Using graph signal processing and deep learning for crime forecasting

• Grant number: 23/04868-7
• Support Opportunities: Scholarships in Brazil - Doctorate (Direct)
• Start date: June 01, 2023
• End date: May 31, 2027
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Agreement: MCTI/MC
• Principal Investigator: Jorge Luis Poco Medina
• Grantee: Juanpablo Andrew Heredia Parillo
• Host Institution: Escola de Matemática Aplicada (EMAp). Fundação Getúlio Vargas (FGV)
• Associated research grant: 21/07012-0 - Data-driven intelligence for urban crime analysis and perception, AP.JP

Abstract

This research aims to understand the frequency behavior of crime in Brazilian cities by studying past crime, urban, and social events to predict future crime events. The plan is to take into account a graph structure and temporal data of the city to predict how crime "flows", using methods such as graph signal processing and deep learning. A PhD fellowship will enable the selection of a student to explore complex methods for creating and adapting alternative models for crime prediction, which is essential for data-driven security planning. (AU)