A public health epidemic prevention and control system based on big data analysis

Abstract

The present invention discloses a public health epidemic prevention and control system based on big data analysis, which relates to the technical field of public health epidemic prevention and control. By integrating the data sources of hospitals, social media and meteorological departments, the system can collect and fuse epidemic-related information in real time, reducing the problem of data lag in traditional monitoring systems. The integration of social media data and weather and environmental data makes early warning of epidemics more accurate, thereby being able to predict the risk of potential epidemic outbreaks in advance, greatly shortening the response time. Through the data acquisition and preprocessing module and the data fusion and feature extraction module, this system can extract useful features from raw data from different sources and perform intelligent analysis. The risk assessment and hotspot identification module generates a comprehensive risk index RI based on the predicted results of epidemic spread, combined with multiple factors such as the environmental pollution index EPI and population mobility data Rk, to provide decision support for government departments.

Description

Technical Field

[0001] The present invention relates to the technical field of public health epidemic prevention and control, and specifically to a public health epidemic prevention and control system based on big data analysis. Background Art

[0002] With the acceleration of globalization and increased population mobility, the risk of infectious disease transmission is increasing, making timely and effective epidemic prevention and control crucial. Against this backdrop, the application of big data technologies offers new solutions for epidemic prevention and control. By integrating diverse data sources, such as hospitals, social media, weather, and environmental monitoring, an epidemic monitoring platform can be effectively constructed. This platform leverages big data analytics to track and predict epidemic spread in real time.

[0003] Currently, traditional epidemic prevention and control systems typically rely on single-source data, primarily relying on hospital case reports, public health agency investigation reports, and some government-issued epidemic bulletins. The biggest problem with this model is the timeliness and comprehensiveness of the data. Case information reported by hospitals is often delayed, while epidemic-related information on social media and the internet is often overlooked, even though this data can provide early warning signals.

[0004] With the widespread use of social media, smartphones, and other IoT devices, the public is updated with real-time updates on epidemic-related dynamics, often revealing the spread of an epidemic earlier than traditional epidemic reporting systems. However, when responding to an epidemic, the lag and incompleteness of data often lead to misguided decisions or delayed responses, missing the best opportunities for prevention and control. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a public health epidemic prevention and control system based on big data analysis, which solves the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a public health epidemic prevention and control system based on big data analysis, including a data acquisition and preprocessing module, a data fusion and feature extraction module, an epidemic spread prediction module, a risk assessment and hotspot identification module, a decision support and emergency response module, and a dynamic feedback and optimization module;

[0007] The data collection and preprocessing module is responsible for collecting raw data from hospitals, social media, and meteorological departments, and preprocessing it to obtain the epidemic prevention data group FYW;

[0008] The feature extraction module extracts features from the epidemic prevention data set FYW, including the emotional fluctuation features related to the epidemic in social media and the correlation features between air pollution and respiratory diseases, to form a health feature set F;

[0009] The epidemic spread prediction module uses the susceptible-infected-recovered model to predict the trend of epidemic spread through the epidemic prevention data set FYW and the health feature set F, analyzes the impact of population mobility data Rk, air quality data AQ and sentiment analysis data SQ on the spread of the epidemic, and obtains the prediction results;

[0010] The risk assessment and hotspot identification module identifies the current epidemic risk areas based on the prediction results, combined with the epidemic density Iden, population mobility data Rk and environmental pollution index EPI, and evaluates them to obtain the comprehensive risk index RI;

[0011] The decision support and emergency response module provides decision support to public health agencies based on the comprehensive risk index RI and formulates prevention and control measures;

[0012] During the epidemic prevention and control process, the dynamic feedback and optimization module monitors the effectiveness of prevention and control measures in real time and adjusts strategies through data feedback.

[0013] Preferably, the data acquisition and preprocessing module includes a data acquisition unit and a data cleaning and standardization unit;

[0014] The collection unit collects raw data from multiple data sources such as hospitals, social media, and meteorological departments, including electronic health records EHR, population mobility data Rk, sentiment analysis data SQ, temperature data TE, humidity data HU, and air quality data AQ;

[0015] Among them, the electronic health record EHR is obtained through the hospital information management system, including the incidence rate of respiratory diseases DiseaseRate; the population flow data Rk is obtained by the ratio of the number of population flows in the region to the total area of ​​the region;

[0016] Sentiment analysis data SQ includes capturing posts related to the epidemic from social media platforms, performing sentiment analysis, and using NLP technology to classify each post into positive, negative, and neutral emotions. The positive, negative, and neutral emotions are then assigned numerical values: positive emotions are +1, negative emotions are -1, and neutral emotions are 0.

[0017] Sentiment analysis data SQ is obtained using the following formula:

[0018] ;

[0019] Where Si represents the i-th social media post, and n represents the total number of posts analyzed;

[0020] The temperature data TE and humidity data HU are obtained through the temperature sensor and humidity sensor respectively;

[0021] Air quality data AQ is obtained through the following formula:

[0022] ;

[0023] Wherein, cPM2.5, cCO and cNO2 represent the concentrations of air particulate matter PM2.5, carbon monoxide CO and nitrogen dioxide NO2, respectively; IPM2.5, ICO and INO2 represent the pollutant normalization indexes of air particulate matter concentrations PM2.5, carbon monoxide CO and nitrogen dioxide NO2, respectively; max represents the peak value;

[0024] The data cleaning and standardization unit cleans and standardizes the original data collected by the data collection unit to obtain the epidemic prevention data group FYW;

[0025] Cleaning involves removing noise and invalid information from the raw data, including missing values ​​and outliers, by using filters;

[0026] Standardization is the process of standardizing the original data into a uniform dimension.

[0027] Preferably, the feature extraction module includes a social media emotion fluctuation feature extraction unit and an air pollution and respiratory disease correlation feature extraction unit;

[0028] The social media emotion fluctuation feature extraction unit extracts emotion fluctuation features based on the sentiment analysis data SQ, analyzes the changing trend of epidemic-related public opinion, and calculates and obtains the emotion fluctuation index EVI and the emotion change trend ET;

[0029] The emotional volatility index EVI is obtained by the following formula:

[0030] ;

[0031] Where N represents the total time period, SQti represents the value of the sentiment analysis data at time point ti, and PSQ represents the average value of the sentiment analysis data;

[0032] The emotion change trend ET is obtained by the following formula:

[0033] ;

[0034] Where SQ(qi) represents the value of the sentiment analysis data at time qi, and SQ(pi) represents the value of the sentiment analysis data at time pi.

[0035] Preferably, the respiratory disease association feature extraction unit calculates and obtains the pollution and health association index PHI based on the association between the air quality data AQ and the electronic health record EHR, specifically by combining the air quality data AQ and the respiratory disease incidence rate DiseaseRate;

[0036] The pollution and health index PHI is obtained by the following formula:

[0037] ;

[0038] Where, ln represents the natural logarithm function;

[0039] The obtained emotion fluctuation index EVI, emotion change trend ET and pollution and health association index PHI are combined to obtain the health feature set F.

