NeuroFramework: A package based on neuroevolutionary algorithms to estimate the melting temperature of ionic liquids

Abstract

In this paper, a Neuroevolutionary framework is presented for training and testing neuroevolutionary algorithms. This algorithm offers flexibility in its design by the stacking of operators to form an algorithm prototype. This approach allows a faster algorithm design where operators can be stacked in the training phase and can even be managed by a dynamic controller. Neural Networks are represented on a graph without layers. This software was designed in modularity and new features can be easily added. This software includes a wrapper for Python for machine learning tasks, especially regression; however, the user can adapt this package for classification.

Keywords

• Neuroevolutionary algorithms
• Ionic liquids
• Neural networks