Output details
11 - Computer Science and Informatics
University College London
Mining mobility data to minimise travellers' spending on public transport.
<15> We mine near 20 million journeys done using Oyster RFID cards in the London's public transport network and discover an overspending, due to incorrect fare purchase, of over 2 million GBP in just 3 months. We propose a data mining technique to dynamically learn travelers’ movement patterns and predict what ticket fare to purchase, obtaining an accuracy of 98% (quantifiable in millions of pounds savings). The work has been covered by BBC News (http://www.bbc.co.uk/news/uk-england-london-13389363). I gave one workshop keynote (LBSN 2011) and one Advanced Distinguished Lecture at the China Computer Federation based on this work. Conference acceptance rate: 17.5%