Aerospace Instrument-Making Annotation << Back Processing Weakly Structured Information in Risk Management Tasks for Complex Technical Systems Rodionov D.G., Konnikov E.A., Polyakov P.A. Digitalization and the increasing complexity of man-made systems are leading to a rise in the volume of operational data and the number of text messages about events and deviations. A significant portion of this data remains poorly structured. This hinders the comparable assessment and prioritization of risk factors. This paper proposes a method for processing loosely structured information for interpretable risk modeling. The method translates factors into the space of infl uence (I), stochastic stability (S), and prevalence (P) indices. The approach is implemented by extracting factors from texts, constructing an accident rate model, and calculating indices followed by clustering. The method has been tested on a corpus of NRC messages with three accident levels: non-accidental, low, and high. A Monte Carlo test has been performed, and the division of factors into critical and routine has been shown, as well as the inverse relationship between I and S. The proposed approach is applicable to event logs, incident notifications, and reports in the nuclear power, electric power, oil and gas, chemical, and transportation industries. Keywords: stochastic stability index, risk factors, criticality, proactive safety management, Monte Carlo simulation, interpretable analysis, weakly structured data, ICE sensitivity analysis. DOI: 10.25791/aviakosmos.3.2026.1546 Pp. 53-61.