Answers are grounded in this paper and cite their sources.
Geological mapping relies heavily on expert interpretation, guided by tacit conceptual models that are rarely formalised.As a result, rock-type classifications often carry significant yet unquantified subjective uncertainty.This study proposes a method to capture and structure such uncertainty in the form of a relational uncertainty matrix which represents a human perception-based rock identification confusion (RIC) matrix.As part of a regional geological mapping project in Trøndelag, Norway, 11 geologists independently assessed their perceived uncertainty in discriminating between rock pairs among 35 rocktype classes found in the mapping area.The resulting pairwise ratings were aggregated into a rock identification confusion matrix that reflects the collective interpretive ambiguity across the expert group.The matrix offers a scalable and reproducible format for encoding expert knowledge about classification uncertainty without requiring consensus formation or relying on qualitative self-assessment.This fills a critical gap for uncertainty-aware geological modelling and loss-sensitive classification workflows.By encoding subjective geological knowledge in a structured, machine-readable format, this approach lays the groundwork for improving transparency and comparability in geological interpretation workflows from localised field surveys to regional and transnational geological mapping programs.
No takes yet. Share an insight, caveat, or question.
Ksienzyk et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: