Study of the time-variation of the interstellar extinction with a machine learning method - Application to the variability analysis for future LSST data

par Mme Johanna Itam-Pasquet (LUPM)

mercredi 16 mars 2016, 14:00 → 15:00 Europe/Paris

Amphithéâtre (CPPM)

Light passing through interstellar clouds is significantly obscured by the dust they contain. One of the interesting questions in the study of the Interstellar Medium is the size of the smallest structure of these clouds. This structure could be hierarchical or fractal, extending down to small-sized clumps. Pfenniger et al. (1994) proposed that the baryonic dark matter associated with galaxies might consist of tiny gas globules. Such clumps would be extremely difficult to detect directly with traditional means (radio observations) due to their cold temperature and small size. However we can search for the transient obscuration of background stars. Such events are expected to be very rare, with much less than 1% of stars in any given direction being obscured at any time. Therefore it is important to search them in big databases by using automatic detection method. In this presentation, I will present the results of an analysis of the SLOAN STRIPE 82 light-curve data set to detect light curves consistent with extinction due to passing dark clouds. Our work is based on a machine learning method. Finally, I will show that the automatic classification methods could be relevant to the variability analysis of future LSST data, especially to the supernovae detection.

Slides seminaire_CPPM.pdf