Author Identification via a Distributed Neural-Evolutionary Hybrid (DiNEH)

In this paper, we propose a non-traditional Genetic & Evolutionary Feature Selection (GEFeS) method for Author Identification. This method evolves a distributed feature vector for each author and it's therefore distributed. We refer to this new approach as a distributed neural evolutionary hybrid (DiNEH). We compare the performance of DiNEH with a number of well-known Authorship Attribution Techniques (AATs) from the literature. DiNEH was able to outperform all of the AATs on one dataset and was second on the second dataset by a narrow margin.