Machine learning identification of massive young stellar objects in Local Group galaxies

Kinson, David Andrew

Abstract

This thesis presents the development and implementation of a machine learning classification of massive Young Stellar Objects (YSOs) in two Local Group galaxies, NGC6822 and M33. Using archival near- and far-IR data, point sources in both galaxies are classified into multiple stellar classes using a Probabilistic Random Forest classifier (PRF) trained on objects of known types. The spatial distributions of all classes are discussed. YSOs are classified with a high level of confidence (up to 97 per cent) in both galaxies. In NGC6822, 125 YSOs are confirmed and 199 are newly identified. All major star forming regions (SFRs) in NGC6822 are recovered and, additionally smaller SFRs are newly identified. In M33 4985 YSOs were identified across the disk of M33 and, applying a density-based clustering analysis 68 SFRs were identified primarily in the galaxy’s spiral arms. SFRs associated with known Hii regions were recovered, with ∼91 per cent of SFRs spatially coincident with giant molecular clouds identified in the literature. Using photometric measurements, as well as SFRs in NGC6822 with an established evolutionary sequence as a benchmark, I employed a novel approach combining, into one metric, ratios of [Hα]/[24μm] and [250μm]/[500μm] to estimate the relative evolutionary status of all M33 SFRs. By comparing the YSOs identified in M33 with model grids for mass determination, a star formation rate is estimated for the first time from direct YSO counts; (1.42±0.16M⊙ yr−1) that is lower than that of the more massive Milky Way as expected. This project for the first time identifies massive YSOs on galactic scales in a Local Group spiral galaxy, extending such analysis beyond the nearby star-forming dwarf galaxies (LMC, SMC and NGC6822). The techniques developed offer an invaluable tool for classifying large data sets.

Citation

Kinson, D. A. (2023). Machine learning identification of massive young stellar objects in Local Group galaxies. (Thesis). Keele University

• Thesis Type: Thesis
• Deposit Date: Jul 11, 2023
• Publicly Available Date: Jul 11, 2023
• Award Date: 2023-06