Analysis of IoT Big Weather Data For Early Flood Forecasting System

Abstract

Due to technological advancements, the Internet of Things (IoT) has been extensively used in a number of environments during this period. The IoT is being utilized successfully, particularly in the field of weather monitoring. As a result, IoT weather sensors generate massive amounts of weather data on a regular basis. This research aims to efficiently analyze the massive amounts of data generated by IoT climate sensors to develop an effective early flood forecasting system. Numerous methods for forecasting floods using historical data have been invented. However, all of these methodologies are becoming inefficient as a result of climate change and the volume of the data. This research provides a methodology for extracting strongly correlated weather features in order to reduce the error in weather forecasts caused by data volume and climate change. The Feed-Forward Artificial Neural Network (FFANN) is used to forecast early rainfall and floods. Furthermore, Chennai has been chosen as the study area for this research. Finally, two experiments are conducted to demonstrate this early flood forecasting system's prediction accuracy and training efficiency. The experimental results demonstrate that the proposed flood forecasting system outperforms recently developed systems in terms of accuracy and training efficiency.