Application of Neural Networks for Landslide Susceptibility Mapping in Turkey

摘要: Landslides are a major natural hazard in many areas of the world, and globally they cause hundreds billions dollars damage, thousands deaths injuries each year. second most common Turkey, Black Sea region that country is particularly affected. Therefore, landslide susceptibility mapping one important issues for urban rural planning Turkey. The reliability these maps depends mostly on amount quality available data used, as well selection robust methodology. Although statistical methods generally have been implemented used evaluating risk medium scale studies, distribution-based cannot handle multi-source commonly collected from nature. These drawbacks responsible on-going investigations into slope instability. To overcome weaknesses, desired technique must be able to multi-type its superiority should increase dimensionality and/or non-linearity problem increases - which when traditional regression often fails produce accurate approximations. neural networks some problems with creation architectures, processing time, negative “black box” syndrome, still an advantage over can deal comprehensively insensitive uncertain measurement errors. it expected application will bring new perspectives assessment In this paper, examined their performance component spatial decision support systems discussed.