Social Activism Analysis: An Application of Machine Learning in the World Values Survey

EasyChair Preprint 1390

Abstract

The study of social sciences is essential to understand different dimensions of human society. Different researches are done to understand human development and its relationships in the community. Given this, in this paper, we have developed a methodology to use typical social science metrics and resources in conjunction with Artificial Intelligence techniques. The goal is to collaborate with research and visualize patterns to help explain human behavior. In this way, we use the World Values Survey's fifth wave data to apply self-learning methods and contribute to the advancement of social science research. We use algorithms to perform classifications, such as Random Forest, Stochastic Gradient Descent and Support Vector Machine in data collected in 58 countries, to verify if there are social patterns that can explain political participation. Thus, we identified that there is a stronger relationship in the results found in the so-called advanced democracies (USA and Europe) compared to those in other societies. From this, we can consider that eventual adjustments in the theory underlying the WVS research or in the instruments of data collection could be made and that more studies are needed to analyze other dimensions.

Keyphrases: Artificial Intelligence, WVS, political participation, social sciences

@booklet{EasyChair:1390,
  author    = {Francielle M. Nascimento and Dante A. C. Barone and Henrique Carlos de Castro},
  title     = {Social Activism Analysis: An Application of Machine Learning in the World Values Survey},
  howpublished = {EasyChair Preprint 1390},
  year      = {EasyChair, 2019}}