A method and system for visualizing analysis of data trends of residual disease

Abstract

The application provides a kind of evaluation of residual disease regression data trend visualization analysis method and system, it is related to data processing technical field, method includes: acquisition evaluation of residual disease regression quantification data and external environment quantification data;Perform time alignment, abnormality rejection, denoising smoothing, missing compensation and numerical standardization processing, obtain standardized evaluation of residual disease regression quantification data and standardized external environment quantification data;By periodically evaluating the overall fluctuation degree of residual disease regression trend, mark trend mutation point, and cluster to form abnormal change interval;Quantify the environmental disturbance driving effect of each trend mutation point in the abnormal change interval, identify and reject environmental false positive anomaly;Abnormal confidence assessment is carried out to the retained trend mutation point, and the key suspected interval is identified;According to the key suspected interval, the authenticity and risk level of abnormal fluctuation are quantified by combining group, behavior and structure characteristics;Perform hierarchical response strategy, and trigger visual warning.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for visual analysis of disability, illness and retirement data trends. Background Technology

[0002] With the rapid development of informatization and intelligentization, data-driven visualization analysis methods have been widely applied in various fields such as healthcare, traffic management, and education management. By collecting, modeling, and graphically displaying multidimensional data, management departments can promptly identify potential risks and trend changes, thereby improving the scientific nature and accuracy of decision-making.

[0003] Currently, functional status assessments for people with disabilities commonly employ periodic, fixed-location testing or single-dimensional threshold alarm systems based on wearable devices. Typically, indicators such as cadence and grip strength are collected at fixed time points (e.g., monthly), their statistical mean is calculated, and compared with norms or previous baselines; alternatively, fixed thresholds are set for continuously monitored signals such as heart rate and acceleration, triggering an alarm when these limits are exceeded. Such systems simplify dynamic, continuous, and non-stationary physiological processes into discrete, static snapshots or isolated out-of-range events, relying entirely on expert experience and universally accepted standards.

[0004] However, existing technologies cannot effectively capture the dynamic evolution of functional states, especially struggling to identify slow, continuous degradation trends and sudden, severe anomalies. Due to the lack of multi-signal collaborative analysis capabilities, the system is prone to false alarms or missed alarms, and its fixed and rigid warning thresholds cannot adapt to individual differences and state changes. This leads to untimely risk identification and delayed prevention and control responses, resulting in fundamental problems such as low sensitivity, poor specificity, and delayed warnings in disability risk identification. Summary of the Invention

[0005] To address the limitations of existing technologies in effectively capturing the dynamic evolution of functional states, particularly in identifying slow, continuous degradation trends and sudden, severe anomalies, and to address the fundamental technical problems of low sensitivity, poor specificity, and delayed early warning in disability assessment risk identification, this invention provides a method and system for visualizing and analyzing disability assessment and deterioration data trends. These problems stem from the lack of multi-signal collaborative analysis capabilities, leading to false alarms or missed alarms, and the rigid, fixed warning thresholds that fail to adapt to individual differences and state changes.

[0006] The technical solutions provided by the embodiments of the present invention are as follows:

[0007] The first aspect of this invention provides a method for visualizing and analyzing trends in disability assessment, illness, and retirement data, comprising:

[0008] S1: Real-time collection of quantitative data on disability, illness, and deterioration assessments, as well as quantitative data on the external environment;

[0009] S2: Perform time alignment, anomaly removal, noise reduction and smoothing, missing data compensation and numerical standardization on the quantitative data of disability assessment and external environment to obtain standardized quantitative data of disability assessment and standardized quantitative data of external environment.

[0010] S3: Based on standardized quantitative data on disability, illness and deterioration assessments and standardized quantitative data on the external environment, the overall fluctuation of the disability, illness and deterioration trend is periodically assessed, trend abrupt change points are marked, and abnormal change intervals are clustered.

[0011] S4: Based on the assessment results of disability, illness and retirement trends, quantify the environmental disturbance driving effect of each trend change point within the abnormal change range, and identify and eliminate environmental false positive abnormalities.

[0012] S5: Assess the confidence level of the retained trend abrupt change points to identify key suspected intervals;

[0013] S6: Based on key suspected intervals, quantify the authenticity and risk level of abnormal fluctuations by combining group, behavioral and structural characteristics;

[0014] S7: Based on authenticity and risk level, implement a tiered response strategy and trigger visual warnings.

[0015] A second aspect of the present invention provides a data trend visualization and analysis system for disability assessment, disability, and retirement, comprising:

[0016] processor;

[0017] The memory stores computer-readable instructions, which, when executed by the processor, implement the method for visualizing and analyzing trends in disability assessment and retirement data as described in the first aspect.

[0018] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for visualizing and analyzing trends in disability assessment and retirement data as described in the first aspect.

[0019] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0020] In this invention, by simultaneously extracting and quantifying rate indicators reflecting the long-term functional degradation trend and response intensity of short-term sudden anomalies, and integrating the synergistic change characteristics among multiple signals for comprehensive risk assessment, continuous and accurate perception of the nonlinear evolution and phased fluctuations of functional state is achieved. This overcomes the limitations of traditional methods in recognizing complex functional evolution patterns, significantly improves the sensitivity and specificity of disability risk identification, and enables early and accurate warning of potential degradation and sudden events. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating a method for visualizing and analyzing trends in disability assessment and retirement data, provided in an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram illustrating the anomaly authenticity assessment values ​​and corresponding risk levels for five trend mutation points provided in an embodiment of the present invention.

[0024] Figure 3 This is a schematic diagram of the structure of a data trend visualization and analysis system for disability assessment, illness, and retirement provided in an embodiment of the present invention. Detailed Implementation

[0025] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0026] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0027] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0028] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0029] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0030] Reference manual attached Figure 1The diagram shows a flowchart of a method for visualizing and analyzing the trends of disability assessment and retirement data provided by an embodiment of the present invention.

[0031] This invention provides a method for visualizing and analyzing trends in disability and retirement data. This method can be implemented using a device for visualizing and analyzing trends in disability and retirement data, which can be a terminal or a server. The processing flow of this method may include the following steps:

[0032] S1: Real-time collection of quantitative data on disability, illness, and deterioration assessments, as well as quantitative data on the external environment.

[0033] Among them, the quantitative data for disability assessment refers to the structured numerical sequence collected through wearable devices, medical sensors, etc., which can objectively and continuously reflect changes in an individual's physiological function and degree of disability.

[0034] Among them, external environmental quantitative data refers to objective environmental parameters that may affect an individual's physiological function or recovery process, which are obtained through environmental sensors or system interfaces.

[0035] Specifically, the system collects quantitative data on disability and retirement assessments, as well as quantitative data on the external environment, in real time. The quantitative data includes the number of retirement applications due to illness, the number of disability assessment applications, the number of people who passed the assessment, the approval time, the number of cases requiring review, the total number of employees on duty, the average length of service, the monthly overtime hours, and the number of sick leave applications per month. The quantitative data on the external environment includes the average daily temperature, the average daily sunshine duration, the daily precipitation, and the average daily PM2.5 concentration.

[0036] Furthermore, quantitative data on disability and medical retirement assessments and external environmental data are collected in real time. Specifically, the number of medical retirement applications is cumulatively counted based on the number of applications recorded within the same statistical period; the number of disability assessment applications is cumulatively counted based on the number of submissions recorded within the same statistical period; the number of approved assessments is cumulatively counted based on the number of approved assessments recorded within the same statistical period; the approval time is calculated as the average of the actual processing time for both medical retirement and disability assessment approval processes within the same statistical period; the number of review cases is cumulatively counted based on the number of cases filed in the review acceptance records within the same statistical period; the total number of employees on duty is summarized and counted based on the number of employees recorded in the on-duty roster within the same statistical period; the average length of service is calculated based on the length of service distribution corresponding to the total number of employees on duty within the same statistical period; the monthly overtime hours are cumulatively counted based on the overtime hours recorded in the overtime registration records within the same statistical period; and the number of sick leave requests per month is cumulatively counted based on the number of sick leave requests recorded within the same statistical period. The quantitative data of the external environment includes daily average temperature, daily average sunshine duration, daily precipitation, and daily average PM2.5 concentration. The daily average temperature is calculated based on the daily temperature observation values ​​for the same statistical period. The daily average sunshine duration is calculated based on the daily sunshine duration observation values ​​for the same statistical period. The daily precipitation is calculated based on the daily precipitation observation values ​​for the same statistical period. The daily average PM2.5 concentration is calculated based on the daily PM2.5 monitoring concentration for the same statistical period.

