Last update: : Monday 1 August 2016, by
Par : Victor Planas-Bielsa, Centre Scientifique de Monaco, Laboratoire International Associé (LIA 647 BioSensib)
Date : Jeudi 6 octobre 2016 à 13h
Lieu : IPHC, Amphithéâtre Grünewald, bâtiment 25
One of our main goals is to develop a method to identify automatically and obtain new information on the breeding cycle of the Adélie penguin (Pygoscelis adeliae), using a statistical and learning machine approach. So far, more than 80% of the breeding couples from the study colony in Pétrels Island (Dumont D’Urville, Terre Adélie) are equipped with a RFID transponder. This colony is only accessible through 3 gateways equipped with automatic identification and weighing systems. The data generated by our systems allow us to reconstruct individually not only the exact timing of the foraging trips at sea during the breeding cycle, but also the body reserves before and after of those trips.
Instead of focusing in a few individuals, we perform a fully statistical approach at the population level using the whole database of approximately 1 million lines.
In this talk we will describe the system, and explain how our statistical approach can provide us with new information that would be otherwise much harder to obtain at the whole colony scale.