When: February 22, 2019 @ 11:00am
Where: PAB, 3rd floor, B305
Disentangling Overlapping Astronomical Sources using Spatial, Spectral, and Temporal Information
We present a powerful new algorithm that combines spatial and spectral information to separate photons from overlapping X-ray sources. We use a Bayesian hierarchical modeling approach to simultaneously infer the number of overlapping sources, to probabilistically separate the photons among the sources, to fit the parameters describing the individual sources, and to coherently quantify the associated uncertainties. The advantages and utility of combining spatial and spectral information are demonstrated through a simulation study and a data analysis. Since many sources vary in intensity over time, our most recent work additionally aims to model the arrival times of the detected photons. We illustrate that this temporal information can further separate sources and allow us to detect transient phenomena.
Dr. Jones’ research interests include astrostatistics, hierarchical Bayesian modeling, machine learning methods, Monte Carlo methodology, and more. He has contributed work in exoplanet detection, light curve classification, and disentangling overlapping sources in spatial and spectral analysis.