Parallelization of ensemble neural networks for spatial land-use modeling

摘要: Artificial neural networks have been widely applied to spatial modeling and knowledge discovery because of their high-level intelligence flexibility. Their highly parallel distributed structure makes them inherently suitable for computing. As the technology high-performance computing evolves resources become more available, new opportunities exist network models benefit from this advancement in terms better handling computational data intensity associated with problems. In study, we present a hybrid ensemble approach land-use change. Our combines shared-memory paradigm embarrassingly method by leveraging power multicore computer clusters. The efficacy is demonstrated parallelization Fuzzy ARTMAP models, which extensively used applications. We adopt an train multiple make use entire dataset simultaneously. evaluate proposed examining performance variation training datasets alternative sizes. Experimental results reveal great potential higher achievement when our large