Metaheuristic Ant Lion and Moth Flame Optimization-Based Novel Approach for Automatic Detection of Hate Speech in Online Social Networks

In the online social networks, blogs, microblogs, bookmarking services and sharing sites, various web forum pages; of knowledge, opinions, ideas, etc. are spreading very quickly. This situation brings dangerous problems in networks. One these is hate speech detection (HSD) problem which covering issues such as insults, swearing, humiliation, discrimination, exclusion, detest, abhor, blast, damn, intolerance. These can be reactions to a person, group, an organization, order, or event. Although few machine learning methods have been used literature solve this important media, performance HSD models terms many metrics needs increased. study, automatic system based on metaheuristic methodology was proposed for better results new problem. optimization approach, Ant Lion Optimization (ALO) algorithm Moth Flame (MFO) were designed first attempt use algorithms solution search strategies HSD. An efficient representation scheme flexible fitness function purpose. Many easily embedded into order simultaneously optimized. Firstly, basic natural language processing (NLP) steps carried out. Feature extraction performed using Bag Words (BoW), Term Frequency (TF), document vector (Word2Vec). Then, performances novel approaches analyzed detail three different real-world data. The obtained also checked against eight popular supervised algorithms, Social Spider (SSO) algorithm, state-of-the-art Tunicate Swarm Algorithm (TSA). Considering evaluation criteria sets experiments, it observed that accuracy, sensitivity, precision, f-score ALO MFO superior methods. As result experimental studies, highest accuracy value 92.1% ALO, while 90.7% MFO. Other numerical values study given experiments section with tables graphics detail. Due promising approaches, they anticipated media networking problems.