[0040] Preferably, the epidemic spread prediction module includes an SIR model construction and parameter estimation unit and an influencing factor analysis and spread trend adjustment unit;

[0041] The SIR model construction and parameter estimation unit constructs a mathematical model of epidemic spread based on the susceptible-infected-recovered model according to the epidemic prevention data set FYW and the health feature set F, analyzes the changing trends of the number of susceptible people S, the number of infected people I, and the number of recovered people R, predicts the trend of future epidemic spread, and obtains the number of susceptible people S(t), the number of infected people I(t), and the number of recovered people R(t) at the future time point t;

[0042] By analyzing population mobility data Rk, air quality data AQ, and sentiment analysis data SQ, we estimate the parameters of the epidemic spread mathematical model, including the infection rate β and recovery rate γ;

[0043] Specifically, population mobility data Rk reflects the flow of people between different regions, affecting the speed and scope of virus transmission; areas with high population mobility lead to an increase in the infection rate β, because more people come into contact with each other, increasing the possibility of transmission; by analyzing the frequency, pattern and amount of population mobility, the infection potential of the area can be inferred, thereby affecting the estimation of the β value.

[0044] Fine particulate matter and air pollutants affect people's respiratory health; when the air quality is poor, people's immune systems are easily affected, thereby exacerbating the spread of disease; areas with higher pollutant concentrations will lead to an increase in the infection rate β, so β ​​can be inferred by analyzing the correlation between the air quality index and the infection rate.

[0045] Sentiment analysis data SQ in social media can reveal people's emotional reactions to the epidemic. When people panic about the epidemic, it leads to more frequent social contacts and unnecessary gatherings, thereby increasing the chance of virus transmission. By analyzing the relationship between emotional fluctuations and the infection rate, the β value can be further adjusted.

[0046] The speed and frequency of population mobility affect the speed of recovery from the epidemic; in places with less population mobility, prevention and control measures are easier to implement, which may lead to a faster recovery; while in areas with frequent population mobility, infected people may not be easy to isolate in a timely manner, and the recovery speed may slow down; therefore, population mobility conditions can help estimate the recovery rate γ.

[0047] The impact of air quality on the health of the population will affect the recovery process. Poor air quality makes patients with chronic diseases more susceptible and prolongs recovery time. Areas with severe air pollution may affect the recovery rate γ, causing it to decrease.

[0048] Sentiment analysis data SQ reflects people's cognition and reaction to the epidemic. When emotions tend to be negative, people will ignore prevention and control measures and delay treatment, which will affect the speed of recovery. Therefore, emotional fluctuations can also have a certain impact on the recovery rate γ.

[0049] The number of susceptible people S(t) at the time point t is obtained by the following formula:

[0050] ;

[0051] Where, Represents the influence coefficient of temperature data TE and humidity data HU on the spread of the epidemic, represents the influence coefficient of the j-th feature in the health feature set F on the spread of the epidemic, Fj represents the j-th feature in the health feature set F, and m represents the total number of features in the health feature set F;

[0052] The number of infected people I(t) at the time point t is obtained by the following formula:

[0053] ;

[0054] The number of recovered people R(t) at the time point t is obtained by the following formula:

[0055] .

[0056] Preferably, the influencing factor analysis and transmission trend adjustment unit analyzes the impact of population mobility data Rk, air quality data AQ, and sentiment analysis data SQ on the spread of the epidemic, and brings them into the epidemic spread mathematical model to adjust the epidemic spread trend, obtain the adjusted infection rate βadj and the adjusted recovery rate γadj, bring them into the epidemic spread mathematical model, and re-predict the number of susceptible people S(t), the number of infected people I(t), and the number of recovered people R(t) at the future time point t;

[0057] The adjusted infection rate βadj is obtained by the following formula:

[0058] ;

[0059] Where, They represent the influence coefficients of population mobility data Rk, air quality data AQ and sentiment analysis data SQ on the infection rate β respectively;

[0060] The adjusted recovery rate γadj is obtained by the following formula:

[0061] ;

[0062] Where, They represent the influence coefficients of population flow data Rk, air quality data AQ and sentiment analysis data SQ on the recovery rate γ respectively.

[0063] Preferably, the risk assessment and hotspot identification module includes a risk assessment unit and a hotspot identification and dynamic update unit;

[0064] The risk assessment unit calculates the comprehensive risk index RI based on the number of infected people I(t) at the future time point t provided by the epidemic spread prediction module, combined with the epidemic density Iden, population mobility data Rk and environmental pollution index EPI, and compares it with the preset risk threshold TRI to assess the epidemic risk of the region;

[0065] Estimate the severity of the epidemic in a region by calculating the epidemic density Iden of the infected population;

[0066] The epidemic density Iden is obtained by the following formula:

[0067] ;

[0068] Where, mA represents the area of ​​the region;

[0069] Calculate the environmental pollution index EPI through air quality data AQ;

[0070] The environmental pollution index EPI is obtained by the following formula:

[0071] ;

[0072] Where M represents the total amount of pollutants, represents the weight coefficient of the iath pollutant, and AQia represents the air quality data of the iath pollutant;

[0073] The comprehensive risk index RI is obtained by the following formula:

[0074] ;

[0075] Where, They represent the preset weight values ​​of epidemic density Iden, population mobility data Rk and environmental pollution index EPI, respectively, and ;

[0076] The epidemic risk of the area is obtained by matching in the following ways:

[0077] When the comprehensive risk index RI ≤ the risk threshold TRI, it means that the epidemic risk in the region is normal;

[0078] When the comprehensive risk index RI is greater than the risk threshold TRI, it means that the epidemic risk in the region is abnormal.

[0079] Preferably, the hotspot identification and dynamic update unit identifies epidemic hotspots based on the comprehensive risk index RI of the region, and updates the comprehensive risk index RI of each region in real time, and dynamically adjusts the identification of epidemic hotspots;

[0080] The obtained comprehensive risk index RI of each region is sorted, and the region with comprehensive risk index RI> risk threshold TRI is marked as hot spot K, indicating that the risk of the region is abnormal;

[0081] Fit all hotspot areas K to form a hotspot set HArea={Area(K)|RI(K)>TRI};

[0082] Where RI(K) represents the comprehensive risk index of the Kth hotspot area, and Area(K) represents the Kth hotspot area, specifically indicating the area with abnormal risk.

[0083] The comprehensive risk index RI of the hotspot area K is updated at fixed intervals to track the changing trend of the epidemic; ensuring the timely identification of hotspot areas and the timely response of prevention and control measures.

