Steps toward a classifier for the Virtual Observatory. I. Classifying the SDSS photometric archive

TL;DR

This paper introduces a new automated classification method combining clustering and agglomeration techniques to categorize objects in large photometric datasets like SDSS, enhancing data analysis efficiency.

Contribution

The paper presents a novel classification approach tailored for large multi-parametric photometric archives, demonstrated on the SDSS data set.

Findings

Effective classification of SDSS objects achieved

Method shows good generalization to complex data

Potential for application to other large surveys

Abstract

Modern photometric multiband digital surveys produce large amounts of data that, in order to be effectively exploited, need automatic tools capable to extract from photometric data an objective classification. We present here a new method for classifying objects in large multi-parametric photometric data bases, consisting of a combination of a clustering algorithm and a cluster agglomeration tool. The generalization capabilities and the potentialities of this approach are tested against the complexity of the Sloan Digital Sky Survey archive, for which an example of application is reported.