Searching for and categorizing emission line stars using hierarchical clustering algorithms and machine learning techniques
Supervisor: Dr. Carol Jones
Can extend to MSc?: Yes
Project Description (Abstract):
The project focuses on emission-line massive stars (Be stars) that have material distributed in a form of a disk surrounding the central star. A common characteristic of Be stars is their rapid rotation and this property must play a role in launching stellar material into orbit to form a disk. The formation, dissipation, and variability of these circumstellar disks are long-standing puzzles in this field of research.
Disks are ubiquitous in astrophysical phenomenon – our galaxy is a disk, planets form in disks, black holes are associated with disks and there are many more examples. Be star disk systems offer a valuable probe to better understand disk physics, to potentially to help answer some of these long-standing important questions described above and may offer clues about other systems with disks.
The OGLE (Optical Gravitational Lensing Experiment) data base was originally designed to look for dark matter in the Galactic bulge and Magellanic Clouds however stars between Earth and these targets were observed by OGLE and there is potentially a wealth of information within this database. In addition to this database the BeSS (Be Star Spectra) or WEBDA (Web version of Base Données Amas) databases could also be searched for emission stars.
The project will use hierarchical clustering algorithms and machine learning techniques in order to find and categorize emission lines based on their shapes and variability. The results of this work will be twofold: the development of tools to search astronomical databases and insight regarding the frequency and variability of massive stars with emission lines.
Minimal experience with computer programming is required.
Published on and maintained in Cascade CMS.