Climatic Temperature Forecasting with Regression Approach

The research study centers on the use of regression technique to predict climate change variation. Living beings on the planet are on a high risk due to the abnormal changes in the climate happening every now and then and, therefore, we need robust forecasting technique for making climate predictions. The key issue faced in prediction of future climate is the selection of appropriate prediction technique for prediction since every technique uses different algorithm and a slight change in forecasting technique can make a tremendous change in the predicted results. The study predicts future temperature using the standard ‘Jet Propulsion Laboratory’, NASA dataset. The raw data is preprocessed to remove any inconsistencies and missing values which is then subjected to polynomial regression model to predict future temperature which leads to climatic variation. This prediction will help environment agencies to take necessary steps as to minimize the impact of drastic climate changes and also to forecast regions with rough weather conditions. The study, through the results, exhibits the effectiveness of the polynomial regression technique as far as climate prediction is concerned. In future the experiment can be conducted with more intrinsic features affecting climate change. The outcome was very productive. Polynomial regression achieved the highest training accuracy of 93.31% and testing accuracy of 91.01%.