[0037] S2: Perform time alignment, anomaly removal, noise reduction and smoothing, missing data compensation and numerical standardization on the quantitative data of disability assessment and external environment to obtain standardized quantitative data of disability assessment and standardized quantitative data of external environment.

[0038] In one possible implementation, S2 specifically refers to:

[0039] A time series alignment algorithm is used to perform time alignment and resampling on quantitative data of disability assessment, illness and degeneration and quantitative data of external environment.

[0040] Among them, time series alignment algorithm refers to the algorithm used to adjust time series data from different sources and with different sampling frequencies to a unified time reference.

[0041] The local anomaly factor algorithm is used to identify and remove anomalies caused by data entry errors and abnormal submissions in the quantitative data of disability assessment, disability assessment, and external environment assessment.

[0042] Among them, the local anomaly factor algorithm refers to a density-based unsupervised anomaly detection algorithm used to identify outliers in a dataset that are significantly different from most samples.

[0043] The Holtwinters exponential smoothing algorithm is used to denoise the quantitative data on disability assessment, illness and degeneration, as well as the quantitative data on the external environment, thereby smoothing out short-term fluctuations in the data.

[0044] The Holt-Winters exponential smoothing algorithm is a forecasting and smoothing method suitable for time series with trend and seasonal components. This algorithm uses a weighted average of historical data and dynamically updates the level, trend, and seasonal components to preserve long-term trends and suppress short-term random fluctuations in the series.

[0045] Using spline interpolation algorithms, missing data caused by data loss and data collection interruption are repaired in quantitative data on disability, illness and retirement assessments and quantitative data on external environment.

[0046] Spline interpolation is a numerical method that uses a piecewise polynomial function to fit known data points, thereby constructing a smooth interpolation curve within the data gap interval. Cubic spline interpolation is commonly used, ensuring that the interpolation function has continuous first and second derivatives at the nodes. It is suitable for repairing time-series data gaps caused by temporary equipment downtime, signal interruptions, etc., maintaining the physiological rationality of the data curve.

[0047] The min-max normalization algorithm is used to normalize the numerical values ​​of quantitative data on disability, illness, and degeneration assessments and quantitative data on the external environment with different scales, thereby eliminating scale differences.

[0048] Among them, the min-max normalization algorithm refers to a linear normalization method that maps the original data to a specified interval (usually [0,1]) according to the following formula.

[0049] Specifically, the collected quantitative data on disability and retirement assessments and external environmental data are aligned and resampled using a time series alignment algorithm. This ensures a one-to-one correspondence between the number of retirement applications, disability assessment applications, number of approved assessments, approval time, number of review cases, total number of employees, average length of service, monthly overtime hours, monthly sick leave, average daily temperature, average daily sunshine duration, daily precipitation, and average daily PM2.5 concentration on the same statistical time period boundary, thus avoiding time misalignment and omissions caused by different sampling frequencies. The Local Anomaly Factor (LOF) algorithm identifies and eliminates outliers caused by data entry errors and abnormal submissions. Based on the local density level of the above indicators in the multidimensional feature space within the same statistical period, the LEF algorithm calculates anomaly scores, thereby eliminating false high and false low values ​​in the number of sick leave applications, disability assessment applications, number of people who have passed assessments, approval time, number of cases reviewed, total number of employees, average length of service, monthly overtime hours, monthly sick leave, average daily temperature, average daily sunshine hours, daily precipitation, and average daily PM2.5 concentration that deviate significantly from historical distributions. This ensures that subsequent analysis only retains valid fluctuations caused by real business behavior and objective environment. The Holt-Winters exponential smoothing algorithm is used for denoising and smoothing. This algorithm considers horizontal, trend, and periodic terms simultaneously. It performs exponential weighted updates on the number of applications for sick leave, disability assessment, number of approved assessments, approval time, number of cases reviewed, total number of employees, average length of service, monthly overtime hours, monthly sick leave, average daily temperature, average daily sunshine hours, daily precipitation, and average daily PM2.5 concentration. This suppresses spike interference caused by occasional abnormal reporting in a single statistical period, and makes the changes in adjacent statistical periods reflect a continuous trend rather than instantaneous noise jumps.

[0050] Furthermore, spline interpolation algorithm is used to repair missing data caused by data loss and collection interruption. The spline interpolation algorithm fits a smooth curve based on the number of sick leave applications, number of disability assessment applications, number of people who passed the assessment, approval time, number of review cases, total number of employees, average length of service, monthly overtime hours, number of sick leave applications, daily average temperature, daily average sunshine duration, daily precipitation and daily average PM2.5 concentration corresponding to adjacent existing data points in time sequence. The data of the missing period is restored along the time axis as a continuous curve to avoid non-realistic jumps at the missing points. The minimum-maximum standardization algorithm is used to normalize the quantitative data of disability and retirement assessments and external environmental data of different dimensions. The minimum-maximum standardization algorithm linearly maps the number of retirement applications, disability assessment applications, number of people who have passed the assessment, approval time, number of review cases, total number of employees, average length of service, monthly overtime hours, number of sick leave applications, daily average temperature, daily average sunshine hours, daily precipitation, and daily average PM2.5 concentration to their respective minimum and maximum values ​​in the historical statistical period. This results in normalized results within a unified dimension range, making the subsequent calculation of assessment values ​​for disability and retirement trends, environmentally driven false positive assessment values, and abnormal authenticity assessment values ​​comparable, and avoiding any high-amplitude indicator from disproportionately amplifying its influence in subsequent judgments.

[0051] In this embodiment of the invention, by performing a full-link collaborative preprocessing process on multi-source heterogeneous data, including time-series alignment, anomaly cleaning, noise smoothing, missing data repair, and scale normalization, a standardized analysis dataset with high consistency, low noise, no missing data, and uniform scale is constructed. This fundamentally eliminates pseudo-fluctuations and misjudgments caused by data quality issues, providing a reliable data foundation for subsequent accurate trend assessment, anomaly identification, and risk quantification.

[0052] S3: Based on standardized quantitative data on disability, illness and deterioration assessments and standardized quantitative data on the external environment, the overall fluctuation of disability, illness and deterioration trends is periodically assessed, trend abrupt change points are marked, and abnormal change intervals are clustered.

[0053] Among them, the trend inflection point refers to the specific time point at which the trend of disability, illness, or decline changes significantly in direction or intensity.

[0054] Among them, the abnormal change interval refers to the time range that represents a continuous abnormal state, which is formed by aggregating multiple "trend change points" that are consecutive or close in time and the data segments between them through a clustering algorithm.

[0055] In one possible implementation, based on standardized quantitative data on disability and disability assessment and standardized quantitative data on the external environment, the overall fluctuation of the disability and disability trend is periodically assessed, specifically including sub-steps S301 to S307:

[0056] S301: Based on standardized quantitative data on disability and medical retirement assessments, extract the number of applications for medical retirement, the number of applications for disability assessment, the number of people who passed the assessment, and the approval time to construct a dataset for disability and medical retirement review.

[0057] S302: Set a fixed time window as an assessment cycle, calculate the average and standard deviation of various disability assessment, sick leave review data and external environment quantitative data within each assessment cycle, and extract the minimum and maximum values ​​of the corresponding data.

[0058] S303: Subtract the absolute value of each disability assessment / retirement audit data from the corresponding average value, and then divide the absolute value by the corresponding standard deviation to obtain the standardized deviation value of each disability assessment / retirement audit data.

[0059] S304: Take the arithmetic mean of the four standardized deviation values ​​to obtain the disease-related decline trend fluctuation term.

[0060] S305: Subtract the corresponding average value from each external environment quantitative data point and take the absolute value. Then, take the exponential function value with the natural constant as the base, add one, and take the natural logarithm. Divide the resulting logarithmic value by the difference between the maximum and minimum values ​​of the corresponding external environment quantitative data point and add one to obtain the normalized fluctuation value of each external environment quantitative data point.

