|
The main aims of the
project can be described as follows:
-
compare two different
clustering algorithms (Snob and Single-Pass) and how features derived
from their clusters improve/degrade the performance of Support Vector
Machines (SVMs)
-
Investigate whether the
success of this method depends on the different characteristics of the
data sets (size, feature type, number of features etc.)
-
Try to determine other
factors that have an affect on the performance of this particular
method
|