| dc.contributor | 統計系 | |
| dc.creator (作者) | 吳漢銘;陳逸瑄;王鴻龍 | |
| dc.creator (作者) | Wu, Han-ming;Chen, Yi-hsuan;Wang, Hong-long | |
| dc.date (日期) | 2023-06 | |
| dc.date.accessioned | 24-May-2024 11:00:41 (UTC+8) | - |
| dc.date.available | 24-May-2024 11:00:41 (UTC+8) | - |
| dc.date.issued (上傳時間) | 24-May-2024 11:00:41 (UTC+8) | - |
| dc.identifier.uri (URI) | https://nccur.lib.nccu.edu.tw/handle/140.119/151244 | - |
| dc.description.abstract (摘要) | 電信信令資料記錄行動裝置的使用者於某時間點的地理位置資訊，具有空間分佈及非線性移動軌跡等多重特徵。相較於戶籍資料或旅運調查資料，運用電信信令資料的分析結果，可提供更即時的人流資訊。本研究結合矩陣視覺化之技術與象徵性資料分析方法於電信信令資料，針對行動裝置使用者的移動軌跡進行二階段的分析與探索，以了解人群與人群間的時空交互作用與群聚移動關係。我們首先採用動態時間校正法作為使用者間軌跡距離之度量指標，計算出使用者距離矩陣後，以矩陣視覺化的技術呈現分群的結果，初步找出移動模式相似的人群。為了解決大數據計算上的困難，我們應用區間型象徵性資料分析法，將相似人群之經緯度數值資料摘要成區間型經緯度資料，並以二維色階作為區間型經緯度資料矩陣視覺化的依據，呈現出實際地理位置的遠近關係與人群移動範圍之大小。藉由本研究針對電信信令資料所提出的人群移動視覺化分析程序，相信可以有效地分析處理大規模即時或長期時空資料，協助專家更精確的預測人流，了解人群移動特徵及其關聯性，並有助於各場域的應用。 | |
| dc.description.abstract (摘要) | Mobile phone data consist of the geographic locations of mobile device users at specific times, which can be considered spatiotemporal data. A variety of characteristics are present in these data, including spatial distribution and non-linear movement trajectories. Data from telecommunication signaling can provide more real-time information about human movements than household registration data or travel surveys. In this study, the matrix visualization (MV) and symbolic data analysis (SDA) are integrated and applied to mobile phone data to explore the movement patterns of mobile device users as well as the spatiotemporal interaction of crowds and the cluster movement relationship between them. First, we calculate the distance between the users' trajectories by using dynamic time warping (DTW). Following that, the distance matrix is subjected to clustering algorithms so that similar movement patterns can be presented using the MV technique. Rather than attempting to compute big data as a whole, the numerical longitudes and latitudes of a similar cluster are aggregated to create an interval-valued longitude and latitude. By utilizing a two-dimensional color spectrum, interval-based MV can visually display distance between the geographical location and the range of crowd movements. We believe the proposed method can effectively be applied to analyze and process large-scale real-time and long-term spatiotemporal data. Using these results, the experts are able to predict crowd flow more accurately, understand crowd movement characteristics and their correlation, and contribute to applications in various fields. | |
| dc.format.extent | 115 bytes | - |
| dc.format.mimetype | text/html | - |
| dc.relation (關聯) | 中國統計學報, Vol.61, No.2, pp.128-151 | |
| dc.subject (關鍵詞) | 資料視覺化; 時空資料; 軌跡分群; 人群移動樣態 | |
| dc.subject (關鍵詞) | Data visualization; Human mobility patterns; Spatiotemporal data; Trajectory clustering | |
| dc.title (題名) | 象徵性資料分析法於電信信令資料的矩陣視覺化與分群 | |
| dc.title (題名) | Matrix Visualization and Clustering of the Mobile Phone Data Based on Symbolic Data Analysis | |
| dc.type (資料類型) | article | |