[0061] S306: Take the arithmetic mean of the four normalized fluctuation values ​​to obtain the external environment correction term.

[0062] S307: Add the fluctuation term of the disability and retirement trend to the external environment correction term to obtain the disability and retirement trend assessment value, and obtain the disability and retirement trend assessment result.

[0063] Specifically, based on the preprocessed quantitative data on disability and medical retirement assessments, the number of applications for medical retirement, the number of applications for disability assessment, the number of approved assessments, and the approval time are extracted to construct a dataset for disability and medical retirement review. The number of applications for medical retirement reflects the concentrated health pressure on the population during the current period; the number of applications for disability assessment characterizes the proactive changing trend of the population's application behavior during this period; the number of approved assessments depicts the actual intensity of the objective disability risk during this period; and the approval time reflects the squeezing effect of the approval load level on the medical retirement processing flow during this period. These factors—business application perspective, objective assessment perspective, population willingness perspective, and process capacity perspective—together constitute a comprehensive business background for the trend of disability and medical retirement assessments. A fixed time window is set as an assessment cycle. This fixed time window serves as a unified statistical caliber, allowing for synchronous analysis of data from different sources within the same time span. This ensures that the four types of data are comparable and consistent across time scales, thereby enabling a complete characterization of the fluctuation intensity of the trend of disability and medical retirement assessments within this assessment cycle. The mean and standard deviation of all disability and medical retirement review data and external environmental quantitative data for each assessment period are calculated. The mean is used to construct the statistical benchmark for that assessment period, and the standard deviation is used to characterize the dispersion of the data within that period. The minimum and maximum values ​​of the corresponding data are extracted and used to establish the data range boundaries for that period. The absolute value of each disability and medical retirement review data is subtracted from its corresponding mean. This absolute value is used to measure the deviation magnitude without directionality. This is then divided by the corresponding standard deviation, which is used to scale data of different magnitudes, resulting in the standardized deviation value for each disability and medical retirement review data. The standardized deviation value characterizes the deviation of the number of medical retirement applications, the number of disability applications, the number of approved assessments, and the approval time from the statistical context of the disability and medical retirement review data within that assessment period, and is used to reduce the absolute value differences between indicators of different magnitudes. The arithmetic mean of the four standardized deviation values ​​is used to achieve an equal-weighted synthesis of the influence of the four audit data, resulting in the retirement trend fluctuation term. This term measures the comprehensive fluctuation intensity of the number of retirement applications, disability assessment applications, number of approved assessments, and approval time within the assessment period, reflecting the overall abnormally active level of current business application and approval behaviors. The absolute value is obtained by subtracting the corresponding average value from each external environmental quantitative data point. This absolute value is used to characterize the magnitude of environmental disturbances without directionality. An exponential function value is then taken with the natural constant e as the base. The exponential function amplifies the impact of sudden environmental changes. One is added to this value, and the natural logarithm is taken. The natural logarithm is used to compress the exponential amplification effect caused by extreme environmental fluctuations. The resulting logarithmic value is divided by the difference between the maximum and minimum values ​​of the corresponding external environmental quantitative data, plus one. The difference between the maximum and minimum values ​​reflects the periodic span of the environmental data, and adding one avoids instability caused by the denominator approaching zero. This yields the normalized fluctuation value for each external environmental quantitative data point.

[0064] Furthermore, the normalized fluctuation value is used to map the intensity of environmental disturbances such as daily average temperature, daily average sunshine duration, daily precipitation, and daily average PM2.5 concentration to a unified scale across different assessment periods, highlighting abrupt environmental changes while suppressing stable backgrounds. The arithmetic mean of the four normalized fluctuation values ​​is taken, and this mean is used to construct a comprehensive measure of environmental impact, yielding an external environment correction term. This external environment correction term quantifies the potential driving impact of external objective factors on the number of applications for medical retirement, disability assessment, the number of approved cases, and approval time within the assessment period. The medical retirement trend fluctuation term is added to the external environment correction term. This addition is used to integrate the intensity of internal business fluctuations and the intensity of external environmental disturbances within a unified numerical space, yielding the disability and medical retirement trend assessment value. This assessment value serves as a comprehensive indicator for the assessment period, characterizing the combined intensity of business fluctuations reflected in the medical retirement trend fluctuation term and the intensity of environmental background reflected in the external environment correction term within the assessment period. This provides a quantitative basis for subsequent trend abrupt change point marking and abnormal change interval identification.

[0065] Optionally, the specific formula for calculating the disability, illness, and decline trend assessment value is as follows:

[0066]

[0067] in, This indicates the assessment value for disability, illness, and retirement trends. This represents the data from the review of the i-th disability / illness retirement assessment. This represents the average value of the disability, illness, and retirement assessment data for the i-th item within the period. This represents the standard deviation of the data for the i-th disability / illness retirement assessment. This represents the j-th external environment quantitative data. This represents the average value of the j-th external environment data. This represents the maximum value of the j-th external environment data item. This represents the minimum value of the j-th external environment data item. This represents the natural logarithm function.

[0068] In one possible implementation, identifying trend abrupt change points and clustering them into abnormal change intervals specifically involves:

[0069] The disability, illness and retirement trend assessment values ​​within a continuous preset assessment period are arranged in chronological order to construct a trend sequence. The difference sequence of the disability, illness and retirement trend assessment values ​​in adjacent assessment periods is calculated, and the difference sequence is processed by moving average to obtain the rate of change of disability, illness and retirement trend.

[0070] The rate of change of disability, illness and retirement trends and the trend change threshold are compared in real time. When the rate of change of disability, illness and retirement trends is greater than the trend change threshold, the corresponding assessment period is marked as a trend inflection point.

[0071] The density-based mutation point aggregation algorithm is used to cluster adjacent trend mutation points. When the time interval between any two trend mutation points is less than the clustering distance threshold and the number of cluster points exceeds the minimum number of clusters, the clustering interval is determined to be an abnormal change interval.

[0072] It should be noted that those skilled in the art can set the trend change threshold, clustering distance threshold, and preset evaluation period according to actual needs, and this invention does not limit these settings.

[0073] Specifically, after obtaining the assessment value of the disability, illness and retirement trend, the assessment values ​​of the disability, illness and retirement trend in N consecutive assessment periods are constructed into a trend sequence in chronological order. The difference sequence of the assessment values ​​of the disability, illness and retirement trend in adjacent assessment periods is calculated so that each difference value corresponds to the net change between two adjacent assessment periods. The difference sequence is then processed by moving average. The moving average is used to suppress the peak response caused by the instantaneous fluctuation of a single assessment period, thereby obtaining the rate of change of the disability, illness and retirement trend. The rate of change of the disability, illness and retirement trend is used to characterize the continuous change speed of the assessment value of the disability, illness and retirement trend on the time axis.

[0074] Where N is a positive integer greater than one.

[0075] Furthermore, the rate of change in disability and retirement assessment trends and the trend change threshold are compared in real time. When the rate of change in disability and retirement assessment trends exceeds the trend change threshold, the corresponding assessment period is marked as a trend inflection point. A trend inflection point indicates that the disability and retirement assessment trend value in that assessment period shows a significant jump compared to the previous assessment period and has a rapid upward trend. A density-based inflection point aggregation algorithm is used to cluster adjacent trend inflection points. The density-based inflection point aggregation algorithm uses temporal proximity as a distance metric, grouping trend inflection points that are close to each other in time into the same cluster unit to identify continuous upward surge behavior. When the time interval between any two trend inflection points is less than the clustering distance threshold and the number of clusters exceeds the minimum number of clusters, the clustering interval is determined to be an abnormal change interval. An abnormal change interval indicates that there are high-frequency and high-amplitude jumps in disability and retirement assessment trend values ​​within that time range, suggesting that the number of retirement applications, disability assessment applications, number of approved assessments, and approval time show abnormally concentrated activity in a short period of time, rather than being a single, occasional fluctuation.

[0076] In this embodiment of the invention, by performing multi-dimensional feature fusion and periodic fluctuation assessment on disability assessment and retirement business data and external environmental data, a comprehensive assessment index capable of sensitively capturing nonlinear jumps in business trends is constructed. Based on a density clustering mechanism, isolated mutation points are aggregated into persistent abnormal time intervals, solving the problem of slow identification of phased and cumulative risk signals by traditional monitoring methods, and realizing accurate positioning and interval-based early warning of the early accumulation of potential systemic risks.

