Landslide susceptibility assessment by using a neuro-fuzzy model: a case study in the Rupestrian heritage rich area of Matera

A novel neuro-fuzzy model successfully mapped landslide susceptibility in the UNESCO-listed Natural Archaeological Park of Matera, Italy. This approach combined neural network training with fuzzy logic to analyze slope instability parameters—such as curvature, elevation, lithology, and fracture density—and produced a high-accuracy susceptibility map validated through ROC analysis, confusion matrix, and SCAI methods. The model outperformed traditional techniques by integrating sensitivity analysis to address uncertainties in membership functions, offering a robust tool for heritage risk assessment in complex, rupestrian landscapes.

Sdao, F., Lioi, D. S., Pascale, S., Caniani, D., & Mancini, I. M. (2013). Landslide susceptibility assessment by using a neuro-fuzzy model: a case study in the Rupestrian heritage rich area of Matera. Natural Hazards and Earth System Sciences, 13(2), 395–407. https://doi.org/10.5194/nhess-13-395-2013