2017 Volume 10 Issue 3 Pages 120-129
In the present study, we aim to assess landslide susceptibility and optimize causative factors using artificial neural network method in Ambon, Indonesia. Elevation, slope angle, slope aspect, lithology, geological density, proximity to river, proximity to faults, and proximity to road networks were chosen as the causative factors. Based on the obtained results, proximity to river and slope aspect were the least influential causative factors in the study, these two causative factors were then eliminated for the optimized landslide model. Proximity to road and geological density were proved to be the most influential causative factors. The six causative factors landslide susceptibility model returned better accuracy when compared to the eight causative factors landslide susceptibility model. The output susceptibility maps were reclassified into five classes ranging from very low to very high susceptibility using Jenks natural break method. 20% of all mapped landslides were used as the validation of the susceptibility models. Receiver operating curves (ROCs) were calculated, the areas under the curve (AUC) for the success rate curves of six factors landslide susceptibility map and eight factors landslide susceptibility maps were 0.770 and 0.734, respectively. The AUC for the prediction rate curve for the six factors and eight factors landslide susceptibility maps were 0.777 and 0.717, respectively.
