Visual
Decision Tree [VisualDT software]
Abstract: Data mining is intended to
extract hidden useful knowledge from large datasets in a given
application. This usefulness relates to the user goal, in other words
only the user can determine whether the resulting knowledge answers his
goal. Therefore, data mining tool should be highly interactive and
participatory. This paper presents an interactive decision tree
algorithm using visualization methods to gain insight into a model
construction task. We show how the user can interactively use
cooperative tools to support the construction of decision tree models.
The idea here is to increase the human participation through
interactive visualization techniques in a data mining environment. The
effective cooperation can bring out some progress towards reaching
advantages like, the user can be an expert of the data domain and can
use this domain knowledge during the whole model construction, the
confidence and comprehensibility of the obtained model are improved
because the user was involved in its construction, we can use the human
pattern recognition capabilities. The experimental results on Statlog
and UCI datasets show that our cooperative tool is comparable to the
automatic algorithm C4.5, but the user has a better understanding of
the obtained model.
