The role of network topological structure and semantic features in creating verbal aggression

Abstract

In this article, we studied the role of the topological structure of semantic networking in creating verbal aggression. It is shown that centralities such as degree, betweenness and closeness play an important role in the activation of verbal aggression in the network. We have also shown that aggressive labelled nodes with spectral clustering in different spectra are often divided into two groups, with the larger group activating more aggressive labelled nodes. In addition, the parameter of eccentric distribution from the origin is introduced to study the dispersion of aggressive nodes around the specific nodes. Hence, studying two networks with different contexts shows that the dispersion of nodes with aggressive labelling around the network's hub, as the centre of the network with political context, is much more than artistic context. In addition, different clusters of verbal aggression in the political and artistic context have the same pattern of frequency. In addition, we investigated semantic features in creating verbal aggression, showing that non-aggressive words are prone to create verbal aggression as much as aggressive words.