ECONOMETRIC MODELING OF THE INVESTMENT ATTRACTIVENESS OF IT PROJECTS USING LOGISTIC REGRESSION

Abstract

The purpose of this study is to develop an econometric toolkit for identifying and quantitatively assessing the factors influencing the probability of attracting external financing for IT projects in the context of the digital economy.

The methodological framework is based on econometric modeling techniques, in particular binary logistic regression (logit model), as well as system analysis, statistical data processing methods, and approaches to handling imbalanced datasets (reweighting and undersampling). Model validation was performed using the Wald test, likelihood ratio test, and the Hosmer-Lemeshow goodness-of-fit test.

Based on the modeling of a sample of Ukrainian IT projects, key determinants of successful fundraising were identified. The presence of early-stage investors was found to have a decisive impact, significantly increasing the likelihood of obtaining financing. A statistically significant positive effect of project commercial orientation and digital presence was established. Furthermore, a synergistic effect between technological innovation, particularly the use of artificial intelligence, and operational flexibility – manifested in reduced time-to-market – was substantiated.

The proposed approach can be applied as a decision-support tool in IT project management, particularly for assessing investment attractiveness, optimizing development strategies, and enhancing the efficiency of commercialization processes for innovative products. The practical significance of the developed models lies in their ability to provide IT managers and venture investors with an analytical toolkit for auditing initiatives at early stages, thereby minimizing the «valley of death» risks and optimizing financial resource allocation. The results contribute to risk reduction at the early stages of the project lifecycle and improve the robustness of managerial decision-making. Future research directions involve the use of recurrent neural networks to forecast the long-term evolutionary trajectories of IT products in a global digital market.