Machine learning of social dynamics in complex networks

Abstract

The interest in learning how rumors spreads in society has led researches to develop theoretical computational models to investigate the underlying mechanisms behind the spread of fake news in our society. Currently, literature on the topic provides a variety of epidemiological and cultural models. While each type of modeling adopts a different approach, all of them aim to resemble some aspects of the collective behavior. These models can be studied in different complex network topologies, and have been widely used to interpret the result of a given dynamic process. Social networks have become an important mean of exchanging information in contemporary society, increasing and diversifying the volume of data. Thus, data mining and machine learning techniques have become viable and advantageous means for the study of the spread of rumors. In this context, this project proposes the development of an approach based on machine learning algorithms to predict the outcome of a given dynamic process as a function of the structural characteristics of complex social networks. (AU)