of each two consecutive time points, and is calculated as follows:

[0167]

[0168] Suppose the weight coefficient λ is adjusted to be proportional to the fluctuation range, that is, λ = 0.1·N, and λ is calculated as follows:

[0169] λ = 0.1·7.07 = 0.707;

[0170] The above results are substituted into the formula to calculate the pressure level score Y:

[0171]

[0172] The result shows that the pressure level score is 7.07, indicating that the fluctuation of the pressure point of the node is large, belonging to a higher abnormal level range.

[0173] The level judgment result of the pressure point is summarized, the data set is classified and statistically analyzed according to the number and distribution characteristics of the differentiated level pressure points, the pressure range and typical fluctuation characteristics are marked, and the pressure anomaly level classification monitoring result is generated;

[0174] First, the pressure point abnormality grade values are sorted, the pressure grades are divided into three categories of low, medium and high, the number of pressure points in each category is re-counted according to the classification results, the mean value and the fluctuation range of each category of pressure points are calculated by counting the pressure value range of each category of pressure points, the data are further compared with the abnormal fluctuation range in the time interval, the correlation characteristics of the pressure value and the fluctuation range in each grade classification are analyzed, and finally the pressure point characteristic labels of each grade classification are obtained, and the pressure abnormality grade classification monitoring result is generated.

[0175] The above is only a preferred embodiment of the present application, and does not limit the present application in other forms. Any skilled person in the art can modify or change the above disclosed technical content to equivalent embodiments applied to other fields, but any simple modification, equivalent change and modification made to the above embodiments without departing from the technical solution content of the present application, according to the technical essence of the present application, still belongs to the protection scope of the technical solution of the present application.

Claims

1. A cancer survivor work stress monitoring system based on big data, characterized in that, The system comprises: The physiological parameter analysis module performs difference analysis on the heart rate variation range and skin conductance fluctuation amplitude in each time period based on the heart rate fluctuation, skin conductance and blood pressure range of the cancer survivor, and filters and classifies the data points with large fluctuations to generate a physiological parameter fluctuation feature result; The acquisition step of the physiological parameter fluctuation feature result is specifically: Based on the heart rate fluctuation, skin conductance and blood pressure range of the cancer survivor, the range of heart rate in each time period is analyzed, the standard deviation is calculated through the skin conductance data, and the heart rate variation range and skin conductance fluctuation amplitude feature data table is obtained; The heart rate variation range and skin conductance fluctuation amplitude feature data table are subjected to difference analysis, and the formula is: ； The difference index value is calculated; wherein, representing a difference indicator value, representing a maximum value of heart rate over the monitoring time period, representing a minimum value of heart rate over the monitoring time period, representing a standard deviation of the electrodermal conductance fluctuation amplitude, and is a weight parameter; Based on the difference index value, a threshold parameter is set, a logical judgment is performed on the difference index, the fluctuation data points that satisfy the condition that the difference index is greater than the threshold value are filtered, the data points that meet the condition are counted and classified, and a physiological parameter fluctuation feature result is generated; The stress inducement analysis module compares the heart rate fluctuation amplitude, skin conductance difference and task intensity in the data points one by one based on the physiological parameter fluctuation feature result, generates a stress inducement difference set, sorts the parameter fluctuation range in the stress inducement difference set according to sensitivity, and generates a stress inducement key distribution result; The work load distribution module classifies and groups the task intensity, work duration and stress peak of the cancer survivor one by one based on the stress inducement key distribution result, obtains work load classification data, adjusts the classification interval load intensity in the work load classification data in priority, and generates a load adjustment path result; The abnormal stress monitoring module analyzes the task intensity distribution and stress fluctuation amplitude distribution of the node pressure based on the load adjustment path result, generates an abnormal stress distribution data set, and performs hierarchical judgment on the distribution and fluctuation amplitude of the abnormal stress points to generate a stress abnormality level classification monitoring result.

2. The big data based cancer survivor work stress monitoring system as claimed in claim 1, wherein, The acquisition step of the stress inducement difference set is specifically: Based on the physiological parameter fluctuation feature result, the heart rate fluctuation amplitude value in the data points is extracted, the values are sorted and classified into a set, the skin conductance difference value and task intensity value are extracted and sorted, and a physiological parameter arrangement result is obtained; Based on the physiological parameter arrangement result, the heart rate fluctuation amplitude is selected, the difference value of adjacent data points is analyzed, the skin conductance difference is processed according to the difference value calculation rule of adjacent data points, the task intensity value change is counted, and a parameter change difference set is obtained; Based on the parameter change difference set, the heart rate fluctuation and skin conductance difference value set are read, the data points in the region are counted, the associated data region is filtered, and a stress inducement difference set is obtained.

