Early detection of systemic financial risks using neural networks based on blockchain market signals

Анотація

Traditional methods of analysis based on aggregated macroeconomic indicators are characterized by time lags and a limited ability to detect crises early. This necessitates the use of alternative sources of information and analysis tools that can provide early signals of financial instability. The purpose of the study is to substantiate a theoretical and methodological approach to the early detection of systemic financial risks using blockchain data and machine learning methods. The study used methods of generalization, analysis, comparison, and conceptual modeling, which allowed for the systematization of modern approaches to risk assessment, the identification of their limitations, and the justification of the feasibility of using innovative analytical tools. Results. The work structures systemic financial risks and identifies their key groups. The analysis results show that the most significant impact on financial stability is exerted by geopolitical and macroeconomic risks, which form the main sources of systemic instability in modern conditions. The analysis indicates the decisive influence of the state of economic activity, geopolitical threats, and inflationary processes on the formation of the most significant sources of systemic financial risks. It is substantiated that blockchain indicators, in particular, address activity, transaction volumes, exchange asset flows, and volatility indicators, allowing for the real-time reflection of market participants' behavioral reactions. A comparative analysis of traditional approaches and blockchain analytics has shown the advantages of using primary transactional data over aggregated macroeconomic indicators, due to greater objectivity, greater efficiency in obtaining information, and greater predictive potential. Conclusions. A conceptual model for early detection of systemic financial risks is proposed, integrating on-chain data and neural network algorithms to perform phased data processing, construct an integrated risk index, and classify it by levels, thereby providing a basis for informed management decisions. The practical significance of the results lies in their potential application by banks, regulators, and investors for monitoring financial stability, early detection of systemic risks, and crisis forecasting.