地理加權卜瓦松迴歸於台灣各鄉鎮區登革熱資料之分析

機構典藏 > College of Business and Management > Graduate Institute & Department of Statistics > Thesis > Item 987654321/111098

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/111098

Title:	地理加權卜瓦松迴歸於台灣各鄉鎮區登革熱資料之分析
Other Titles:	A geographically weighted Poisson regression analysis of township-level Dengue fever data in Taiwan
Authors:	王一安;Wang, Yi an
Contributors:	淡江大學統計學系碩士班陳怡如;Chen, Yi-Ju
Keywords:	登革熱;發生率;過度離散;卜瓦松迴歸;地理加權迴歸;地理加權卜瓦松迴歸;空間非平穩性;Dengue fever;Incidence rate;overdispersion;Poisson regression;Geographically weighted regression;Geographically Weighted Poisson Regression;Spatial Nonstationary
Date:	2016
Issue Date:	2017-08-24 23:43:54 (UTC+8)
Abstract:	近幾年登革熱疫情持續延燒，2015年台灣登革熱的病例數甚至突破四萬筆病例數。過去對於登革熱疫情空間分布上的研究主要著重於高雄、台南和屏東三個疫情較嚴重的區域，較少探討台灣整個區域疫情的狀況。一般來說，登革熱的研究通常以發生率來描述疫情之嚴重性，且在流病領域中較常使用卜瓦松迴歸模型在疾病分析中，特別是罕見疾病。本研究主要針對2015年分析台灣各鄉鎮區的登革熱資料，利用地理加權卜瓦松迴歸分析之技術探討不同的社經特性、人口結構與環境因子是否對於疫情的發生率存在空間非平穩的效果。但是由於登革熱屬於稀有疾病，不少鄉鎮區都無病例發生，造成登革熱病例分布呈現嚴重傾斜，存在過度離散的問題，若使用一般卜瓦松模型來分析資料會提升估計參數的型一誤差做出錯誤推論，因此為了解決過度離散的問題，本研究將使用Empirical變異數估計法對資料做調整，以期望藉由分析地區疫情的發生率及各種影響因子，可以擬定出不同的防疫措施，提前預防或者避免大規模的疫情發生。 Recent years the Dengue fever has continued to spread. In 2015, the number of cases exceeded 40,000. In the past, researchers mainly focused on researching the Dengue fever in Kaohsiung, Tainan and Pingtung, where the epidemic was more serious. They rarely discussed the epidemic all over Taiwan. In general, incidence usually described the severity of the epidemic in the study of Dengue fever. Also, in the field of epidemiology the Poisson Regression Model is commonly used in disease analysis, especially in the rare disease. In this thesis, we focused on analyzing the Dengue fever throughout Taiwan in 2015, and used Geographically Weighted Poisson Regression Analysis to explore whether socioeconomics status, population structure and environmental factors could have the spatial nonstationary effect on the incidence of the disease. However, because the Dengue fever is a rare disease, no cases occurred in some counties and towns.It makes the distributions of the Dengue fever cases unbalanced and overdispersed. Using the general Poisson model could raise the type I error and resulted in erroneous inferences. Therefore, in order to solve this problem, we used Empirical variance estimation to adjust the data. Moreover, through analyzing the incidence and the impact factors, we expected to develop different prevention measures which can prevent Dengue fever from breaking out.
Appears in Collections:	[Graduate Institute & Department of Statistics] Thesis

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	150	View/Open

Loading...