APPLICATION OF ARTIFICIAL NEURAL NETWORK AND LOGISTIC REGRESSION METHODS TO LANDSLIDE SUSCEPTIBILITY MAPPING AND COMPARISON OF THE RESULTS FOR THE ULUS DISTRICT, BARTIN

Eker A. M., Dikmen M., Cambazoglu S., Duzgun S. H. S. B., AKGÜN H.

JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY, cilt.27, sa.1, ss.163-173, 2012 (SCI-Expanded, Scopus)

Özet

The purpose of this study is to prepare a landslide susceptibility map for the Ulus district in Bartin, in the Black Sea region of Turkey, by utilizing logistic regression (LR) and artificial neural network (ANN) analyses. In this study, the landslide classification map prepared by The General Directorate of Mineral Research and Exploration (MTA) was used as a base map for landslide occurrence. The entire analyses were implemented with respect to active landslides. Fourteen explanatory variables were digitized, compiled and manipulated within a GIS environment. The study area was divided into 250 m x 250 m grid cells and Kernel density estimation technique was applied to obtain more meaningful population distribution over the area for the landslide inventory information. All variables were incorporated into this developed inventory data (dependent variable). The dependent variable was divided into two data sets as calibration and validation. Two different methods, LR and ANN analyses were utilized to find the relationship between the dependent and independent variables and also to compare the results of these techniques for the production of landslide susceptibility zonation, and to evaluate the optimum susceptibility zonation method.