[0077] S4: Based on the assessment results of disability, illness and retirement trends, quantify the environmental disturbance driving effect of each trend change point within the abnormal change range, and identify and eliminate environmental false positive abnormalities.

[0078] Among them, the environmental disturbance-driven effect refers to the short-term, non-pathological or policy-related shocks or fluctuations in the data related to disability, illness, and retirement caused by external environmental factors.

[0079] Among them, environmental false positive anomalies refer to those abnormal disability and retirement data that are mainly driven by environmental disturbances rather than reflecting the actual functional decline of individuals or policy and institutional changes.

[0080] In one possible implementation, based on the assessment results of disability, illness, and retirement trends, the environmental disturbance driving effect of each trend abrupt change point within the abnormal change range is quantified, specifically including sub-steps S401 to S404:

[0081] S401: Calculate the average and standard deviation of the disability, disability and retirement trend assessment values ​​and external environmental quantitative data for all assessment periods within the abnormal change interval.

[0082] S402: For each trend inflection point, take the absolute value of the difference between the disability, illness and retirement trend assessment value and the average value of the disability, illness and retirement trend assessment value within the abnormal change interval, and then divide it by the standard deviation of the disability, illness and retirement trend assessment value within the abnormal change interval. Take the hyperbolic tangent function value of the obtained ratio and add one to obtain the trend correction term.

[0083] S403: Take the absolute value of the difference between the daily average temperature, daily average sunshine duration, daily precipitation, and daily average PM2.5 concentration at the current trend abrupt change point and the corresponding average value within the abnormal change range, divide each by the corresponding standard deviation plus one, and sum them up after taking the exponential function value with the natural constant as the base. This yields the environmental fluctuation response term.

[0084] S404: Divide the environmental fluctuation response term by four times the trend correction term, add one to the resulting ratio, and take the natural logarithm to obtain the environmental-driven false positive assessment value.

[0085] Specifically, the mean and standard deviation of the disability, illness, and retirement trend assessment values ​​and external environmental quantitative data for all assessment periods within the abnormal change interval are calculated. The mean is used as the benchmark level for this interval, and the standard deviation is used to define the dispersion of fluctuations within this interval. For each trend inflection point, the absolute value of the difference between the disability, illness, and retirement trend assessment value and the mean of the disability, illness, and retirement trend assessment values ​​within the abnormal change interval is taken. This absolute value operation is used to measure the deviation of the trend inflection point relative to the benchmark level. This value is then divided by the standard deviation of the disability, illness, and retirement trend assessment values ​​within the abnormal change interval. This division is used to normalize the deviation on a uniform scale, making the deviation levels of different assessment periods comparable. The resulting ratio is then taken as the hyperbolic tangent function value and incremented by one. The hyperbolic tangent function is used to compress the disturbance of extreme deviation values ​​on the overall judgment, making the trend correction term able to suppress strong noise points. The increment operation is used to avoid zero values ​​that would cause instability in subsequent logarithmic calculations, resulting in the trend correction term. The trend correction term is used to describe the degree of deviation of the business intensity of the trend inflection point relative to the overall level of the abnormal change interval, and is used to suppress excessive amplification caused by the inherent noise of a single assessment period.

[0086] Furthermore, the absolute values ​​of the differences between the daily average temperature, daily average sunshine duration, daily precipitation, and daily average PM2.5 concentration at the current trend abrupt change point and the corresponding average values ​​within the abnormal change interval are taken. This absolute value operation is used to quantify the meteorological deviation intensity of the period in which the trend abrupt change point occurs. Then, each is divided by the corresponding standard deviation plus one. This division by the standard deviation operation is used to map the disturbance amplitude of different external environmental quantitative data to a unified scale. The addition of one is used to avoid the excessive amplification of values ​​caused by low variance variables. The exponential function values ​​are taken with the natural constant e as the base. The exponential function is used to enhance the response gradient caused by meteorological abrupt changes, so that the significantly deviated meteorological conditions are highlighted in the environmental fluctuation response term. Then, the four exponential function values ​​are summed. The summation operation is used to aggregate the comprehensive disturbance intensity of the four meteorological factors into a single index space to obtain the environmental fluctuation response term. The environmental fluctuation response term is used to quantify whether the trend abrupt change point is in a period of meteorological state change or air quality change, thereby reflecting whether the surge in the number of applications for medical retirement, disability assessment, number of approved assessments, and approval time is mainly driven by the external environment. The environmental fluctuation response term is divided by four times the trend correction term. This four-times trend correction term serves as a normalization scaling factor to limit the amplification effect of the environmental fluctuation response term under extreme environmental conditions and to suppress the disproportionate impact of the external environment on the judgment result when the business deviation intensity is low. The resulting ratio is then incremented by one and its natural logarithm is taken. This increment is used to avoid zero input, which would render the logarithm uncomputable. The logarithmic function is used to further compress extreme values ​​and enhance numerical stability, yielding the environmentally driven false positive assessment value. This value characterizes whether the abnormal fluctuation at the trend abrupt change point primarily originates from the combined disturbances of daily average temperature, daily average sunshine duration, daily precipitation, and daily average PM2.5 concentration, rather than from abnormally active disability / retirement assessment behavior itself.

[0087] Optionally, the specific formula for calculating the environmentally driven false positive assessment value is as follows:

[0088]

[0089] in, This indicates the environmentally driven false positive assessment value. This represents the disability, illness, and retirement trend assessment value, indicating the current point of abrupt change in the trend. This represents the average value of the disability, illness, and retirement trend assessment within the abnormal change range. This represents the standard deviation of the assessment of disability, illness, and retirement trends within the range of abnormal changes. This represents the average daily temperature at the point where the current trend abruptly changes. This represents the average daily temperature within the range of abnormal changes. This represents the standard deviation of daily average temperature within the range of abnormal changes. This represents the average daily sunshine duration at the current trend inflection point. This represents the average daily sunshine duration within the range of abnormal changes. It represents the standard deviation of the average daily sunshine duration within the range of abnormal changes. This represents the average daily precipitation at the point where the current trend changes abruptly. This represents the average daily precipitation within the range of abnormal changes. This represents the standard deviation of daily precipitation within the range of abnormal changes. This represents the average daily PM2.5 concentration at the current point of trend inflection. This represents the average daily PM2.5 concentration within the range of abnormal changes. This represents the standard deviation of daily average PM2.5 concentration within the range of abnormal changes. This represents the hyperbolic tangent function.

[0090] In one possible implementation, identifying and eliminating environmental false positives specifically involves:

[0091] The system compares the environmental-driven false positive assessment value with the false positive determination threshold in real time. When the environmental-driven false positive assessment value is greater than the false positive determination threshold, the abnormal trend change point is identified as a false positive fluctuation, automatically marked as an environmental anomaly, and removed from the abnormal change range. When the environmental-driven false positive assessment value is less than or equal to the false positive determination threshold, the corresponding trend change point is retained.

[0092] Specifically, the system compares the environmental-driven false positive assessment value and the false positive judgment threshold in real time. The false positive judgment threshold is a judgment boundary set for the intensity of environmental driving factors, used to distinguish between a concentrated increase in cases dominated by environmental background and a suspected abnormal increase in business. When the environmental-driven false positive assessment value is greater than the false positive judgment threshold, the abnormal trend change point is judged as a false positive fluctuation, automatically marked as an environmental anomaly, and removed from the abnormal change range. After removal, it will not enter the subsequent manual risk monitoring stage, thus avoiding the misclassification of large-scale reporting peaks formed by seasonal peaks in illness or concentrated periods of job retirement as high-risk behaviors. When the environmental-driven false positive assessment value is less than or equal to the false positive judgment threshold, the corresponding trend change point is retained. The retained trend change point is regarded as having an increase in abnormal intensity unrelated to environmental factors, and enters the subsequent in-depth analysis of the anomaly confidence and authenticity, used to focus on identifying high-risk situations such as suspected deliberate fraud, active abnormal reporting, and concentrated avoidance of review.

[0093] It should be noted that those skilled in the art can set the false positive threshold according to actual needs, and this invention does not limit it.

