Integrated processing method and platform based on GIS

Abstract

This invention discloses an integrated processing method and platform based on GIS, comprising: collecting multi-source spatial data related to epidemics through a GIS geographic information system, and storing the multi-source spatial data after unifying spatial coordinates and standardizing format; acquiring historical public health data and extracting risk characteristic parameters from historical event data; constructing a multi-dimensional early warning model through multi-source spatial data and risk characteristic parameters, and using real-time data from multiple monitoring points as input to the multi-dimensional early warning model for fusion analysis; generating early warning information when a potential public health risk event meets preset threshold conditions, and displaying the spatial distribution and temporal changes of the early warning information through a GIS visualization interface. The beneficial effect of this invention is that by using a multi-scale spatiotemporal coupling anomaly detection mechanism, multi-level spatial radii and corresponding time windows are systematically combined, achieving full-coverage anomaly identification from macro-regions to micro-localities.

Description

Technical Field

[0001] This invention relates to the field of public health emergency management technology, specifically to an integrated processing method and platform based on GIS. Background Technology

[0002] With the increasing frequency of global public health emergencies, establishing efficient early warning and emergency command systems has become an urgent need for governments worldwide. Most existing public health early warning systems are based on Geographic Information System (GIS) technology, integrating multi-source data for risk identification and early warning dissemination. These systems typically employ historical data analysis, real-time monitoring, and risk assessment, playing a crucial role in epidemic prevention and control and response to public health emergencies. However, traditional early warning systems still have shortcomings in key areas such as data fusion, anomaly detection, and risk assessment, affecting the accuracy and timeliness of early warnings.

[0003] The main shortcomings of existing technologies are as follows: First, spatiotemporal anomaly detection mostly employs single-scale analysis, making it difficult to simultaneously capture large-scale trend changes and localized sudden anomalies, easily leading to false alarms or missed alarms; second, anomaly verification mechanisms are relatively simple, lacking multi-level cross-validation, resulting in insufficient reliability of early warnings; third, early warning thresholds are mostly fixed settings, unable to be dynamically adjusted according to different regional characteristics and time patterns, resulting in poor adaptability. These technological limitations restrict further improvement in the effectiveness of public health early warning systems. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is: how to provide an integrated processing method and platform based on GIS to improve the accuracy, reliability and adaptability of the public health early warning system, reduce false alarms and missed alarms, and enhance emergency decision support capabilities.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an integrated processing method based on GIS, comprising,

[0007] Collect multi-source spatial data related to the epidemic through a GIS geographic information system, and then store the multi-source spatial data related to the epidemic after unifying spatial coordinates and standardizing the format.

[0008] Obtain historical public health data, store the historical public health data in a structured manner, and extract time distribution characteristics, spatial diffusion characteristics, and influencing factor characteristics from the historical event data as risk characteristic parameters;

[0009] By using the multi-source spatial data and the risk characteristic parameters, a multi-dimensional early warning model is constructed. Real-time data from multiple monitoring points are input into the multi-dimensional early warning model for fusion analysis to identify potential public health risk events.

[0010] When a potential public health risk event meets a preset threshold condition, an early warning message is generated, and the spatial distribution and temporal changes of the early warning message are displayed through a GIS visualization interface.

[0011] As a preferred embodiment of the GIS-based integrated processing method described in this invention, the step of extracting the risk feature parameters includes:

[0012] The historical public health data is segmented according to time series, and the frequency and intensity distribution of public health events in different time periods are statistically analyzed to extract time distribution feature parameters.

[0013] Based on the geographic coordinate information of historical events, the starting point, propagation path and scope of influence of the events are analyzed, and spatial diffusion characteristic parameters are extracted through spatial clustering algorithms;

[0014] Identify environmental, demographic, transportation, and socioeconomic factors associated with historical public health events, and extract characteristic parameters of these factors through correlation analysis.

[0015] As a preferred embodiment of the GIS-based integrated processing method described in this invention, the step of fusion analysis of the multi-dimensional early warning model includes:

[0016] Perform spatiotemporal coupling anomaly detection on real-time data from each monitoring point and output the spatiotemporal coupling anomaly area;

[0017] Perform neighbor-to-neighbor cross-validation on the spatiotemporal coupling anomaly region and output the reliable anomaly region that has passed the validation.

[0018] Adjust the early warning judgment threshold according to the credible anomaly region;

[0019] Based on the early warning judgment threshold, the risk level of the credible abnormal area is assessed, and potential public health risk events are output.

[0020] The beneficial effects of this preferred technical solution are as follows: by adopting a four-step fusion analysis process of spatiotemporal coupling anomaly detection, neighbor point cross-validation, threshold dynamic adjustment and risk level assessment, a complete early warning judgment chain is formed, which can effectively reduce false alarms and missed alarms and improve the reliability and accuracy of early warning.

