A User-Friendly Framework for Multilanguage ANN Generation: Real Case Applications

Abstract

This article presents a framework, developed in JAVA, that facilitates the design and implementation of artificial neural networks and provides automatic multilanguage code generation (ANSI C and JAVA). The framework consists of three systems: a library, in which we factorize network-related functionalities, and two tools, an interpreter and a Web application, that provide ANN-related features. The functioning of this tool was tested by applying it to two problems in two different fields.