Identification of public procurement contracts with a high risk of non-performance based on neural networks

摘要

The article is devoted to the use of machine learning methods, neural networks in particular, to predict the results of the execution of public procurement contracts. In the framework of the study, a classification of public contracts in pipe industry was carried out in order to identify contracts with high risk of non-performance, a Python application had been developed for the solution of the problem, which implemented compositions of neural network models of the architecture proposed in this work. The results of mathematical models application are described and their accuracy is analyzed. It is noted that the developed software product can be used for solving the described classification problem using neural network models with proper selection of training data.