SUPERNOVAE AUTOMATIC CLASSIFICATION METHOD BY MODELING HUMAN ANALYSIS USING ARTIFICIAL NEURAL NETWORKS
Marcelo Módolo1; Lamartine Guimarães2; Reinaldo Rosa3
1 National Institute for Space Research (INPE) and Methodist University of São Paulo (UMESP); 2 Institute for Advanced Studies of the Department of Aerospace Science and Technology (IEAv/DCTA); 3 National Institute for Space Research (INPE)
Resumo
Classify a recently discovered supernova is not trivial and only a few experts astronomers are able to perform it. The existing automatic classifiers did not do the modeling of the human way of analyzing the spectrum to classify supernovas. They only compares the spectrum similarity of discovered supernova with spectra of supernovae have already been classified. The automatic method proposing in this paper models the human way of classification using Neural Networks Multilayer Perceptron to analyze the supernovae spectra. The experiments performed obtained significant results indicating the viability of using this method in places that require an automatic analysis or that have no specialist.
Palavras-chave: Celestial mechanics and astronomy dynamics; modeling, numerical simulation and optimization; nonlinear systems and neural dynamics; supernovae automatic classification; supernovae spectrum analysis