The Interpolation Of Incidence Of Notifiable Disease In The Missing Report Area

The worldwide outbreak of the infectious disease has fast replaced the local area as the result of the global environmental changing and the perfect large-scale traffic accessibility, and gradually becomes a burning issue for the international community. Currently, the infectious disease earns the world’s attention. Governments have established Disease Control&Prevention Center to take practical and feasible measures in infectious disease monitoring and managing. Even though, there are still some missing reports existed in surveillance system due to some unpredictable factors. Often, this irregular action have a severe impact on report quality. It is generally known that the low level of data quality can have a negative effect on analyzing the trend of disease, completing the disease risk assessment and making the rule to control of infectious disease. The dissertation focuses on the interpolation of the incidence of notifiable disease in the missing report area. The result intends to work out an agenda for the health department to supervise, guide and examine the system.In fact, the characteristic of spatial data causes that the work of spatial analysis may be affected by the nonhomogeneity in every situation. The better understanding of the nonhomogeneity, the more efficiency of the spatial sampling. The author tackled some problems in study of surfaces with stratified nonhomogeneity by proposing the novel spatial models-MSN and Geographical detector, which are then applied to the example of hand, food and mouth disease issue. Specifically, this paper include the following:1) Exploratory Spatial Data Analysis is used for revealing the spatial structure and a certain relationship between spatial data, viewed as a precursor to a wider spatial analysis. With the purpose of measuring the correlation between the two group spatial data, Pearson coefficient of correlation is introduced in this part. Geographical detector is proposed to assess the risks of health. According to the regional variations, the factor detector identifies3factors that are responsible for the disease, including the number of hospital beds, the density of population and the GDP per person.2) In geosciences, a surface represents a spatially varying attribute, which means that the surfaces are interpreted according to the attributes categories being represented. The dissertation uses K-means as the clustering method to analyze the attributes over the spatial region of interest and the best minute class number is obtained by parameter of Geographical Detector.3) The uneven spatial distribution of sampling and the spatial nonhomogeneity of the actual interest surface can generate a considerable bias between the naive sample mean and the population mean. The surface is decomposed into a set of spatially substrata within which the random field is spatially homogeneous. The model Means of Surfaces With Nonhomogeneity (MSN) is introduced to generate the best linear unbiased estimator. Based on the math model, the author determine the theoretical model that is fitted to the samples of the local specific stratum and determine the global model fitted to the samples of all different strata.Finally, under the main content summarized in the last part of this dissertation, the problems for further studied in this field are put forward.