[0084] Preferably, the decision support and emergency response module includes a decision support unit and an emergency response unit;

[0085] The decision support unit divides the regional epidemic risk into the first and second levels according to the preset risk threshold TRI, and provides prevention and control measures for each hotspot area K;

[0086] When the comprehensive risk index RI ≤ the risk threshold TRI, it is marked as the first level, indicating that the area does not belong to the hotspot area K, and the epidemic risk in the area is normal. Normal prevention and control measures are maintained, and routine monitoring and reporting are carried out;

[0087] When the comprehensive risk index RI is greater than the risk threshold TRI, it is marked as the second level, indicating that the epidemic risk in the hotspot area K is abnormal, and prevention and control measures are provided, including implementing blockade measures, isolating infected people, and adjusting the allocation of medical resources;

[0088] The emergency response unit formulates an emergency response plan based on the prevention and control measures provided by the decision support unit, including resource deployment, personnel scheduling and emergency measures;

[0089] Among them, resource allocation includes allocation of hospital beds, ambulances, testing equipment and personal protective equipment;

[0090] Personnel dispatch includes assigning relevant personnel to areas with abnormal epidemic risks to conduct epidemic monitoring and prevention and control;

[0091] Emergency measures include setting up temporary epidemic prevention and control points, strengthening the dissemination of epidemic information and health education for residents.

[0092] Preferably, the dynamic feedback and optimization module includes a prevention and control effect monitoring unit and a strategy adjustment unit;

[0093] The prevention and control effect monitoring unit obtains real-time epidemic data and the implementation of prevention and control measures, including the number of infected people I, the number of recovered people R, the number of deaths and the implementation intensity of prevention and control measures, including the degree of implementation of blockade, testing and isolation, and measures the prevention and control effect through the infection rate β and recovery rate γ to determine whether the prevention and control measures are effective; if the infection rate does not decrease significantly or the cure rate does not increase significantly, it means that there may be problems with the prevention and control measures and they need to be adjusted.

[0094] The strategy adjustment unit calculates and identifies whether the current prevention and control measures have achieved the expected results based on the feedback provided by the prevention and control effect monitoring unit, obtains the epidemic control index YZ, and compares it with the preset epidemic control threshold TYZ to determine the effectiveness of the prevention and control measures;

[0095] The epidemic control index YZ is obtained by using a weighted algorithm on the infection rate β and the recovery rate γ;

[0096] The effects of the control measures are obtained by matching the following methods:

[0097] When the epidemic control index YZ ≥ the epidemic control threshold TYZ, it means that the prevention and control measures are effective;

[0098] When the epidemic control index YZ is less than the epidemic control threshold TYZ, it means that the prevention and control measures are ineffective, and the prevention and control measures should be readjusted and a new emergency response plan should be formulated.

[0099] The present invention provides a public health epidemic prevention and control system based on big data analysis, which has the following beneficial effects:

[0100] (1) When the system is running, by integrating data sources from hospitals, social media, and meteorological departments, the system can collect and integrate epidemic-related information in real time, reducing the problem of data lag in traditional monitoring systems. The integration of social media data and weather and environmental data makes early warning of the epidemic more accurate, thereby being able to predict the risk of potential epidemic outbreaks in advance and greatly shortening the response time. Through the data collection and preprocessing module and the data fusion and feature extraction module, this system can extract useful features from raw data from different sources and perform intelligent analysis. This approach can help prevention and control agencies fully understand the development dynamics of the epidemic, identify potential factors that cause the epidemic, and provide strong support for the formulation of more accurate and flexible prevention and control measures.

[0101] The risk assessment and hotspot identification module accurately identifies high-risk areas for the epidemic based on the predicted results of the epidemic spread, combined with multiple factors such as the environmental pollution index EPI and population mobility data Rk, and generates a comprehensive risk index RI to provide decision-making support for government departments.

[0102] (2) The system uses the air pollution and respiratory disease association feature extraction unit, combined with the pollution and health index (PHI), to analyze the relationship between air quality and the occurrence of respiratory diseases. Through regression analysis, the system can reveal the correlation between pollutant concentrations and respiratory diseases, thereby providing a more accurate health risk assessment for epidemic prevention and control. Especially in the context of increased population mobility and intensified environmental pollution, the system can predict the potential threat of environmental factors to health and adjust prevention and control strategies in a timely manner to reduce the risk of epidemics caused by intensified pollution.

[0103] The system uses the epidemic spread prediction module and the risk assessment and hotspot identification module, combined with multi-dimensional data such as sentiment analysis, air quality, and population mobility, to accurately predict the epidemic spread trend and identify high-risk areas.

[0104] (3) By constructing the SIR model and combining it with the epidemic prevention data set FYW and the health feature set F, the module can accurately predict the number of susceptible, infected, and recovered people at different time points, providing detailed data on the epidemic spread trend. Compared with traditional empirical prediction methods, this prediction method based on the SIR model can more scientifically reflect the dynamic changes in the spread of the epidemic and avoid the errors caused by overly simple linear models. By predicting the number of susceptible, infected, and recovered people at future time points, the government and public health agencies can adjust prevention and control strategies in a timely manner to ensure the accuracy and timeliness of prevention and control measures.

[0105] (4) Through the calculated comprehensive risk index RI, the module can identify and mark areas with higher epidemic risks in real time. When the RI value exceeds the preset risk threshold TRI, it means that there is a high risk of epidemic spread in the area, and prevention and control measures must be taken as a priority. The risk assessment process combines multi-dimensional factors such as epidemic spread data, population mobility, and environmental pollution to improve the accuracy of risk assessment. The introduction of the environmental pollution index EPI, especially the impact of air quality on the epidemic, provides a quantitative analysis of potential environmental factors for epidemic spread. Through this type of comprehensive analysis, the system can better evaluate areas with heavy pollution and dense population mobility, which are often prone to becoming high-risk areas for epidemic spread. By deeply combining comprehensive environmental and epidemic data, the spread of the epidemic can be more accurately predicted and controlled. BRIEF DESCRIPTION OF THE DRAWINGS

[0106] Figure 1 This is a flowchart of a public health epidemic prevention and control system based on big data analysis in the present invention. DETAILED DESCRIPTION

[0107] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0108] Example 1

[0109] The present invention provides a public health epidemic prevention and control system based on big data analysis, please refer to Figure 1 , including data collection and preprocessing module, data fusion and feature extraction module, epidemic spread prediction module, risk assessment and hotspot identification module, decision support and emergency response module and dynamic feedback and optimization module;

[0110] The data collection and preprocessing module is responsible for collecting raw data from hospitals, social media, and meteorological departments, and preprocessing it to obtain the epidemic prevention data group FYW;

[0111] The feature extraction module extracts features from the epidemic prevention data set FYW, including the emotional fluctuation features related to the epidemic in social media and the correlation features between air pollution and respiratory diseases, to form a health feature set F;

[0112] The epidemic spread prediction module uses the susceptible-infected-recovered model to predict the trend of epidemic spread through the epidemic prevention data set FYW and the health feature set F, analyzes the impact of population mobility data Rk, air quality data AQ and sentiment analysis data SQ on the spread of the epidemic, and obtains the prediction results;

[0113] The risk assessment and hotspot identification module identifies the current epidemic risk areas based on the prediction results, combined with the epidemic density Iden, population mobility data Rk and environmental pollution index EPI, and evaluates them to obtain the comprehensive risk index RI;

[0114] The decision support and emergency response module provides decision support to public health agencies based on the comprehensive risk index RI and formulates prevention and control measures;

[0115] During the epidemic prevention and control process, the dynamic feedback and optimization module monitors the effectiveness of prevention and control measures in real time and adjusts strategies through data feedback.

