Desenvolvimento de um mecanismo detector de anomalias em licitações municipais

Abstract

The bidding process is the method used by the public administration to make purchases and sales of resources consumed by an institution, for example, in the purchase of inputs for the municipality such as paper, pens, envelopes, even machines for asphalt revitalization. To assist in these bidding processes there are systems that monitor and detect these cases of anomalies, but these systems are, in general, geared towards applications in the federal and/or state public spheres. However, there have been increasingly frequent cases and indications of fraud in bidding processes, be it through overbilling of prices, directional fraud, document forgery, among others. When it comes to municipal public administration, there are no tools capable of identifying any kind of anomalies in this process, which arouses the practical and social relevance of this type of innovation. Thus, in this work we implemented intelligent computational resources capable of finding prices of similar items that, by comparing the prices of bid items and the prices found, track bidding frauds. With the results found it was possible to identify possible indications of fraud, in such a way that one of the items had a price 79% higher than the one found by the automated searcher, while the other item had a price 43% higher than the one found.