Predicting Climatic Parameters Using the ARIMA Time Series and LSTM Deep Learning Models for Vidarbha Region

Abstract

A daily increase in atmospheric temperatures causes changes in rainfall patterns. It has a substantial impact on environmental problems worldwide. Understanding these variations is essential, especially concerning how changes in rainfall affect natural calamities like droughts, floods, and global warming. The Vidarbha region of Maharashtra, India, is selected as a subject of this work. Climate change impacts groundwater levels because the temperature rise reduces it. After considering all the factors, it is a must to apply any forecasting model so that future disasters can be predicted. The study evaluates dataset accuracy and visualises data using RStudio. The work tries forecasting rainfall, humidity, and groundwater data using the ARIMA Time Series model and LSTM machine-learning algorithm. Also, a comparative analysis of both approaches is given here. This thorough investigation has consequences for the area's hydrological planning, disaster preparedness, and economic development.