Spectral Classification of Asteroids by Random Forest

Abstract

With the increasing spectral and photometric data of asteroids, a variety of classification methods for asteroids have been proposed. This paper classifies asteroids based on the observations in the Sloan Digital Sky Survey (SDSS) Moving Object Catalogue (MOC) by using the random forest algorithm. In combination with the present taxonomies of Tholen, Bus, Lazzaro, and DeMeo, and the principal component analysis, we have classified 48642 asteroids according to their SDSS magnitudes at the g, r, i, and z wavebands. In this way, these asteroids are divided into 8 (C, X, S, B, D, K, L, and V) classes.