Astronomy and Astrophysics – Astronomy
Scientific paper
Jan 2010
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2010aas...21547703d&link_type=abstract
American Astronomical Society, AAS Meeting #215, #477.03; Bulletin of the American Astronomical Society, Vol. 42, p.564
Astronomy and Astrophysics
Astronomy
Scientific paper
In modern synoptic sky surveys, reliable and robust real-time classification of transient events is essential in order to exploit the full scientific potential of these surveys, and conduct optimized follow-up observations and additional measurements. This generally has to be done using sparse and heterogeneous data for individual events, both from the survey pipelines and existing archives. Besides, in order to catch some of the more interesting objects, the classification may need to be done using a minimal number of observations available. Here we describe a fusion network that combines the results from the different kinds of classifiers we are building (eg, Bayesian Networks, GPR, Neural Networks) that can be triggered by a specific event. Two of such classification methods involve Bayesian Networks with colors and contextual information, such as distance from a galaxy and galactic latitude and Gaussian Process Regression with lightcurves. Both are fast enough to be used in a real time pipeline and can deal with missing data and work with a small number of epochs. The high order fusion module can help to improve the classification, narrowing down the number of competing classes and leading to optimal follow-ups and new discoveries.
Djorgovski Stanislav G.
Donalek Ciro
Drake Andrew J.
Graham Mary Jane
Mahabal Ashish A.
No associations
LandOfFree
Real Time Classification of Transient Events in Synoptic Sky Surveys does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Real Time Classification of Transient Events in Synoptic Sky Surveys, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Real Time Classification of Transient Events in Synoptic Sky Surveys will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1887817