Detecção de casos suspeitos de fraudes em licitações realizadas nos municípios da Paraíba: uma aplicação de técnicas de mineração de dados

Abstract

This dissertation aims to detect collusive bidding schemes in government procurement auctions during 2005 to 2016 in the municipalities of Paraíba state (Brazil). We use data from Court of Audit of the State of Paraíba and data mining techniques to identify association rules among bidders. The association rules were evaluated considering patterns of cooperative strategies in repeated games, taking account the bid winner likelihood and the bidders competition on average. The finds suggest many cases of biddercollusion,includingmealssuppliers,cleaningservices,rentalofvehicles,copiers, stages, chemical toilets, fuel suppliers, automotive parts, sports equipments, medicines and hospital material, consulting in accounting, civil engineering and advertising. We also find several patterns of association between bidders with evidence of competition simulation and concentrated activity in a few municipalities. In general, the results provide important guidelines to support the work of government audition.