A machine learning framework for Astronomical Light Curve Analysis and Classification applied to the STILT database
Transient and Time Domain Astronomy: Robotic telescopes, surveys and the evolution of transient phenomena
Paul R. McWhirter
Liverpool John Moores University
Paul R. McWhirter (1,2), Iain A. Steele (2), Dhiya Al-Jumeily (1), Abir Hussain (1) and Paul Fergus (1)
Modern time-domain astronomy is capable of collecting a staggeringly
large amount of data on millions of objects in real time. Therefore
the production of methods and systems for the automated classification
of time-domain astronomical objects is of great importance. The
Liverpool Telescope has a number of wide-field image gathering
instruments mounted upon its structure named the Small Telescopes
Installed at the Liverpool Telescope (STILT). These instruments have
been in operation since March 2009 gathering data of large areas of
sky around the current field of view of the main telescope. We applied
a method designed to extract time-translation invariant features from
the light curves of each object for future input into a classification
system. These efforts were met with limited success due to noise and
uneven sampling within the data. Therefore we propose a methodology
capable of distributing the data analysis for improved signal
detection resolution. Simultaneously an intelligence service extracts
features from the resulting light curves for the production of learned
models for high accuracy classification. This framework will be highly
scalable whilst maintaining the production of accurate features based
on the fitting of harmonic models to the light curves and the
generation of abstracted feature maps within the STILT database.


09:00 - 10:30
BS - Lecture Theatre A25 (121)