On Machine-Learned Classification of Variable Stars with Sparse and Noisy Time-Series Data

Abstract

With the coming data deluge from synoptic surveys, there is a growing need for frameworks that can quickly and automatically produce calibrated classification probabilities for newly-observed variables based on a small number of time-series measurements. In this paper, we introduce a methodology for variable-star classification, drawing from modern machine-learning techniques. We describe how to homogenize the information gleaned from light curves by selection and computation of real-numbered metrics ("feature"), detail methods to robustly estimate periodic light-curve features, introduce tree-ensemble methods for accurate variable star classification, and show how to rigorously evaluate the classification results using cross validation. On a 25-class data set of 1542 well-studied variable stars, we achieve a 22.8% overall classification error using the random forest classifier; this represents a 24% improvement over the best previous classifier on these data. This methodology is effective for identifying samples of specific science classes: for pulsational variables used in Milky Way tomography we obtain a discovery efficiency of 98.2% and for eclipsing systems we find an efficiency of 99.1%, both at 95% purity. We show that the random forest (RF) classifier is superior to other machine-learned methods in terms of accuracy, speed, and relative immunity to features with no useful class information; the RF classifier can also be used to estimate the importance of each feature in classification. Additionally, we present the first astronomical use of hierarchical classification methods to incorporate a known class taxonomy in the classifier, which further reduces the catastrophic error rate to 7.8%. Excluding low-amplitude sources, our overall error rate improves to 14%, with a catastrophic error rate of 3.5%.

23 pages, 9 figures

References in corpus (14)

• The generalised Lomb-Scargle periodogram. A new formalism for the floating-mean and Keplerian periodograms
• Stellar SEDs from 0.3-2.5 Microns: Tracing the Stellar Locus and Searching for Color Outliers in SDSS and 2MASS
• SDSS Standard Star Catalog for Stripe 82: the Dawn of Industrial 1% Optical Photometry
• Automated supervised classification of variable stars I. Methodology
• Results from the Supernova Photometric Classification Challenge
• Optimal Time-Series Selection of Quasars
• Robust Machine Learning Applied to Astronomical Datasets I: Star-Galaxy Classification of the SDSS DR3 Using Decision Trees
• Variable stars across the observational HR diagram
• How to Find More Supernovae with Less Work: Object Classification Techniques for Difference Imaging
• Automated supervised classification of variable stars II. Application to the OGLE database
• On the co-existence of chemically peculiar Bp stars, slowly pulsating B stars and constant B stars in the same part of the H-R diagram
• HD 97048's Circumstellar Environment as Revealed by a HST/ACS Coronagraphic Study of Disk Candidate Stars
• Binarity and multiperiodicity in high-amplitude delta Scuti stars
• Exploring the Variable Sky with the Sloan Digital Sky Survey

Cited by in corpus (193)

