TimeCAP repurposes Large Language Models from text generators to powerful time series interpreters by contextualizing temporal data for improved event prediction.

• Enhances prediction accuracy by incorporating contextual information in time series analysis
• Extends LLMs beyond their typical usage as predictors to become contextualizers of complex time data
• Creates a framework applicable across climate modeling, healthcare, and financial analytics
• Demonstrates particularly valuable applications in medical monitoring systems

For healthcare applications, TimeCAP enables more accurate patient monitoring and disease progression tracking by incorporating contextual information that traditional time series models might miss.

TimeCAP: Learning to Contextualize, Augment, and Predict Time Series Events with Large Language Model Agents