piSVM1.2.1 Analytics Training (10x-CV optimized) Indian Pines Images Processed 30 Features 52 Classes


piSVM version 1.2.1 Deployment JUDGE cluster of the Juelich Supercomputing Centre in Germany Parallel Support Vector Machine (SVM) classification runs to analyse remote sensing hyperspectral images (52 land cover classes, 30 features). Indian processed: 1417x614x30 (training 10% and test) Data preprocessed with PCA, ESDAP, NWFE. Parameters have been optimized using 10-fold cross-validation, more information available at: http://hdl.handle.net/11304/5bba8e36-fe63-11e4-8a18-f31aa6f4d448 Supplemental material for paper study. Corresponding dataset (processed) available at: http://hdl.handle.net/11304/7e8eec8e-ad61-11e4-ac7e-860aa0063d1f