Abstract
This article explores the conceptual and theoretical intersections between Punctuated Equilibrium Theory (PET) and artificial neural networks (NNs) within the context of policy change analysis. Despite some similarities between PET and NNs, limited systematic research has been conducted to bridge the gap between political science and computer science. The paper addresses this conceptual gap by presenting a theory-oriented, explorative examination, focusing on the commonalities in their principles, such as information processing, dynamic modeling, and adaptation. The study contributes to methodology- and theory-oriented research on policy agendas by extending PET through the incorporation of NNs. The article employs a conceptual lens to establish parallels between PET and NNs, emphasizing their shared features in dealing with complex, dynamic, and adaptive systems. The exploration of anomalies and outliers in policy time-series data serves as a case study to illustrate the potential synergy between political science and STEM sciences (science, technology, engineering, and mathematics). The paper concludes by proposing avenues for future research that can further integrate these allegedly separate disciplines and enhance our understanding of policy dynamics.