Human Mobility Prediction

Human mobility prediction aims to forecast individuals' movements, leveraging data like GPS traces and socio-demographic information to understand travel patterns and inform applications like urban planning and public health. Recent research heavily utilizes large language models (LLMs) and transformer architectures, often within agentic frameworks that decompose the prediction task into sub-problems focusing on spatial-temporal patterns, individual routines, and the influence of contextual factors like events or urban structure. These advancements improve prediction accuracy and offer interpretability, addressing limitations of previous methods reliant on extensive, private datasets. The resulting insights have significant implications for optimizing transportation systems, managing public events, and responding to crises like epidemics.