連續型變數之記數過程在傳染病資料上之應用; Counting process approach to infectious disease data with continuous covariates

NCU Institutional Repository > 理學院 > 數學研究所 > 博碩士論文 > Item 987654321/7837

請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/7837

題名:	連續型變數之記數過程在傳染病資料上之應用;Counting process approach to infectious disease data with continuous covariates
作者:	陳中興;Chung-Hsing Chen
貢獻者:	數學研究所
關鍵詞:	記數過程;傳染病;counting process;infectious disease
日期:	2008-06-04
上傳時間:	2009-09-22 11:06:45 (UTC+8)
出版者:	國立中央大學圖書館
摘要:	這篇論文提出一個分析傳染病資料的新模型與演算法。利用點過程（point process）並且考慮連續型的解釋變數（continuous explanatory variable）以及更廣泛的感染函數（infectivity function）去建構出家庭中每一個人被感染與被移除（removal）的條件機率。我們定義兩個記數過程（counting process），各自代表著在一個家庭中每一個人何時被感染與何時被移除。這些發生的條件機率可以用來描述傳染病的擴散速度；同時，這些條件機率也受到一些個人的特徵與我們設計的函數所影響。我們利用貝氏分析（Bayesian inference）裡常用的馬可夫鏈蒙地卡羅（Markov Chain Monte Carlo）演算法發展出一種特別的演算法¬，並且用它分析傳染病的特性與人的特徵對傳染病的影響；包括分析模擬的結果以及分析真實資料的結果。 This paper proposes a point process model for infectious disease data that take into consideration continuous explanatory variables regarding infectivity, susceptibility to infection and removal rate and allow parametric family of baseline infectivity functions. For each individual in a closed community, we define two counting processes; one jumps when this individual gets infected and the other jumps when this individual gets removed. The intensities of these counting processes are used to describe the spread of the infectious disease. These intensities have one component describing the way that individual covariates may affect infectivity, susceptibility to infection or removal; these intensities also have a baseline infectivity function, belonging to a parametric family of functions. Customized MCMC algorithms are developed for Bayesian inference based on removal times and covariates of each individual. Simulation studies and analysis of real infectious data are provided to illustrate the numerical performance of the methods.
顯示於類別:	[數學研究所] 博碩士論文

文件中的檔案:

檔案	大小	格式	瀏覽次數

在NCUIR中所有的資料項目都受到原著作權保護.

社群 sharing

資料載入中.....