• The Zwicky Transient Facility: System Overview, Performance, and First Results
• The SDSS-III Baryon Oscillation Spectroscopic Survey: Quasar Target Selection for Data Release Nine
• Sagittarius II, Draco II and Laevens 3: three new Milky Way satellites discovered in the Pan-STARRS 1 3pi Survey
• The ASAS-SN Catalog of Variable Stars II: Uniform Classification of 412,000 Known Variables
• Kepler Eclipsing Binary Stars. III. Classification of Kepler Eclipsing Binary Light Curves with Locally Linear Embedding
• Machine-Learned Identification of RR Lyrae Stars from Sparse, Multi-band Data: the PS1 Sample
• Automating Discovery and Classification of Transients and Variable Stars in the Synoptic Survey Era
• Automated Transient Identification in the Dark Energy Survey
• Machine Learning for the Zwicky Transient Facility
• Alert Classification for the ALeRCE Broker System: The Light Curve Classifier
• Exploring the Variable Sky with LINEAR. III. Classification of Periodic Light Curves
• A recurrent neural network for classification of unevenly sampled variable stars
• Using Machine Learning for Discovery in Synoptic Survey Imaging
• VAST: An ASKAP Survey for Variables and Slow Transients
• Gaia Data Release 3: All-sky classification of 12.4 million variable sources into 25 classes
• K2 Variable Catalogue II: Machine Learning Classification of Variable Stars and Eclipsing Binaries in K2 Fields 0-4
• The many-faceted light curves of young disk-bearing stars in Upper Sco and $\rho$ Oph observed by $K2$ Campaign 2
• Automatic Classification of Kepler Planetary Transit Candidates
• Comparative performance of selected variability detection techniques in photometric time series
• A comparison of period finding algorithms
• Machine Learning-based Brokers for Real-time Classification of the LSST Alert Stream
• Construction of a Calibrated Probabilistic Classification Catalog: Application to 50k Variable Sources in the All-Sky Automated Survey
• The EPOCH Project: I. Periodic variable stars in the EROS-2 LMC database
• OGLE-III Microlensing Events and the Structure of the Galactic Bulge
• The Ĝ Search for Extraterrestrial Civilizations with Large Energy Supplies. IV. The Signatures and Information Content of Transiting Megastructures
• Computational Intelligence Challenges and Applications on Large-Scale Astronomical Time Series Databases
• The VMC survey - XXVI. Structure of the Small Magellanic Cloud from RR Lyrae stars
• Theoretical Models of Optical Transients. I. A Broad Exploration of the Duration-Luminosity Phase Space
• SDSS J0159+0105: A Radio-Quiet Quasar with a Centi-Parsec Supermassive Black Hole Binary Candidate
• Active Learning to Overcome Sample Selection Bias: Application to Photometric Variable Star Classification
• Highly Variable Extinction and Accretion in the Jet-driving Class I Type Young Star PTF 10nvg (V2492 Cyg, IRAS 20496+4354)
• GalaxyNet: Connecting galaxies and dark matter haloes with deep neural networks and reinforcement learning in large volumes
• Alert Classification for the ALeRCE Broker System: The Real-time Stamp Classifier
• Machine Learning Techniques for Stellar Light Curve Classification
• Deep-Learnt Classification of Light Curves
• Agatha: disentangling periodic signals from correlated noise in a periodogram framework
• A comparison of quasar emission reconstruction techniques for $z\ge 5.0$ Lyman-$\alpha$ and Lyman-$\beta$ transmission
• Mapping the outer bulge with RRab stars from the VVV Survey
• Scalable End-to-end Recurrent Neural Network for Variable star classification
• Systematic Serendipity: A Test of Unsupervised Machine Learning as a Method for Anomaly Detection
• An analysis of feature relevance in the classification of astronomical transients with machine learning methods
• Semi-supervised Learning for Photometric Supernova Classification
• Deep multi-survey classification of variable stars
• Automated Classification of Periodic Variable Stars detected by the Wide-field Infrared Survey Explorer
• Characterizing the local relation between star formation rate and gas-phase metallicity in MaNGA spiral galaxies
• A machine learned classifier for RR Lyrae in the VVV survey
• Automated classification of Hipparcos unsolved variables
• On Neural Architectures for Astronomical Time-series Classification with Application to Variable Stars
• Discovery of 36 eclipsing EL CVn binaries found by the Palomar Transient Factory
• The Extremely Luminous Quasar Survey (ELQS) in the SDSS footprint I.: Infrared Based Candidate Selection
• A Machine Learning Method to Infer Fundamental Stellar Parameters from Photometric Light Curves
• Machine-assisted discovery of relationships in astronomy
• Automatic Classification of Variable Stars in Catalogs with missing data
• Search for high-amplitude Delta Scuti and RR Lyrae stars in Sloan Digital Sky Survey Stripe 82 using principal component analysis
• Unsupervised machine learning for transient discovery in Deeper, Wider, Faster light curves
• Robust period estimation using mutual information for multi-band light curves in the synoptic survey era
• The VVV Templates Project. Towards an Automated Classification of VVV Light-Curves. I. Building a database of stellar variability in the near-infrared
• Comparing Multi-class, Binary and Hierarchical Machine Learning Classification schemes for variable stars
• An improved quasar detection method in EROS-2 and MACHO LMC datasets
• The Outer Halo of the Milky Way as Probed by RR Lyr Variables from the Palomar Transient Facility
