Erosion Susceptibility mapping using Machine Learning and GIS: A case study of Kelantan

The issue of soil erosion in Kelantan resulted in mudflow, river bank degradation and drinking water pollution. Therefore, this project is focused on predicting the location of soil erosion by using logistic regression Machine Learning algorithm and GIS. Erosion causative factors such as DEM, curvature, slope, rainfall, landuse, soil erodibility and geology were evaluated. Based on the selected causative factors (CF), the map of CFs was being produced by using the ArcGIS software. Then, the data of causative factor from the ArcGIS was being used in training in machine learning (ML). There are 175-point location of soil erosion was used as to validate the soil erosion map.The weighted value of each factor was calculated according to the logistic regression (LR) and soil erosion susceptibility map was created.