Data-driven classification of animal trajectories

Data science is revolutionizing our understanding in a wide variety of fields, such as speech recognition, autonomous driving and protein folding. This has been possible due to dramatic improvements in computational power and data collection and storage. In the case of animal movement, thanks to the development and deployment of tracking devices during the past decades, there exists now a considerable number of animal trajectories, susceptible of being analyzed via data-driven methods. However, its potential remains generally unexplored under these novel techniques. Our goal will be to assess the performance and adequateness of data-based tools in marine animal movement. Here, we will consider the species classification task as a proxy. Specifically, we will evaluate the effectiveness of several algorithms representative of the state of the art, as well as propose modifications in the input that could improve their ability to gain insight from the data.