| Title | : | Mining Trajectory Data for Smart City Applications |
| Speaker | : | Nandani Garg (IITM) |
| Details | : | Fri, 20 Oct, 2017 4:00 PM @ A M Turing Hall |
| Abstract: | : | Better public transportation is the major factor for underpinning economic growth. We study two problems in the area of public transportation. Firstly, we study the problem of route recommendation to idle taxi drivers such that the distance between the taxi and an anticipated customer request is minimized. To anticipate when and where future customer requests are likely to come from and accordingly recommend routes, we develop a route recommendation engine called MDM: Minimizing Distance through Multi-armed Bandit. In contrast to existing techniques, MDM employs a continuous learning platform where the underlying model to predict future customer requests is dynamically updated. Extensive experiments on real taxi data from New York and San Francisco reveal that MDM is up to 75% better than the state of the art and robust to anomalous events such as concerts, sporting events, etc. Secondly, to design an efficient bus transit system, a fundamental requirement is the complete list of all bus stops in the city. Often bus stop information is incomplete, erroneous, and outdated leading to sub-optimal planning and operations and consequent reduction in the transit agency’s ridership. For this, we propose an algorithm to mine bus stops automatically from bus GPS trajectories. The proposed technique is powered by a novel combination of feature mining with classification algorithms to predict bus stops in a city. Our technique negates the need for manual field visits to annotate bus stops and saves time and cost for transit agencies. We perform extensive empirical analysis on real datasets from Chennai, India and firmly establish the reliability of our technique by achieving an accuracy of almost 100%. |
