FLOOD PREDICTION AND ASSESSMENT PLATFORM A MULTI-MODEL APPROACH

Floods are the most destructive natural disasters, causing huge damage to life and socioeconomic system. Early flood warning systems are good countermeasures against flood hazards. So here we propose a flood prediction system that uses machine learning models to deliver chances of flood based on location and time queries. The system can be integrated with multi-model deep learning methods such as Support Vector machines, Neural Networks and Extreme Learning Model to predict the chances of flood based on various environmental, climatic factors and geographical conditions. Based on the results, actions that need to be taken can also be predicted. The model is trained on a comprehensive data set expanding over years. The system has the ability to compare the results of various sub models such that the overall accuracy can be modified. It is helpful for the people living in the flood affected areas as well as to the ones who are related to the organizations in this field. Keywords—flood, multi-model, neural network, support vector machine, Extreme learning machines, Decision Tree Classifier