|Abstract:||The problem of feature selection is fundamental in a number of different
tasks like classification, data mining, image processing, conceptual learning,
etc. . . In recent times, the growing inportance of knowledge discovery and
data-mining approaches in practical applications has made the feature selection
problem a quite hot topic, especially when considering the mining of
knowledge from real-world databases or warehouses, containing not only a
huge amount of records, but also a significant number of features not always
relevant for the task at hand.
Looking at the literature, there are essentially two main fields where the
feature selection problem has been extensively studied: Statistical Pattern Recognition, Machine Learning.
In the first field, feature selection is considered from the classification point
of view, i.e. the problem is approached having the construction of an efficient
classifier (i.e. a pattern recognizer) as a target.
In the Machine Learning community, more emphasis is given to the more
general problem of concept learning, even if classification still remain an
important issue. However, this difference has provided different approaches
aimed at solving the problem, addressing different perspectives and points of
In the paPER, we will report on the main proposals and approches
developed inside the two communities.|