Benjamin KLOTZ - Data Science
Date: May 4th 2017 Location: Eurecom - Eurecom
Abstract: Connected vehicles and people generate a lot of location-based data, which generally amounts to timestamp GPS information. In the literature, some researchers have looked into trajectory mining: the extraction of knowledge from a dataset of trajectories such as the main patterns and trajectory density. If a semantically annotated map is available and mapped to the GPS traces, it is possible to do semantic trajectory mining and have many more information to mine. This allows a number of applications in various domains from urban planning to ecology but this also raises a number of challenges. First, a pre-processing step is necessary to have clean and usable data. Second, an adapted mining algorithm must be chosen, and finally one must adapt the data to fit the mining method. In this talk, we will present the most commonly used methods in semantic trajectory mining as well as their applications. Bio : Benjamin Klotz got a Master on ARIA (Control Engineering, Robotics and Applied Informatics) from the Ecole Centrale of Nantes, France in 2015. Since April 2016, Benjamin is a PhD student in BMW Research and EURECOM working on Semantic Technologies for Vehicle Data.