A Peer Recommendation Model to Avoid Hate Speech Engagements in Multiplex Social Networks

Multiplex social network relationships are quite strong in most occurrences, especially within a strong peer network (a cluster of near engaging friends). Moreover, hate speech is found on most online social media platforms. Hence, this study aims to identify hate speech discussions among peer networks. This paper discusses a novel model to recommend a peer under the context of multiplex social networks to minimize the hate speech engagements; Facebook, Twitter, and YouTube social media networks (SMN) were used in this experiment. Collaborative filtering defines an interest-based recommendation model. Under the context of user engagements, some topics become of more user interest. Hence, some social media posts drastically spread over multiplex layers rapidly, initiating a high social impact on a specific topic. The research gap is identifying the peer network that reduces hate speech in multiplex social networks. Hence, this study provides a social innovation platform for peer recommendations to avoid social splits. First, this research contributes by proposing a novel methodology for identifying user engagements on online social networks by mining interactive social network graphs. Secondly, it provides an algorithm for recommending a multi-dimensional recommendation model by using collaborative filtering. Upon the proposed algorithm, a system that recommends engagements in any given online social network to minimize hate speech was implemented. Accordingly, the novel algorithm evaluates by using recommendation precision. The results show that the novel algorithm is highly applicable for peer recommendation in multiplex social networks to avoid hate speech discussions.