The Data Analysis of Determining Potential Flood-Prone Areas in DKI Jakarta Using Classification Model Approach

The study aims to analyze the flood data in the five Jakarta areas. The focus is to determine areas that are potentially prone to flooding. Currently, there is no identification of potential flood areas in Jakarta based on actual parameters such as total rainfall, altitude, and population density. Based on this problem, a data-driven method is applied using the C4.5 algorithm. Decision trees are used to assist the classification process in determining areas that have the potential to be prone to flooding have potential flooding in the DKI Jakarta area. The 10-fold cross validation is employed along with the confusion matrix to evaluate the model between the actual and predicted results. The study shows that this algorithm can model the Jakarta potential flood-prone areas with an accuracy value of 87.20% with precision and recall values of 90.62% and 94.84%. Based on the model, the predicted flooding area can be identified utilizing the parameters.