TimeRAG: BOOSTING LLM Time Series Forecasting via Retrieval-Augmented Generation
This addresses the challenge of improving time series forecasting with LLMs for domains like finance or weather, though it is incremental as it builds on existing RAG methods.
The paper tackles the problem of limited transferability and excessive training in LLM-based time series forecasting by proposing TimeRAG, a framework that incorporates Retrieval-Augmented Generation (RAG) to retrieve similar historical patterns and use them as prompts, resulting in an average prediction accuracy improvement of 2.97% across various datasets.
Although the rise of large language models (LLMs) has introduced new opportunities for time series forecasting, existing LLM-based solutions require excessive training and exhibit limited transferability. In view of these challenges, we propose TimeRAG, a framework that incorporates Retrieval-Augmented Generation (RAG) into time series forecasting LLMs, which constructs a time series knowledge base from historical sequences, retrieves reference sequences from the knowledge base that exhibit similar patterns to the query sequence measured by Dynamic Time Warping (DTW), and combines these reference sequences and the prediction query as a textual prompt to the time series forecasting LLM. Experiments on datasets from various domains show that the integration of RAG improved the prediction accuracy of the original model by 2.97% on average.