Adapting Wavelet-based Techniques to Photometrically Classify Cosmological Samples of Supernova

The Pan-STARRS Medium Deep Survey hosted at the Space Telescope Science Institute is the best source of archival multi-band photometry prior to the Large Synoptic Survey Telescope (LSST). Working with a multi-faceted classification pipeline, called SNMachine, we classified different supernovae based on their light curves. We first trained a classifier, named ‘Random Forest’, on simulated data, for which the different types of the supernovae were known, and then applied the classifier to real data, using the largest homogeneous sample of photometric SNIa. This made it possible to determine the probability of each supernova to be classified as type Ia. The classified sample was then used to construct a Hubble diagram in order to measure the rate of expansion of the universe, as well as the equation of state of dark energy.