Dynamic Insights into Dengue: Leveraging Spatio-Temporal Graph Convolution Networks

Abstract

Examining the nexus between dengue fever and weather variables, this study proposes the innovative application of Spatio-Temporal Graph Convolution Network (STGCN) models to unravel complex patterns. Through extensive experimentation, our STGCN model demonstrates exceptional proficiency in capturing synthetic data patterns generated by the SEIR-SEI model, indicating its potential for real-world applications. Integrating human mobility networks and weather attributes into our approach enables more precise predictions of dengue cases. Leveraging these advanced machine learning techniques, particularly STGCN, we significantly advance our comprehension of the nuanced dynamics of disease transmission. This interdisciplinary approach represents a seminal advancement in combating vector-borne diseases, with profound implications for public health policy and decision-making.