[0094] In this embodiment of the invention, by constructing a quantitative assessment model that integrates the intensity of business trends and the intensity of environmental disturbances, abnormal data fluctuations driven by external environmental factors are accurately separated, and environmental false positive signals are automatically identified and eliminated based on a threshold judgment mechanism. This solves the problem of high false alarm rate and poor warning specificity caused by the inability of traditional risk monitoring methods to distinguish between internal and external influencing factors, and significantly improves the ability of subsequent anomaly analysis to focus on real high-risk events and the accuracy of identification.

[0095] S5: Assess the confidence level of the retained trend abrupt change points to identify key suspected intervals.

[0096] Among them, the anomaly confidence level is a quantitative indicator used to measure the strength and credibility of the correlation (i.e., the fact that they together constitute a persistent anomaly) between any two adjacent trend change points that have been retained after S4 screening.

[0097] Among them, the key suspected interval refers to a time period that the system has determined to have a high probability and persistent real anomaly risk, requiring the initiation of the highest priority investigation and manual review.

[0098] In one possible implementation, S5 specifically includes sub-steps S501 to S503:

[0099] S501: For trend mutation points retained after removing environmental false positive anomalies, construct abnormal residual sequences in chronological order.

[0100] S502: Based on the abnormal residual sequence, calculate the difference between the time interval between adjacent trend change points and the assessment value of disability, illness and retirement trend. Take the absolute value of the difference, take the exponential function value with the natural constant as the base, and then multiply it by the reciprocal of the time interval to obtain the abnormality confidence value.

[0101] S503: When the anomaly confidence value is greater than the first confidence threshold, adjacent trend abrupt change points are aggregated into a hidden anomaly cluster. When a series of hidden anomaly clusters with values ​​greater than a preset value appear and their duration reaches the minimum duration threshold, the corresponding time range is marked as a key suspected interval. The key suspected interval is only de-marked when all anomaly confidence values ​​within a subsequent consecutive preset evaluation period are lower than the second confidence threshold.

[0102] It should be noted that those skilled in the art can set preset values ​​and preset evaluation periods according to actual needs, and this invention does not limit these settings.

[0103] Specifically, for the trend abrupt changes retained after false positive removal, an abnormal residual sequence is constructed in chronological order. This abnormal residual sequence records the retention distribution of trend abrupt changes on the time axis where the environment-driven false positive assessment value did not trigger removal. The difference between the time interval between adjacent trend abrupt changes and the disability / retirement trend assessment value is calculated. The absolute value of the difference is then used to take an exponential function value with the natural constant e as the base, which amplifies the impact of a sudden increase in the disability / retirement trend assessment value. This value is then multiplied by the reciprocal of the time interval, so that the closer the two trend abrupt changes are in time, the higher their contribution to the anomaly confidence. This anomaly confidence value is used to characterize whether adjacent trend abrupt changes exhibit suspicious behavior of a high-amplitude continuous surge in a short period of time.

[0104] Furthermore, when the anomaly confidence value exceeds the first confidence threshold, adjacent trend abrupt change points are aggregated into hidden anomaly clusters. Hidden anomaly clusters represent high-intensity, continuous fluctuations that cannot be explained by environmental factors within a short period. When M hidden anomaly clusters appear consecutively and their duration reaches the minimum duration threshold, the corresponding time range is marked as a key suspected interval. This key suspected interval is considered a high-concern period requiring further investigation, used to indicate potential behavioral tendencies such as coordinated reporting, concentrated avoidance of review, and concentrated creation of peak numbers in disability retirement and disability assessment applications. After a key suspected interval is marked, it is only demarked if all anomaly confidence values ​​are below the second confidence threshold for the subsequent L consecutive assessment periods. This avoids frequent and repeated marking and demarking of key suspected intervals due to short-term fluctuations and ensures that only persistent anomalies receive high priority attention. Here, M and L are both positive integers greater than one.

[0105] It should be noted that those skilled in the art can set the size of the first confidence threshold, the second confidence threshold, and the minimum duration threshold according to actual needs, and this invention does not limit these settings.

[0106] In this embodiment of the invention, by performing spatiotemporal correlation quantitative analysis on the suspected risk points retained after filtering, a confidence index capable of characterizing high-density, high-intensity continuous abnormal patterns is constructed. Based on the rules of persistence and aggregation, key suspicious periods that require priority intervention are automatically identified. This solves the shortcomings of traditional early warning methods that overreact to scattered, low-frequency risk signals but fail to respond adequately to systemic and persistent risks, and achieves precise focusing and dynamic tracking of concealed and organized high-risk behaviors.

[0107] S6: Based on key suspected intervals, the authenticity and risk level of abnormal fluctuations are quantified by combining group, behavioral and structural characteristics.

[0108] In one possible implementation, S6 specifically includes sub-steps S601 to S605:

[0109] S601: For trend abrupt change points within key suspected intervals, divide the number of re-examination cases at the current trend abrupt change point by the total number of on-duty personnel plus one, add one to the resulting ratio, and take the natural logarithm to obtain the group abnormality intensity item.

[0110] S602: Divide the number of sick leave requests per month by the total overtime hours per month plus one, add one to the resulting ratio, and take the natural logarithm to obtain the behavioral abnormality intensity item.

[0111] S603: Add the group abnormality intensity term to the behavior abnormality intensity term and divide by the square root of the environmental-driven false positive assessment value plus one to obtain the comprehensive abnormality intensity term.

[0112] S604: Calculate the difference between the current trend inflection point and the average length of service in the previous assessment period, divide it by the average length of service in the previous assessment period plus one, and obtain the absolute value of the ratio to get the structural change term.

[0113] S605: Multiply the comprehensive abnormality intensity item, structural change item, disability and retirement trend assessment value by one and the abnormality amplification coefficient in sequence to obtain the abnormality authenticity assessment value.

[0114] Specifically, for trend abrupt change points within key suspected intervals, the number of cases reviewed at the current trend abrupt change point is divided by the total number of on-duty personnel plus one. The addition of one is used to avoid excessive amplification of the ratio when the total number of on-duty personnel is small. The obtained ratio is then increased by one and its natural logarithm is taken. The natural logarithm is used to compress extreme high values ​​and enhance the comparability between different orders of magnitude, resulting in a group anomaly intensity term. The group anomaly intensity term is used to reflect whether the number of reviewed cases within the scope of on-duty personnel shows an abnormally concentrated upward trend, thereby measuring whether the anomaly shows group clustering characteristics.

[0115] Furthermore, the monthly number of sick leave applications is divided by the monthly overtime hours plus one. Adding one avoids instability in the denominator when overtime hours are zero. The resulting ratio is then rounded to its natural logarithm. This natural logarithm transforms drastic deviations into a stable logarithmic scale, yielding a behavioral anomaly intensity term. This term reflects the degree of deviation between sick leave and overtime behavior within the same time period, quantifying whether a significant increase in sick leave applications occurs while overtime hours do not, a behavior mismatched with normal workload. The group anomaly intensity term is added to the behavioral anomaly intensity term, then divided by the square root of the environmentally driven false positive assessment value plus one. The square root weakens the amplification effect of the environmentally driven false positive assessment value under extreme conditions, and adding one avoids instability caused by the denominator approaching zero, resulting in a comprehensive anomaly intensity term. This comprehensive anomaly intensity term measures the extent to which the surge in behavioral intensity at this trend inflection point cannot be explained by average daily temperature, average daily sunshine duration, daily precipitation, and average daily PM2.5 concentration, thus highlighting the risk of artificially inflated reporting unrelated to environmental background. Calculate the difference between the current trend inflection point and the average length of service in the previous assessment period. This difference is used to characterize the instantaneous magnitude of the change in personnel structure. Then divide this by the average length of service in the previous assessment period plus one. Adding one is used to avoid the deviation being infinitely magnified when the average length of service is low. Take the absolute value of the resulting ratio. The absolute value is used to characterize the magnitude of the change in length of service without direction. This yields the structural change term. The structural change term is used to describe whether there has been a sudden change in the average length of service of on-the-job personnel, indicating whether there is a tendency for changes in the declared structure caused by a sudden decrease in the number of young or old personnel.

