The weirdest SDSS galaxies: results from an outlier detection algorithm

Abstract

How can we discover objects we did not know existed within the large datasets that now abound in astronomy? We present an outlier detection algorithm that we developed, based on an unsupervised Random Forest. We test the algorithm on more than two million galaxy spectra from the Sloan Digital Sky Survey and examine the 400 galaxies with the highest outlier score. We find objects which have extreme emission line ratios and abnormally strong absorption lines, objects with unusual continua, including extremely reddened galaxies. We find galaxy-galaxy gravitational lenses, double-peaked emission line galaxies, and close galaxy pairs. We find galaxies with high ionisation lines, galaxies which host supernovae, and galaxies with unusual gas kinematics. Only a fraction of the outliers we find were reported by previous studies that used specific and tailored algorithms to find a single class of unusual objects. Our algorithm is general and detects all of these classes, and many more, regardless of what makes them peculiar. It can be executed on imaging, time-series, and other spectroscopic data, operates well with thousands of features, is not sensitive to missing values, and is easily parallelisable.

MNRAS accepted. Weirdness score for all SDSS galaxies at http://www.wise-obs.tau.ac.il/~dovip/weird-galaxies/ and code available at https://github.com/dalya/WeirdestGalaxies

References in corpus (17)

• The Host Galaxies and Classification of Active Galactic Nuclei
• Astrophysics Source Code Library
• The MAPPINGS III Library of Fast Radiative Shock Models
• Evolutionary Stellar Population Synthesis with MILES. Part I: The Base Models and a New Line Index System
• Kiloparsec-scale outflows are prevalent among luminous AGN: outflows and feedback in the context of the overall AGN population
• The Sloan Lens ACS Survey. V. The Full ACS Strong-Lens Sample
• Photometric and Spectroscopic Properties of Type II-P Supernovae
• Collisional heating as the origin of filament emission in galaxy clusters
• The chemical diversity of exo-terrestrial planetary debris around white dwarfs
• A Catalog of Local E+A(post-starburst) Galaxies selected from the Sloan Digital Sky Survey Data Release 5
• Estimation of stellar atmospheric parameters from SDSS/SEGUE spectra
• A unified explanation for the supernova rate-galaxy mass dependency based on supernovae discovered in Sloan galaxy spectra
• Machine learning for transient discovery in Pan-STARRS1 difference imaging
• A sample of Seyfert-2 galaxies with ultra-luminous galaxy-wide NLRs -- Quasar light echos?
• Multivariate Classification with Random Forests for Gravitational Wave Searches of Black Hole Binary Coalescence
• Active Galactic Nuclei with Double-Peaked Balmer Lines: I. Black Hole Masses and the Eddington ratios
• A Chandra Look at Five of the Broadest Double-Peaked Balmer-Line Emitters

Cited by in corpus (14)

• Surveying the reach and maturity of machine learning and artificial intelligence in astronomy
• An automatic taxonomy of galaxy morphology using unsupervised machine learning
• Extragalactic Radio Continuum Surveys and the Transformation of Radio Astronomy
• Black hole mass estimation for Active Galactic Nuclei from a new angle
• Discovering the Unexpected in Astronomical Survey Data
• Systematic Serendipity: A Test of Unsupervised Machine Learning as a Method for Anomaly Detection
• The use of convolutional neural networks for modelling large optically-selected strong galaxy-lens samples
• Evidence of ongoing AGN-driven feedback in a quiescent post starburst E+A galaxy
• Automated novelty detection in the WISE survey with one-class support vector machines
• Quasar Lenses and Galactic Streams: Outlier Selection and GAIA Multiplet Detection
• Neural network-based anomaly detection for high-resolution X-ray spectroscopy
• NuSTAR hard X-ray data and Gemini 3D spectra reveal powerful AGN and outflow histories in two low-redshift Lyman-$\alpha$ blobs
• Dimension Estimation Using Autoencoders
• ConvoSource: Radio-Astronomical Source-Finding with Convolutional Neural Networks