Catalog for: Machine and Deep Learning Applied to Galaxy Morphology - A Comparative Study

Published: 5 November 2019| Version 1 | DOI: 10.17632/tg7c985c2n.1
Contributors:
Paulo Barchi, Camila de Sá Freitas, Thiago Gonçalves, Tatiana Moura, Reinaldo Rosa, Rubens Sautter, Esteban Clua, Bruno Marques, Reinaldo de Carvalho, Marcelle Soares-Santos

Description

Machine and Deep Learning morphological classification for 670,560 galaxies from Sloan Digital Sky Survey Data Release 7 (SDSS-DR7). Classifications are provided for 2 classes problem (0: elliptical; or, 1: spiral galaxy) and 3 classes problem (0: elliptical, 1: non-barred spiral, or 2: barred spiral galaxy). ML2classes classification is obtained by Traditional Machine Learning Approach, using Morphological non-parametric parameters and Decision Tree. Classifications using Deep Learning are obtained using a Convolutional Neural Network (CNN). Morphological non-parametric parameters are provided as well: Concentration (C), Asymmetry (A), Smoothness (S), Gradient Pattern Analysis (G2) parameter and Entropy (H). We also provide the Error from CyMorph processing. All error flags are mapped as follows: Error = 0: success (no errors); Error = 1: many objects of significant brightness inside 2 Rp of the galaxy; Error = 2: not possible to calculate the galaxy's Rp; Error = 3: problem calculating GPA; Error = 4: problem calculating H; Error = 5: problem calculating C; Error = 6: problem calculating A; Error = 7: problem calculating S.

Files

Categories

Astrophysics, Machine Learning, Galaxy, Deep Learning

Licence