WSEAS Transactions on Applied and Theoretical Mechanics
Print ISSN: 1991-8747, E-ISSN: 2224-3429
Volume 14, 2019
Rough Set-based Rule Set Generation to see the Cause of the Occurrence of a Series of Grooves in Gravure Printing
Author:
Abstract: In the domain of gravure printing the phenomenon of grooves or bands that appear during printing have been hard to solve problems for decades. Such phenomenon causes delays and extra costs for possible maintenance. Therefore, in this paper we want find more accurate knowledge model in rule form to see what could be the main causes by analyzing the related data set. Because rough set theory-based data mining algorithms analyze solely based on data, it is highly possible that they may find more hidden knowledge than other heuristic-based data mining algorithms. But, this good point can be an obstacle when target data sets have many attributes, and the data sets have key-like attributes due to the limitation of the rough set-based algorithms. Because our target data set of gravure printing has such characteristics, we first applied a heuristic-based rule discovery algorithm called RIPPER, after that select attributes only were supplied as an input to rough set theory-based algorithm called MODLEM. Experiments showed very good result in accuracy.
Search Articles
Pages: 118-128
WSEAS Transactions on Applied and Theoretical Mechanics, ISSN / E-ISSN: 1991-8747 / 2224-3429, Volume 14, 2019, Art. #12