[0116] Furthermore, the comprehensive anomaly intensity item, structural change item, disability / retirement trend assessment value plus one, and an anomaly amplification coefficient are multiplied sequentially. This multiplication is used to integrate the intensity of behavioral deviation, the magnitude of personnel structure changes, and the business trend background on a unified scale. Adding one ensures that the disability / retirement trend assessment value still contributes stably when it is near zero, resulting in the anomaly authenticity assessment value. The anomaly amplification coefficient is set based on the characteristic intensity of historically confirmed high-risk trend abrupt changes, reflecting the typical amplification level of such risks. Introducing the anomaly amplification coefficient amplifies high-risk trend abrupt changes in the anomaly authenticity assessment value, facilitating the direct triggering of major anomaly alarms.

[0117] Optionally, the specific formula for calculating the anomaly authenticity assessment value is as follows:

[0118]

[0119] in, This indicates the value used to assess the authenticity of anomalies. This indicates the number of cases reviewed at the current trend inflection point. This represents the total number of employees currently on duty at the point where the trend suddenly changes. This indicates the number of sick leave applications in a month at which the current trend changes abruptly. This indicates the monthly overtime hours at the current trend inflection point. This represents the average length of service at the point where the current trend changes abruptly. This indicates the average length of service in the previous assessment period. The assessment value for disability, illness, and retirement trends indicates the current point of abrupt change in the trend. The environmental-driven false positive assessment value represents the current trend inflection point. This indicates the amplification factor.

[0120] For example, the abnormal authenticity assessment values ​​under five trend inflection points and the input data used to calculate these assessment values ​​are presented. Specifically, the abnormality amplification factor is uniformly set to 1000. The number of review cases for trend inflection point 1 is 8, the total number of employees is 320, the number of sick leave requests per month is 45, the monthly overtime hours are 120, the average length of service at the current trend inflection point is 9.20, the average length of service in the previous assessment period is 9.00, the disability / illness / retirement trend assessment value is 1.35, the environment-driven false positive assessment value is 0.40, and under the effect of the abnormality amplification factor, the corresponding abnormality authenticity assessment value is 9.81. Trend inflection point 2 has 12 reviewed cases, a total of 310 employees, 32 sick leave requests per month, 140 overtime hours per month, an average length of service of 9.10 years in the current trend inflection point, an average length of service of 9.20 years in the previous assessment period, a disability / illness / retirement trend assessment value of 0.80, an environment-driven false positive assessment value of 0.10, and a corresponding anomaly authenticity assessment value of 3.25. Trend inflection point 3 has 15 reviewed cases, a total of 305 employees, 60 sick leave requests per month, 110 overtime hours per month, an average length of service of 8.90 years in the current trend inflection point, an average length of service of 9.00 years in the previous assessment period, a disability / illness / retirement trend assessment value of 1.80, an environment-driven false positive assessment value of 0.30, and a corresponding anomaly authenticity assessment value of 8.68. Trend inflection point 4 has 7 reviewed cases, a total of 298 employees, 53 sick leave requests per month, 150 overtime hours per month, an average length of service of 9.05 years in the current trend inflection point, an average length of service of 9.10 years in the previous assessment period, a disability / illness / retirement trend assessment value of 1.30, an environment-driven false positive assessment value of 0.40, and a corresponding anomaly authenticity assessment value of 2.26. Trend inflection point 5 has 20 reviewed cases, a total of 290 employees, 75 sick leave requests per month, 100 overtime hours per month, an average length of service of 8.70 years in the current trend inflection point, an average length of service of 8.90 years in the previous assessment period, a disability / illness / retirement trend assessment value of 2.10, an environment-driven false positive assessment value of 0.50, and a corresponding anomaly authenticity assessment value of 22.81.

[0121] Reference manual attached Figure 2 The diagram illustrates the anomaly authenticity assessment values ​​and corresponding risk levels for five trend mutation points provided in an embodiment of the present invention.

[0122] Appendix Figure 2 The graph displays the anomaly authenticity assessment values ​​(S) and corresponding risk levels for five trend inflection points. Different colored bars are used to distinguish different risk levels, clearly indicating the degree of anomaly at each trend inflection point. Green bars represent normal fluctuations, yellow bars represent general anomalies, and red bars represent major anomalies. Furthermore, multi-level anomaly thresholds S1 and S2 are plotted as dashed lines in the graph, serving as baselines for anomaly level classification and indicating when different levels of response strategies should be implemented. As shown in the graph, the anomaly authenticity assessment values ​​for trend inflection points 2 and 4 are both below S1, and are therefore classified as normal fluctuations, requiring only archiving and inclusion in baseline updates. The anomaly authenticity assessment values ​​for trend inflection points 1 and 3 are both between S1 and S2, and are classified as general anomalies, requiring lightweight monitoring. The anomaly authenticity assessment value for trend inflection point 5 exceeds S2, and is classified as a major anomaly, triggering a red alert and being pushed to the audit stage for joint review. (Appendix) Figure 2 This intuitively demonstrates the ability of this invention to classify and determine trend change points based on anomaly authenticity assessment values, proving that the system can simultaneously quantify risk magnitude, classify levels, and trigger actions in the same time series, thereby providing targeted intervention priorities for manual review departments.

[0123] In this embodiment of the invention, by performing multi-dimensional fusion analysis on the group activity, individual behavior patterns and organizational structure characteristics within key suspected intervals, an anomaly authenticity assessment model that can eliminate environmental interference and comprehensively quantify the real risk level is constructed. This solves the problem that traditional methods rely on single-dimensional indicators and cannot effectively distinguish between real malicious behavior and occasional environmental fluctuations, and significantly improves the accuracy and reliability of high-risk event identification and characterization.

[0124] S7: Based on authenticity and risk level, implement a tiered response strategy and trigger visual warnings.

[0125] The tiered response strategy refers to a set of management actions and resource allocation schemes that are automatically executed by the system based on the different numerical ranges of the anomaly authenticity assessment value (a quantitative risk score that integrates business anomalies and environmental impacts). These actions are progressive and layered, from light to heavy.

[0126] Among them, visual early warning refers to the process by which the system proactively presents abstract risk data and system decisions to managers in an intuitive and easy-to-understand visual form through a graphical user interface.

[0127] In one possible implementation, S7 specifically refers to:

[0128] When the abnormality authenticity assessment value is less than or equal to the first abnormality threshold, the current trend mutation point is determined to be a normal fluctuation. The corresponding disability, illness and retirement trend assessment value, environment-driven false positive assessment value and abnormality authenticity assessment value data are automatically archived and associated for storage, and updated to the disability, illness and retirement quantitative data baseline library.

[0129] The baseline database for disability, illness, and degeneration assessment refers to a dynamically updated database or data model used to store historical normal behavioral patterns and data distributions of individuals or groups. It stores not only raw data but also statistical characteristics learned from historical data.

[0130] When the abnormality authenticity assessment value is greater than the first abnormality threshold but less than the second abnormality threshold, the current trend change point is determined to be a general abnormality, and a lightweight monitoring mechanism is activated. Specifically, the lightweight monitoring mechanism tracks the changing trends of subsequent disability, illness, and retirement trend assessment values ​​and subsequent environment-driven false positive assessment values. If the abnormality authenticity assessment value remains greater than the first abnormality threshold within a consecutive preset assessment period, it automatically escalates to a key monitoring status, and a yellow risk marker is generated on the risk warning interface to alert the manual review department.

[0131] When the anomaly authenticity assessment value is greater than or equal to the second anomaly threshold, the current trend abrupt change point is determined to be a major anomaly, and the anomaly review process is initiated. Specifically, the anomaly review process involves: triggering a red risk alarm, generating a risk visualization report containing the anomaly authenticity assessment value, risk level, and recommended measures, and pushing the risk visualization report to the review execution end via API interface in the form of an alarm payload, triggering immediate review.

[0132] API refers to Application Programming Interface. It is a communication channel and protocol between two independent software systems for exchanging data in a predetermined format and making function calls.

[0133] It should be noted that the multi-level anomaly threshold includes a first anomaly threshold and a second anomaly threshold.

[0134] It should be noted that those skilled in the art can set the size of the first abnormal threshold and the second abnormal threshold according to actual needs, and the present invention does not limit them.

