A grid data mining architecture for learning classifier systems
Manuel Filipe Santos, Wesley Mathew, Tim Kovacs, Henrique Santos · WSEAS Transactions on Computers archive · 2009
Recently, there is a growing interest among the researchers and software developers in exploring Learning Classifier System (LCS) implemented in parallel and distributed grid structure for data mining, due to its practical applications. The paper highlights the some aspects of the LCS and studying the competitive data mining model with homogeneous data. In order to establish more efficient distributed environment, in the current work, Grid computing architecture is considered a better distributed framework in Supervised Classifier System (UCS). The fundamental structure of this work allows each site of the distributed environment to manage independent UCS and such local sites transmit learning models to the global model for making complete knowledge of the problem. The Boolean 11-multiplexer problems are used for the execution. Hence, the main objective of this work is to keep the average accuracy of distributed mode without loosing accuracy rate compared to models. The experimental results showed that the testing accuracy of distributed mode is higher than other models.