AOM-MPA: Arabic Opinion Mining using Marine Predators Algorithm based Feature Selection
Diaa Salama AbdElminaam, Nabil Neggaz, Ibrahim A Gomaa, Fatma Helmy Ismail, Ahmed El-Sawy
Abstract
Recently, the automatic process of Arabic opinion mining (AOM) attracts the intention of scientific researchers, especially with the real challenges of the arabic language as dialectal language, Arabizi format, which combine multilingual forms. The prepossessing aims to realize a filtering operation on text, then TF-IDF is applied in order to transform the text to numerical vector. The main drawback of TF-IDF arises on the dimensionality. Thus, the feature selection plays a vital role for selecting the relevant words using swarm intelligence algorithms. The primary objective of this paper is to use the Marine Predators Algorithm (MPA) to determine the nature of opinions in the Arabic language in order to select the most pertinent terms. The experimental study is conducted on OCA data-set, and the results indicate that the proposed algorithm MPA outperforms other algorithms inspired by swarm intelligence and others from the literature in terms of accuracy and selection ratio.