Monday

We present a supervised machine learning methodology to classify stellar populations in the Local Group dwarf-irregular galaxy NGC 6822 (1”≈2.4pc). Given its isolation and sub-solar metallicity, NGC6822 is an undisturbed analogue to the Small Magellanic Cloud.

Near-IR colours (J-H, H-K, J-K), K-band magnitudes and far-IR surface brightness (70 and 160 μm) measured from Spitzer and Herschel images are used as features to train a Probabilistic Random Forest classifier (PRF). Point-sources are classified into eight target classes: young stellar objects (YSOs), oxygen- and carbon-rich asymptotic giant branch stars, red giant branch and red super-giant stars, massive main-sequence stars, Galactic foreground stars and active galactic nuclei. The PRF identifies sources with an accuracy ∼90 per cent across all target classes rising to ∼96 per cent for YSOs.

We confirm the nature of 129 out of 270 literature YSO candidates with enough feature information, and identify 172 new YSOs and candidates, mostly but not exclusively located in known star forming regions. These YSOs have mass estimates between ∼15 − 50 M☉, representing the most massive YSO population in NGC 6822. Another 69 out of 270 candidates are definitively classified as non-YSOs by the PRF analysis. We characterise the star formation environment by comparing the spatial distribution of YSOs to those of gas and dust using archival images.

We also explore the potential of using (unsupervised) t-distributed stochastic neighbour embedding maps for the identification of the same stellar population classified by the PRF.

The techniques used in NGC6822 are being applied in the Local Group spiral galaxy M33 (1”≈4pc) with the aims of identifying star forming regions and resolving YSOs for the first time across the whole disc of another spiral galaxy, and characterising their properties in the context of the galaxy structure and environment. We present preliminary results of this analysis.