李文宗2006-07-252018-06-292006-07-252018-06-292000-07-31http://ntur.lib.ntu.edu.tw//handle/246246/4679疾病的時空變異是流行病學的重 要課題。本計畫第二年（即本年度）針 對改善邊緣效應進行研究，提出了「補 點式卜瓦松趨勢面分析法」。電腦模擬 研究發現此新方法可以大幅改善邊緣 效應，並且有很精確的定位死亡率聚集 中心的能力。此新方法亦可很容易地加 入時間維度。本計畫以臺灣本島民國71 年至民國85 年口腔癌死亡率為例進行 研究。結果發現新方法可以清楚的顯示 口腔癌死亡率的時空動態變化Spatiotemporal variation in the occurrence of diseases is an important issue in epidemiology. In this second-year project, we proposed a new method, the “imputed Poisson trend surface analysis”, which can reduce the impacts of “edge effects”. The method can locate the mortality hot-spot quite accurately. The method can also be easily extended to deal with the time dimension. We analyzed oral cancer mortality in the Taiwan island proper between 1982 and 1996 using the new method. And we found that the new method can provide a clear portray of spatiotemporal interaction of oral cancer mortality in a dynamic way.application/pdf18664 bytesapplication/pdfzh-TW國立臺灣大學公共衛生學院流行病學研究所趨勢面分析卜瓦松迴歸口腔癌Trend Surface AnalysisPoisson RegressionOral Cancer[SDGs]SDG3reporthttp://ntur.lib.ntu.edu.tw/bitstream/246246/4679/1/892320B002108.pdf