Research On Spatiotemporal Modeling And Large-screen Interaction Methods For Epidemiological Data

The recent developments in the pandemic situation have been severe,with spatiotemporal analysis in epidemiological studies and large-screen visualization interaction becoming focal points.As epidemic transmission shows strong spatiotemporal correlation,understanding the spatiotemporal distribution features of epidemiological data and thereby discovering hidden knowledge to achieve spatiotemporal modeling is of great importance.The epidemiological data presents the characteristics of large scale and complex data types,making the demand for visualization interaction in large-screen scenarios even more intensive,and supporting precise interaction with visualization objects while effectively conveying data content faces significant challenges.Considering the complex factors affecting epidemics and the higher degree of interaction freedom in large-screen scenarios,the following issues remain unresolved in relevant research on spatiotemporal modeling and large-screen interaction methods for epidemiological data.1.The spatiotemporal analysis of epidemiological data holds important research value.Ignoring the spatiotemporal distribution of data and hidden spatiotemporal correlation attribute features may lead to poor risk prediction results.Thus,it is necessary to consider building a risk prediction model based on spatiotemporal correlation attributes,analyze the historical changes and spatial aggregation patterns of epidemiological data,and introduce these into the model features to enhance prediction accuracy.2.Visualization interaction methods of spatiotemporal analysis of epidemics focus on the presentation of large-scale data content,but face difficulties in acquiring effective spatiotemporal correlation information.There is a lack of effective visualization interaction strategies in large-screen visualization scenarios,capable of implementing precise target interaction under the limitations of spatial interaction scope and available interaction devices,to support the spatiotemporal correlation interaction of epidemiological data.3.In the scenario of spatiotemporal analysis of epidemics,existing interaction methods cannot meet the requirements for dynamic adjustment of interaction areas and complete display of data.The interaction functions are single,and interaction flexibility is limited.It is necessary to design interactive strategies with diversity and effectiveness,expand input channels,enrich interaction forms,and achieve rich and flexible interaction while ensuring interaction speed and accuracy.Based on these problems,this study mainly focuses on the spatiotemporal analysis and modeling of epidemiological data,large-screen interaction strategies for spatiotemporal analysis of epidemics,and the extension of multimodal interaction strategies for spatiotemporal analysis of epidemics.The specific contributions are as follows:1.A spatiotemporal component fusion model(STCFM)based epidemic risk prediction is proposed.Dengue fever epidemic is used as the research case,and mosquito density and spatiotemporal correlation feature factors are introduced to separately establish component models for time trends and spatial distribution.Multiscale modeling of time dependency is performed in the time dimension,and multivariate spatial correlation analysis is conducted in the spatial dimension to enhance feature representation.Finally,spatiotemporal component models are fused through ensemble learning methods.The validation on a real dengue fever epidemic dataset shows that STCFM is superior to other comparative models,proving the importance of spatiotemporal analysis in epidemiological data.2.Design visualization interaction methods and high-precision target interaction strategies suitable for large-screen display devices for spatiotemporal analysis scenarios of epidemics.Smart handheld devices are used as interaction devices.Interaction spaces are designed based on control modes that include rotation posture and touch operations,and cursor forms that include area cursor and point cursor.A hierarchical architecture is used to solve the demand for high-precision target interaction,providing an effective solution for target selection tasks in the spatiotemporal analysis scenarios of epidemics,to support the visualization of spatiotemporal correlation in epidemiological data.3.A multimodal extended interaction strategy is proposed,introducing pressure input to extend control modes,enriching target acquisition mechanisms through regional coverage context transmission,and implementing multi-level gain interaction strategies and cursor projection interaction strategies to enrich the interaction dimensions in large-screen scenarios.In the spatiotemporal analysis scenario of epidemics,it supports dynamic adjustment of interaction areas and cross-device data interaction functions,enhancing data expression while still satisfying the smoothness and diversity of interaction with visualization objects in large-screen scenarios.