Forecasting temperature anomalies of planet Earth: A Comparative Analysis of AI Models

Natural disasters are growing more severe than ever before. Therefore, monitoring and analyzing variations in the temperature of planet Earth is essential to raise public awareness on climate change. This paper provides insight into the annual variations in temperature of planet Earth from 1901 to 2020 using various Artificial Intelligence learning models such as LSTM, GRU, SARIMAX. Post investigation suggested that Machine Learning model, SARIMAX outperforms the Deep RNN models such as LSTM and GRU variants for predicting Earth's temperature. The results of the study can be utilized for forecasting the average temperature over the next few decades. It can also be employed for policy decisions by the concerned authorities to mitigate the adverse effects of rising Earth temperature.