Abstract
Existing empirical stability graphs are inadequate for narrow-vein stope stability prediction. This study explores the use of Fuzzy Inference Systems (FISs) as an alternative to conventional empirical methods for evaluating stope overbreak in narrow-vein mining. A FIS is a rule-based computational model that converts inputs into outputs using fuzzy logic, linguistic rules, and membership functions, making it well suited for handling epistemic uncertainty associated with narrow-vein mining problems. Stope performance data from an Australian mine was used to establish the FIS. The input parameters included stope walls, hydraulic radius, rock mass domain and stability number. The FIS predictions were in close agreement with the observed stope dilution. Overall, the FIS model demonstrated good performance, with accuracy, sensitivity and specificity values of 87%, 67% and 91%, respectively. These findings suggest that FIS can serve as a valuable tool for assessing stope overbreak in narrow-vein environments and can complement existing empirical methods to help reduce dilution and improve stope design outcomes.