Human Rabies Analysis And Prediction And Visualization Platform Construction For Hubei Province Based On Data Mining Technology

Nowadays,with the continuous updating and improvement of theory and techniques recently,data mining has been well applied in biomedicine domain,which transforms complex biomedical data into useful information for scientific exploration by the help of multidisciplinary ideas such as statistics,information science and visualization.Rabies is a zoonotic disease which can cause fatal injury after being infected with the virus.The incidence of rabies in China is very high,ranking only after India.Therefore,it will be meaningful to mine the epidemiological characteristics of human rabies and explore its development pattern to inspire rabies prevention work.On the basis of the current work of human rabies data analysis,the following three scientific questions are proposed:(1)How to explore the causality among rabies surveillance data and predict their future trends?(2)How to quantify and investigate the impact of key factors on the number of rabies incidence?(3)How to investigate the dynamic epidemiological characteristics of human rabies in Hubei by three dimensions: time,space,and crowd,and to build a visual analysis platform?Addressing the aforesaid problems,this dissertation is based on data mining methods,choosing human rabies in Hubei Province and performing data analysis on it.The work is summarized as follows:(1)Based on the time series data of the number of human rabies incidence,exposure and vaccination in Hubei Province,a modified Granger causality test was utilized to mine the causality among the three series.The results that the number of rabies incidence and rabies exposure were Granger cause of the number of rabies vaccination,while the remaining series combinations were not causally related was obtained.Then,in order to predict the future situation of local rabies incidence and rabies exposure,SARIMA prediction model was built based on their data characteristics,respectively.(2)Collect main influencing factors associated with socio-economics and climate from various types of relevant impact data,and the influence of local resident population,GDP,temperature,precipitation on rabies incidence was obtained by building a panel regression model.Meanwhile,geographic environmental data features were quantitatively extracted from remote sensing satellite images and digital elevation model data which came from complicated sources in Hubei Province using geo-information processing methods.And the difference significance test showed that the elevation and slope of regions had influence on the rabies incidence.(3)Based on the descriptive research method and visualization techniques,data analysis was performed by organically combining three dimensions of time,space and crowd.The characteristics of season of rabies incidence and rabies exposure and the dynamic characteristics of time-space distribution and crowd distribution of rabies incidence in Hubei Province were obtained.In addition,a provincial-level visualization and analysis platform of the historical human rabies surveillance data was built for researchers to conduct interactive data query and control and supervision reference through Web development technology.