Combining Clustering and Classification methods for Galaxy Morphology Identification

Dias, Ian Alton (2024) Combining Clustering and Classification methods for Galaxy Morphology Identification. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

Galaxy Morphology is the study of galaxies based on their shapes and structures. Traditional research primarily uses classification or clustering techniques to categorize galaxies into distinct groups such as Elliptical and Spiral Galaxies. Currently as more telescopic surveys are planned to be launched, the challenge that is faced by astronomical scientists, is to classify these huge amounts of data for further research, although supervised techniques do work well, scientists have shown concern when working with human labelled data due to potential biases. Thus, this research proposes a new machine learning framework to potentially reduce human annotations in data by combining Image Classification and Clustering techniques. The framework combines a Hierarchical Clustering technique using the Entropy, Gini and Gradient Moment (EGG) coefficients, with Neural Networks. The research will conduct two tests, by implementing the Hierarchical Clustering using HDBSCAN, paired with the classification model EfficientNetB0, and for the second test combining Self Organising Maps along with a CNN architecture. The SDSS catalog containing approximately 670,000 galaxy jpeg images and FITS data, out of which approximately 13,452 of both, comprising nearly 50% of each class, will be used to conduct this research, pertaining to the limited availability of resources and time constraints. The results show the hdbscan+efficientNetb0 and Som+CNN frameworks giving an accuracy of 90% and 93% respectively.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Stynes, Paul UNSPECIFIED Jilani, Musfira UNSPECIFIED
Uncontrolled Keywords:	Galaxy Morphology; Classification; Clustering; HDBSCAN; EfficientNetB0; SOM; CNN
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QB Astronomy Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	15 Aug 2025 17:42
Last Modified:	15 Aug 2025 17:42
URI:	https://norma.ncirl.ie/id/eprint/8554

Actions (login required)

View Item