AN INTEGRAL CRITERION FOR SUBSTANTIATING IT-PROJECT SELECTION DECISIONS BASED ON NEURO-FUZZY MODELLING

Abstract

In the digital economy, decisions on the selection and scaling of information technology projects are made under pronounced uncertainty and a high risk of capital loss. The economic nature of digital products – their intangibility, knowledge intensity, network effects and rapid technological obsolescence – sits uneasily with the traditional investmentanalysis toolkit, which presumes deterministic cash flows and fully specified input data. At the same time, investment practice is shifting towards responsible investing, in which environmental, social and governance considerations are weighed alongside financial ones, and conformity with sustainability requirements increasingly conditions access to institutional funding. This raises the question of how to combine the financial and non-financial dimensions of a project within a single, transparent and reproducible measure of its attractiveness, and how to do so when the available data are incomplete and heterogeneous. The article discusses an approach to constructing such an integral decision-making criterion for IT-project selection on the basis of neuro-fuzzy modelling. It considers how heterogeneous project characteristics can be ordered into a multi-level hierarchy that preserves their economic interpretation; how the parameters of the membership functions of linguistic terms can be derived from neural-network clustering rather than from expert judgement, thereby reducing subjectivity; and how a compositional convolution rule can aggregate qualitative and quantitative indicators without a rule base or expert weighting coefficients. Attention is also given to the design of a rating scale that expresses the sustainability of a project – its capacity to remain viable and to keep developing under uncertainty and crisis-driven fluctuations of the market – and to the problem of choosing an evaluation model as a compromise between predictive accuracy and computational complexity. The article thus foregrounds the economic meaning of sustainable, resourceaware project management and the role of a reproducible attractiveness measure as a precondition for transparent responsible investing.