Public Policies and Affirmative Actions at Public Universities: defining an algorithm to predict drop outs of entrance students with racial quotes

• Grant number: 23/10240-0
• Support Opportunities: Research Grants - Research in Public Policies
• Duration: April 01, 2024 - March 31, 2028
• Field of knowledge: Humanities - Political Science - Public Policies
• Principal Investigator: Sueyla Ferreira da Silva dos Santos
• Grantee: Sueyla Ferreira da Silva dos Santos
• Host Institution: Faculdade de Ciências e Tecnologia (FCT). Universidade Estadual Paulista (UNESP). Campus de Presidente Prudente. Presidente Prudente , SP, Brazil
• Associated researchers: Danillo Roberto Pereira ; Divino José da Silva ; Gustavo Bizarria Gibin ; LUCAS SILVESTRE DOS SANTOS ; Mônica Paiva Quast ; Murilo Rodolfo Cândido ; Nara Torrecilha Ferreira ; Vanda Moreira Machado Lima
• Associated scholarship(s): 24/09322-5 - Development of Public Policies for the Retention of Affirmative Action Entrants: An Innovative Approach with Predictive Algorithms and Artificial Intelligence, BP.TT; 24/09688-0 - Development of Public Policies for the Retention of Affirmative Action Entrants: An Innovative Approach with Predictive Algorithms and Artificial Intelligence., BP.TT

Abstract

Public Policies of Affirmative Actions (PPAA) emerged as a result of a historical struggle by social movements in pursuit of equity and inclusion. The enactment of Law No. 12,711/2012, known as the Law of Quotas, established the reservation of half of the vacancies in public universities for candidates coming from public schools and in situations of socioeconomic vulnerability. At São Paulo State University, Júlio Mesquita Filho (UNESP), initiatives were created to collaborate in the implementation of affirmative actions, the Coordinatorship for Affirmative Actions, Diversity, and Equity (CAADI), and the Black Oriented Nucleus of Research and Extension at UNESP (NUPE). Therefore, the retention of these incoming students has been guided by educational management. The objective is to analyze the markers to predict the potential risk of dropouts among university students who were admitted through reserved seats and quotas in the undergraduate courses at FCT/UNESP. The aim is to provide information to support the improvement of public policies for promoting equity in higher education. Machine learning techniques will be applied through predictive modeling to develop an artificial intelligence product that provides a risk classifier for dropout among students admitted through affirmative actions. The research team will be responsible for executing and monitoring the project's goals, in collaboration with working groups composed of technical scholarship holders, researchers from the host, associated, and partner institutions. Data extraction will be performed through academic information systems, socioeconomic records, and administrative databases. For privacy assurance, the data will be anonymized. To build the predictive model, a set of temporal data will be generated using neural networks for data classification. Consensus techniques and strategies for result dissemination will be applied to ensure the translation of scientific knowledge into a language applicable to public policies. Therefore, the results of this investigation will be disseminated in both scientific and political spheres, aiming to contribute to the definition of a public agenda in support of PPAA (Public Policies of Affirmative Actions) and a draft with propositional actions to enhance these initiatives in the state of São Paulo. The research aims to build a high-precision and valid classifier for the risk of dropout in higher education, enabling the development of preventive strategies that assist in the retention of university students admitted through affirmative actions. (AU)