[0021] As a preferred embodiment of the GIS-based integrated processing method described in this invention, the spatiotemporal coupling verification mechanism includes:

[0022] A multi-scale detection system is constructed by setting multi-level spatial radii R1, R2, R3 and corresponding time windows T1, T2, T3;

[0023] Anomaly intensity is calculated at each scale level. When the rate of change of data at a monitoring point within a time window T exceeds α times the historical average for the same period, it is marked as a time anomaly.

[0024] When the number of time anomalies within different spatial radii reaches the preset numbers X1, X2, and X3 respectively, the multi-scale coupling strength coefficient C is calculated.

[0025] Adjust the coupling strength determination threshold βt based on the multi-scale anomaly pattern characteristics of the current time period;

[0026] When the coupling strength coefficient C exceeds the dynamic threshold βt, the spatial radius R range is marked as a spatiotemporal coupling anomalous region.

[0027] The beneficial effects of this preferred technical solution are as follows: by constructing a multi-scale detection system through multi-level spatial radii and time windows, it is possible to simultaneously capture large-scale trend changes and local sudden anomalies. The multi-scale coupling strength coefficient and threshold adjustment mechanism improve the accuracy and adaptability of anomaly detection.

[0028] As a preferred embodiment of the GIS-based integrated processing method described in this invention, the step of performing neighbor-to-neighbor cross-validation on the spatiotemporally coupled abnormal region includes:

[0029] For the spatiotemporal coupling anomaly region, construct the network topology of the monitoring points in the region and identify the boundary monitoring points located at the boundary of the spatiotemporal coupling anomaly region;

[0030] Calculate the data correlation coefficients between the boundary monitoring point and its adjacent monitoring points at different distance levels within the normal area, including the close distance correlation coefficient R1, the medium distance correlation coefficient R2, and the long distance correlation coefficient R3, and obtain the correlation coefficients of the corresponding levels in the same historical period as the evaluation benchmark value.

[0031] When the change pattern of the multi-level correlation coefficient of at least Y monitoring points in the boundary monitoring points is consistent with the data change trend of the monitoring points in the spatiotemporal coupling anomaly region, the spatiotemporal coupling anomaly region is output as a reliable anomaly region.

[0032] The beneficial effects of this preferred technical solution are as follows: by adopting a multi-level correlation coefficient verification mechanism, through hierarchical cross-validation at near, medium and far distances, it is possible to identify real abnormal areas and eliminate data noise interference, thereby improving the credibility of early warning information.

[0033] As a preferred embodiment of the GIS-based integrated processing method described in this invention, the risk level assessment includes:

[0034] A comprehensive risk score is calculated for the credible anomaly region, and the comprehensive risk score is calculated based on the coupling strength coefficient, the scope of influence, and the development trend.

[0035] When the comprehensive risk score exceeds the early warning threshold, the credible abnormal area is identified as a potential public health risk event.

[0036] As a preferred embodiment of the GIS-based integrated processing method described in this invention, the step of adjusting the early warning judgment threshold according to the credible anomaly area includes:

[0037] Based on the distribution characteristics of the coupling strength coefficient of the credible anomaly region, calculate the benchmark value of the anomaly strength for the current time period;

[0038] Based on the seasonal variation patterns in the aforementioned time distribution characteristics, the benchmark value of the anomaly intensity is time-corrected.

[0039] By utilizing the regional propagation characteristics in the spatial diffusion features, the warning judgment threshold is spatially differentiated and adjusted so that different regions can use corresponding warning judgment thresholds.

[0040] The beneficial effects of this preferred technical solution are: by combining time correction and spatial differentiation adjustment to achieve dynamic optimization of the early warning threshold, it can adapt to the characteristics of different regions and time periods, and improve the early warning system's adaptability to complex environments.

[0041] As a preferred embodiment of the GIS-based integrated processing method described in this invention, the early warning information generation adopts a multi-level information fusion mechanism:

[0042] Basic early warning elements are generated based on the spatiotemporal coupling anomaly detection results;

[0043] Reliability assessment elements are generated based on the results of the adjacent point cross-validation.

[0044] The risk level assessment results are communicated to generate response strategy elements;

[0045] The basic early warning elements, reliability assessment elements, and response strategy elements are fused and encoded to generate early warning information containing complete decision support information.

[0046] This invention provides an integrated processing platform based on GIS.

[0047] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an integrated processing platform based on GIS, comprising:

[0048] The data acquisition module is used to collect multi-source spatial data related to the epidemic through a GIS geographic information system, and to perform spatial coordinate unification and format standardization processing.

[0049] The historical data analysis module is used to acquire and structure public health historical data and extract risk characteristic parameters;

[0050] The early warning model construction module is used to build multi-dimensional early warning models based on multi-source spatial data and risk characteristic parameters.

[0051] The real-time monitoring module is used to receive real-time data from multiple monitoring points and input it into the early warning model for fusion analysis;

[0052] The anomaly detection module is used to perform spatiotemporal coupling anomaly detection and neighbor-point cross-validation.

[0053] The risk assessment module is used to assess risk levels and adjust early warning thresholds.

