Table of Contents
The idea of machine learning (ML) is not new, but ML has become a tool that many astronomers are using at their analysis.
It has been used to:
- Predict the photometric redshifts of distant galaxies
- Determine HI richness of galaxies
- Classify galaxy morphologies, i.e. ellipticals, disky, barred, etc.
- and many other applications.
Machine learning has definitely started to change the field of astronomy and astrophysics. This is the reason for why it is important to properly understand how to get the best results for a ML project.
There are two main types of ML algorithms, i.e. supervised and unsupervised. The first type is mainly used for classification problems, while the second one is mainly used for clustering problems.
Supervised machine learning refers to the prediction of
classification) or values (
regression) based on a
previous knowledge of the data, i.e. the user know the truth of
each observation, and tries to make an educated guess of for
unobserved data based on a training process with data.