CINTIA 2: uma hierarquia de redes neurais artificiais binárias para classificação inteligente de supernovas

Abstract

Supernovae are catastrophic events in which some stars explode. The classification of supernovas is done by specialists by means of the analysis of the light spectrum that have lines of absorption and emission in certain regions of the wavelength. The supernovae light spectrua present patterns that can be used in machine learning algorithms, thus enabling automatic and intelligent classification of supernovae. Automatic classification is essential for the processing of large amounts of data in equipment installed in remote locations, where it is not always possible to have a specialist. The aim of this work is to present CINTIA 2, an enhancement of the Intelligent Classifier of Type Ia supernovae (CIntIa), which uses a hierarchy of binary neural networks of the Perceptron kind to classify supernovas in types Ia, Ib, Ic and II. We present the architecture of CINTIA 2 and the tool derived from it, developed in the programming languages Python and C ++. The results obtained presented excellent performance, mainly in the classification of types Ia and II. A comparison with works found in the literature shows that CINTIA 2 is superior in quantity and diversity of data and reaches classification indexes comparable to the other classifiers.