[0054] The early warning information generation module is used to generate early warning information containing complete decision support information;

[0055] The GIS visualization module is used to display the spatial distribution and temporal changes of early warning information in the GIS interface.

[0056] As a preferred embodiment of the GIS-based integrated processing platform described in this invention, the GIS visualization module includes a multi-level map display interface;

[0057] The early warning information generation module has a multi-channel information push function, which selects the push method and push target according to the early warning level;

[0058] The real-time monitoring module supports access from multiple data sources, including medical institution data, environmental monitoring data, personnel flow data, and social media data.

[0059] The beneficial effects of this invention are:

[0060] By employing a multi-scale spatiotemporal coupling anomaly detection mechanism, multi-level spatial radii and corresponding time windows are systematically combined to achieve full-coverage anomaly identification from macroscopic regions to microscopic localities. The calculation of the multi-scale coupling strength coefficient not only integrates anomaly point distribution information at each scale level, but more importantly, it achieves adaptive optimization of detection sensitivity through a threshold adjustment mechanism. This multi-scale collaborative detection overcomes the limitations of traditional single-scale analysis, accurately identifying local sudden anomalies while capturing large-scale trend changes, fundamentally solving the technical challenge of both false positives and false negatives in existing technologies.

[0061] The multi-level cross-validation mechanism identifies boundary monitoring points by constructing a network topology and calculates correlation coefficients for near, medium, and long distances, forming a three-dimensional validation system. When the changing patterns of the multi-level correlation coefficients are consistent with the changing trends of monitoring point data within the abnormal area, it can accurately distinguish between real abnormal signals and data noise interference. This hierarchical validation mechanism, combined with spatiotemporal coupled detection, produces a synergistic enhancement effect, enabling the early warning system to maintain high sensitivity while improving reliability, effectively avoiding the validation blind spot problem of traditional validation methods.

[0062] The deep integration of risk feature parameter extraction and multi-dimensional early warning models achieves an organic unity between historical patterns and real-time monitoring. Seasonal correction is performed using temporal distribution characteristics, and regional differentiation is achieved by combining spatial diffusion characteristics, enabling the early warning threshold to be accurately adapted to different spatiotemporal conditions. This dynamic adjustment mechanism based on historical features, together with real-time anomaly detection, forms a positive feedback loop, continuously optimizing the accuracy of the early warning system and achieving a sustained improvement in early warning precision. Attached Figure Description

[0063] To more clearly illustrate the technical solutions of 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.

[0064] Figure 1 The present invention provides an overall flowchart of a GIS-based integrated processing method according to an embodiment of the present invention. Detailed Implementation

[0065] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0066] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides an integrated processing method based on GIS, including the following steps S1 to S4:

[0067] S1. Collect multi-source spatial data related to the epidemic through a GIS geographic information system, and store the multi-source spatial data related to the epidemic after unifying spatial coordinates and standardizing the format.

[0068] In this embodiment, the multi-source spatial data related to the epidemic includes data on the distribution of medical institutions, population density distribution, transportation network data, environmental monitoring point data, and administrative division boundary data. This spatial data is collected from different data sources, such as the Health Commission, the Statistics Bureau, the Transportation Department, and the Environmental Protection Department, through a GIS data interface.

[0069] Because the various data sources use different coordinate systems, such as the Beijing 54 coordinate system, the Xi'an 80 coordinate system, and the WGS84 coordinate system, all spatial data was first uniformly converted to the National 2000 coordinate system to ensure spatial coordinate consistency. Format standardization processing included converting vector data of different formats to Shapefile format, raster data to GeoTIFF format, and storing attribute data using UTF-8 encoding. The processed data was stored in a spatial database, and a spatial index was created to improve query efficiency.

[0070] S2. Obtain historical public health data, store the historical public health data in a structured manner, and extract time distribution characteristics, spatial diffusion characteristics, and influencing factor characteristics from the historical event data as risk characteristic parameters.

[0071] In this embodiment, the public health historical data covers records of public health events such as influenza outbreaks, hand-foot-mouth disease outbreaks, and food poisoning incidents from the past decade. Each record includes basic information such as the time of occurrence, geographical location, number of people affected, and duration. The system stores this historical data in a structured manner according to a standardized data model, establishing a relational database table structure with fields such as event number, timestamp, spatial coordinates, event type, and severity.

[0072] The steps for extracting the risk characteristic parameters include:

[0073] S2.1. The public health historical data is segmented according to time series, and the frequency and intensity distribution of public health events in different time periods are statistically analyzed to extract time distribution feature parameters.

[0074] Historical data was segmented by month to analyze the frequency and average number of people affected by public health events each month. Analysis revealed that influenza outbreaks primarily occurred from November to March of the following year, exhibiting a peak frequency distribution characteristic of winter and spring; hand-foot-and-mouth disease outbreaks mainly occurred from April to September, showing a spring-summer high incidence pattern. The system extracted temporal distribution characteristic parameters, including seasonality coefficients, monthly probability distributions, and intensity trend coefficients, providing a temporal reference benchmark for subsequent early warning models.

