Citation

Ghorbanzadeh, O., Rostamzadeh, H., Blaschke, T., Gholaminia, K., Aryal, J. (2018). A new GIS-based data mining technique using an adaptive neuro-fuzzy inference system (ANFIS) and k-fold cross-validation approach for land subsidence susceptibility mapping. Natural Hazards, 94: 497-517. Link to paper

Abstract

This study combines geographic information system (GIS) technology with an adaptive neuro-fuzzy inference system (ANFIS) for land subsidence susceptibility mapping in the Marand plain, northwest Iran. The research evaluated the predictive performance of ANFIS using six different membership functions to determine the most effective configuration. The receiver operating characteristic (ROC) analyses for the six land subsidence susceptibility maps yielded very high prediction values for two methods: the difference of sigmoid membership function (DsigMF) with an AUC of 0.958 and the Gaussian membership function (GaussMF) with an AUC of 0.951. A k-fold cross-validation approach was implemented to ensure robust model evaluation and reduce overfitting. The methodology demonstrates that ANFIS can effectively integrate multiple conditioning factors including groundwater levels, geology, soil properties, and land use patterns to predict areas susceptible to land subsidence. This research contributes to advancing geohazard susceptibility assessment by combining fuzzy logic's ability to handle uncertainty with neural networks' learning capabilities, providing a powerful tool for land use planning and subsidence risk management in regions threatened by ground deformation.

AI Visual Interpretation (Beta)