SPATIAL MODELLING OF FLOOD SUSCEPTIBILITY IN PERLIS USING GEOGRAPHIC INFORMATION SYSTEM (GIS) AND SUPPORT VECTOR MACHINE (SVM)

Flooding had caused serious damage to the environment around the world. Managing flood event in a proper way could minimize the impact of the flood disaster to the economy and environment. This project assesses the flood susceptibility of Perlis to determine the risk of flooding area in that state. This assessment is one of the mitigation planning to prevent the flooding from damaging the public or private properties of Perlis. In this research, Geographic Information System (GIS) is integrated with Machine Learning model which is Support Vector Machine (SVM) to develop the Perlis’s flood susceptibility map. Linear Kernel function are used in this project as this function provide efficient two-class classifier by separating the class features linearly. Determination and justification of the datasets are made in methodology section to identify the flooding area. Then, Area Under Curve (AUC) method are used to test the validity and accuracy of the susceptibility map. This procedure is used to produce the flood susceptibility map of Perlis with high accuracy which is 79.86%.