[0135] Specifically, the system compares the anomaly authenticity assessment value S with the multi-level anomaly thresholds S1 and S2 in real time to determine the anomaly level and execute a graded response strategy. The anomaly authenticity assessment value S quantifies the actual anomaly risk intensity of a single trend inflection point, while the multi-level anomaly thresholds S1 and S2 are grading boundaries determined based on the distribution of historical disability and deterioration quantitative data, used to classify trend inflection points into normal fluctuations, general anomalies, and major anomalies.

[0136] Further, when S ≤ S1, it is determined that the current trend mutation point is a normal fluctuation, and the corresponding disability assessment trend evaluation value, environmental driving false positive evaluation value, and abnormal authenticity evaluation value data are automatically archived for associated storage and updated to the disability assessment quantitative data baseline library. The disability assessment quantitative data baseline library refers to a data benchmark set constructed by archiving, accumulating, updating the disability assessment quantitative data, external environment quantitative data, and corresponding evaluation indicators formed under historical normal fluctuation states, and using them as a reference for subsequent judgments. This archiving behavior forms a comparison background for subsequent judgments by accumulating normal fluctuation samples into the disability assessment quantitative data baseline library, reducing meaningless repeated alarms.

[0137] Further, when S1 < S < S2, it is determined as a general anomaly, and a lightweight monitoring mechanism is started: track the change trends of the subsequent disability assessment trend evaluation value and environmental driving false positive evaluation value, and monitor whether the risk corresponding to this trend mutation point continues to amplify. If S is still greater than S1 for K consecutive evaluation cycles, it is automatically upgraded to the key monitoring state, so that this trend mutation point is no longer treated as a single anomaly, but is regarded as a continuously active risk source. At the same time, a yellow risk mark is generated on the risk warning interface to prompt the manual review department to pay attention. The yellow risk mark visually displays the current risk level and corresponding data source, facilitating the manual review department to prioritize the investigation of key objects without intervening in all trend mutation points. Here, K is a positive integer greater than one.

[0138] Further, when S ≥ S2, it is determined as a major anomaly, and an anomaly review process is started: trigger a red risk alarm, which is used to indicate that the risk level of the current trend mutation point has reached the highest disposal level, and generate a risk visualization report containing the abnormal authenticity evaluation value, risk level, and recommended measures. The abnormal authenticity evaluation value is used to quantify the actual abnormal intensity of this trend mutation point, the risk level is used to indicate the disposal level that needs to be executed, and the recommended measures are used to indicate the subsequent manual verification direction. And the risk visualization report is pushed to the review execution end in the form of an alarm payload through the API interface. The API interface push is used to synchronize the key quantitative results of this trend mutation point in terms of time position, group abnormal intensity item, behavior abnormal intensity item, structure change item, and disability assessment trend evaluation value to the review execution end, triggering an immediate review, avoiding major anomalies from lagging into the manual verification link and ensuring that this trend mutation point is prioritized for investigation after being marked as a major anomaly.

[0139] In this embodiment of the invention, based on the results of a detailed quantitative assessment of the authenticity of anomalies, a multi-level intelligent response is achieved, from baseline learning and continuous tracking to immediate handling. Combined with a visual interface and system interface, a closed-loop flow of risk information is realized, which solves the problems of rigidity, delayed response and lack of adaptability in traditional risk handling processes. This constructs an intelligent risk management closed loop that integrates automated monitoring, hierarchical early warning and collaborative handling.

[0140] In practical applications, a high-quality standardized dataset is constructed by performing time-series alignment, noise cleaning, and scale normalization on multi-source data for disability, illness, and retirement assessments. Based on this, the system integrates multi-dimensional features such as business trends, environmental disturbances, and group behavior. Through a series of algorithmic models including periodic assessment, false positive filtering, confidence clustering, and authenticity quantification, it achieves a layer-by-layer refinement from raw data to risk assessment. Based on the quantified risk level, a tiered response strategy is implemented, from baseline learning and continuous tracking to immediate review, and a closed-loop early warning system is completed through a visual interface and system interface. This invention not only significantly improves the early identification capability and early warning accuracy of hidden and systemic risks, but also achieves a fundamental transformation in disability, illness, and retirement risk management from passive response to intelligent prediction, and from single-point handling to systematic prevention and control through full-process automation and adaptive optimization.

[0141] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0142] In this invention, by simultaneously extracting and quantifying rate indicators reflecting the long-term functional degradation trend and response intensity of short-term sudden anomalies, and integrating the synergistic change characteristics among multiple signals for comprehensive risk assessment, continuous and accurate perception of the nonlinear evolution and phased fluctuations of functional state is achieved. This overcomes the limitations of traditional methods in recognizing complex functional evolution patterns, significantly improves the sensitivity and specificity of disability risk identification, and enables early and accurate warning of potential degradation and sudden events.

[0143] Reference manual attached Figure 3 The diagram shows a structural schematic of a data trend visualization and analysis system for disability assessment, illness, and retirement provided by the present invention.

[0144] This invention also provides a data trend visualization and analysis system 20 for disability assessment and retirement due to illness, applied to the above-mentioned data trend visualization and analysis method for disability assessment and retirement due to illness, comprising:

[0145] Processor 201.

[0146] The memory 202 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 201, they implement the method for visual analysis of disability assessment, disability and retirement data trends as described in the method embodiment.

[0147] The disability assessment and retirement data trend visualization analysis system 20 provided by the present invention can execute the above-mentioned disability assessment and retirement data trend visualization analysis method and achieve the same or similar technical effects. To avoid duplication, the present invention will not elaborate further.

[0148] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0149] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0150] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0151] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0152] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0153] It should be understood that, in various embodiments of the present invention, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0154] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0155] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0156] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0157] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0158] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0159] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0160] This invention provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method for visualizing and analyzing trends in disability assessment, illness, and retirement data as described in the method embodiments.

[0161] The present invention provides a computer-readable storage medium that can realize the steps and effects of the data trend visualization analysis method for disability assessment, illness and retirement described in the above method embodiments. To avoid repetition, the present invention will not repeat the details.

[0162] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0163] The following points need to be explained:

[0164] (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0165] (2) For clarity, the thickness of layers or regions is enlarged or reduced in the drawings used to describe embodiments of the invention, i.e., these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element or there may be intermediate elements.

[0166] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0167] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for visualizing and analyzing trends in disability, illness, and retirement data, characterized in that, include: S1: Real-time collection of quantitative data on disability, illness, and deterioration assessments, as well as quantitative data on the external environment; S2: Perform time alignment, anomaly removal, noise reduction and smoothing, missing data compensation and numerical standardization on the quantitative data of disability assessment and the quantitative data of external environment to obtain standardized quantitative data of disability assessment and standardized quantitative data of external environment. S3: Based on the standardized quantitative data of disability assessment and disability and the standardized quantitative data of external environment, the overall fluctuation of the disability assessment and disability trend is periodically evaluated, the trend change points are marked, and abnormal change intervals are clustered. S4: Based on the assessment results of disability, illness and retirement trends, quantify the environmental disturbance driving effect of each trend mutation point within the abnormal change range, and identify and eliminate environmental false positive anomalies. S5: Assess the confidence level of the retained trend abrupt change points to identify key suspected intervals; S6: Based on the key suspected intervals, the authenticity and risk level of abnormal fluctuations are quantified by combining group, behavioral and structural characteristics; S7: Based on the stated authenticity and risk level, implement a tiered response strategy and trigger a visual warning.

2. The method for visualizing and analyzing trends in disability, illness, and retirement data according to claim 1, characterized in that, Specifically, S2 is: The time series alignment algorithm is used to perform time alignment and resampling on the quantitative data of disability assessment and the quantitative data of external environment. The local anomaly factor algorithm is used to identify and remove anomalies caused by data entry errors and abnormal submissions in the quantitative data of disability assessment and the quantitative data of external environment. The Holt-Winters exponential smoothing algorithm is used to denoise the quantitative data on disability assessment and the quantitative data on the external environment, thereby smoothing out short-term fluctuations in the data. The missing data caused by data loss and data collection interruption in the quantitative data of disability assessment and the quantitative data of external environment are repaired by spline interpolation algorithm. The minimum-maximum standardization algorithm is used to normalize the numerical values ​​of the disability assessment data and the external environment data of different dimensions, thereby eliminating scale differences.