[0116] In this embodiment, by integrating data sources from hospitals, social media, and meteorological departments, the system can collect and fuse epidemic-related information in real time, reducing the problem of data lag in traditional monitoring systems. The integration of social media data and weather and environmental data makes early warning of epidemics more accurate, thereby being able to predict the risk of potential epidemic outbreaks in advance and greatly shortening the response time. Through the data acquisition and preprocessing module and the data fusion and feature extraction module, this system can extract useful features from raw data from different sources and perform intelligent analysis. This approach can help prevention and control agencies fully understand the development trends of the epidemic, identify potential factors that trigger the epidemic, and provide strong support for the formulation of more accurate and flexible prevention and control measures.

[0117] The risk assessment and hotspot identification module uses epidemic spread predictions, combined with factors such as the Environmental Pollution Index (EPI) and population mobility data (Rk), to accurately identify high-risk areas for the epidemic and generate a comprehensive risk index (RI) to support government decision-making. This refined risk assessment allows for more precise deployment of prevention and control measures, avoiding waste of resources and excessive prevention and control measures. Furthermore, real-time risk assessment and hotspot identification help public health agencies efficiently allocate resources in response to the epidemic, focusing medical and testing resources on the most urgent areas and further improving the effectiveness of the epidemic response.

[0118] Through its dynamic feedback and optimization module, this system monitors the effectiveness of prevention and control measures in real time and adjusts strategies based on real-time data feedback. This mechanism promptly identifies issues with prevention and control measures, prevents lags in prevention and control strategies, and ensures their effectiveness at different stages of the epidemic. Based on different data feedback, prevention and control measures can be flexibly adjusted, improving the accuracy and timeliness of prevention and control responses and reducing the risk of epidemic spread.

[0119] Through its feature extraction module and epidemic spread prediction model, the system dynamically adjusts prevention and control strategies and resource allocation based on different epidemic trends and characteristics. This strategy optimization, powered by big data analysis, enables prevention and control measures to adapt to different epidemic stages and emergencies, improving their adaptability and sustainability, and ensuring long-term effectiveness.

[0120] Example 2

[0121] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the data acquisition and preprocessing module includes a data acquisition unit and a data cleaning and standardization unit;

[0122] The collection unit collects raw data from multiple data sources such as hospitals, social media, and meteorological departments, including electronic health records EHR, population mobility data Rk, sentiment analysis data SQ, temperature data TE, humidity data HU, and air quality data AQ;

[0123] Among them, the electronic health record EHR is obtained through the hospital information management system, including the incidence rate of respiratory diseases DiseaseRate; the population flow data Rk is obtained by the ratio of the number of population flows in the region to the total area of ​​the region;

[0124] Sentiment analysis data SQ includes capturing posts related to the epidemic from social media platforms, performing sentiment analysis, and using NLP technology to classify each post into positive, negative, and neutral emotions. The positive, negative, and neutral emotions are then assigned numerical values: positive emotions are +1, negative emotions are -1, and neutral emotions are 0.

[0125] Sentiment analysis data SQ is obtained using the following formula:

[0126] ;

[0127] Where Si represents the i-th social media post, and n represents the total number of posts analyzed;

[0128] The temperature data TE and humidity data HU are obtained through the temperature sensor and humidity sensor respectively;

[0129] Air quality data AQ is obtained through the following formula:

[0130] ;

[0131] Wherein, cPM2.5, cCO and cNO2 represent the concentrations of air particulate matter PM2.5, carbon monoxide CO and nitrogen dioxide NO2, respectively; IPM2.5, ICO and INO2 represent the pollutant normalization indexes of air particulate matter concentrations PM2.5, carbon monoxide CO and nitrogen dioxide NO2, respectively; max represents the peak value;

[0132] The data cleaning and standardization unit cleans and standardizes the original data collected by the data collection unit to obtain the epidemic prevention data group FYW;

[0133] Cleaning involves removing noise and invalid information from the raw data, including missing values ​​and outliers, by using filters;

[0134] Standardization is the process of standardizing the original data into a uniform dimension.

[0135] The feature extraction module includes a social media emotion fluctuation feature extraction unit and an air pollution and respiratory disease correlation feature extraction unit;

[0136] The social media emotion fluctuation feature extraction unit extracts emotion fluctuation features based on the sentiment analysis data SQ, analyzes the changing trend of epidemic-related public opinion, and calculates and obtains the emotion fluctuation index EVI and the emotion change trend ET;

[0137] The emotional volatility index EVI is obtained by the following formula:

[0138] ;

[0139] Where N represents the total time period, SQti represents the value of the sentiment analysis data at time point ti, and PSQ represents the average value of the sentiment analysis data;

[0140] The emotion change trend ET is obtained by the following formula:

[0141] ;

[0142] Where SQ(qi) represents the value of the sentiment analysis data at time qi, and SQ(pi) represents the value of the sentiment analysis data at time pi.

[0143] The respiratory disease correlation feature extraction unit calculates and obtains the pollution and health correlation index PHI based on the correlation between the air quality data AQ and the electronic health record EHR, specifically by combining the air quality data AQ and the respiratory disease incidence rate DiseaseRate;

[0144] The pollution and health index PHI is obtained by the following formula:

[0145] ;

[0146] Where, ln represents the natural logarithm function;

[0147] The obtained emotion fluctuation index EVI, emotion change trend ET and pollution and health association index PHI are combined to obtain the health feature set F.

[0148] In this embodiment, the system integrates data from multiple sources, including electronic health records (EHRs), population mobility data (Rk), sentiment analysis data (SQ), temperature data (TE), humidity data (HU), and air quality data (AQ). This allows epidemic prevention and control to no longer rely solely on a single data source. By timely collecting sentiment analysis data from social media and real-time meteorological and environmental data, the system can capture abnormal fluctuations and potential risks of the epidemic earlier, overcoming the problems of single data and delayed updates in traditional epidemic prevention and control systems, and improving the timeliness and response capabilities of epidemic prevention and control decisions.

[0149] Thanks to the data cleaning and standardization unit, the system effectively removes noise and invalid information from collected data, handles missing values ​​and outliers, and ensures data accuracy and reliability. Furthermore, standardization unifies data of varying scales, avoiding analytical errors caused by scale differences between data sources and ensuring the accuracy of subsequent analysis and modeling. Compared to traditional data processing methods, this automated and systematic data cleaning and standardization process not only improves data quality but also provides reliable data support for subsequent intelligent analysis and prediction.

