Akademik Veri Yönetim Sistemi
Iranian Journal of Science and Technology - Transactions of Civil Engineering, 2024 (SCI-Expanded)
This study focuses on landslide susceptibility assessment of the area between Güzelyalı and Lapseki (Çanakkale, Türkiye) by using logistic regression, artificial neural network (ANN) and support vector machine methods. Nine input parameters such as topographic elevation, lithology, slope, land use, aspect, curvature, distance to streams, TWI, and NDVI were selected as the landslide conditioning parameters. The frequency ratio values were also calculated for the parameters and their subclasses and were assigned to express all continuous and categorical input parameters in the same scale for the considered prediction models. In addition, sensitivity (Recall), accuracy, precision, kappa indexes, F1-score and receiver operating characteristic based on area under curve approach were calculated to assess the performances of the so produced landslide susceptibility maps. Considering all performance indicators, the most successful model was revealed as the map produced by ANN model. Producing such maps, testing their performances and using them into the practice, sustainability can be achieved in regional planning, land use and urban development stages. More importantly, a fundamental step will be taken for future works such as hazard and risk assessments in the region.