• Automatic vetting of planet candidates from ground based surveys: Machine learning with NGTS
• Clustering Based Feature Learning on Variable Stars
• Unsupervised Classification of Variable Stars
• Into the Darkness: Classical and Type II Cepheids in the Zona Galactica Incognita
• Automatic classification of time-variable X-ray sources
• TESS Data for Asteroseismology (T'DA) Stellar Variability Classification Pipeline: Set-Up and Application to the Kepler Q9 Data
• The QUEST-La Silla AGN Variability Survey: selection of AGN candidates through optical variability
• AutoRegressive Planet Search: Methodology
• Five New Outbursting AM CVn Systems Discovered by the Palomar Transient Factory
• The M33 Synoptic Stellar Survey. II. Mira Variables
• Discovery of New Dipper Stars with K2: A Window into the Inner Disk Region of T Tauri Stars
• Data challenges of time domain astronomy
• Sky Surveys
• A Mid-infrared Study of RR Lyrae Stars with the WISE All-Sky Data Release
• Deep Neural Network Classifier for Variable Stars with Novelty Detection Capability
• Feature Selection Strategies for Classifying High Dimensional Astronomical Data Sets
• An irregular discrete time series model to identify residuals with autocorrelation in astronomical light curves
• A random forest-based selection of optically variable AGN in the VST-COSMOS field
• Variability, periodicity and contact binaries in WISE
• A Machine Learning Classifier for Microlensing in Wide-Field Surveys
• Meta Classification for Variable Stars
• PTF1 J191905.19+481506.2 - A Partially Eclipsing AM CVn System Discovered in the Palomar Transient Factory
• Classifying Image Sequences of Astronomical Transients with Deep Neural Networks
• Photometric Classification of quasars from RCS-2 using Random Forest
• Computational Tools for the Spectroscopic Analysis of White Dwarfs
• A method for finding anomalous astronomical light curves and their analogs
• Identifying Tidal Disruption Events via Prior Photometric Selection of Their Preferred Hosts
• Identification of RR Lyrae stars in multiband, sparsely-sampled data from the Dark Energy Survey using template fitting and Random Forest classification
• Real-Time Data Mining of Massive Data Streams from Synoptic Sky Surveys
• Astraea: Predicting Long Rotation Periods with 27-Day Light Curves
• The frequency of Kozai-Lidov disc oscillation driven giant outbursts in Be/X-ray binaries
• Automatic Survey-Invariant Variable Star Classification
• Mid-infrared Period-Luminosity Relations of RR Lyrae Stars Derived from the WISE Preliminary Data Release
• Finding Black Holes with Black Boxes -- Using Machine Learning to Identify Globular Clusters with Black Hole Subsystems
• MeerCRAB: MeerLICHT Classification of Real and Bogus Transients using Deep Learning
• Classification of OGLE eclipsing binary stars based on their morphology type with Locally Linear Embedding
• New Insights into Time Series Analysis II -- No Correlated Observations
• Preparing for advanced LIGO: A Star-Galaxy Separation Catalog for the Palomar Transient Factory
• Beyond the Local Volume. I. Surface Densities of Ultracool Dwarfs in Deep HST/WFC3 Parallel Fields
• New Insights into Time Series Analysis - I - Correlated observations
• Discovery of Bright Galactic R Coronae Borealis and DY Persei Variables: Rare Gems Mined from ACVS
• Classification of Periodic Variable Stars with Novel Cyclic-Permutation Invariant Neural Networks
• Discrete-time autoregressive model for unequally spaced time-series observations
• Modeling Light Curves for Improved Classification
• The VVV Infrared Variability Catalog (VIVA-I)
• ANTARES: A Prototype Transient Broker System
• Rapid, Machine-Learned Resource Allocation: Application to High-redshift GRB Follow-up
• Uncertain classification of Variable Stars: handling observational GAPS and noise
• Classification of Multiwavelength Transients with Machine Learning
• Flares hunting in hot subdwarf and white dwarf stars from Cycles 1-5 of TESS photometry
• Optimizing Automated Classification of Periodic Variable Stars in New Synoptic Surveys
• The Structure of the Young Star Cluster NGC 6231. I. Stellar Population
• Density Based Outlier Scoring on Kepler Data
• A classification algorithm for time-domain novelties in preparation for LSST alerts: Application to variable stars and transients detected with DECam in the Galactic Bulge
• The variability of blazars throughout the electromagnetic spectrum
• MANTRA: A Machine Learning reference lightcurve dataset for astronomical transient event recognition
• The effect of phased recurrent units in the classification of multiple catalogs of astronomical lightcurves
• Identification of Stellar Flares Using Differential Evolution Template Optimization
• Everything we'd like to do with LSST data, but we don't know (yet) how
• The Highly-Eccentric Detached Eclipsing Binaries in ACVS and MACC
• New methods to assess and improve LIGO detector duty cycle
• Streaming Classification of Variable Stars
• Gaia eclipsing binary and multiple systems. Supervised classification and self-organizing maps
• Calibrating Long Period Variables as Standard Candles with Machine Learning
• The Solar System Notification Alert Processing System (SNAPS): Design, Architecture, and First Data Release (SNAPShot1)
• The Most Predictive Physical Properties for the Stellar Population Radial Profiles of Nearby Galaxies
• Novel bivariate autoregressive model for predicting and forecasting irregularly observed time series
• Flashes in a Star Stream: Automated Classification of Astronomical Transient Events
