| Abstract : | The discovery and classification of Young Stellar Objects (YSOs) are essential for understanding the star formation process. Traditional approaches to identification of YSOs have relied primarily on broadband color criteria derived using color-color (CCD) and color-magnitude diagrams (CMD). However, the classifications from these methods are quite ambiguous due to the overlap of different source types in CCDs and CMDs and do not state the confidence in classification. Here, we employ machine and deep learning algorithms to probabilistically classify YSOs, Active Galactic Nuclei (AGN), Asymptotic Giant Branch (AGB) and main-sequence (MS) stars. We simultaneously use photometric colors and magnitudes in three near-infrared (2MASS J, H, and Ks) and four mid-infrared bands (AllWISE W1, W2, W3, and W4) for classification. Our approach also sub-classifies YSOs into Class I, II, III and flat spectrum YSO and AGB stars into carbon-rich (C AGB) and oxygen-rich (O AGB) AGB stars. We adopt an ensemble approach to account for the uncertainty in photometric measurements during classification. We apply our ensemble of 300 classifiers to nearby spatially isolated targets compiled in preparation for NASA’s upcoming space mission Spectro-Photometer for the History of the Universe, Epoch of Reionization and Ices Explorer (SPHEREx), i.e., infrared-bright sources likely to include a large proportion of YSOs. Our classification indicates that out of 8,426,242 sources, 1,046,344 have class predictions with probability exceeding 90%, amongst which ∼58% are YSOs. Amongst the rest ∼38% are AGB stars, ~3% are (reddened) MS stars, and ~2,000 sources are AGNs whose red broadband colors mimic YSOs. We validate our classification by analyzing the spatial distribution of identified candidate YSOs towards the Cygnus-X star-forming complex. We estimate the distance to the Cygnus-X star-forming complex using Gaia DR3 parallax measurements of candidate YSOs attributed to Cygnus-X and obtain a value of 1.47±0.48 kpc in-line with previous estimates. |
