›School of Astrostatistics 2015: Clustering and Classification

Objectives

The primary goal of the École d'Astrostatistique is to train astronomers to the use of modern statistical techniques. It also aims at bridging the gap between the two communities by emphasizing on the practice during concrete works in common, to give firm grounds to the theoretical lessons, and to initiate works on problems brought by the participants.

The School of Statistics for Astrophysics 2015 will be devoted to classification and clustering.

Grouping objects obtained from the astronomical observations into distinct categories has always been a necessity imposed by their vast diversity. This is the case for stars, galaxies, asteroids, supernova, active galactic nuclei, gamma-ray bursts and many others. This clustering by similarities (unsupervised classification) is a prerequisite to any physical modelisation. In this purpose, astronomers have always used heuristic, simple and subjective techniques, based on only one or two parameters, most often with the help of a visual examination. Consequently, for a given type of astrophysical objects, may exist a large number of classifications, sometimes overlapping each other.

However, the primary role of a classification is to describe the diversity with a simple scheme while showing the relationships between the classes. In addition, this traditional practice cannot be transposed in the era of huge databases. Astronomers are now obliged to consider using objective and multivariate clustering methods that allow for automatic (supervised) classifications.

The objective of the School of Astrostatistics 2015 is to train astronomers to the use of a large range of statistical and algorithmic approaches of unsupervised and supervised classification, from the most basic to the state-of-the-art ones, including the spatial clustering, the clustering of variables and the graphical methods.

This school is a École Thématique du CNRS 2015, supported by INSU and INSMI, from the Formation Permanente that fully supports the participation of CNRS personnel.

This school is of course opened to any astronomer.

No specific knowledge of statistics is required, and the mathematical background of all astronomers is sufficient to attend the school.