Investigação da aplicação de algoritmos de agrupamento para o problema astrofísico de classificação de galáxias

Abstract

The emergence oflarge databases streamlined storage and search ofrecords so there is a large quantity information that have to be analyzed. For the exploration of such data in a timely manner it’s used the process of data mining to analyze data from different perspectives and allows automatic discovery of patterns and information. Besides it provides the ability to predict a future observation. It is intended to explore astronomical databases with morphometric parameters of galaxies. Clustering algorithms are used to identify natural clusters and patterns as a previous step of galaxies classification. The algorithms: Expectation Maximization (EM) and K-means will be applied to synthetic and real data from EFIGI survey (Extraction de Formes Idealisées de Galaxies en Imagerie) with galaxies of all morphologycal types as measured by MORFOMETRYKA. After the data is grouped by the algorithms the silhouette is used as a method of results validation. In this morphometric space ofparameters the galaxies classes will be detected. With these results it is possible to study the morphometric continuity in populations ofspiral and elliptical galaxies.