[0150] Using the social media sentiment fluctuation feature extraction unit, the system extracts the sentiment fluctuation index (EVI) and sentiment change trend (ET) from sentiment analysis data (SQ), enabling real-time analysis of public sentiment regarding the epidemic. This feature extraction mechanism can deeply mine epidemic-related social media content, promptly capturing trends in public sentiment fluctuations and providing crucial insights for prevention and control decisions. By monitoring sentiment fluctuations in real time, prevention and control agencies can identify public opinion risks early, prevent the spread of panic, and improve public sentiment management and prevention effectiveness.

[0151] The system uses a feature extraction unit for the association between air pollution and respiratory diseases, combined with the Pollution Health Index (PHI), to analyze the relationship between air quality and the occurrence of respiratory diseases. Through regression analysis, the system can reveal the correlation between pollutant concentrations and respiratory diseases, thereby providing more accurate health risk assessments for epidemic prevention and control. Especially in the context of increased population mobility and intensified environmental pollution, the system can predict potential health threats posed by environmental factors and promptly adjust prevention and control strategies to reduce the risk of epidemics caused by increased pollution.

[0152] The system uses its epidemic spread prediction module, risk assessment module, and hotspot identification module, combining sentiment analysis, air quality, population mobility, and other multi-dimensional data to accurately predict epidemic spread trends and identify high-risk areas. By comprehensively analyzing the impact of sentiment fluctuations, environmental pollution, and population mobility, the system can identify potential hotspots for the development of the epidemic, providing decision-making support for governments and public health agencies and enabling timely response measures. Improved prediction and risk assessment enable governments to implement effective prevention and control measures before an epidemic breaks out, avoiding the lags inherent in traditional emergency response.

[0153] Example 3

[0154] This embodiment is explained in Example 2, please refer to Figure 1 ,Specifically: the epidemic spread prediction module includes an SIR model construction and parameter estimation unit and an ,influence factor analysis and spread trend adjustment unit;

[0155] The SIR model construction and parameter estimation unit constructs a mathematical model of epidemic spread based on the susceptible-infected-recovered model according to the epidemic prevention data set FYW and the health feature set F, analyzes the changing trends of the number of susceptible people S, the number of infected people I, and the number of recovered people R, predicts the trend of future epidemic spread, and obtains the number of susceptible people S(t), the number of infected people I(t), and the number of recovered people R(t) at the future time point t;

[0156] By analyzing population mobility data Rk, air quality data AQ, and sentiment analysis data SQ, we estimate the parameters of the epidemic spread mathematical model, including the infection rate β and recovery rate γ;

[0157] The number of susceptible people S(t) at the time point t is obtained by the following formula:

[0158] ;

[0159] Where, Represents the influence coefficient of temperature data TE and humidity data HU on the spread of the epidemic, represents the influence coefficient of the j-th feature in the health feature set F on the spread of the epidemic, Fj represents the j-th feature in the health feature set F, and m represents the total number of features in the health feature set F;

[0160] The number of infected people I(t) at the time point t is obtained by the following formula:

[0161] ;

[0162] The number of recovered people R(t) at the time point t is obtained by the following formula:

[0163] .

[0164] The influencing factor analysis and transmission trend adjustment unit analyzes the impact of population mobility data Rk, air quality data AQ, and sentiment analysis data SQ on the spread of the epidemic, and brings them into the epidemic spread mathematical model to adjust the epidemic spread trend, obtain the adjusted infection rate βadj and the adjusted recovery rate γadj, bring them into the epidemic spread mathematical model, and re-predict the number of susceptible people S(t), the number of infected people I(t), and the number of recovered people R(t) at the future time point t;

[0165] The adjusted infection rate βadj is obtained by the following formula:

[0166] ;

[0167] Where, They represent the influence coefficients of population mobility data Rk, air quality data AQ and sentiment analysis data SQ on the infection rate β respectively;

[0168] The adjusted recovery rate γadj is obtained by the following formula:

[0169] ;

[0170] Where, They represent the influence coefficients of population flow data Rk, air quality data AQ and sentiment analysis data SQ on the recovery rate γ respectively.

[0171] In this example, by constructing a SIR model and combining it with the epidemic prevention data set FYW and the health feature set F, the module can accurately predict the number of susceptible, infected, and recovered people at different time points, providing detailed data on the spread of the epidemic. Compared with traditional empirical prediction methods, this SIR model-based prediction method can more scientifically reflect the dynamic changes in the spread of the epidemic and avoid the errors caused by overly simple linear models. By predicting the number of susceptible, infected, and recovered people at future time points, governments and public health agencies can promptly adjust prevention and control strategies to ensure the accuracy and timeliness of prevention and control measures.

[0172] The influencing factor analysis and transmission trend adjustment unit combines multi-dimensional data such as population mobility data Rk, air quality data AQ, and sentiment analysis data SQ with SIR model parameters, including the transmission rate β and recovery rate γ, to adjust the epidemic transmission trend. By dynamically adjusting the transmission rate β and recovery rate γ, the system can optimize the epidemic transmission trend in real time based on actual conditions, improving forecasting flexibility. This is particularly important in rapidly changing epidemic situations. By introducing adjusted transmission rates βadj and γadj, the forecast results better reflect the epidemic transmission trends under different environmental and social factors. This adjustment not only improves the adaptability of the model but also optimizes the effectiveness of prevention and control measures, ensuring efficient resource allocation. By comprehensively analyzing factors such as temperature data TE, humidity data HU, sentiment analysis data SQ, and air quality data AQ, the system can comprehensively assess the potential impact of different environmental and social factors on epidemic transmission.

[0173] By combining the SIR model with analysis of multiple influencing factors, the system can provide decision-makers with more accurate epidemic forecasts and risk assessments. This allows governments and public health agencies to take timely countermeasures. The spread of an epidemic is often influenced by multiple factors, such as climate conditions, population density, social behavior, and public sentiment. The parameters of the epidemic spread model may change significantly in different social environments and at different stages of the epidemic. Therefore, a dynamic adjustment mechanism allows the system to flexibly respond to these changes and quickly adjust to new transmission trends. This flexibility makes the system applicable not only to the current epidemic, but also to different types of infectious disease transmission that may emerge in the future, providing continuous support to public health agencies.

[0174] Example 4

[0175] This embodiment is explained in Example 3, please refer to Figure 1 ,Specifically: the risk assessment and hotspot identification module includes a risk assessment unit and a hotspot identification and dynamic ,update unit;

[0176] The risk assessment unit calculates the comprehensive risk index RI based on the number of infected people I(t) at the future time point t provided by the epidemic spread prediction module, combined with the epidemic density Iden, population mobility data Rk and environmental pollution index EPI, and compares it with the preset risk threshold TRI to assess the epidemic risk of the region;

[0177] Estimate the severity of the epidemic in a region by calculating the epidemic density Iden of the infected population;

[0178] The epidemic density Iden is obtained by the following formula:

[0179] ;

[0180] Where, mA represents the area of ​​the region;

[0181] Calculate the environmental pollution index EPI through air quality data AQ;

[0182] The environmental pollution index EPI is obtained by the following formula:

[0183] ;

[0184] Where M represents the total amount of pollutants, represents the weight coefficient of the iath pollutant, and AQia represents the air quality data of the iath pollutant;

[0185] The comprehensive risk index RI is obtained by the following formula:

[0186] ;

[0187] Where, They represent the preset weight values ​​of epidemic density Iden, population mobility data Rk and environmental pollution index EPI, respectively, and ;

[0188] The epidemic risk of the area is obtained by matching in the following ways:

[0189] When the comprehensive risk index RI ≤ the risk threshold TRI, it means that the epidemic risk in the region is normal;

[0190] When the comprehensive risk index RI is greater than the risk threshold TRI, it means that the epidemic risk in the region is abnormal.

