Learning to predict the duration of an automobile trip

摘要: In this paper, we explore the use of machine learning and data mining to improve prediction travel times in an automobile. We consider two formulations problem, one that involves predicting speeds at different stages along route another relies on direct transit time. focus second formulation, which apply collected from San Diego freeway system. report experiments these with k-nearest neighbour combined a wrapper select useful features normalization parameters. The results suggest 3-nearest neighbour, when using information sensors, substantially outperforms predictions available existing digital maps. Analyses also reveal some surprises about usefulness other like time day trip.