Measuring topic divergence in document networks and its possible applications

The paper proposes a new step to promote interdisciplinary propagation of such a popular in the fields of applied statistics and machine learning approach, as the topic modeling (TM). We introduce TM-based concept of «topical divergence» to account for the document (or its author) individuality with regard to its local network neighborhood. We propose several possible interdisciplinary applications of topical divergence related to sociology and management of science.