Neural network classification of technological development projects in Russian companies: Perspectives of extreme project management

Abstract

To achieve the goals of sustainable socio-economic development and ensure the technological sovereignty of the Russian Federation, it is critically important to choose effective methods of technological project management. The article develops a novel toolkit for neural network classification of technological development projects in Russian companies and justifies the use of extreme project management to such endeavors. The methodological framework resides in the concepts of project lifecycle management and data mining. The study employed the following research methods: textual analysis of project documentation using large language models, and intelligent project classification based on two-dimensional projection of multidimensional clusters using Orange Data Mining. The research draws upon regulatory documents in the field of Russia’s scientific and technological development and open databases on Russian companies’ projects. The work proposes a new neural network classification toolkit based on a large language model. We have found that most development projects related to critical and end-to-end technologies are characterized by low certainty of goals and solutions, which necessitates applying extreme approaches to manage them. The findings can be used by Russian companies for reasoning the choice of a project management model, as well as by experts when evaluating technological development projects.