Developing a Flood Risk Assessment Using Support Vector Machine and Convolutional Neural Network: A Conceptual Framework

Abstract

Flooding is one of the most devastating natural hazards that affect not only to infrastructures and agriculture but also to human lives. The prominent effect of global warming boasted its danger and impact in a wider range. In order to address and provide more effective measures to lessen the impact of flood hazards, it would be better to identify first the areas with such flood vulnerability. The proposed study aims to exploit the data available from the Geographical Information System (GIS) and the technology advancement in the modern world in producing a reliable flood susceptibility and probability map. Fusing ConvNet, a feedforward neural networks that specialize in image processing and prediction with SVM, a supervised machine learning for classification and regression analysis for a better image map results. Distinct image prediction output from dilated convolution and deconvolution network will be used as an input to SVM in producing its final output.