Adaptive Neuro-Fuzzy Inference System (ANFIS) Method for Developing a Decision Support System for Determining Landslide Susceptibility

Abstract

Landslide catastrophes are one of the disasters that frequently occur in Indonesia owing to the weather and climatic features, regional terrain, and geological formations that make this nation prone to landslides. The primary goal of this research is to compare the application of the fuzzy logic technique and the adaptive neuro-fuzzy inference system (ANFIS) approach to landslide detection sensors based on prior research in order to identify landslide-prone locations more easily. The Adaptive Neuro-Fuzzy Inference System (ANFIS) technique analyzes the landslide area using three factors. Rainfall, land slope, and soil moisture are examples of these factors. This variable is used to assess the area's level of vulnerability to landslides: very safe, relatively safe, relatively potential, potential, and very potential. In the study, each piece of data is subjected to a training and testing procedure to identify landslide vulnerability, with the factors and weighting methods aligned with current government standards. This study compares the rules outcomes to those of past studies as well as the system results. Based on the studies findings, it can be stated that the decision support system for the degree of landslide vulnerability utilizing the ANFIS approach is superior to the fuzzy logic method, with an accuracy rate of 86.21%.