Taxi passenger hotspot prediction using automatic ARIMA model

As a transportation mode, taxis facing low occupancy rate problems at certain time and over-demand at another. This issue is due to imbalance of supply and demand of taxi service. According to some studies the imbalance caused by inefficient taxis distribution. Recently, there are data sources available by utilizing GPS technology that can be used to obtain spatio-temporal information. This spatio-temporal information is valuable to develop intelligent taxi system. There are a lot of previous research related to intelligent taxi system in several research topics, one of which on the analysis hotspot area. This study utilizes Automatic ARIMA Model to perform time-series analysis to predict the passenger's hotspot area based on the spatio-temporal data provided from local taxi firm in Bandung. The challenge of this research is to use the right method in the data-preprocessing phase so the Automatic ARIMA Model can process the spatio-temporal data. Some of the alternatives proposed for preprocessing phase and experimental results show that grid mapping method can be used well in the preprocessing phase. Research results show that the Automatic ARIMA can be used to conduct an analysis of the spatio-temporal data. This is indicated by Mean Absolute Scaled Error (MASE) value 0.8797 for New York City dataset and 0.6338 for Bandung dataset. Cross-validation analysis also showed satisfactory results when the actual demand is quite large. However, if the actual demand is close to zero, the result of the analysis becomes less reliable. This can be understood as a lack in the quality of the data, not in the prediction model.

[1]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1971 .

[2]  Dale G. Bails.  Business fluctuations : forecasting techniques and applications / Dale g. Bails, Larry C. Peppers , 1982 .

[3]  Walter Vandaele,et al.  Applied Time Series and Box-Jenkins Models , 1983 .

[4]  John C. Hoff,et al.  A practical guide to Box-Jenkins forecasting , 1983 .

[5]  Alan Pankratz,et al.  Forecasting with univariate Box-Jenkins models : concepts and cases , 1983 .

[6]  Billy M. Williams,et al.  Modeling and Forecasting Vehicular Traffic Flow as a Seasonal ARIMA Process: Theoretical Basis and Empirical Results , 2003, Journal of Transportation Engineering.

[7]  Guoqiang Peter Zhang,et al.  Time series forecasting using a hybrid ARIMA and neural network model , 2003, Neurocomputing.

[8]  Bo Zhou,et al.  Network Traffic Modeling and Prediction with ARIMA / GARCH , 2005 .

[9]  Rob J Hyndman,et al.  Another look at measures of forecast accuracy , 2006 .

[10]  Gyung-Leen Park,et al.  Analysis of the Passenger Pick-Up Pattern for Taxi Location Recommendation , 2008, 2008 Fourth International Conference on Networked Computing and Advanced Information Management.

[11]  Rob J Hyndman,et al.  Automatic Time Series Forecasting: The forecast Package for R , 2008 .

[12]  Gaetano Valenti,et al.  Traffic Estimation And Prediction Based On Real Time Floating Car Data , 2008, 2008 11th International IEEE Conference on Intelligent Transportation Systems.

[13]  Yu-chin Tai,et al.  iTaxi : Context-Aware Taxi Demand Hotspots Prediction Using Ontology and Data Mining Approaches , 2010 .

[14]  Jane Yung-jen Hsu,et al.  Context-aware taxi demand hotspots prediction , 2010, Int. J. Bus. Intell. Data Min..

[15]  Yue Wang,et al.  Hotspot District Trajectory Prediction , 2010, WAIM Workshops.

[16]  Kung-Sik Chan,et al.  Time Series Analysis: With Applications in R , 2010 .

[17]  Lin Sun,et al.  Hunting or waiting? Discovering passenger-finding strategies from a large-scale real-world taxi dataset , 2011, 2011 IEEE International Conference on Pervasive Computing and Communications Workshops (PERCOM Workshops).

[18]  Xing Xie,et al.  Where to find my next passenger , 2011, UbiComp '11.

[19]  João Gama,et al.  A predictive model for the passenger demand on a taxi network , 2012, 2012 15th International IEEE Conference on Intelligent Transportation Systems.

[20]  Gang Pan,et al.  Mining the semantics of origin-destination flows using taxi traces , 2012, UbiComp '12.

[21]  Michel Ferreira,et al.  On Predicting the Taxi-Passenger Demand: A Real-Time Approach , 2013, EPIA.

[22]  João Gama,et al.  Predicting Taxi–Passenger Demand Using Streaming Data , 2013, IEEE Transactions on Intelligent Transportation Systems.

[23]  Guntur C Purba.  ANALISA SUPPLY DAN DEMAND ANGKUTAN TAKSI DI KOTA MEDAN BERDASARKAN TINGKAT OKUPANSI DAN BIAYA OPERASIONAL KENDARAAN , 2014 .