Journal paper by Sinčić et al. has been published

In Remote Sensing Journal, a scientific paper developed in the scope of the LandSlidePlan project titled „A Comprehensive Comparison of Stable and Unstable Sampling Strategies in Large-Scale Landslide Susceptibility Models Using Machine Learning Methods“ was published. Authors of the paper are Marko Sinčić, Sanja Bernat Gazibara, Mauro Rossi, Martin Krkač and Snježana Mihalić Arbanas. The paper presents eight different strategies for using stable and unstalbe terrain samples in landslide susceptibility modelling. The research was conducted as a landslide susceptibility assesment in large scale, on a LandSlidePlan project pilot area, i.e. Hrvatsko Zagorje. By using four machine learning methods, i.e. Logistic Regression, Neural Network, Random Forests and Support Vector Machine, 32 landslide susceptibility maps were developed which were classified into zones of very low, low, medium, high and very high susceptibility. Comparison of parameters that characterize each landslide susceptibility map resulted in advantages and disadvantages of each sampling strategy and machine learning method which were carried out as conclusions.