-
Notifications
You must be signed in to change notification settings - Fork 0
/
abstract.tex
7 lines (5 loc) · 1.11 KB
/
abstract.tex
1
2
3
4
5
6
7
\begin{resumo}[Abstract]
Data mining has become an important strategic factor in the business environment as it enables non-trivial information to be identified in order to leverage business or even promote product and service innovations. The data classification is one of the most important, and recurring, topics found in data mining and in this context, the following work aims to propose a multiobjective algorithm based on PSO for data classification through rule extraction. In this approach, each particle (candidate solution) represents an IF-THEN rule, which is converted into a logical predicate in the construction of a selection operation in the SQL language for performance evaluation of the algorithm. During the computational experiments, it was observed that the proposed approach was competitive, with promising results when compared to other classic classification methods, recognized in the literature, especially in unbalanced datasets.
\vspace{1.5ex}
\noindent \textbf{Palavras-chave}: Particle Swarm Optimization, Swarm Intelligence, Multi-objective Approach, Rule Mining and Data Classification.
\end{resumo}