臺灣博碩士論文加值系統

臺北市自98年推動公共租賃自行車系統以來，自行車租賃次數已突破1億人次，平均每日租借次數更高達5萬次以上，對於高使用率的公共自行車租賃系統，租賃站缺車問題成了民眾最關心的潛在問題，如果無法在預期時間及地點租車，可能會造成金錢、時間等各方面的成本浪費。
本研究收集了臺北市公共租賃自行車系統107年1月下旬約922萬筆的自行車租賃站即時監測資料，再配合氣候、時間、空間等各項環境資料，運用資料探勘的決策術、羅吉斯迴歸、隨機森林及類神經網路等方法，預測未來一定時間內自行車租賃站是否會發生缺車問題，結果顯示預測效能較佳的是以隨機森林所建立的模型，針對未來5分鐘租賃站是否發生缺車的問題AUC達0.8485。此外，為能正確得知實際影響自行車缺車問題的原因，本研究亦將各變項進行卡方統計量分析，結果顯示影響最大的變項為租賃站即時週轉率、時段、降水量及鄰近租賃站租借率。透過本研究結果可大幅降低未來缺車問題的不確定性，亦可依照即時預測的缺車狀況，針對預定行程進行適度的調整。

There are more than 100 millions rental records since Taipei City adapted Public bicycle system (PBS) in 2009, and with 50 thousand daily ridership on average.
In terms of the increasing occupation coefficient of PBS, i.e., growing bike rental demand, the lacks of bicycles in each rental station became a concerned issue. This research aims to define models that can well describe the patterns of bike scarcity in various conditions.
This study collected 9.22 million records from Taipei public bike system in 2018.01. The data were compared with the weather condition, time slot, and spatial distribution. Various analysis theories such as Decision Tree, Logistic regression, random forest, and neural network were applied in this study to model the bike scarcity patterns. In our results, random forest revealed a better reliability on forecasting bike unavailable situation within the next five minutes, the AUC exceeded 0.8485. Chi-square was also adapted to evaluate the influence of this experiment. Turnover rate, time slot, rainfall, and adjacent station are the most significant criteria. In summary, this study developed the bike scarcity forecasting model that provides an acceptable explanatory ability. Furthermore, this model can be expected to apply for various applications such as bike tour suggestion system, administrative management system.

誌謝 i
摘要 i
Abstract ii
目錄 iii
圖目錄 vi
表目錄 vii
第1章 緒論 1
1.1 研究背景 1
1.2 研究動機 2
1.3 研究目的 3
第2章 文獻探討 4
2.1 臺北市公共自行車 4
2.1.1 租借費率說明 4
2.1.2 自行車租賃次數及租賃站歷程 5
2.1.3 租賃站分佈 6
2.2 影響使用者租賃行為因素相關研究 7
2.2.1 租賃站租借、環境因素 7
2.2.2 氣候因素 8
2.2.3 時間因素 8
2.2.4 空間因素 9
2.3 自行車使用行為預測相關研究 12
2.4 總結 18
第3章 研究方法 19
3.1 資料收集 20
3.1.1 租賃站即時車輛數資料(站位資料) 21
3.1.2 臺北市YouBike租借資料 21
3.1.3 全台路網資料(shp.) 21
3.1.4 氣象觀測資料(csv) 22
3.1.5 臺北捷運站點位資料(shp.) 23
3.1.6 臺北市公車站牌位置圖(shp.) 24
3.1.7 臺北市各級學校位置(csv.) 24
3.1.8 臺北市景點位置資料(xml.) 25
3.1.9 臺北市公司設立登記清冊(csv.) 25
3.1.10 各鄉鎮市區人口密度(csv.) 26
3.2 資料前處理 26
3.2.1 結構化資料前處理 27
3.2.2 空間資料前處理 28
3.2.3 資料綜合處理 31
3.3 研究變項說明 32
3.3.1 租賃站位址相關 33
3.3.2 租賃站租借相關 33
3.3.3 時間相關變項 33
3.3.4 氣候相關變項 33
3.3.5 空間鄰近相關變項 34
3.4 機器學習技術 35
3.4.1 決策樹(Decision Tree) 35
3.4.2 羅吉斯迴歸(Logistic regression) 36
3.4.3 類神經網路(Neural network) 36
3.5 研究工具說明 36
3.5.1 Orange 36
3.5.2 ArcGIS 37
3.6 實驗設計與評估 37
3.6.1 實驗設計 37
3.6.2 驗證方式 40
3.6.3 評估指標 41
第4章 實驗結果與分析 43
4.1 實驗資料 43
4.2 實驗結果 44
4.3 變項重要性排序 47
4.4 綜合討論 49
第5章 結論與建議 51
5.1 研究結論與貢獻 51
5.2 研究限制 52
5.3 未來研究方向與建議 52
參考文獻 54
附錄 57

