An Effective Taxi Recommender System Based on a Spatiotemporal Factor Analysis Model

An Effective Taxi Recommender System Based on a Spatiotemporal Factor Analysis Model

陳郁婷(Chen, Yuting)

指導教授 ： 黃仁竑

國立中正大學/工學院/資訊工程學系/碩士(2013年)

摘要

利用一星期的真實的GPS資料建立出一個計程車推薦系統，並以模擬一台虛擬計程車的方式評估推薦系統的效能。從我們的模擬數據可以發現用原模型建立的資料來進行模擬，根據系統推薦出的地點路線去等待客人的虛擬計程車司機的一天平均收入，在週間時，可以比99%的計程車司機都來得高，在週末時優於96%的計程車司機。在用非原模型建立的新的一星期的資料進行模擬時，虛擬計程車司機的一天平均收入在週間時，可以比94%的計程車司機都來得高，在週末時優於82%的計程車司機。故計程車司機使用本推薦系統可以得到比採用自身經驗來得更好的收入。

關鍵字

計程車 ； 資料探勘

並列摘要

We built a taxi recommender system based on a whole week real GPS data, and simulated a virtual taxi to evaluate the performance of the recommender system. We found that the revenue of the virtual taxi drivers who used the recommender system were 99% better than other taxi drivers in weekday, and also got 96% better than other taxi drivers in weekend. We built a taxi recommender system by a real GPS data in another week, and simultaneously simulated a virtual taxi to evaluate the stability and performance of our recommender system. We found that the revenue is 94% better than other taxi drivers in weekday and 82% better than other taxi drivers in weekend. As a
result, taxi drivers who use our taxi recommender system get more revenue than other taxi drivers which drive by their experience.

並列關鍵字

Taxi trajectories ； Spatiotemporal data mining

參考文獻

[2] Yang Yue, Yan Zhuang, Qingquan Li, and Qingzhou Mao. Mining Time-dependent Attractive Areas and Movement Patterns from Taxi Trajectory Data. In Geoinformatics, 2009.

[3] Y. Zheng, L. Zhang, X. Xie, and W.-Y. Ma. Mining interesting locations and travel sequences from GPS trajectories. In Proc. WWW, pp. 791–800, 2009.

[4] X. Cao, G. Cong, and C. Jensen. Mining Significant Semantic Location From GPS Data. In VLDB, pages 1009–1020, 2010.

[13] H. Cao, N. Mamoulis, and D. W. Cheung. Discovery of periodic patterns in spatiotemporal sequences. In TKDE’07

[14] M. Mun et. al.,Personal Data Vaults: A Locus of Control for Personal Data Streams, ACM CoNEXT, 2010