[0075] S2.2 Based on the geographic coordinate information of historical events, analyze the starting point, propagation path and scope of influence of the events, and extract spatial diffusion characteristic parameters through spatial clustering algorithms.

[0076] The DBSCAN spatial clustering algorithm was used to cluster the geographic coordinates of historical events, identifying high-incidence areas and transmission hotspots of public health events. By analyzing the spread of a particular influenza outbreak, the system identified the origin of the event as a transportation hub in the city center, with the transmission path spreading along major transportation arteries to surrounding districts and counties. The affected area expanded from a 5-kilometer radius at the origin to a 25-kilometer radius within 14 days. Extracted spatial diffusion characteristic parameters included the origin density coefficient, transmission speed coefficient, and diffusion range growth rate; these parameters reflect the spatial transmission patterns of different types of public health events.

[0077] S2.3 Identify environmental, demographic, transportation, and socioeconomic factors related to historical public health events, and extract characteristic parameters of these factors through correlation analysis.

[0078] Correlation analysis identified influencing factors closely related to the occurrence of public health events. Environmental factors included temperature, humidity, and air quality index, with temperature showing a negative correlation coefficient of -0.72 with influenza incidence. Population factors included population density and age structure ratio, with population density showing a positive correlation with the speed of event transmission. Transportation factors included road density and passenger station density, with areas surrounding transportation hubs having a higher risk of events. Socioeconomic factors included per capita income level and density of medical resources. The system extracted characteristic parameters of the influencing factors, including the weight coefficients and threshold ranges of each factor, providing a quantitative basis for the construction of a multi-dimensional early warning model.

[0079] S3. Construct a multi-dimensional early warning model using the multi-source spatial data and the risk characteristic parameters. Input real-time data from multiple monitoring points into the multi-dimensional early warning model for fusion analysis to identify potential public health risk events.

[0080] The steps for fusion analysis of multi-dimensional early warning models include S3.1~S3.4:

[0081] S3.1 Detect spatiotemporal coupling anomalies in the real-time data of each monitoring point and output the spatiotemporal coupling anomaly area;

[0082] In this embodiment, the system establishes a city-wide monitoring network, including 120 medical institution monitoring points, 85 pharmacy monitoring points and 50 school monitoring points, to collect key indicator data in real time, such as the number of people visiting fever clinics, the sales volume of antipyretic analgesics, and the student absenteeism rate.

[0083] The spatiotemporal coupling verification mechanism includes A1~A5:

[0084] A1. Set multi-level spatial radii R1, R2, R3 and corresponding time windows T1, T2, T3 to construct a multi-scale detection system;

[0085] Based on the characteristics of urban spatial structure, three levels of spatial radii are set: R1=2 km corresponding to the community scale, R2=5 km corresponding to the street scale, and R3=10 km corresponding to the district / county scale. Correspondingly, time windows are set: T1=3 days for capturing short-term fluctuations, T2=7 days for identifying periodic changes, and T3=14 days for detecting long-term trends. This multi-scale detection system can simultaneously monitor abnormal patterns across different spatial ranges and time spans, ensuring the detection of both localized emergencies and the identification of large-scale diffusion trends.

[0086] A2. Calculate the anomaly intensity at each scale level. When the data change rate of a monitoring point within the time window T exceeds α times the historical average for the same period, it is marked as a time anomaly.

[0087] Taking fever clinic monitoring sites as an example, we calculate the rate of change of data for each monitoring site within different time windows. A certain monitoring site has an average daily number of patients of 45 within a T1=3 day window, while the historical average for the same period is 30, with a standard deviation of 5. We set α=2.0, and calculate the rate of change... If the value exceeds α × standard deviation, the monitoring point is marked as a time outlier. Similar calculations are performed at three different scales to generate outlier identification results for different time spans.

[0088] A3. When the number of time anomalies within different spatial radii reaches the preset number X1, X2, and X3 respectively, calculate the multi-scale coupling strength coefficient C.

[0089] Preset threshold numbers X1=3, X2=5, and X3=8 are set, corresponding to the minimum number of anomalies required within different spatial radii. When a region has 4 anomalies within R1=2 km, 7 anomalies within R2=5 km, and 12 anomalies within R3=10 km, the quantity requirements for each level are met. At this point, the multi-scale coupling strength coefficient C is calculated. This coefficient comprehensively considers the spatial density, temporal concentration, and data variation amplitude of anomalies at each scale level. The specific calculation formula is C = (w1×D1 + w2×D2 + w3×D3) / (w1+w2+w3), where D1, D2, and D3 are the anomaly density indices for each scale, and w1, w2, and w3 are the corresponding weights.

[0090] A4. Adjust the coupling strength determination threshold βt based on the multi-scale anomaly pattern characteristics of the current time period;

[0091] Analysis indicates that the current period coincides with the peak season for winter influenza. Historical data shows that the abnormal pattern during this period is characterized by high frequency and moderate intensity of sporadic outbreaks from multiple locations. Based on this characteristic, the base threshold β0=0.6 was adjusted to βt=0.55, lowering the decision threshold to improve detection sensitivity. Simultaneously, considering the influence of environmental factors during the current period, such as lower-than-usual temperatures and poorer air quality, the threshold was further fine-tuned to βt=0.52, achieving dynamic threshold optimization based on real-time environmental conditions.

