A0662
Title: Machine learning models for predicting the socio-economic impacts of climate change
Authors: Angela Maria DUggento - University of Bari Aldo Moro (Italy) [presenting]
Margaret Antonicelli - IULM University Milano (Italy)
Abstract: Climate change has caused enormous environmental damage through forest fires and floods. These extreme events also have social, political, and economic consequences. To understand these delayed effects of climate change, it is important to consider some phenomena as environmental problems. Climate justice focuses on climate change as an ethical and political issue by highlighting how the most vulnerable populations, despite contributing the least to the problem, suffer the most from its impacts, making it a human rights issue. Climate migration is largely determined by the availability and accessibility of resources and exposure to environmental hazards. Gender inequality could also be exacerbated by climate change, as women are more vulnerable due to the division of labor, restrictive gender norms, and underrepresentation in climate-related decision-making. Finally, economic impacts such as the rise in coffee and cocoa prices also have a direct effect on people's lives. To analyze climate change trends, an analysis of anomalous temperature data in European countries from 1850 to 2023 is carried out using machine learning prediction models. Conventional linear models have struggled to accurately capture climate trends, even with SARIMA. A long short-term memory (LSTM) neural network optimized with a modified Adam algorithm showed superior prediction performance.