• Weighted statistical parameters for irregularly sampled time series
• Semi-Supervised Classification and Clustering Analysis for Variable Stars
• Understanding of the properties of neural network approaches for transient light curve approximations
• Confirmation of Monoperiodicity Above 20 Seconds for Two Blue Large-Amplitude Pulsators
• Selection of Burst-like Transients and Stochastic Variables Using Multi-Band Image Differencing in the Pan-STARRS1 Medium-Deep Survey
• Selection of optically variable active galactic nuclei via a random forest algorithm
• The Synthetic-Oversampling Method: Using Photometric Colors to Discover Extremely Metal-Poor Stars
• New Insights into Time Series Analysis III - Setting constraints on period analysis
• Variable Star Classification Using Multi-View Metric Learning
• A Detection Metric Designed for O'Connell Effect Eclipsing Binaries
• Hunting Gravitational Waves with Multi-Messenger Counterparts: Australia's Role
• On the optical counterparts of radio transients and variables
• VVV catalog of ab-type RR Lyrae in the inner Galactic bulge
• Comparative Analysis of a Transition Region Bright Point with a Blinker and Coronal Bright Point Using Multiple EIS Emission Lines
• Long-Period High-Amplitude Red Variables in the KELT Survey
• Connecting the time domain community with the Virtual Astronomical Observatory
• A Catalog of LAMOST Variable Sources Based on Time-domain Photometry of ZTF
• Stop&Hop: Early Classification of Irregular Time Series
• Investigation of stellar magnetic activity using variational autoencoder based on low-resolution spectroscopic survey
• Automated classification of eclipsing binary systems in the VVV Survey
• A Fast Approximate Approach to Microlensing Survey Analysis
• Statistical Study of 2XMMi-DR3/SDSS-DR8 Cross-correlation Sample
• Time-domain study of the young massive cluster Westerlund 2 with the Hubble Space Telescope. I
• Variability search in M 31 using Principal Component Analysis and the Hubble Source Catalog
• Automatic Catalog of RRLyrae from $\sim$ 14 million VVV Light Curves: How far can we go with traditional machine-learning?
• Classifying High-cadence Microlensing Light Curves I; Defining Features
• Characterizing the Best Cosmic Telescopes with the Millennium Simulations
• Online classification for time-domain astronomy
• How machine learning conquers the unitary limit
• A Neural Network Perturbation Theory Based on the Born Series
• On the use of machine learning algorithms in the measurement of stellar magnetic fields
• Finding radio transients with anomaly detection and active learning based on volunteer classifications
• X-ray Sources Classification Using Machine Learning: A Study with EP-WXT Pathfinder LEIA
• Automated Real-Time Classification and Decision Making in Massive Data Streams from Synoptic Sky Surveys
• On the use of variability time-scales as an early classifier of radio transients and variables
• The impact of Gaia and LSST on binary stars and exo-planets
• An Algorithm for the Visualization of Relevant Patterns in Astronomical Light Curves
• An Information Theory Approach on Deciding Spectroscopic Follow Ups
• Estimating a Common Period for a Set of Irregularly Sampled Functions with Applications to Periodic Variable Star Data
• YOUNG Star detrending for Transiting Exoplanet Recovery (YOUNGSTER) II: Using Self-Organising Maps to explore young star variability in Sectors 1-13 of TESS data
• Informative Bayesian model selection for RR Lyrae star classifiers
• Data Driven Discovery in Astrophysics
• Multiwavelength study of radio galaxy Pictor A: detection of western hotspot in far-UV and possible origin of high energy emissions
• Leveraging pre-trained vision Transformers for multi-band photometric light curve classification
• Generalised learning of time-series: Ornstein-Uhlenbeck processes
• Unsupervised learning for variability detection with Gaia DR3 photometry. The main sequence-white dwarf valley
• GRAPE: Genetic Routine for Astronomical Period Estimation
• Navigating AGN variability with self-organizing maps
• Realizing the potential of astrostatistics and astroinformatics
• NGTS clusters survey IV. Search for Dipper stars in the Orion Nebular Cluster
• Informative regularization for a multi-layer perceptron RR Lyrae classifier under data shift
• Identifying Radio Active Galactic Nuclei with Machine Learning and Large-Area Surveys
• Automated all-sky detection of γ Doradus / δ Scuti hybrids in TESS data from positive unlabelled (PU) learning
• Margin-free classification and new class detection using finite Dirichlet mixtures
• Multivariate time series transformer embeddings for light curves of periodic variable stars
• QPOML: A Machine Learning Approach to Detect and Characterize Quasi-Periodic Oscillations in X-ray Binaries
• Periodic Variable Stars Modulated by Time-Varying Parameters
• Advanced Astroinformatics for Variable Star Classification
• Machine Learning Identification of Gravitationally Microlensed Gamma-Ray Bursts
• Designing Test Information and Test Information in Design
• Classification of blazars based on data-driven approaches
• GPU-Enabled Searches for Periodic Signals of Unknown Shape
• Finding Fast Transients in Real Time Using Novel Light Curve Analysis Algorithm
• Exploring Late Stellar Evolution in the Era of Large Surveys: Machine Learning Prospects for Hot Subdwarfs and White Dwarfs
• Light curve fingerprints: an automated approach to the extraction of X-ray variability patterns with feature aggregation -- an example application to GRS 1915+105