Stacking for machine learning redshifts applied to SDSS galaxies

Abstract

We present an analysis of a general machine learning technique called 'stacking' for the estimation of photometric redshifts. Stacking techniques can feed the photometric redshift estimate, as output by a base algorithm, back into the same algorithm as an additional input feature in a subsequent learning round. We shown how all tested base algorithms benefit from at least one additional stacking round (or layer). To demonstrate the benefit of stacking, we apply the method to both unsupervised machine learning techniques based on self-organising maps (SOMs), and supervised machine learning methods based on decision trees. We explore a range of stacking architectures, such as the number of layers and the number of base learners per layer. Finally we explore the effectiveness of stacking even when using a successful algorithm such as AdaBoost. We observe a significant improvement of between 1.9% and 21% on all computed metrics when stacking is applied to weak learners (such as SOMs and decision trees). When applied to strong learning algorithms (such as AdaBoost) the ratio of improvement shrinks, but still remains positive and is between 0.4% and 2.5% for the explored metrics and comes at almost no additional computational cost.

13 pages, 3 tables, 7 figures version accepted by MNRAS, minor text updates. Results and conclusions unchanged

References in corpus (32)

• Scikit-learn: Machine Learning in Python
• The UKIRT Infrared Deep Sky Survey (UKIDSS)
• The 2.5 m Telescope of the Sloan Digital Sky Survey
• New evolutionary models for pre-main sequence and main sequence low-mass stars down to the hydrogen-burning limit
• From filamentary clouds to prestellar cores to the stellar IMF: Initial highlights from the Herschel Gould Belt survey
• The Tenth Data Release of the Sloan Digital Sky Survey: First Spectroscopic Data from the SDSS-III Apache Point Observatory Galactic Evolution Experiment
• The distance to the Orion Nebula
• Rotation-invariant convolutional neural networks for galaxy morphology prediction
• The Importance of Radiative Feedback for the Stellar Initial Mass Function
• TPZ : Photometric redshift PDFs and ancillary information by using prediction trees and random forests
• Mapping the Galaxy Color-Redshift Relation: Optimal Photometric Redshift Calibration Strategies for Cosmology Surveys
• Photometric redshift analysis in the Dark Energy Survey Science Verification data
• Redshift distributions of galaxies in the DES Science Verification shear catalogue and implications for weak lensing
• SOMz: photometric redshift PDFs with self organizing maps and random atlas
• Analytical theory for the initial mass function: III time dependence and star formation rate
• The Mass Distributions of Starless and Protostellar Cores in Gould Belt Clouds
• CO Abundance Variations in the Orion Molecular Cloud
• Orion Revisited - I. The massive cluster in front of the Orion Nebula Cluster
• Using neural networks to estimate redshift distributions. An application to CFHTLenS
• Anomaly detection for machine learning redshifts applied to SDSS galaxies
• Feature importance for machine learning redshifts applied to SDSS galaxies
• Exhausting the Information: Novel Bayesian Combination of Photometric Redshift PDFs
• Towards Precise Ages and Masses of Free Floating Planetary Mass Brown Dwarfs
• A deep survey of brown dwarfs in Orion with Gemini
• New Panoramic View of ${}^{12}$CO and 1.1 mm Continuum Emission in the Orion A Molecular Cloud. I. Survey Overview and Possible External Triggers of Star Formation
• A Hybrid Ensemble Learning Approach to Star-Galaxy Classification
• A Wide-Field Survey of the Orion Nebula Cluster in the Near-Infrared
• GAz: A Genetic Algorithm for Photometric Redshift Estimation
• Data augmentation for machine learning redshifts applied to SDSS galaxies
• Tuning target selection algorithms to improve galaxy redshift estimates
• Measuring photometric redshifts using galaxy images and Deep Neural Networks
• QSO Selection and Photometric Redshifts with Neural Networks

Cited by in corpus (15)

• Separation of pulsar signals from noise with supervised machine learning algorithms
• Morpho-z: improving photometric redshifts with galaxy morphology
• A review of unsupervised learning in astronomy
• A cooperative approach among methods for photometric redshifts estimation: an application to KiDS data
• Inferring physical properties of galaxies from their emission line spectra
• SKA Science Data Challenge 2: analysis and results
• Euclid preparation: XXII. Selection of Quiescent Galaxies from Mock Photometry using Machine Learning
• Preliminary results of using k-Nearest Neighbours Regression to estimate the redshift of radio selected datasets
• Comparison Analysis of Tree Based and Ensembled Regression Algorithms for Traffic Accident Severity Prediction
• Improving machine learning-derived photometric redshifts and physical property estimates using unlabelled observations
• Selection of powerful radio galaxies with machine learning
• Identifying type II quasars at intermediate redshift with few-shot learning photometric classification
• Machine learning applications in astrophysics: Photometric redshift estimation
• Automated quasar continuum estimation using neural networks: a comparative study of deep-learning architectures
• Euclid preparation. Estimating galaxy physical properties using CatBoost chained regressors with attention