A comparative study for landslide susceptibility mapping using GIS-based multi-criteria decision analysis (MCDA), logistic regression (LR) and association rule mining (ARM)

摘要: Abstract Landslide susceptibility mapping is one of the crucial stages landslide hazard and risk assessment. Moreover, maps assist planners, local administrations, decision makers in disaster planning. Various approaches have been applied literature to increase accuracy maps. The determination suitable method plays critical role for obtaining required accuracy. In this study, performances three quantitative methods are evaluated. logistic regression (LR) analysis typical example statistical methods, while GIS-based multi-criteria analyses (MCDA) a ssociation rule mining (ARM) examples heuristic data respectively. based on obtained Şavsat Artvin province (NE Turkey ) where region has intense landslides. influencing parameters study area lithology, land use/land cover, soil type, erosion, altitude, slope, aspect, distance drainage, depth, fault, road. models then compared evaluated by using pixel-based evaluation metrics. Results showed that ARM provides higher quality percent (QP), whereas damage detection (PDD) LR method. lowest QP MCDA. I t found better than MCDA modeling they can be integrated obtain performance.