Research on the spatial relationship among dengue incidence rate and social economic factors

• Main Authors: LIU, WEI-KANG, 劉維剛
• Other Authors: Chou, Tien-Yin
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/62468d

碩士 === 逢甲大學 === 都市計畫與空間資訊學系 === 105 === The epidemic of dengue fever worldwide has become much more severe under the recent trend of climate change resulted from global warming. Warmer climate enables the distribution of dengue vectors expands from tropical zone to temperate zone. More than 40% of the world's populations are living in areas where are at risk for infection, and Taiwan is no exception. In this study, we attempt to explore the spatial relation of incidence rate of dengue fever in Tainan city, by using social economic data collected in 2015 with basic statistical area (BSA) as data format. First we calculate Moran’s I to identify the existence of clusters in raw dengue incidence rate, then use spatial empirical Bayes smoothing method to adjust its spatial distribution. By correlation analysis we identify four social economic variables having highest correlation coefficient with dengue incidence rate, namely population density, proximity to provincial road, proximity to train station and education level below junior high school. We apply these variables to construct an ordinary least squares (OLS) model and a Geographically Weighted Regression (GWR) model as a comparison. By comparing analytic results from two models, we found that GWR is more suitable than OLS for modeling social economic variables in respect to higher R2 and AICc. This result shows that the social economic data collected from local area is able to reflect infection risk of dengue fever, which gives a clue to public health policy making. However, interpretation of the GWR modeling results requires the use of contextual and other underlying spatial information. More efforts are definitely needed for further study in this topic.