Real time location prediction with taxi-GPS data streams

Abstract

The prediction of the destination location at the time of pickup is an important problem withpotential for substantial impact on the efficiency of a GPS-enabled taxi service. While this pro-blem has been explored earlier in the batch data set-up, we propose in this paper new solutions inthe streaming data set-up. We examine four incremental learning methods using a dampedwindow model namely, Multivariate multiple regression, Spherical-spherical regression,Randomized spherical K-NN regression and an Ensemble of these methods for their effectivenessin solving the destination prediction problem. The performance of these methods on several largedatasets are evaluated using suitably chosen metrics and they were also compared with someother existing methods. We found that the Multivariate multiple regression method has the bestperformance in terms of prediction accuracy but the Spherical-spherical regression method is thebest performer when we take into account the accuracy time trade-offcriterion. The next pickuplocation problem, where we are interested in predicting the next pickup location for a taxi giventhe dropofflocation coordinates of the previous trip as input is also considered and the afore-mentioned methods are examined for their suitability using real world datasets. As in the case ofdestination prediction problem, here also wefind that the Multivariate multiple regressionmethod gives better performance than the rest when we consider prediction accuracy but theSpherical-spherical regression method is the best performer when the accuracy-time trade-offcriterion is taken into account.