Algoritmos de aprendizado de máquina no contexto da confiança em governos latino-americanos

Abstract

Machine learning has advanced significantly in recent decades and its impact is felt in a variety of cases, enabling computers to process, learn, and extract knowledge from the vast amount of data generated daily. Trust is the foundation of numerous interac-tions essential for healthy societies and the level of trust in government can be influ-enced by different factors. Therefore, it is of common societal interest to understand the landscape of trust in government and the factors that affect it. However, there is still a lack of research analyzing trust in government with a data mining and machine learning approach, with studies focusing on contexts such as social trust, citizen polit-ical engagement, political trust, and better understanding populist voting. Thus, the objective of this study is to apply machine learning techniques to classify the level of trust in national governments in Latin America, based on respondents’ self-perception, using public data. To achieve this, open data from a public opinion survey containing questions about society, democracy, economy, and trust in government were used. The data were provided by the Latinobarómetro Corporation, a non-profit organization based in Santiago, Chile, dedicated to conducting public opinion surveys in Latin America.