We present the Distributed Executor for RSES (DIXER) which is a supplementary software for the Rough Set Exploration System (RSES). It takes an advantage of grid computing paradigm and allows to shorten the time necessary for experiments by employing all available workstations to run a scenario of experiments. DIXER software includes most important rough set classification algorithms from RSES and also other algorithms for distributed machine learning. It creates an easy to operate and utilize platform for grid computations and provides a robust and fault tolerant environment for computation-heavy experiments. We provide also experimental evaluation of DIXER that proves at least 96% efficiency in parallelization.
Jan G. Bazan, Rafal Latkowski, Marcin S. Szczuka