Semantic technologies on the mission: Preventing corruption in public procurement

Serbia recently decided to organize the country as a democracy, with major political goal to join European Union. This initiated adoption of new laws according to the best practice from European countries regulations. The area of public procurement was one of them, with major aim to increase the transparency and concurrency in the procurement process in order to make it more efficient. During last 10 years, we saw the rise and fall of procurement process in Serbia. Initially, new rules gave positive results, with constantly increasing amount of realized procurements and 7.2 as an average number of bidders per procurement in 2004. In order to support the process and make it more visible, Internet portal for procurement was deployed. Unfortunately, in the next period number of irregular procurements constantly increased while number of bidders decreased to an average of 2.7 in 2011. Analysts agree that major problems in the process are poor application of regulations and high level of corruption. Further, they identified three phases of procurement where corruption usually happen: planning phase, setting up estimation criteria and realization phase. Our proposal is to introduce semantic technologies in the process, in order to enable data manipulation by machines. We propose meta-model for the definition of procurement documents, and another one meta-model for the specification of alert rules, together with domain-specific language. This would enable an expert to set up rules regarding specific conditions, in order to be alerted about possibly irregular procurements. Application of proposed solution should enable earlier recognition of potentially irregular procurements, which can be prevented before realization or sanctioned prior to obsolescence. 2014 Elsevier B.V. All rights reserved.