More Info
									                DISCOVERING FUZZY RULES
                    VARIABLE ELIMINATION
                                        Jan Bohacik∗

Abstract: A group of fuzzy IF-THEN rules is belonging to one of the most popu-
lar, most effective, and user-friendliest knowledge representations. For this reason,
extraction of these rules is becoming a more-and-more important part of the Data
Mining stage in the Knowledge Discovery in Databases Process. In this paper,
a direct algorithm for extracting fuzzy IF-THEN rules on the basis of linguistic
variable elimination is described. The algorithm is implemented within a designed
object-oriented software library Fuzzy Rule Miner. Besides the introduced algo-
rithm, it implements two algorithms for fuzzy rule extraction based on using fuzzy
decision trees of ID3 kind. An essential precondition for comparing the imple-
mented algorithms and for verifying the legitimacy of the introduced algorithm is
performance of experiments. The goal of experiments is to take in the behavior
of algorithms on testing databases from the UCI Repository of Machine Learning
Databases and to make comparisons of algorithms with one another. According to
the conducted experiments, the introduced algorithm achieves high accuracy levels
of discovered knowledge. The paper also contains a classification of rules and a
specification of the Fuzzy Rule Discovery in Databases Process.

Key words: Fuzzy rules, fuzzy decision trees, linguistic variable elimination,
           classification, fuzzy rule extraction, data mining, knowledge
           discovery in databases
Received: July 15, 2009
Revised and accepted: January 29, 2010

1.     Introduction
At present, people have to manage more and more data of diverse kind. Large
amount of existing data causes a lot of problems of various kind. One of the most
visible of them is a person’s or human team’s inability to be able to maintain and
process data in reasonable time. Because information technologies are used on a
mass scale, large amount of other data arises. This kind of data can provide the
   ∗ Jan Bohacik (written J´n Boh´ˇik with all Slovak diacritics)
                           a     ac
Department of Informatics, Faculty of Management Science and Informatics, University of Zilina,
Slovak Republic,

c ICS AS CR 2010                                                                           45
                         Neural Network World 1/10, 45-61

runners or the owners of these systems useful information. As a result of the ex-
ponential growth of commercially utilizable data, database systems are incessantly
developed. Database systems are a tried tool for manipulation with large data
files, i.e. with large databases. Nowadays, most freq
To top