Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/93635| Title: | 基於 DTW 的分割時間序列模型進行 Python/Dash 可視化分析台灣空氣盒子數據 Using DTW-based Partitioning Time Series Models with Python/Dash Visualization Analysis AirBox Data in Taiwan |
| Authors: | 陳柏伍 Po-Wu Chen |
| Advisor: | 潘建興 Frederick Kin Hing Phoa |
| Co-Advisor: | 楊鈞澔 Chun-Hao Yang |
| Keyword: | DTW(動態時間規則),時間序列,SARIMA,空氣盒子,PM2.5,聚類,DASH應用程式, DTW (Dynamic Time Warping),Time series,SARIMA,AirBox,PM2.5,Clustering,DASH app, |
| Publication Year : | 2024 |
| Degree: | 碩士 |
| Abstract: | 鑒於台灣嚴重的空氣污染問題,我們的研究目標是準確地捕捉和預測 PM2.5 濃度,這對於監測空氣質量並協助專家們確定 PM2.5 源頭至關重要。我們利用基於 DTW 的時間序列模型來實現這一目標,該模型首先使用 DTW 計算每個時間序列之間的距離,並根據這些距離將類似模式的時間序列分組為不同的群集。我們採用了常見的分群方法,如階層式分群,以實現這一步驟。接著,我們將 SARIMA 模型應用於每個群集,以捕捉趨勢並預測未來的值。此外,我們還利用 Python/Dash 可視化工具進行視覺分析,以促進對 AirBox 數據的理解。透過這些分析工具,我們能夠觀察 PM2.5 的趨勢,並識別影響濃度水平的關鍵特徵,進而更好地了解台灣的空氣污染情況。
本研究的背景涉及在台灣部署 AirBox 設備的相關信息,以及先前對空氣質量監測站和創新方法(如 MOB 或 MTSC)的研究。這些既有研究為我們的工作奠定了基礎,並提供了寶貴的參考。我們將這些信息結合,以更好地理解台灣空氣污染問題的本質,並為改善環境質量提供重要的指引。 通過本研究的分析方法,我們將能夠更好地理解時間序列數據的模式和趨勢,從而提高對 PM2.5 濃度數據的分類和分析能力。通過分析 AirBox 數據,我們可以更好地識別影響 PM2.5 濃度的因素,進而針對這些因素制定更有效的空氣質量管理策略。我們的研究成果將為未來的工作提供重要的研究方向和參考依據 Due to the serious air pollution issue in Taiwan, we aim to accurately capture and predict PM2.5 concentrations. This is crucial for monitoring air quality and helps experts determine the source of PM2.5. This study utilizes DTW-based Partitioning Time Series Models, which first use DTW to calculate the distance between each time series, grouping those with similar patterns into clusters based on their DTW-distances. Common clustering methods such as hierarchical clustering, are employed. Subsequently, SARIMA models are applied to capture trends and forecast future values. In addition, we use Python/Dash visualization tools to promote visual analysis of the AirBox data, allowing us to observe PM2.5 trends and identify key features affecting concentration levels. This helps us better understand air pollution in Taiwan. The background involves the deployment of AirBox devices across Taiwan, and previous research on air quality monitoring stations and innovative methods like MOB or MTSC has laid the foundation for this study. |
| URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/93635 |
| DOI: | 10.6342/NTU202402061 |
| Fulltext Rights: | 同意授權(限校園內公開) |
| metadata.dc.date.embargo-lift: | 2025-07-22 |
| Appears in Collections: | 應用數學科學研究所 |
Files in This Item:
| File | Size | Format | |
|---|---|---|---|
| ntu-112-2.pdf Access limited in NTU ip range | 10.36 MB | Adobe PDF |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
