Metadata Matters for Time Series: Informative Forecasting with Transformers

TL;DR

This paper introduces MetaTST, a Transformer-based model that incorporates metadata encoded by large language models to improve the accuracy and interpretability of time series forecasting across diverse scenarios.

Contribution

The paper proposes a novel approach to include structured metadata into time series forecasting models using natural language templates and LLMs, enhancing performance and interpretability.

Findings

MetaTST achieves state-of-the-art results on multiple benchmarks.

Incorporating metadata improves forecasting accuracy across scenarios.

The approach effectively handles large-scale, diverse datasets.

Abstract

Time series forecasting is prevalent in extensive real-world applications, such as financial analysis and energy planning. Previous studies primarily focus on time series modality, endeavoring to capture the intricate variations and dependencies inherent in time series. Beyond numerical time series data, we notice that metadata (e.g.~dataset and variate descriptions) also carries valuable information essential for forecasting, which can be used to identify the application scenario and provide more interpretable knowledge than digit sequences. Inspired by this observation, we propose a Metadata-informed Time Series Transformer (MetaTST), which incorporates multiple levels of context-specific metadata into Transformer forecasting models to enable informative time series forecasting. To tackle the unstructured nature of metadata, MetaTST formalizes them into natural languages by pre-designed templates and leverages large language models (LLMs) to encode these texts into metadata tokens as a supplement to classic series tokens, resulting in an informative embedding. Further, a Transformer encoder is employed to communicate series and metadata tokens, which can extend series representations by metadata information for more accurate forecasting. This design also allows the model to adaptively learn context-specific patterns across various scenarios, which is particularly effective in handling large-scale, diverse-scenario forecasting tasks. Experimentally, MetaTST achieves state-of-the-art compared to advanced time series models and LLM-based methods on widely acknowledged short- and long-term forecasting benchmarks, covering both single-dataset individual and multi-dataset joint training settings.