A Genetic Algorithm For Discovering Interesting Fuzzy Prediction Rules Applications To Science And Technology Data.pdf

A ic Algorithm for Discovering Interesting Fuzzy Prediction
Rules: applications to science and technology data



Wesley Romão Alex A. Freitas Roberto C. S. Pacheco
UEM, DIN-CTC PUC-PR, ET UFSC, PPGEP-CTC
Av. Colombo, 5790 Rua Imaculada Conceição, 1155 C. P. 476
Maringá – PR. Curitiba – PR. Florianópolis, SC.
87020-900 - Brazil 80215-901 – Brazil 88040-900 – Brazil
******@ ******@ ******@
/~wesley .br/~alex (55) (48) 331-7016
(55) (44) 263-2479 (55) (41) 330-1347

Abstract It should be noted that this task can be regarded as a
generalization of the well-known classification task of
Data mining consists of extracting knowledge
from data. This paper addresses the discovery of data mining. In classification there is a single goal
attribute to be predicted, whereas we allow more than one
knowledge in the form of prediction IF-THEN
goal attribute to be defined by the user.
rules, which are a popular form of knowledge
representation in data mining. In this context, we Note that, although there are several goal attributes to be
propose a new ic Algorithm (GA) designed predicted, each rule predicts the value of a single goal
specifically for discovering interesting fuzzy attribute in its consequent. However, different rules can
prediction rules. The GA searches for prediction predic