Development of a transactional multifactor econometric platform for the integral assessment of the fundamental value of a public energy corporation in the context of geo-economic fragmentation, sanctions regimes and stochastic volatility of commodity

In the context of modern geo-economic turbulence, characterized by the escalation of sanctions restrictions, the structural transformation of global energy markets, the monetary restriction of leading central banks and the increasing stochasticity of price indicators for hydrocarbons, traditional paradigms for assessing the investment attractiveness of public companies in the oil and gas sector demonstrate systemic insufficiency. This study is devoted to the development of a comprehensive transactional methodology that integrates quantitative analysis of a wide range of endogenous and exogenous determinants of corporate value, from global price targets and macroeconomic aggregates to indicators of corporate governance, technological sovereignty, ESG factors and force majeure. Using an ensemble of advanced machine learning algorithms, including gradient boosting on decision trees, recurrent neural network architectures with a selective attention mechanism and regularized regression models with an elastic network, a multifactorial model has been built that allows quantitative decomposition of the contribution of each factor to the dynamics of market capitalization and identification of key cost drivers in various phases of the economic cycle. Based on the identified correlation and regression dependencies, an end-to–end predictive platform has been developed that converges three classical assessment approaches – multiplier, cash flow discounting, and dividend discounting - each of which is modified through dynamic ML adaptation of key parameters to the current macroeconomic and geopolitical environment, taking into account nonlinear effects, interfacial interactions, and time lags. The final block of the methodological pipeline includes a multivariate scenario analysis with the generation of five development trajectories – from extremely pessimistic to extremely optimistic - based on the joint variation of key external parameters, the construction of probabilistic confidence intervals using multidimensional logit transformation and filtering of the estimates obtained through multi-technical indicators, dividend protection and qualitative adjustments that take into account premiums for ESG rating, corporate governance and technological sovereignty. Practical testing of the developed toolkit on a retrospective dashboard for 2010-2026 has allowed us to determine a reasonable fair value price range structured across five investment zones – from aggressive accumulation to full profit–taking - which provides institutional and private investors with a clear navigational tool for decision-making in the face of fundamental uncertainty. The proposed methodology, due to its modular architecture, multifactor saturation and the possibility of regular recalibration based on updated data, is a universal platform for evaluating companies in the commodity sector in turbulent economic conditions and can serve as a theoretical and methodological basis for further research in the field of financial engineering and investment analysis.