Enhanced Arabic Feature Selection using an Improved Whale Optimization Algorithm

Research in the field of Natural Language Processing for Arabic documents that incorporate metaheuristic techniques remain limited, primarily due to the intricate nature of the language, which encompasses challenges related to its structure, morphology, syntactic analysis, grammatical conventions, and derivational processes. This study introduces a novel approach for the processing of Arabic documents, which aims to enhance the whale optimization algorithm. This approach functions as a supervised learning technique, emulating the behaviors of whales, including bubble netting and spiral movements toward their targets. The proposed methodology represents an advancement of the traditional whale optimization algorithm, characterized by its adaptability and efficiency. It effectively integrates exploration and exploitation strategies, drawing inspiration from the natural behaviours of humpback whales while also incorporating the beneficial attributes of chaotic functions. This adaptability positions the method to address complex optimization challenges across various fields. Beside, the experiment show that enhanced method has a better result by using Firefly Algorithm classifier and the TF-IDFC-RF feature extraction.