[0092] A5. When the coupling strength coefficient C exceeds the dynamic threshold βt, the spatial radius R range is marked as a spatiotemporal coupling anomaly region.

[0093] The calculated multi-scale coupling strength coefficient C=0.58 exceeds the adjusted threshold βt=0.52. Therefore, an area centered on this region with a radius of R2=5 kilometers is designated as a spatiotemporal coupling anomaly region. This anomaly region encompasses 15 monitoring points, including 3 community health service centers, 8 pharmacies, and 4 schools, providing a clear validation scope for subsequent neighbor-point cross-validation. Simultaneously, the identification time, spatial boundary coordinates, and relevant monitoring point information of the anomaly region are recorded, forming a complete spatiotemporal coupling anomaly detection result.

[0094] S3.2 Perform neighbor-to-neighbor cross-validation on the spatiotemporal coupling anomaly region and output the validated reliable anomaly region;

[0095] In this embodiment, the spatiotemporal coupling anomaly region identified in step S3.1 is subjected to neighbor-point cross-validation. This anomaly region contains 15 monitoring points and is a circular area with a radius of 5 kilometers centered on a community health service center.

[0096] The steps for performing neighbor-to-neighbor cross-validation on the spatiotemporal coupling anomaly region include B1 to B3:

[0097] B1. Construct a network topology for monitoring points within the spatiotemporal coupling anomaly region and identify boundary monitoring points located at the boundary of the spatiotemporal coupling anomaly region;

[0098] A Delaunay triangulation topology is constructed based on the geographic coordinates of the monitoring points. Boundary monitoring points are identified by calculating the shortest distance between each monitoring point and the boundary of the anomaly area. In this embodiment, the boundary of the anomaly area is a circular boundary with a radius of 5 kilometers. Six monitoring points within 500 meters of this boundary are identified as boundary monitoring points, including two community health service stations, three pharmacies, and one primary school. These boundary monitoring points are located in the transition zone between the anomaly area and the normal area.

[0099] B2. Calculate the data correlation coefficients between the boundary monitoring point and its adjacent monitoring points at different distance levels within the normal area, including the close distance correlation coefficient R1, the medium distance correlation coefficient R2, and the long distance correlation coefficient R3, and obtain the correlation coefficients of the corresponding levels in the same historical period as the evaluation benchmark value.

[0100] Taking one of the boundary monitoring points (community health service station A) as an example, its close neighbors (two pharmacies within 1 km), medium neighbors (one community health center within 2-3 km), and distant neighbors (one hospital within 4-5 km) were identified within the normal area. By calculating the Pearson correlation coefficients between this boundary monitoring point and its neighbors at each level over the past 7 days of fever clinic data, the correlation coefficients were found to be R1=0.75 for close neighbors, R2=0.68 for medium neighbors, and R3=0.45 for distant neighbors. Simultaneously, the system extracted the correlation coefficients for the corresponding levels from the same historical period (same month and week last year) as benchmark values: close benchmark = 0.85, medium benchmark = 0.78, and distant benchmark = 0.52. Comparison revealed that the current correlation coefficients at each level are lower than the historical benchmark values, indicating an abnormal decrease in the data correlation between the boundary monitoring point and the monitoring points in the normal area.

[0101] B3. When the multi-level correlation coefficient change pattern of at least Y monitoring points in the boundary monitoring points is consistent with the data change trend of the monitoring points in the spatiotemporal coupling abnormal area, the spatiotemporal coupling abnormal area will be output as a reliable abnormal area.

[0102] Setting Y=4 means that at least four boundary monitoring points are required to meet the verification conditions. Analysis revealed that the multi-level correlation coefficients of five out of the six boundary monitoring points showed a decreasing trend, and this decreasing pattern formed a clear inverse relationship with the generally increasing trend observed in monitoring points within the abnormal area.

[0103] Specifically, the fever clinic data at monitoring points within the abnormal area increased by an average of 45% compared to the same period in history, while the correlation coefficient between boundary monitoring points and adjacent points in the normal area decreased by an average of 18%. This "inner increase and outer decrease" pattern is consistent with the transmission characteristics of real public health events. Since the number of boundary monitoring points meeting the verification conditions (5) exceeded the preset threshold Y=4, this spatiotemporally coupled abnormal area was output as a credible abnormal area.

