fairfolio
Saved Projects

Identification of Satellite Galaxies around Milky Way Galactic Analogs Using Machine Learning Algorithms

CSEF · 2014 Physics & Astronomy

Vikas C. Bhetanabhotla · Harker (The) School, San Jose, Santa Clara County

Overview

Objectives/Goals The Milky Way (MW) is host to a population of two dozen satellite galaxies. However, observed properties of these MW satellites differ significantly from those predicted by the Lambda Cold Dark Matter Theory (LCDM), constituting the discrepancy at the center of the Dwarf Galaxy Problem. To investigate this discrepancy, we extend our cosmological context and study satellite populations of MW analog host galaxies within redshift < 0.01, selected to resemble the MW in luminosity and environment. Studying such analogous satellite populations and their properties will yield insight into their correlation with the LCDM model, which can be used to determine where and why the MW satellites break off with the LCDM model. A significant obstacle is the detection of these MW analog satellite galaxies against a large background of faint galaxies and stars: spectroscopic follow-up is expensive, and so accurate classification of satellite candidates based on photometric and morphological data is important. Methods/Materials We compile a machine learning training set containing photometric and morphological data for the MW analog satellite populations, and train a Neural Network (NN) and Support Vector Machine (SVM) classifier on a pre-processed version of the data set. We measure effectiveness of the algorithms in terms of purity and completeness of classification. Results Through a series of experiments with the classifiers, we find the SVM to achieve 93% completeness and the NN to achieve 97% completeness for a fixed 10% purity. We also find the photometric data to be the most informative in classification. Conclusions/Discussion We thus present a more accurate method of identifying MW analog satellite galaxies among a large sample of background galaxies, a major stepping-stone in solving the Dwarf Galaxy Problem.

Summary statement

My research presents a new method to identify pure samples of Milky Way analog satellite populations with accuracy in identification of 97%; this method can be used to solve the Dwarf Galaxy Problem.

Help received

This research was conducted at the Stanford University astrophysics department, where I worked under the guidance of a professor and two postdoctoral fellows. I received guidance from my professor throughout the project and help from the postdoctoral fellows in learning necessary techniques.

Competition history

  • CSEF 2014 Physics & Astronomy · Entry S1803

Resources

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

Source: California Science & Engineering Fair public projects

Save projects to your library

Sign in with Google to keep track of projects you find interesting, organized into folders. An account also raises your daily allowance for “Has this been done?”, and lets you create a key for the MCP server with a much higher limit than anonymous use. Browsing stays public.

Continue with Google