SHS Web of Conferences (Jan 2023)
Prediction of Carbon Dioxide Level Using Statistical Learning and Its Potential Correlation With Global Warming
Abstract
The Industrial Revolution caused a huge change in the climate of our planet. Since the 19th century, a high level of atmospheric carbon dioxide has contributed to global warming and other environmental problems. We first acknowledge the substantial correlations between the CO2 levels or temperatures and the years before creating our models. In this situation, we propose that the ARIMA model, which combines the auto-regression and moving average models, is essential for issue analysis. In order to estimate CO2 concentrations and land-ocean temperatures, we create polynomial models as well as an ARIMA model with seasonality. Following these hypotheses, we discover that the CO2 concentrations and temperatures have a significant direct link. In order to forecast the future relationships between CO2 concentrations and temperatures, we also attempt to employ polynomial function. We constantly reflect on and reexamine the issues as we construct these models in order to have a greater grasp of the circumstances. Each of our models is also evaluated, and the most precise one is used to make forecasts. Based on Matlab, we can quickly calculate the data, utilize iterations to determine the ideal model parameters, and then display our findings in diagrams.