Prediction of links and vertices in scientific communities

• Grant number: 18/07444-5
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: August 01, 2018
• End date: July 31, 2019
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Alneu de Andrade Lopes
• Grantee: Gustavo Sutter Pessurno de Carvalho
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

Complex networks is a novel research area in artificial intelligence to deal with real systems in which different parts interact. This systems can be social, biological, technological, or related to information represented as graphs with non-trivial topology and, in general, large-scale. Such graph, which we refer to as network, is represented by a set of vertices and a set of edges (or links). The study of network evolution is important in many domains, for example in social networks where scientific collaboration networks stands out. Many effort has been made in this context, among which in link prediction techniques, methods capable of predicting the appearance of a possible relation between two researchers not yet connected. However, despite such advances in link prediction, to the extent of our knowledge there is no technique capable of predicting new vertex in the network. Namely, there are methods capable of predicting the appearance of new relations between researchers already in the network. This projects aims to study solutions for the problem, proposing a new strategy for predicting the appearance of vertices in a social network of scientific collaboration. With such advance it would be possible, for example, to better estimate the evolution of networks and to give more accurate recommendation for the cooperation between different research groups focusing on their evolution. (AU)