[0191] The hotspot identification and dynamic update unit identifies epidemic hotspots based on the comprehensive risk index RI of the region, updates the comprehensive risk index RI of each region in real time, and dynamically adjusts the identification of epidemic hotspots;

[0192] The obtained comprehensive risk index RI of each region is sorted, and the region with comprehensive risk index RI> risk threshold TRI is marked as hot spot K, indicating that the risk of the region is abnormal;

[0193] Fit all hotspot areas K to form a hotspot set HArea={Area(K)|RI(K)>TRI};

[0194] Where RI(K) represents the comprehensive risk index of the Kth hotspot area, and Area(K) represents the Kth hotspot area, specifically indicating the area with abnormal risk.

[0195] The comprehensive risk index RI of the hotspot area K is updated at fixed intervals to track the changing trend of the epidemic.

[0196] In this example, the epidemic risk in each region was quantitatively assessed by calculating the Comprehensive Risk Index (RI). This multi-factor risk assessment approach more accurately reflects the complexity of the epidemic's development than traditional single-factor assessments, especially in the context of diverse environments, social behaviors, and population flows. The Comprehensive Risk Index (RI) provides a dynamic, quantifiable risk value for each region, enabling decision-makers to quickly identify potential high-risk areas and implement more targeted prevention and control measures.

[0197] Using the calculated comprehensive risk index (RI), the module can identify and flag areas with higher epidemic risk in real time. When the RI value exceeds the preset risk threshold (TRI), the area faces a high risk of epidemic spread and prevention and control measures must be prioritized. The risk assessment process integrates multiple factors, including epidemic spread data, population mobility, and environmental pollution, improving the accuracy of risk assessments. The introduction of the Environmental Pollution Index (EPI), particularly the impact of air quality on epidemics, provides a quantitative analysis of potential environmental factors contributing to epidemic spread. Through this comprehensive analysis, the system can better assess areas with high pollution and dense population mobility, which are often high-risk areas for epidemic spread. By deeply integrating comprehensive environmental and epidemic data, the spread of the epidemic can be more accurately predicted and controlled.

[0198] By promptly assessing and responding to regional risks, the module can mitigate the spread of epidemics caused by information lags. Dynamically adjusting the identification of hotspots effectively ensures the timeliness of public health decision-making, enabling effective control of key areas in the early stages of an epidemic to prevent its spread. Real-time monitoring of changes in regional risk indices provides timely and accurate decision-making support to decision-makers, ensuring that prevention and control measures can be adjusted as quickly as possible to avoid unnecessary delays.

[0199] Example 5

[0200] This embodiment is explained in Example 4. Please refer to Figure 1 ,Specifically: the decision support and emergency response module includes a decision support unit and an emergency response unit;

[0201] The decision support unit divides the regional epidemic risk into the first and second levels according to the preset risk threshold TRI, and provides prevention and control measures for each hotspot area K;

[0202] When the comprehensive risk index RI ≤ the risk threshold TRI, it is marked as the first level, indicating that the area does not belong to the hotspot area K, and the epidemic risk in the area is normal. Normal prevention and control measures are maintained, and routine monitoring and reporting are carried out;

[0203] When the comprehensive risk index RI is greater than the risk threshold TRI, it is marked as the second level, indicating that the epidemic risk in the hotspot area K is abnormal, and prevention and control measures are provided, including the implementation of blockade measures, isolation of infected persons, and adjustment of medical resource allocation. Adjustment of medical resource allocation specifically involves the implementation of a preset plan to increase medical resource allocation, including proportionally increasing the amount of medical resource allocation;

[0204] The emergency response unit formulates an emergency response plan based on the prevention and control measures provided by the decision support unit, including resource deployment, personnel scheduling and emergency measures;

[0205] Among them, resource allocation includes allocation of hospital beds, ambulances, testing equipment and personal protective equipment;

[0206] Personnel dispatch includes assigning relevant personnel to areas with abnormal epidemic risks to conduct epidemic monitoring and prevention and control;

[0207] Emergency measures include setting up temporary epidemic prevention and control points, strengthening the dissemination of epidemic information and health education for residents.

[0208] The dynamic feedback and optimization module includes a prevention and control effect monitoring unit and a strategy adjustment unit;

[0209] The prevention and control effect monitoring unit obtains real-time epidemic data and the implementation of prevention and control measures, including the number of infected people I, the number of recovered people R, the number of deaths, and the implementation intensity of prevention and control measures, and measures the prevention and control effect through the infection rate β and the recovery rate γ to determine whether the prevention and control measures are effective; the strategy adjustment unit calculates and identifies whether the current prevention and control measures have achieved the expected effect based on the feedback provided by the prevention and control effect monitoring unit, obtains the epidemic control index YZ, and compares it with the preset epidemic control threshold TYZ to determine the effectiveness of the prevention and control measures;

[0210] The epidemic control index YZ is obtained by using a weighted algorithm on the infection rate β and the recovery rate γ;

[0211] The effects of the control measures are obtained by matching the following methods:

[0212] When the epidemic control index YZ ≥ the epidemic control threshold TYZ, it means that the prevention and control measures are effective;

[0213] When the epidemic control index YZ is less than the epidemic control threshold TYZ, it means that the prevention and control measures are ineffective, and the prevention and control measures should be readjusted and a new emergency response plan should be formulated.

[0214] In this embodiment, the decision support unit compares the comprehensive risk index RI with the preset risk threshold TRI to accurately divide the epidemic area into different risk levels. Such a grading system can help public health departments take different levels of prevention and control measures according to actual conditions:

[0215] Low-risk areas (RI ≤ TRI): Maintain normal prevention and control measures, conduct routine monitoring and reporting, and avoid excessive intervention and waste of resources. High-risk areas (RI > TRI): Implement more stringent measures, such as lockdowns, isolation of infected individuals, and expanded allocation of medical resources, to effectively control the spread of the epidemic. This data-driven, tiered approach to prevention and control avoids excessive or insufficient prevention and control measures under a unified standard, allowing for flexible adjustments based on the actual risk situation in different regions to maximize prevention and control efficiency.

