RE‐IMPLEMENTACIÓN DEL PAQUETE AMORE

Abstract

The AMORE package for neural network training and simulation has been developed by the EDMANS research group as a natural consequence of the research projects developed by the group. The virtues and usefulness of the package in its current state of development have been recognized not only by the EDMANS group itself, where it has become a widely used tool, but by the scientific community abroad our borders, as the citing data evidences. Nevertheless, the consolidation of the PMML specification, currently in version 4.0, as a de facto standard, as well as the recent extension of the R language to the object oriented programming field, claims for an upgrade of the package. This paper shows some of the main features that the new version of the AMORE package will display once these improvements are implemented following this new approach. Keywords: EDMANS, data mining, neural networks, quality, industrial processes.