Searching time period-based longest frequent path in big trajectory data

Trajectories contain considerable routing information, and trajectory-based routing gains the attention of researchers recently. This paper argues the problem of searching time period Longest Frequent Path(TPLFP). The TPLFP, as one of the reasonable alternative Most Frequent Path (MFP), is close to MFP enough and maximizes the number of frequent route segments. First, we define TPLFP formally. To acquire accurate path frequencies, we describe the footmark, the footmark graph notations and design the Linear Sketch Footmark Index (LSFI) to speed up extracting proper footmark and constructing related footmark graph. Next, we develop a Best-First search algorithm with four pruning strategies. Next, we give an advanced footmark graph and its building algorithm. The extensive experiments demonstrate the effectiveness and efficiency of our index schemes and algorithms, which can find TPLFP results in expected response time.