Using a Local Surrogate Model to Interpret Temporal Shifts in Global Annual Data

arXiv:2404.11874v1 Announce Type: new
Abstract: This paper focuses on explaining changes over time in globally-sourced, annual temporal data, with the specific objective of identifying pivotal factors that contribute to these temporal shifts. Leveraging such analytical frameworks can yield transformative impacts, including the informed refinement of public policy and the identification of key drivers affecting a country's economic evolution. We employ Local Interpretable Model-agnostic Explanations (LIME) to shed light on national happiness indices, economic freedom, and population metrics, spanning variable time …

abstract arxiv cs.ai cs.lg data frameworks global impacts paper pivotal policy public public policy temporal type