Banca de DEFESA: ANDRÉIA CRISTINA DOS SANTOS GUSMÃO
Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.STUDENT : ANDRÉIA CRISTINA DOS SANTOS GUSMÃO
DATE: 22/05/2023
TIME: 14:00
LOCAL: Remoto, via GoogleMeet https://meet.google.com/ckq-uymh-kbw
TITLE:
Identification of Cliques in Scientific Collaboration Networks: an algorithm based on chordal graphs
PAGES: 108
BIG AREA: Ciências Exatas e da Terra
AREA: Ciência da Computação
SUBÁREA: Metodologia e Técnicas da Computação
SPECIALTY: Sistemas de Informação
SUMMARY:
This work proposes an algorithm for identifying cliques in scientific collaboration networks based on chordal graphs, where vertices are represented by co-authors, edges by co-authorship, and the weight of each edge is the number of publications existing for each pair of vertices. Enumerating all possible combinations of a researcher in a co-authorship network is a difficult problem, as the size of the co-author group and the number of co-authors with whom this researcher has published over their academic career increase. The proposed algorithm aims to identify all relevant co-author cliques, i.e., those who have published together at least once without clique duplication, allowing for analysis of the behavior of individual-centered co-authorship collaborations and the hierarchy of academic collaboration. One possible application of the algorithm addressed in this paper is to measure the longevity (time) and size of co-authorship (number of co-authors), enabling analysis of researcher trajectories.
We conducted experiments using the DBLP database, filtering publications by type: event articles, books, and journal articles, and the results obtained reveal that smaller groups have greater longevity when compared to larger groups, for all types of publications and also for the entire database without division by type. We used two measures for calculating longevity: (1) considering the difference between the year of the last publication and the year of the first publication; and (2) considering only the number of years in which there were publications. No group with more than 13 co-authors published together for more than 10 years. The largest group size is 449 co-authors and occurred only once, for a journal article, and the longest longevity is a group of two co-authors, with 55 years and 41 years for longevity calculations (1) and (2), respectively.
Regarding the hierarchy of cliques, for the observed cases, heart cliques prevailed as influencers of daughter cliques, as they had the first publication. Regarding heart cliques with the most relevant cliques, the scenario is that heart cliques are consequences of co-authorship collaboration, i.e., larger co-author groups that over time decreased, originating new smaller co-author groups.
This paper contributes (i) with a novel algorithm that allows for analysis of researcher behavior over time; and (ii) a novel characterization of trajectories (in terms of longevity and size), allowing observation of each researcher's behavior not in isolation/independence but in a group/dependent on collaborative and social interaction, as science is perceived today.
COMMITTEE MEMBERS:
Presidente - Interno ao Programa - 1934625 - JESUS PASCUAL MENA CHALCO
Membro Titular - Examinador(a) Interno ao Programa - 1722875 - DAVID CORREA MARTINS JUNIOR
Membro Titular - Examinador(a) Externo ao Programa - 2364326 - ALEXANDRE DONIZETI ALVES
Membro Titular - Examinador(a) Externo à Instituição - FABIO CASTRO GOUVEIA - FIOCRUZ - RJ
Membro Titular - Examinador(a) Externo à Instituição - Ivan Torres Pisa
Membro Suplente - Examinador(a) Externo à Instituição - WASHINGTON LUÍS RIBEIRO DE CARVALHO SEGUNDO
Membro Suplente - Examinador(a) Externo à Instituição - FABIO MASCARENHAS E SILVA - UFPE