3. The big data based cancer survivor job stress monitoring system as claimed in claim 2, wherein, The acquisition step of the stress inducement key distribution result is specifically: According to each parameter in the stress inducement difference set, the standard deviation and mean of the fluctuation amplitude are identified by counting the fluctuation range of the parameter, the parameter fluctuation feature is analyzed, the standard deviation is used to reflect the discreteness of the fluctuation degree, the corresponding size of the fluctuation level is judged, and a fluctuation feature set is generated; Based on the fluctuation feature set, the sensitivity of each parameter is weighted and sorted by combining the weight, and the calculation formula is: ； Generate a sorted stress inducement sensitivity list; wherein, representing a sensitivity ranking value, representing a parameter a weight coefficient, representing a parameter a standard deviation, representing a parameter a mean value, representing a total number of parameters, representing a summation symbol; Based on the sorted stress inducement sensitivity list, by comparing the influence of the sensitivity of each parameter on the distribution of stress inducement, the fluctuation features of the weight parameters in the sensitivity sorting are screened, the sensitivity parameters are selected as the key stress inducement, and the key distribution result of the stress inducement is established.

4. The big data based cancer survivor job stress monitoring system as claimed in claim 3, wherein, The acquisition step of the work load classification data is specifically: Based on the stress inducement key distribution result, the task intensity influence range analysis is performed, the work time and the stress peak value are compared and classified respectively, and the correlation parameter is calculated, the low correlation task intensity data is removed, and the work load distribution parameter is obtained; For the work load distribution parameter, the task intensity, work time and stress peak value data are grouped and classified, the average value, standard deviation and coefficient of variation of each group of data characteristic value are identified, the load distance is calculated, and the formula is: ； Get the work load characteristic distance parameter; wherein, represents the load distance, represents the single task intensity, represents the overall mean of the task intensity, represents the single task work duration, represents the overall mean of the work duration, represents the stress peak of the single task; According to the work load characteristic distance parameter, grouping analysis and load distance screening are performed, the task intensity, work time and stress peak value load distance are sorted and labeled with screening characteristic values, classified according to the load distance, and work load classification data is obtained.

5. The big data based cancer survivor job stress monitoring system as claimed in claim 4, wherein, The acquisition step of the load adjustment path result is specifically: Based on the work load classification data, the load intensity of the classification interval is extracted, the load intensity values are sorted and checked one by one, the corresponding order and value of the load intensity of the classification interval after sorting are recorded, and the load intensity sorting result is generated; Based on the load intensity sorting result, the priority is sequentially distributed, the load intensity value is sequentially set priority number and corresponding, the priority number distribution order is recorded, the mapping relationship between priority number and classification interval load intensity is analyzed, and the load priority distribution result is obtained; Based on the load priority distribution result, the load intensity of the classification interval is adjusted item by item, the adjustment path is determined combined with the priority number, the adjusted load intensity value is compared with that before adjustment, the adjustment path parameter change and adjustment amplitude are recorded, and the load adjustment path result is generated.

6. The big data based cancer survivor job stress monitoring system as claimed in claim 5, wherein, The acquisition step of the abnormal stress distribution data set is specifically: Based on the load adjustment path result, the time series information in the node stress data is called, the task intensity distribution and the value of the stress fluctuation amplitude are normalized, the time series data is divided into fixed intervals, and the fluctuation amplitude is compared to determine whether it exceeds the threshold value, and the stress fluctuation classification result is obtained; The task intensity distribution and stress fluctuation amplitude data in the stress fluctuation classification result are analyzed, and the formula is used: ； The stress abnormal distribution coefficient is calculated; wherein, a coefficient representative of an abnormal distribution of pressure, a maximum value representative of pressure, an average value representative of pressure, a standard deviation representative of pressure, a maximum value representative of task intensity, a minimum value representative of task intensity; According to the stress abnormal distribution coefficient, the screened abnormal node stress data is reclassified, the stress fluctuation amplitude of the abnormal node is analyzed in combination with the corresponding task intensity, the abnormal data is labeled according to the amplitude of the task intensity change, and the abnormal stress distribution data set is generated.

7. The big data based cancer survivor job stress monitoring system as claimed in claim 6, wherein, The acquisition step of the stress abnormal grade classification monitoring result is specifically: Based on the abnormal pressure distribution data set, the pressure value and fluctuation amplitude of the pressure point data are extracted, the data is classified and processed, and the pressure mean and standard deviation are calculated, the data with a standard deviation lower than a set threshold is removed, and a pressure value preliminary screening result is obtained; Based on the pressure value preliminary screening result, a hierarchical judgment of the pressure point is performed, the pressure grade of each pressure point is analyzed, and a formula is used: ； The pressure grade score is calculated, and the grade judgment result of the pressure point is generated; wherein, represents a stress level score, represents a stress value, is an average value of stress, is a standard deviation of stress, represents a stress change rate, is a weight coefficient adjusted according to a data distribution self-matching. The grade judgment result of the pressure point is summarized, the data set is classified and statistically analyzed according to the number and distribution characteristics of the differential grade pressure points, the pressure range and typical fluctuation characteristics are marked, and a pressure abnormal grade classification monitoring result is generated.

Images