[0104] S3.3 Adjust the early warning judgment threshold according to the credible anomaly region;

[0105] The steps for adjusting the early warning threshold include C1 to C3:

[0106] C1. Calculate the baseline value of the anomaly intensity for the current time period based on the distribution characteristics of the coupling strength coefficient of the credible anomaly region;

[0107] Analysis of the coupling strength coefficient distribution of 15 monitoring points within the credible anomaly region revealed a pattern of high coefficients at the center and low coefficients at the edges. The average coupling strength coefficient for the four monitoring points in the central region was 0.72, for the seven monitoring points in the middle region it was 0.58, and for the four monitoring points in the edge region it was 0.45. A weighted average was used to calculate the overall coupling strength coefficient for the region as 0.58. Combining this with historical anomaly intensity distribution data for the current period (December), a baseline value of 0.52 was calculated for the current anomaly intensity. This baseline value reflects the degree of anomaly in the current anomaly region relative to the historical average for the same period.

[0108] C2. Based on the seasonal variation patterns in the aforementioned time distribution characteristics, perform time correction on the benchmark value of the anomaly intensity;

[0109] Based on the temporal distribution characteristic parameters extracted in step S2.1, December is a peak season for winter influenza. Historical data shows that the frequency of public health events during this period is 35% higher than the annual average, and the intensity of these events is generally higher. A seasonal correction coefficient of 1.2 was used to adjust the baseline value of the abnormal intensity, resulting in a corrected baseline value of 0.52 × 1.2 = 0.624. Considering that the current week is a traditional peak week for influenza outbreaks, a periodic fine-tuning coefficient of 1.05 was further applied, ultimately yielding a time-corrected baseline value of 0.624 × 1.05 = 0.655.

[0110] C3. Based on the regional propagation characteristics in the spatial diffusion features, the warning judgment threshold is spatially differentiated and adjusted so that different regions can use corresponding warning judgment thresholds.

[0111] Based on the spatial diffusion feature parameters extracted in step S2.2, the current credible anomaly area is identified as being located in a densely populated urban transportation hub area. The historical propagation velocity coefficient for this area is 1.8, higher than the 0.6 coefficient for suburban areas. Considering the high propagation risk characteristics of transportation hub areas, a stricter warning threshold is applied to this area. The time-corrected baseline value of 0.655 is multiplied by a spatial risk adjustment coefficient of 0.85, resulting in a warning threshold of 0.655 × 0.85 = 0.557 for this area. For surrounding residential areas, a standard adjustment coefficient of 1.0 is used, maintaining the threshold at 0.655; for more distant suburban areas, a more lenient adjustment coefficient of 1.15 is used, adjusting the threshold to 0.655 × 1.15 = 0.753.

[0112] S3.4. Assess the risk level of the credible abnormal area based on the early warning judgment threshold and output the potential public health risk event;

[0113] The risk level assessment includes D1 to D2:

[0114] D1. Calculate a comprehensive risk score for the credible anomaly region. The comprehensive risk score is calculated based on the coupling strength coefficient, the scope of influence, and the development trend.

[0115] The comprehensive risk score of the credible anomaly area is calculated, which takes into account three dimensions: coupling strength coefficient 0.58 (weight 0.4), influence range index 0.75 (calculated based on 15 monitoring points and a 5-kilometer coverage radius, weight 0.3), and development trend index 0.82 (calculated based on the rate of increase of data over the past 7 days, weight 0.3).

[0116] The comprehensive risk score is calculated as follows: Score = 0.58 × 0.4 + 0.75 × 0.3 + 0.82 × 0.3 = 0.232 + 0.225 + 0.246 = 0.703. This score reflects the overall risk level of the current abnormal area; a higher score indicates a greater degree of risk.

[0117] D2. When the comprehensive risk score exceeds the early warning judgment threshold, the credible abnormal area is identified as a potential public health risk event.

[0118] Comparing the calculated comprehensive risk score of 0.703 with the warning threshold of 0.557 for the region, it was found that the score significantly exceeded the threshold, with an excess of (0.703-0.557) / 0.557×100%=26.2%.

[0119] Because the comprehensive risk score exceeded the warning threshold, the area with the credible anomaly was identified as a potential public health risk event, and the event level was set as "medium risk". At the same time, a detailed risk event report was generated, including the time and location of the event, information on the monitoring points involved, a detailed risk score, and recommended response measures, providing comprehensive information support for emergency command and decision-making.

[0120] S4. When the potential public health risk event meets the preset threshold conditions, an early warning information is generated, and the spatial distribution and temporal changes of the early warning information are displayed through a GIS visualization interface.

[0121] In this embodiment, the comprehensive risk score of the potential public health risk event identified in step S3.4 is 0.703, which exceeds the preset threshold of 0.557, thus meeting the conditions for generating early warning information. A multi-level information fusion mechanism is then activated to generate complete early warning information and display it visually.

[0122] The early warning information is generated using a multi-level information fusion mechanism:

[0123] Basic early warning elements are generated based on the spatiotemporal coupling anomaly detection results;

[0124] Basic early warning elements are generated based on the spatiotemporal coupling anomaly detection results. These elements include the spatial boundary coordinates of the anomaly area (center point: 118.25°E, 32.18°N, radius 5 km), detection time (14:30, December 15, 2024), a list of monitoring points involved (detailed information for 15 monitoring points), multi-scale coupling strength coefficient (0.58), and anomaly duration (72 consecutive hours). The basic early warning elements also include anomaly pattern descriptions: a 45% increase in the number of patients visiting fever clinics compared to the same period in previous years, a 38% increase in sales of antipyretic analgesics, and a 25% increase in student absenteeism, exhibiting typical characteristics of the early stages of a respiratory infectious disease outbreak.