中文部分
左宏昌、歐崇明。（民105）。公共自行車租賃站狀態應用馬可夫鏈模型之大數據預測。第十五屆離島資訊技術與應用研討會，高雄。
賴淑芳。（民103）。公共自行車租賃系統之城市行銷與產品定位－以臺北微笑單車為例。都市交通，20-31。
鄭淑麗。（民107）。臺北市公共自行車使用特性。107年統計精進與推展研討會，臺北市。
鍾智林、李舒媛。（民107）。以悠遊卡大數據初探 YouBike 租賃及轉乘捷運行為。都市交通，6-36。
陳宥伃、林承萓、廖邕（民105）。公共自行車使用行為之研究動向。休閒與社會研究，201–210。
黃晏珊、鍾智林 民104）。公共自行車系統營運特性大數據分析-以台北YouBike為例。基於美好生活的交通綜合治理。第二十三屆海峽兩岸都市交通學術研討會，新北市。

英文部分
Alexander, R., & Rixey, R. A. (2013). Station-Level Forecasting of Bike Sharing Ridership: Station Network Effects in Three U.S. Systems. Transportation Research Record: Journal of the Transportation Research Board, (2387), 46–55.
Caulfield, B., O'Mahony, M., Brazil, W., & Weldon, P. (2017). Examining usage patterns of a bike-sharing scheme in a medium sized city. Transportation research part A: policy and practice, 100, 152-161.
Chen, L., Zhang, D., Wang, L., Yang, D., Ma, X., Li, S., Wu, Z., et al. (2016). Dynamic Cluster-Based Over-Demand Prediction in Bike Sharing Systems. Proceedings of the 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing (pp. 841–852).
El-Assi, W., Mahmoud, M. S., & Habib, K. N. (2017). Effects of built environment and weather on bike sharing demand: a station level analysis of commercial bike sharing in Toronto. Transportation, 44(3), 589-613.
Faghih-Imani, A., & Eluru, N. (2016). Incorporating the impact of spatio-temporal interactions on bicycle sharing system demand: A case study of New York CitiBike system. Journal of Transport Geography, 54, 218-227.
Gebhart, K., & Noland, R. B. (2014). The impact of weather conditions on bikeshare trips in Washington, DC. Transportation, 41(6), 1205–1225.
Giot, R., & Cherrier, R. (2014). Predicting Bikeshare System Usage Up to One Day Ahead. IEEE Symposium Series in Computational Intelligence 2014 (SSCI 2014), 2–9.
Kou, Z., & Cai, H. (2019). Understanding bike sharing travel patterns: An analysis of trip data from eight cities. Physica A: Statistical Mechanics and its Applications, 515, 785-797.
Lin, L., He, Z., & Peeta, S. (2018). Predicting station-level hourly demand in a large-scale bike-sharing network: A graph convolutional neural network approach. Transportation Research Part C: Emerging Technologies, 97, 258-276.
Mattson, J., & Godavarthy, R. (2017). Bike share in Fargo, North Dakota: Keys to success and factors affecting ridership. Sustainable Cities and Society, 34(March), 174–182.
Singhvi, D., Singhvi, S., Frazier, P. I., Henderson, S. G., O'Mahony, E., Shmoys, D. B., & Woodard, D. B. (2015, January). Predicting Bike Usage for New York City's Bike Sharing System. In AAAI Workshop: Computational Sustainability.
Tan, P.-N., Steinbach, M., Karpatne, A., & Kumar, V. (2018). Introduction to Data Mining (2nd ed.). Pearson.
Westland, J. C., Mou, J., & Yin, D. (2018). Demand cycles and market segmentation in bicycle sharing. Information Processing and Management, (June), 1–13. Elsevier. Retrieved from https://doi.org/10.1016/j.ipm.2018.09.006
Xu, C., Ji, J., & Liu, P. (2018). The station-free sharing bike demand forecasting with a deep learning approach and large-scale datasets. Transportation Research Part C: Emerging Technologies, 95, 47–60. Retrieved November 23, 2018, from https://www.sciencedirect.com/science/article/pii/S0968090X18306764
Zhang, Y., Thomas, T., Brussel, M., & van Maarseveen, M. (2017). Exploring the impact of built environment factors on the use of public bikes at bike stations: Case study in Zhongshan, China. Journal of Transport Geography, 58, 59–70. Retrieved November 23, 2018, from https://www.sciencedirect.com/science/article/pii/S0966692316300412