USANDO ANÁLISES SOCIAIS NA IDENTIFICAÇÃO DE NÓS RELEVANTES EM UM CENÁRIO MULTIRREDES: OPERAÇÃO LICITANTE FANTASMA, UM ESTUDO DE CASO / USING SOCIAL ANALYSIS TO IDENTIFY RELEVANT NODES IN A MULTI-NETWORK SCENARIO: THE GHOST BIDDER OPERATION, A CASE STUDY

This paper proposes the NDNS (Nodes Detection using Network Science) model that, using complex networks theory, aims to find the most relevant nodes in a multi-network scenario in a more efficient way than well-known and established centrality metrics. The paper uses, as a case study, a Brazillian corruption investigation of public tenders known as Ghost Bidder operation. Considering four years of investigation, the NDNS model, when compared to four metrics of centrality (betweenness, eigenvector, weighted degree, page rank, and its normalized geometric mean), achieved 93% precision and 94% recall in detecting fraudulent values against 38% and 51%, respectively, of the second-best positioned measures.