Machine learning is a branch of Artificial Intelligence
(AI) defined (by Arthur Samuel one of the pioneers of AI) as
the field of study that gives to computers the ability to learn without
being explicitly programmed or in other words systems that can learn
from data. During the last ten years, SVMs have been showing a great
potential for data classification also in satellite data processing
(Mountrakis et al., 2011) and for these reasons has been adopted in
this study. It is expected that the use of SVMs in remote sensing should
enable us to successfully handle small training data sets and outperform
other more traditional classifiers (Mountrakis et al., 2011).