[0125] Reliability assessment elements are generated based on the results of the adjacent point cross-validation.

[0126] Reliability assessment elements were generated based on the results of neighboring point cross-validation, with a validation success rate of 83% (5 out of 6 boundary monitoring points passed validation), and a high level of confidence. The reliability assessment elements detailed the changes in correlation coefficients across multiple levels: short-range correlation decreased by 12%, medium-range correlation decreased by 13%, and long-range correlation decreased by 13%, with the change pattern showing an inverse correlation to the upward trend of data within the anomalous area. Therefore, the reliability assessment concludes that the probability of this anomalous area representing a genuine public health risk event is 89%, ruling out the possibility of data anomalies or systemic false alarms.

[0127] The risk level assessment results are communicated to generate response strategy elements;

[0128] Based on the assessment results of the medium-risk level, corresponding response strategy elements are generated.

[0129] Immediate response includes: activating the Level II emergency response plan, notifying the district CDC to conduct on-site epidemiological investigations, and increasing the monitoring frequency of medical institutions in the abnormal area to twice a day.

[0130] Short-term measures include: setting up three temporary fever detection points in the abnormal area, strengthening the school morning check system, and issuing health tips to the public.

[0131] Long-term monitoring includes: expanding the monitoring range to a radius of 8 kilometers, continuously tracking for 14 days, and establishing an information linkage mechanism with surrounding areas.

[0132] The response strategy also includes a resource needs assessment: 6 epidemiological investigators, 3 sets of testing equipment, and 500 doses of emergency medicine are needed.

[0133] The basic early warning elements, reliability assessment elements, and response strategy elements are fused and encoded to generate early warning information containing complete decision support information.

[0134] The three types of elements are fused and coded to generate structured early warning information in XML format. The early warning code is "ALERT-2024121514-MR-089", which includes a timestamp, risk level identifier, and reliability code. The fused early warning information includes the warning level (medium risk - yellow alert), spatial range (circular area with a radius of 5 kilometers), time validity period (72 hours), affected population (approximately 85,000 people), event type (suspected outbreak of respiratory infectious disease), confidence level (89%), and a complete list of response recommendations.

[0135] Example 2 is an embodiment of the present invention, which provides an integrated processing platform based on GIS, including:

[0136] The data acquisition module is used to collect multi-source spatial data related to the epidemic through a GIS geographic information system, and to perform spatial coordinate unification and format standardization processing.

[0137] The historical data analysis module is used to acquire and structure public health historical data and extract risk characteristic parameters;

[0138] The early warning model construction module is used to build multi-dimensional early warning models based on multi-source spatial data and risk characteristic parameters.

[0139] The real-time monitoring module is used to receive real-time data from multiple monitoring points and input it into the early warning model for fusion analysis;

[0140] The anomaly detection module is used to perform spatiotemporal coupling anomaly detection and neighbor-point cross-validation.

[0141] The risk assessment module is used to assess risk levels and adjust early warning thresholds.

[0142] The early warning information generation module is used to generate early warning information containing complete decision support information;

[0143] The GIS visualization module is used to display the spatial distribution and temporal changes of early warning information in the GIS interface.

[0144] The GIS visualization module includes a multi-level map display interface;

[0145] The early warning information generation module has a multi-channel information push function, which selects the push method and push targets according to the early warning level;

[0146] The real-time monitoring module supports access from multiple data sources, including data from medical institutions, environmental monitoring data, population movement data, and social media data.

[0147] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A GIS-based integrated processing method, characterized in that: include, Collect multi-source spatial data related to the epidemic through a GIS geographic information system, and store the multi-source spatial data related to the epidemic after unifying spatial coordinates and standardizing format. Obtain historical public health data, store the historical public health data in a structured manner, and extract time distribution characteristics, spatial diffusion characteristics, and influencing factor characteristics from the historical event data as risk characteristic parameters; By using the multi-source spatial data and the risk characteristic parameters, a multi-dimensional early warning model is constructed. Real-time data from multiple monitoring points are input into the multi-dimensional early warning model for fusion analysis to identify potential public health risk events. When the potential public health risk event meets the preset threshold conditions, an early warning information is generated, and the spatial distribution and temporal changes of the early warning information are displayed through a GIS visualization interface. The steps for fusion analysis of the multi-dimensional early warning model include: Perform spatiotemporal coupling anomaly detection on real-time data from each monitoring point and output the spatiotemporal coupling anomaly area; Perform neighbor-to-neighbor cross-validation on the spatiotemporal coupling anomaly region and output the reliable anomaly region that has passed the validation. Adjust the early warning judgment threshold according to the credible anomaly region; The risk level of the credible abnormal area is assessed based on the early warning judgment threshold, and potential public health risk events are output; the spatiotemporal coupling verification mechanism includes: A multi-scale detection system is constructed by setting multi-level spatial radii R1, R2, R3 and corresponding time windows T1, T2, T3; Anomaly intensity is calculated at each scale level. When the rate of change of data at a monitoring point within a time window T exceeds α times the historical average for the same period, it is marked as a time anomaly. When the number of time anomalies within different spatial radii reaches the preset numbers X1, X2, and X3 respectively, the multi-scale coupling strength coefficient C is calculated. Adjust the coupling strength determination threshold βt based on the multi-scale anomaly pattern characteristics of the current time period; When the coupling strength coefficient C exceeds the dynamic threshold βt, the spatial radius R range is marked as a spatiotemporal coupling anomalous region.

