Neural network modelling of regional innovation ecosystems

The emerging innovation-based economy, intense competition, rapid technological change combined with the need to perform business analysis of huge data flows, make it in creasingly important to introduce new effective approaches to regional development. The re search aims at creating a unified methodology for forming regional innovation ecosystems. The methodological basis of the study includes strategic management theory and regional governance. The paper proposes a number of methods to apply neural network modelling to gain new knowledge, generalize, and achieve a better understanding of the totality of facts and theo ries in the field of innovation ecosystems development in Russia. The authors perform neural network modelling using artificial neural networks – Kohonen self-organising maps and infor mation technologies; analyse the multidimensional space of innovative development indicators in regions by means of data mining. The research systematises current scholarly viewpoints on innovation ecosystems and offers a new perspective on the entrepreneurial ecosystem ap proach, which key players in the innovation economy, including the government, the business, the science and education, the public sector can adopt. The authors demonstrate the uneven ness of Russian regions’ innovative development and suggest a methodological approach for forming innovation ecosystems. When clustering data set of innovative development indicators, the authors distribute regions between the four clusters, thereby proving significant differences between the innovation ecosystems. The analysis of the results of the neural network modelling provides the practical and sound scientific underpinning for the development of innovation ecosystems. The research findings can be useful while formulating development strategies and programmes designed to stimulate innovative processes in the regions.