Astronomical data mining with neural networks

Abstract

We give a brief overview of artificial neural networks (ANNs) , focusing on Ko honen networks (KNs). The two kinds of KNs will be described in detail: the unsupervised self-organizing map (SOM) and the supervised learning vector quantization (LVQ). We then apply these algorithms to two astronomical clas sification problems: the classification of broad absorption line quasars (BALQ SOs) and of gamma-ray bursts (GRBs). In the context of BALQSOs, we find a BALQSO fraction of 10.4%, and compile a catalogue from the Sloan Digital Sky Survey (SDSS) using the supervised LVQ. This is currently the most com plete BALQSO catalogue. We then apply the unsupervised SOM to GRB light curves obtained from the Burst and Transient Source Experiment (BATSE). Using only shape-dependent variables, we find that two classes are recovered: single-pulsed bursts (SPBs) and multi-pulsed bursts (MPBs). We show that

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata