Application of prediction–area plot in the assessment of MCDM methods through VIKOR, PROMETHEE II, and permutation

One of the most important natural hazards is landslides that after the earthquake and floods cause the highest damage to humans. Nowadays, landslide events are taken into consideration by the experts in this domain and their mapping and detecting always is one of the important concerns. One of the useful methods is multi-criteria decision making that is used in order to prioritize landslide-prone zones. These methods perform based on several quantitative or qualitative criteria. Multi-criteria decision-making methods that have been the focus of researchers in recent decades use several measurement criteria instead of one. However, due to the broadness and diversity of existing methods the selection of a proper approach has become cumbersome; therefore, selecting a reliable technique for landslide evaluation can be of a great interest. In this study MCDM methods (VIKOR, PROMETHEE II, and permutation) were applied in order to detect and map landslide-prone points. Then obtained results were assessed by using prediction area (P–A) technique. Consequently, the intersection points of the P–A plot for PROMETHEE II model show 82 percent of landslide occurrences predicted only in 18 percent of the study area, while the intersection point of the P–A plot for permutation and VIKOR models shows 67 and 73 percent of landslide occurrences predicted only in 33 and 27 percent of the study area, respectively.