Comparing the Generalized Poisson Regression And Negative Binomial Regression Methods In Dealing With Overdispersion

• Main Authors: Yu-Ting Hsu, 徐郁婷
• Other Authors: Yi-Hau Chen
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/y8n823

碩士 === 國立陽明大學 === 公共衛生研究所 === 105 === Understand the characteristics of infectious diseases, contribute to the assessment of the disease epidemic, the future development trend for further study, and health policy makers to take appropriate health policy to prevent the expansion of the disease. Dengue fever is the main exploratory infectious diseases in this research. The number of cases of dengue fever was treated as response variables, which was a discrete type and the number of morbidity was not large. The Poisson regression model, which is a common type of discrete data, was selected as a benchmark and compared with other estimates of discrete data Model(Generalized Poisson regression and Negative binomial regression), to explore its differences.