透過您的圖書館登入
IP:3.131.110.169
  • 學位論文

台灣PM2.5月分量之空間分佈

Spatial distribution of PM2.5 monthly quantiles in Taiwan

指導教授 : 張雅梅

摘要


本研究提出以非平穩空間模型(non-stationary spatial model) 應用於台灣地區懸浮粒子濃度(PM2.5) 的資料,並對各月份懸浮粒子濃度的分量進行模型的建構,此非平穩空間模型為數個基底函數以及若干個平穩過程之線性組合所構成。在此模型的設定下,我們必須估計非常多的參數,也因為必須要估計的參數非常的多,我們會利用最小絕對壓縮與篩選運算法(Least absolute shrinkage and selection operator, Lasso),去同時做到參數估計及模型篩選的效果。在計算上我們也可以透過Efron et al. 在2004 年所提出的最小角度法(lars) 的R 軟體套件,幫助我們更有效的完成模型的建立。本研究將分析結果繪製為空間分佈圖,藉此來觀察台灣地區各月份懸浮粒子濃度分量的分佈情形,以及其空間相關性之討論。根據結果顯示,秋、冬季時,台灣空氣中有較高的懸浮粒子含量。金門及馬祖地區在各個季節有較高的懸浮粒子含量,而台南、高雄、嘉義等中南部沿海地區則在秋冬兩季有著較高的含量。金門地區在懸浮粒子的各個分量下,都有較大的變異程度。金門、大寮、二林與馬祖東引對於懸浮粒子含量相關性的構成有較大的影響。

並列摘要


In this research, a non-stationary spatial model is applied to the monthly quantiles of Taiwan PM2.5 data. In our model setting, there are a lot of parameters to estimate. The Least absolute shrinkage and selection operator (Lasso) is used to estimate the parameters. The method of Lasso can deal with model selection and parameter estimation simultaneously. In estimation, we use the lars package in R language which can solve the Lasso estimates e ciently. The result of the analysis is displayed in plots to observe the PM2.5 distribution. In autumn and winter, the air of Taiwan has higher content of PM2.5. Kimen and Matsu have higher content of PM2.5 in various seasons. Tainan, Kaohsiung and Chiayi have higher content of PM2.5 in autumn and winter. The PM2.5 of Kinmen has a greater variation.

參考文獻


[1] R. Tibshirani, Regression shrinkage and selection via the lasso," Journal
[2] B. Efron, T. Hastie, I. Johnstone, and R. Tibshirani, Least angle re-
posure in early life to polycyclic aromatic hydrocarbons using total sus-
ux and ground temperature," Environmental Geology, vol. 48,
[13] P. M. Anderson, The uses and limitations of trend surface analysis in

延伸閱讀