Detecting The Epidemic Spatio-temporal Evolution Pattern Of Hunan Malaria

Abstract:The study of epidemic spatio-temporal evolution pattern contribute to grasp the epidemic trend and the developing rules, and to make scientific decisions of controlling the spread of the epidemic. The present researches considered only from the independent case samples, ignored the regional difference of case and spatial correlation. Hence, there are some study methods of detecting epidemic spatio-temporal evolution pattern proposed from three aspects of spatial autocorrelation, spatial clustering and regional centroid deviation. The primary contents of the thesis Call be summarized as follows:1) The spatial autocorrelation analysis is used for detecting epidemic spatio-temporal evolution pattern. Firstly, we calculate the spatial correlation coefficient of these cases by the global spatial autocorrelation analysis, to detect cases spatial aggregation. Next, we detect the epidemic disease space hot region by local spatial autocorrelation analysis, and then mining epidemic spatio-temporal evolution procession via results of time series spatial autocorrelation analysis.2) This paper expands an epidemic spatio-temporal evolution model based on spatial clustering analysis method. First, we divide administrative unit into grids on the basis of the grid technology, and cluster grids by using the topological adjacency-based clustering method to gain spaces of disease spatial aggregation. Then, we can detect epidemic deviation procession of disease spatial aggregation spaces via results of time series.3) A centroid transferring curve based spatio-temporal pattern expression-CTCSTP is proposed in this paper. Firstly, the weighted centroids of disease for every year are computed. The centroid transferring curve composited by these weighted centroids. Then, the deviation distance and orientation of centroid transferring curve can be computed to convey intuitively the disease spatio-temporal evolving pattern.