3. The method for visualizing and analyzing trends in disability, illness, and retirement data according to claim 1, characterized in that, The method of periodically assessing the overall fluctuation of the disability, illness, and retirement trend based on the standardized quantitative data of disability assessment and the standardized quantitative data of the external environment specifically includes: S301: Based on the standardized quantitative data of disability and medical retirement assessment, extract the number of medical retirement applications, the number of disability assessment applications, the number of people who passed the assessment, and the approval time to construct a disability and medical retirement review dataset; S302: Set a fixed time window as an evaluation cycle, calculate the average value and standard deviation of the disability assessment and retirement review data and the external environment quantitative data within each evaluation cycle, and extract the minimum and maximum values ​​of the corresponding data; S303: Subtract the absolute value of each disability assessment and retirement review data item from the corresponding average value, and then divide it by the corresponding standard deviation to obtain the standardized deviation value of each disability assessment and retirement review data item. S304: Take the arithmetic mean of the standardized deviation values ​​of the four items to obtain the disease-related decline trend fluctuation term; S305: Subtract the corresponding average value from each of the external environment quantitative data items and take the absolute value. Then take the exponential function value with the natural constant as the base, add one, and take the natural logarithm. Divide the resulting logarithm value by the difference between the maximum and minimum values ​​of the corresponding external environment quantitative data items and add one to obtain the normalized fluctuation value of each of the external environment quantitative data items. S306: Take the arithmetic mean of the normalized fluctuation values ​​of the four items to obtain the external environment correction term; S307: Add the disability and retirement trend fluctuation term to the external environment correction term to obtain the disability and retirement trend assessment value, and obtain the disability and retirement trend assessment result.

4. The method for visualizing and analyzing trends in disability assessment, illness, and retirement data according to claim 3, is characterized in that, The specific steps for marking trend abrupt change points and clustering them into abnormal change intervals are as follows: The disability, illness and retirement trend assessment values ​​within a continuous preset assessment period are arranged in chronological order to construct a trend sequence. The difference sequence of the disability, illness and retirement trend assessment values ​​in adjacent assessment periods is calculated, and the difference sequence is processed by moving average to obtain the rate of change of disability, illness and retirement trend. The rate of change of the disability, illness and retirement trend and the trend change threshold are compared in real time. When the rate of change of the disability, illness and retirement trend is greater than the trend change threshold, the corresponding assessment period is marked as a trend change point. The density-based mutation point aggregation algorithm is used to cluster adjacent trend mutation points. When the time interval between any two trend mutation points is less than the clustering distance threshold and the number of cluster points exceeds the minimum number of clusters, the clustering interval is determined to be an abnormal change interval.

5. The method for visualizing and analyzing trends in disability, illness, and retirement data according to claim 1, characterized in that, The quantification of the environmental disturbance driving effect at each trend abrupt change point within the abnormal change range based on the disability, illness, and retirement trend assessment results specifically includes: S401: Calculate the disability, illness and retirement trend assessment values ​​for all assessment periods within the abnormal change interval and the average and standard deviation of the external environment quantitative data; S402: For each trend abrupt change point, take the absolute value of the difference between the disability, illness and retirement trend assessment value and the average value of the disability, illness and retirement trend assessment value within the abnormal change interval, and then divide it by the standard deviation of the disability, illness and retirement trend assessment value within the abnormal change interval. The resulting ratio is taken as the hyperbolic tangent function value and one is added to obtain the trend correction term. S403: Take the absolute value of the difference between the daily average temperature, daily average sunshine duration, daily precipitation and daily average PM2.5 concentration at the current trend change point and the corresponding average value within the abnormal change interval, divide them by the corresponding standard deviation plus one, and sum them up after taking the exponential function value with the natural constant as the base to obtain the environmental fluctuation response term; S404: Divide the environmental fluctuation response term by four times the trend correction term, add one to the resulting ratio, and take the natural logarithm to obtain the environmental-driven false positive assessment value.

6. The method for visualizing and analyzing trends in disability, illness, and retirement data according to claim 1, characterized in that, The specific steps for identifying and eliminating environmental false positives are as follows: The system compares the environment-driven false positive assessment value with the false positive determination threshold in real time. When the environment-driven false positive assessment value is greater than the false positive determination threshold, the abnormality of the trend change point is determined to be a false positive fluctuation, automatically marked as an environmental anomaly, and removed from the abnormal change range. When the environment-driven false positive assessment value is less than or equal to the false positive determination threshold, the corresponding trend change point is retained.

7. The method for visualizing and analyzing trends in disability, illness, and retirement data according to claim 1, characterized in that, S5 specifically includes: S501: Construct an abnormal residual sequence in chronological order of the trend mutation points retained after removing the environmental false positive anomalies; S502: Based on the abnormal residual sequence, calculate the difference between the time interval between adjacent trend change points and the assessment value of disability and retirement trend. Take the absolute value of the difference, take the exponential function value with the natural constant as the base, and then multiply it by the reciprocal of the time interval to obtain the abnormal confidence value. S503: When the anomaly confidence value is greater than the first confidence threshold, the adjacent trend change points are aggregated into a hidden anomaly cluster; when the hidden anomaly clusters with values ​​greater than a preset value appear consecutively and the duration reaches the minimum duration threshold, the corresponding time range is marked as the key suspected interval; wherein, the key suspected interval is removed from the key suspected interval marking only when all anomaly confidence values ​​are lower than the second confidence threshold in subsequent consecutive preset evaluation periods.

8. The method for visualizing and analyzing trends in disability, illness, and retirement data according to claim 1, characterized in that, S6 specifically includes: S601: For the trend abrupt change points within the key suspected interval, divide the number of review cases at the current trend abrupt change point by the total number of on-duty personnel plus one, add one to the resulting ratio, and take the natural logarithm to obtain the group anomaly intensity item; S602: Divide the number of sick leave requests per month by the total overtime hours per month and add one, add one to the ratio, and take the natural logarithm to obtain the behavioral abnormality intensity item; S603: Add the group anomaly intensity term to the behavior anomaly intensity term and divide by the square root of the environmental-driven false positive assessment value plus one to obtain the comprehensive anomaly intensity term. S604: Calculate the difference between the current trend inflection point and the average length of service in the previous assessment period, divide it by the average length of service in the previous assessment period plus one, and obtain the absolute value of the ratio to get the structural change term. S605: Multiply the comprehensive abnormality intensity item, the structural change item, the disability and retirement trend assessment value by one and the abnormality amplification coefficient in sequence to obtain the abnormality authenticity assessment value.

9. The method for visualizing and analyzing trends in disability, illness, and retirement data according to claim 8, characterized in that, Specifically, S7 is: When the abnormal authenticity assessment value is less than or equal to the first abnormal threshold, the current trend mutation point is determined to be a normal fluctuation. The corresponding disability assessment trend assessment value, environment-driven false positive assessment value and abnormal authenticity assessment value data are automatically archived and associated for storage, and updated to the disability assessment quantitative data baseline library. When the abnormal authenticity assessment value is greater than the first abnormal threshold and less than the second abnormal threshold, the current trend change point is determined to be a general abnormality, and a lightweight monitoring mechanism is activated. Specifically, the lightweight monitoring mechanism tracks the changing trends of subsequent disability, illness, and retirement trend assessment values ​​and subsequent environment-driven false positive assessment values. If the abnormal authenticity assessment value is still greater than the first abnormal threshold within a consecutive preset assessment period, it is automatically upgraded to a key monitoring status, and a yellow risk mark is generated on the risk warning interface to prompt the manual review department to pay attention. When the anomaly authenticity assessment value is greater than or equal to the second anomaly threshold, the current trend mutation point is determined to be a major anomaly, and the anomaly review process is initiated. Specifically, the anomaly review process involves: triggering a red risk alarm, generating a risk visualization report containing the anomaly authenticity assessment value, risk level, and recommended measures, and pushing the risk visualization report to the review execution end via an API interface in the form of an alarm load to trigger immediate review.

10. A data trend visualization and analysis system for disability assessment, characterized in that, include: processor; A memory storing computer-readable instructions, which, when executed by the processor, implement the method for visual analysis of disability, illness, and retirement data trends as described in any one of claims 1 to 9.

Images