[0216] The emergency response unit develops targeted and flexible emergency response plans based on the prevention and control measures provided by the decision support unit. Through dynamic adjustments to resource allocation, personnel scheduling, and emergency measures, the system is able to rapidly respond to changes in the epidemic: Medical resources, including hospital beds, ambulances, and personal protective equipment, are flexibly allocated based on regional risk levels to avoid waste and ensure medical needs in key areas. Medical staff and related personnel are promptly dispatched to high-risk areas to ensure the timeliness and effectiveness of epidemic monitoring and prevention and control. For example, temporary epidemic prevention and control points are established, and efforts are made to increase the dissemination of epidemic information and public health education to ensure a timely and effective public response to the epidemic. This dynamic adjustment and flexible response strategy ensures that public health institutions can respond to sudden changes in the epidemic, reduce the risk of spread, and safeguard public safety.

[0217] The prevention and control effectiveness monitoring unit obtains real-time epidemic data and the implementation of prevention and control measures, allowing rapid assessment of the actual effectiveness of prevention and control measures, including the number of infected, recovered, and deceased people, as well as the intensity of implementation of prevention and control measures. The effectiveness of prevention and control measures is assessed by calculating the infection rate β and the recovery rate γ, helping decision-makers determine whether prevention and control measures have achieved their intended goals. This real-time monitoring system:

[0218] The strategy adjustment unit can judge the effectiveness of prevention and control measures in real time based on the feedback from the prevention and control effect monitoring unit, and optimize the prevention and control strategy based on the changes in the epidemic control index YZ. This optimization mechanism includes:

[0219] Data-driven adjustments: Through real-time feedback, the system can quickly identify shortcomings in prevention and control measures and make adjustments.

[0220] Avoiding strategy lags: This real-time adjustment capability ensures that epidemic prevention and control is not affected by outdated strategies, and avoids the adverse effects of delayed responses on the spread of the epidemic.

[0221] By combining real-time data feedback with a dynamic adjustment mechanism, the system can provide decision makers with accurate and timely decision support. The decision support unit provides precise prevention and control measures for each region, ensuring that decisions are based on reliable data and avoiding the limitations of over-reliance on traditional empirical judgment. The real-time and precise nature of this decision-making process makes epidemic response more efficient and flexible. Through the joint operation of the decision support and emergency response modules, the system can form a multi-level, multi-angle response strategy from macro-epidemic risk assessment to micro-emergency response implementation. Different levels of prevention and control strategies are adopted for epidemics in different regions, avoiding the limitations of a single prevention and control measure and enabling timely adjustment of emergency response plans.

[0222] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A public health epidemic prevention and control system based on big data analysis, characterized by: It includes data collection and preprocessing module, data fusion and feature extraction module, epidemic spread prediction module, risk assessment and hotspot identification module, decision support and emergency response module and dynamic feedback and optimization module; The data collection and preprocessing module is responsible for collecting raw data from hospitals, social media, and meteorological departments, and preprocessing it to obtain the epidemic prevention data group FYW; The feature extraction module extracts features from the epidemic prevention data set FYW, including the emotional fluctuation features related to the epidemic in social media and the correlation features between air pollution and respiratory diseases, to form a health feature set F; The epidemic spread prediction module uses the susceptible-infected-recovered model to predict the trend of epidemic spread through the epidemic prevention data set FYW and the health feature set F, analyzes the impact of population mobility data Rk, air quality data AQ and sentiment analysis data SQ on the spread of the epidemic, and obtains the prediction results; The epidemic spread prediction module includes an SIR model construction and parameter estimation unit and an influencing factor analysis and spread trend adjustment unit; The SIR model construction and parameter estimation unit constructs a mathematical model of epidemic spread based on the susceptible-infected-recovered model according to the epidemic prevention data set FYW and the health feature set F, analyzes the changing trends of the number of susceptible people S, the number of infected people I, and the number of recovered people R, predicts the trend of future epidemic spread, and obtains the number of susceptible people S(t), the number of infected people I(t), and the number of recovered people R(t) at the future time point t; By analyzing population mobility data Rk, air quality data AQ, and sentiment analysis data SQ, we estimate the parameters of the epidemic spread mathematical model, including the infection rate β and recovery rate γ; The number of susceptible people S(t) at the time point t is obtained by the following formula: ； Where, Represents the influence coefficient of temperature data TE and humidity data HU on the spread of the epidemic, represents the influence coefficient of the j-th feature in the health feature set F on the spread of the epidemic, Fj represents the j-th feature in the health feature set F, and m represents the total number of features in the health feature set F; The number of infected people I(t) at the time point t is obtained by the following formula: ； The number of recovered people R(t) at the time point t is obtained by the following formula: ； The influencing factor analysis and transmission trend adjustment unit analyzes the impact of population mobility data Rk, air quality data AQ, and sentiment analysis data SQ on the spread of the epidemic, and brings them into the epidemic spread mathematical model to adjust the epidemic spread trend, obtain the adjusted infection rate βadj and the adjusted recovery rate γadj, bring them into the epidemic spread mathematical model, and re-predict the number of susceptible people S(t), the number of infected people I(t), and the number of recovered people R(t) at the future time point t; The adjusted infection rate βadj is obtained by the following formula: ； Where, They represent the influence coefficients of population mobility data Rk, air quality data AQ and sentiment analysis data SQ on the infection rate β respectively; The adjusted recovery rate γadj is obtained by the following formula: ； Where, They represent the influence coefficients of population mobility data Rk, air quality data AQ and sentiment analysis data SQ on the recovery rate γ respectively; The risk assessment and hotspot identification module identifies the current epidemic risk areas based on the prediction results, combined with the epidemic density Iden, population mobility data Rk and environmental pollution index EPI, and conducts an assessment to obtain a comprehensive risk index RI; The decision support and emergency response module provides decision support to public health agencies based on the comprehensive risk index RI and formulates prevention and control measures; During the epidemic prevention and control process, the dynamic feedback and optimization module monitors the effectiveness of prevention and control measures in real time and adjusts strategies through data feedback.

2. A public health epidemic prevention and control system based on big data analysis according to claim 1, characterized in that: The data acquisition and preprocessing module includes a data acquisition unit and a data cleaning and standardization unit; The collection unit collects raw data from multiple data sources such as hospitals, social media, and meteorological departments, including electronic health records EHR, population mobility data Rk, sentiment analysis data SQ, temperature data TE, humidity data HU, and air quality data AQ; Among them, the electronic health record EHR is obtained through the hospital information management system, including the incidence rate of respiratory diseases DiseaseRate; the population flow data Rk is obtained by the ratio of the number of population flows in the region to the total area of ​​the region; Sentiment analysis data SQ includes capturing posts related to the epidemic from social media platforms, performing sentiment analysis, and using NLP technology to classify each post into positive, negative, and neutral emotions. The positive, negative, and neutral emotions are then assigned numerical values: positive emotions are +1, negative emotions are -1, and neutral emotions are 0. Sentiment analysis data SQ is obtained using the following formula: ； Where Si represents the i-th social media post, and n represents the total number of posts analyzed; The temperature data TE and humidity data HU are obtained through the temperature sensor and humidity sensor respectively; Air quality data AQ is obtained through the following formula: ； Wherein, cPM2.5, cCO and cNO2 represent the concentrations of air particulate matter PM2.5, carbon monoxide CO and nitrogen dioxide NO2, respectively; IPM2.5, ICO and INO2 represent the pollutant normalization indexes of air particulate matter concentrations PM2.5, carbon monoxide CO and nitrogen dioxide NO2, respectively; max represents the peak value; The data cleaning and standardization unit cleans and standardizes the original data collected by the data collection unit to obtain the epidemic prevention data group FYW; Cleaning involves removing noise and invalid information from the raw data, including missing values ​​and outliers, by using filters; Standardization is the process of standardizing the original data into a uniform dimension.