2. The integrated processing method based on GIS as described in claim 1, characterized in that: The steps for extracting the risk characteristic parameters include: The historical public health data is segmented according to time series, and the frequency and intensity distribution of public health events in different time periods are statistically analyzed to extract time distribution feature parameters; Based on the geographic coordinate information of historical events, the starting point, propagation path and scope of influence of the events are analyzed, and spatial diffusion characteristic parameters are extracted through spatial clustering algorithms; Identify environmental, demographic, transportation, and socioeconomic factors associated with historical public health events, and extract characteristic parameters of these factors through correlation analysis.

3. The integrated processing method based on GIS as described in claim 2, characterized in that: The steps for performing neighbor-to-neighbor cross-validation on the spatiotemporal coupling anomaly region include: For the spatiotemporal coupling anomaly region, construct the network topology of the monitoring points in the region and identify the boundary monitoring points located at the boundary of the spatiotemporal coupling anomaly region; Calculate the data correlation coefficients between the boundary monitoring point and its adjacent monitoring points at different distance levels within the normal area, including the close distance correlation coefficient R1, the medium distance correlation coefficient R2, and the long distance correlation coefficient R3, and obtain the correlation coefficients of the corresponding levels in the same historical period as the evaluation benchmark value. When the change pattern of the multi-level correlation coefficient of at least Y monitoring points in the boundary monitoring points is consistent with the data change trend of the monitoring points in the spatiotemporal coupling anomaly region, the spatiotemporal coupling anomaly region is output as a reliable anomaly region.

4. The integrated processing method based on GIS as described in claim 3, characterized in that: The risk level assessment includes: A comprehensive risk score is calculated for the credible anomaly region, and the comprehensive risk score is calculated based on the coupling strength coefficient, the scope of influence, and the development trend. When the comprehensive risk score exceeds the early warning threshold, the credible abnormal area is identified as a potential public health risk event.

5. The integrated processing method based on GIS as described in claim 4, characterized in that: The steps for adjusting the early warning determination threshold based on the credible anomaly region include: Based on the distribution characteristics of the coupling strength coefficient of the credible anomaly region, calculate the benchmark value of the anomaly strength for the current time period; Based on the seasonal variation patterns in the aforementioned time distribution characteristics, the benchmark value of the anomaly intensity is time-corrected. By utilizing the regional propagation characteristics in the spatial diffusion features, the warning judgment threshold is spatially differentiated and adjusted so that different regions can use corresponding warning judgment thresholds.

6. The integrated processing method based on GIS as described in claim 5, characterized in that: The early warning information is generated using a multi-level information fusion mechanism: Basic early warning elements are generated based on the spatiotemporal coupling anomaly detection results; Reliability assessment elements are generated based on the results of the adjacent point cross-validation. The risk level assessment results are communicated to generate response strategy elements; The basic early warning elements, reliability assessment elements, and response strategy elements are fused and encoded to generate early warning information containing complete decision support information.

7. A GIS-based integrated processing platform, employing the GIS-based integrated processing method as described in any one of claims 1 to 6, characterized in that: include: The data acquisition module is used to collect multi-source spatial data related to the epidemic through a GIS geographic information system, and to perform spatial coordinate unification and format standardization processing. The historical data analysis module is used to acquire and structure public health historical data and extract risk characteristic parameters; The early warning model construction module is used to build multi-dimensional early warning models based on multi-source spatial data and risk characteristic parameters. The real-time monitoring module is used to receive real-time data from multiple monitoring points and input it into the early warning model for fusion analysis; The anomaly detection module is used to perform spatiotemporal coupling anomaly detection and neighbor-point cross-validation. The risk assessment module is used to assess risk levels and adjust early warning thresholds. The early warning information generation module is used to generate early warning information containing complete decision support information; The GIS visualization module is used to display the spatial distribution and temporal changes of early warning information in the GIS interface.

8. The integrated processing platform based on GIS as described in claim 7, characterized in that: The GIS visualization module includes a multi-level map display interface; The early warning information generation module has a multi-channel information push function, which selects the push method and push target according to the early warning level; The real-time monitoring module supports access from multiple data sources, including medical institution data, environmental monitoring data, personnel flow data, and social media data.

Images