Finding structures in integral of motion space of the Milky Way halo and proving their statistical significance

Lövdal, Sofie (2021) Finding structures in integral of motion space of the Milky Way halo and proving their statistical significance. Master's Thesis / Essay, Computing Science.

Preview	Text mCS_2021_SSLovdal.pdf Download (31MB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (100kB)

Abstract

Galaxies grow hierarchically by accreting smaller structures and stars in the acquired object will retain similar orbital parameters: the integrals of motion. Clustering is non-trivial in this space. This thesis develops a data-driven and statistically based method for finding clusters in integral of motion space of the Milky Way halo, together with evaluating their significance. Our clustering method is based on an exhaustive use of the single linkage algorithm using four features: energy, angular momentum, angular momentum in z-direction and circularity. We use Poisson statistics to determine the significance in overdensity of each cluster compared to an artificial reference halo and expected density in the same region. Furthermore, we apply the HDBSCAN algorithm in velocity space to get an indication of substructure within a cluster. We also determine a membership probability by modeling the stars and clusters as probability density functions in our clustering space, given the measurement uncertainties in the underlying data. We apply our methods on the Gaia eDR3 RVS sample within 5 kpc. The results assign 55% of the stars in our data set to some significant cluster, with 419 clusters detected in total. The majority of clusters depict clear substructure in velocity space, with clusters being divisible into 1-4 components. Further analysis is needed to determine the precise properties of the detected clusters and their correspondence to previously established structures.

Item Type:	Thesis (Master's Thesis / Essay)
Supervisor name:	Biehl, M. and Helmi, A.
Degree programme:	Computing Science
Thesis type:	Master's Thesis / Essay
Language:	English
Date Deposited:	17 Mar 2021 16:08
Last Modified:	17 Mar 2021 16:08
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/24089

Actions (login required)

View Item