Taxi Demand Forecasting: A HEDGE-Based Tessellation Strategy for Improved Accuracy

A key problem in location-based modeling and forecasting lies in identifying suitable spatial and temporal resolutions. In particular, judicious spatial partitioning can play a significant role in enhancing the performance of location-based forecasting models. In this paper, we investigate two widely used tessellation strategies for partitioning city space, in the context of real-time taxi demand forecasting. Our study compares (1) the Geohash tessellation and (2) the Voronoi tessellation, using two distinct taxi demand data sets, over multiple time scales. For the purpose of comparison, we employ classical time-series tools to model the spatio-temporal demand. Our study finds that the performance of each tessellation strategy is highly dependent on the city geography, spatial distribution of the data, and the time of the day, and that neither strategy is found to perform optimally across the forecast horizon. We propose a combining algorithm that selects the best tessellation strategy at each time step, based on their recent performance. Our algorithm is a non-stationary variant of the well-known HEDGE algorithm for choosing the best advice from multiple experts. We show that the proposed strategy performs consistently better than either of the two tessellation strategies across the data sets considered, at multiple time scales, and with different performance metrics. We achieved an average accuracy of above 80% per km
²
for both data sets considered at 60 min aggregation levels.