Classification of Flood Disaster Predictions using the C5.0 and SVM Algorithms based on Flood Disaster Prone Areas

摘要

Many researchers have been motivated to improve the performance of predictive methods. So that is what prompted researchers to conduct this research in order to find out the object of the Flood Disaster, whether it can be done using the Classification method. The main factor in the occurrence of Flood Disaster is the increasing intensity of rainfall that clogs the river water flow, which further pressures the river water to dike embankments that are no longer strong by carrying materials found in the flow of water from upstream to downstream, such as Wood, Mud, There are even rubbish from homebased industries which are carried away by flood flows which cause many rivers to become clogged. Floods have the meaning of one of the natural disasters that occur due to increased rainfall from normal which can cause casualties and often occur in lowland areas. In this study, a classification will be conducted on how to predict Flood Disasters based on their Prone Areas and this research takes the Bandung area as the object of research. The algorithm used in this study is to compare 2 Classification algorithms namely C5.0 and SVM Algorithms to determine the accuracy value of which algorithm is much higher than other algorithms based on the Prone Areas. C5.0 and SVM algorithms can be used on datasets that have been modeled to produce value accuracy. Data processing in this study uses the Orange application which can be used to create a model of data that has been processed into information that can be given to the public for the early warning of the flood disaster that they will face in the future obtained from past data. Orange is one of the open source software used for processing Data Analytics / Data Mining.