CD-SEIZ: Cognition-Driven SEIZ Compartmental Model for the Prediction of\n Information Cascades on Twitter

Information spreading social media platforms has become ubiquitous in our\nlives due to viral information propagation regardless of its veracity. Some\ninformation cascades turn out to be viral since they circulated rapidly on the\nInternet. The uncontrollable virality of manipulated or disorientated true\ninformation (fake news) might be quite harmful, while the spread of the true\nnews is advantageous, especially in emergencies. We tackle the problem of\npredicting information cascades by presenting a novel variant of SEIZ\n(Susceptible/ Exposed/ Infected/ Skeptics) model that outperforms the original\nversion by taking into account the cognitive processing depth of users. We\ndefine an information cascade as the set of social media users' reactions to\nthe original content which requires at least minimal physical and cognitive\neffort; therefore, we considered retweet/ reply/ quote (mention) activities and\ntested our framework on the Syrian White Helmets Twitter data set from April\n1st, 2018 to April 30th, 2019. In the prediction of cascade pattern via\ntraditional compartmental models, all the activities are grouped, and their\nsummation is taken into account; however, transition rates between compartments\nshould vary according to the activity type since their requirements of physical\nand cognitive efforts are not same. Based on this assumption, we design a\ncognition-driven SEIZ (CD-SEIZ) model in the prediction of information cascades\non Twitter. We tested SIS, SEIZ, and CD-SEIZ models on 1000 Twitter cascades\nand found that CD-SEIZ has a significantly low fitting error and provides a\nstatistically more accurate estimation.\n