3. A public health epidemic prevention and control system based on big data analysis according to claim 2, characterized in that: The feature extraction module includes a social media emotion fluctuation feature extraction unit and an air pollution and respiratory disease correlation feature extraction unit; The social media emotion fluctuation feature extraction unit extracts emotion fluctuation features based on the sentiment analysis data SQ, analyzes the changing trend of epidemic-related public opinion, and calculates and obtains the emotion fluctuation index EVI and the emotion change trend ET; The emotional volatility index EVI is obtained by the following formula: ； Where N represents the total time period, SQti represents the value of the sentiment analysis data at time point ti, and PSQ represents the average value of the sentiment analysis data; The emotion change trend ET is obtained by the following formula: ； Where SQ(qi) represents the value of the sentiment analysis data at time qi, and SQ(pi) represents the value of the sentiment analysis data at time pi.

4. A public health epidemic prevention and control system based on big data analysis according to claim 3, characterized in that: The respiratory disease correlation feature extraction unit calculates and obtains the pollution and health correlation index PHI based on the correlation between the air quality data AQ and the electronic health record EHR, specifically by combining the air quality data AQ and the respiratory disease incidence rate DiseaseRate; The pollution and health index PHI is obtained by the following formula: ； Where, ln represents the natural logarithm function; The obtained emotion fluctuation index EVI, emotion change trend ET and pollution and health association index PHI are combined to obtain the health feature set F.

5. A public health epidemic prevention and control system based on big data analysis according to claim 4, characterized in that: The risk assessment and hotspot identification module includes a risk assessment unit and a hotspot identification and dynamic update unit; The risk assessment unit calculates the comprehensive risk index RI based on the number of infected people I(t) at the future time point t provided by the epidemic spread prediction module, combined with the epidemic density Iden, population mobility data Rk and environmental pollution index EPI, and compares it with the preset risk threshold TRI to assess the epidemic risk of the region; Estimate the severity of the epidemic in a region by calculating the epidemic density Iden of the infected population; The epidemic density Iden is obtained by the following formula: ； Where, mA represents the area of ​​the region; Calculate the environmental pollution index EPI through air quality data AQ; The environmental pollution index EPI is obtained by the following formula: ； Where M represents the total amount of pollutants, represents the weight coefficient of the iath pollutant, and AQia represents the air quality data of the iath pollutant; The comprehensive risk index RI is obtained by the following formula: ； Where, They represent the preset weight values ​​of epidemic density Iden, population mobility data Rk and environmental pollution index EPI, respectively, and ; The epidemic risk of the area is obtained by matching in the following ways: When the comprehensive risk index RI ≤ the risk threshold TRI, it means that the epidemic risk in the region is normal; When the comprehensive risk index RI is greater than the risk threshold TRI, it means that the epidemic risk in the region is abnormal.

6. A public health epidemic prevention and control system based on big data analysis according to claim 5, characterized in that: The hotspot identification and dynamic update unit identifies epidemic hotspots based on the comprehensive risk index RI of the region, updates the comprehensive risk index RI of each region in real time, and dynamically adjusts the identification of epidemic hotspots; The obtained comprehensive risk index RI of each region is sorted, and the region with comprehensive risk index RI> risk threshold TRI is marked as hot spot K, indicating that the risk of the region is abnormal; Fit all hotspot areas K to form a hotspot set HArea={Area(K)|RI(K)>TRI}; Where RI(K) represents the comprehensive risk index of the Kth hotspot area, and Area(K) represents the Kth hotspot area, specifically indicating the area with abnormal risk. The comprehensive risk index RI of the hotspot area K is updated at fixed intervals to track the changing trend of the epidemic.

7. The public health epidemic prevention and control system based on big data analysis according to claim 6, characterized in that: The decision support and emergency response module includes a decision support unit and an emergency response unit; The decision support unit divides the regional epidemic risk into the first and second levels according to the preset risk threshold TRI, and provides prevention and control measures for each hotspot area K; When the comprehensive risk index RI ≤ the risk threshold TRI, it is marked as the first level, indicating that the area does not belong to the hotspot area K, and the epidemic risk in the area is normal. Normal prevention and control measures are maintained, and routine monitoring and reporting are carried out; When the comprehensive risk index RI is greater than the risk threshold TRI, it is marked as the second level, indicating that the epidemic risk in the hotspot area K is abnormal, and prevention and control measures are provided, including implementing blockade measures, isolating infected people, and adjusting the allocation of medical resources; The emergency response unit formulates an emergency response plan based on the prevention and control measures provided by the decision support unit, including resource deployment, personnel scheduling and emergency measures; Among them, resource allocation includes allocation of hospital beds, ambulances, testing equipment and personal protective equipment; Personnel dispatch includes assigning relevant personnel to areas with abnormal epidemic risks to conduct epidemic monitoring and prevention and control; Emergency measures include setting up temporary epidemic prevention and control points, strengthening the dissemination of epidemic information and health education for residents.

8. The public health epidemic prevention and control system based on big data analysis according to claim 5, characterized in that: The dynamic feedback and optimization module includes a prevention and control effect monitoring unit and a strategy adjustment unit; The prevention and control effect monitoring unit obtains real-time epidemic data and the implementation of prevention and control measures, including the number of infected people I, the number of recovered people R, the number of deaths, and the implementation intensity of prevention and control measures, and measures the prevention and control effect through the infection rate β and recovery rate γ to determine whether the prevention and control measures are effective; The strategy adjustment unit calculates and identifies whether the current prevention and control measures have achieved the expected results based on the feedback provided by the prevention and control effect monitoring unit, obtains the epidemic control index YZ, and compares it with the preset epidemic control threshold TYZ to determine the effectiveness of the prevention and control measures; The epidemic control index YZ is obtained by using a weighted algorithm on the infection rate β and the recovery rate γ; The effects of the control measures are obtained by matching the following methods: When the epidemic control index YZ ≥ the epidemic control threshold TYZ, it means that the prevention and control measures are effective; When the epidemic control index YZ is less than the epidemic control threshold TYZ, it means that the prevention and control measures are ineffective, and the prevention and control measures should be readjusted and a new emergency response plan should be formulated.

Images