Research seminar “Using AI to predict the success of the tender, bidders and winners in some of Prozorro markets”


Is it possible to know in advance who will win the Prozorro tender that had just been announced? Any policy maker would answer – it’s impossible. Prior to the launch of Prozorro in 2016, sometimes the winners were determined in the government officials’ officials. Unlike at the times of the “Visnyk of public procurement”, Prozorro is positioned as the place where the name of the successful bidder and whether it will be successful (productive) cannot be known in advance  – this is an unpredictable environment where the result is unknown in advance.


But it is quite realistic to predict the likelihood of a certain event, using machine learning algorithms and historical data produced by the system. We will take a closer look at machine learning algorithms and neural networks that can predict, with 60-65% accuracy, the winner of the tender and the success of the tender on examples of the fuel and medicines markets.