Disaster Management with Sentiment Analysis and Earthquake/Tsunami/Flood Prediction System

Abstract

Earthquakes, floods, and tsunamis are ever-present dangers to people and property, calling for efficient
and prompt disaster management plans. This work presents a hybrid system that combines “Random Forest (RF)
and Long Short-Term Memory (LSTM)” algorithms to forecast natural disasters and categorize social media data
regarding humanitarian aid sentiments. Using datasets of tweets about disasters as training, the LSTM model
classifies people's demands during these times into sentiment classes including pity, urgent pleas for help, and
other similar terms. Predicting the probability of earthquakes, tsunamis, and floods using environmental variables
like rainfall, latitude, longitude, depth, and dam height is done using the Random Forest method.