An Artificial Neural Network Approach to Classification of Galaxy Spectra

Abstract

We present a method for automated classification of galaxies with low signal-to-noise (S/N) spectra typical of redshift surveys. We develop spectral simulations based on the parameters for the 2dF Galaxy Redshift Survey, and with these simulations we investigate the technique of Principal Component Analysis when applied specifically to spectra of low S/N. We relate the objective principal components to features in the spectra and use a small number of components to successfully reconstruct the underlying signal from the low quality spectra. Using the principal components as input, we train an Artificial Neural Network to classify the noisy simulated spectra into morphological classes, revealing the success of the classification against the observed $b_{\rm J}$ magnitude of the source, which we compare with alternative methods of classification. We find that more than 90\% of our sample of normal galaxies are correctly classified into one of five broad morphological classes for simulations at $b_{\rm J}$=19.7. By dividing the data into separate sets we show that a classification onto the Hubble sequence is only relevant for normal galaxies and that spectra with unusual features should be incorporated into a classification scheme based predominantly on their spectral signatures. We discuss how an Artificial Neural Network can be used to distinguish normal and unusual galaxy spectra, and discuss the possible application of these results to spectra from galaxy redshift surveys.