Flood Forecasting System Using Data Analysis
Floods have become one of the biggest calamities in different areas of the world. There are systems designed to predict the rainfall and cyclones with the help of satellites, the need for monitoring them is necessary to prevent flood. This research study focuses on proposing different machine learning approaches to forecast flood with combination of factors affecting floods. The purpose of this system is to carry out data analysis techniques in forecasting floods. Initially, data is collected from different resources and pre-processed. Using exploratory data analysis, features like rainfall, water level of river, water flow of river are considered. Classification techniques like decision tree, support vector machine, and random forest classification algorithms are used to study. Accuracy, recall, precision, F-measures and ROC Area measures are estimated to prove the results. The main goal of this system is to forecasts the occurrence of flood as a ‘yes’ or ‘no ‘binary output. This is achieved using the values of features as an input. Decision tree technique with highest classification accuracy of 86.6% yields the optimal results.