Classification of planetary nebulae by cluster analysis and artificial neural networks

Abstract

According to the chemical composition, a sample of 192 Planetary Nebulae of different types has been re-classified, and 41 others have been classified for the first time, by means of two methods not employed so far in this field: hierarchical cluster analysis and supervised artificial neural network. The cluster analysis reveals itself as a good first guess for grouping Planetary Nebulae, while an artificial neural network provides reliable automated classification of this kind of objects.