細懸浮微粒是指懸浮在空氣中的固體顆粒或液滴，其粒徑小於或等於2.5微米。它是空氣污染的主要來源之一，易深入到人體肺部，並能由肺泡中的微血管進入到血液中對人體造成危害。行政院環境保護署設置之空氣品質監測網中關於細懸浮微粒資料之監測，至今已有長達八年的數據。本研究針對此數據，檢視細懸浮微粒監測資料之品質，並建立其濃度變化之時間序列模型，進而探討各測站之細懸浮微粒濃度之變化趨勢及彼此間的相關性。

此外，空氣品質監測網中除了有細懸浮微粒濃度資料，亦包含其餘十二種變數：懸浮微粒、二氧化硫、二氧化氮、臭氧、大氣溫度、一氧化碳、一氧化氮、氮氧化物、雨量、溼度、風向及風速，我們選取以上十二種變數為解釋變數，將細懸浮微粒資料視為反應變數，然後將解釋變數經由主成分分析後，再以主成分分析所得的分數與反應變數建立迴歸模型，藉此探討其與細懸浮微粒濃度的關係。另外，挑選出高屏地區十五個測站中五個較具代表性的測站，建立管制圖，用以檢視細懸浮微粒監測資料之品質。未來可提供環保署做為細懸浮微粒監測數據品質評估，及未來對細懸浮微粒管制策略之參考。

Fine suspended particulates (PM2.5) are the suspended particulate with particle size less than 2.5 μ m/m^3. They go deep into the lungs easily and through alveolar capillaries enter the blood circulation, then harm the human body. The epidemiology study has discussed that fine suspended particulates have influence to incidence and death rate of lungs disease. The Environmental Protection Administration (EPA) in Taiwan has monitored the PM2.5 for eight years and has setup air quality monitoring network. This study is aimed at investigating the trend of PM2.5 and the relationship among all stations through analyzing the monitoring data.
Besides PM2.5, the air quality monitoring network has also been monitoring other variables, such as: PM10, SO2, NO2, O3, CO, NO, NOx, temperature, rainfall, humidity, direction and velocity of wind. In order to investigate the correlation between PM2.5 and other variables, we refer to PM2.5 as the response variable and the other variables as the predictor or regressor variables. We use the factor analysis to maximize the variance of a linear combination of the regressor variables, and then try to fit a linear regression model between PM2.5 and the factor scores which obtained from the results of factor analysis. Furthermore, for the purpose of controlling quality of PM2.5, we select five more representative stations from fifteen stations in the southern Taiwan to draw the quality control charts. Later the fitted time series model is used to forecast the future trend, which may be useful to the EPA,
as reference to the environment control quality of PM2.5.

論文審定書─────────────────────────ｉ
誌謝────────────────────────────ｉｉ
摘要───────────────────────────ｉｉｉ
Ａｂｓｔｒａｃｔ───────────────────ｉｖ
１前言────────────────────────────１
２資料敘述─────────────────────────１
　２.１資料來源──────────────────────２
　２.２相關名詞──────────────────────２
　　　２.２.１測站類型─────────────────２
　　　２.２.２無效值───────────────────２
　２.３資料選取與處理──────────────────３
３研究方法─────────────────────────４
　３.１時間序列模型───────────────────４
　３.２品質管制──────────────────────５
　３.３時間序列集群分析────────────────６
　３.４因素分析與迴歸分析───────────────６
４實證分析─────────────────────────８
　４.１時間序列模型配適結果─────────────８
　４.２品質管制圖繪製────────────────１１
　４.３集群分析分類結果───────────────１５
　４.４因素分析與迴歸模型配適結果────────１５
５評估與探討───────────────────────２１
參考文獻──────────────────────────２２
附錄一────────────────────────────２３
附錄二────────────────────────────２４
附錄三────────────────────────────２５
附錄四────────────────────────────３１
附錄五────────────────────────────４４
附錄六────────────────────────────４８

[1] Montgomery, D.C., Peck, E.A., and Vining G.G. (2006). Introduction to linear regression analysis, 4th. Springer, New York.
[2] Montgomery, D.C.(2013). Statistical quality control - A modern introduction, 7th. Wiley, New York.
[3] Rencher, A.C. (2002). Methods of Multivariate Analysis, 2nd. Wiley, New York.
[4] Tsay, R.S. (2010). Analysis of Financial Time Series, 3rd. Wiley, New York.
[5] United States Environmental Protection Agency-Particulate Matter <http://www.epa.gov/pm/health.html>
[6] 行政院環境保護署-空氣品質監測網 <http://taqm.epa.gov.tw/taqm/zh-tw/default.aspx>
[7] 黃玉潔(2013)。細懸浮微粒濃度資料之可靠度及趨勢分析，國立中山大學應用數學系碩士論文。