Abstract:In this study, we introduce the disruption coefficient (DIO) and development coefficient (DEO) as basic measures for detecting critical nodes in directed networks with the perspective of innovation. To expand the information dimension, we further propose the 2-step versions of the measures (2stepDIO and 2stepDEO). The theoretical foundations of this study are information theory (Shannon, 1948) and disruption indices (Wu et al., 2019). The new measures combine the focal nodesâ€™ prior information and posterior information, by measuring the extent to which they disrupt the previous network paradigm, which distinguishes them from traditional measures, e.g., degree and betweenness. Moreover, the properties of the measures are elaborated, with an empirical study on citation networks for papers published in Journal of Informetrics from 2010-2017. We also checked the convergent validity and consistency of the metrics. It turns out that DIO and 2stepDIO have significant yet weak correlations with other measures. Based on these findings, we argue that DIO and 2stepDIO are vital and irreplaceable for detecting the critical nodes in directed information networks, as they can reflect the innovation level rather than merely the influence.

Disruption Coefficient, Development Coefficient and Their 2-Step Versions As Basic Measures in Directed Information Networks