BLAST: Balanced Sampling Time Series Corpus for Universal Forecasting Models

TL;DR

BLAST is a new balanced sampling corpus for time series forecasting that improves model generalization and efficiency by enhancing data diversity through innovative sampling and clustering techniques.

Contribution

We introduce BLAST, a large-scale, diversity-enhanced pre-training corpus for universal time series forecasting models, utilizing balanced sampling, clustering, and mixup methods.

Findings

Models trained on BLAST outperform existing datasets in forecasting accuracy.

BLAST reduces training resources needed for state-of-the-art performance.

Data diversity is crucial for effective universal time series forecasting.

Abstract

The advent of universal time series forecasting models has revolutionized zero-shot forecasting across diverse domains, yet the critical role of data diversity in training these models remains underexplored. Existing large-scale time series datasets often suffer from inherent biases and imbalanced distributions, leading to suboptimal model performance and generalization. To address this gap, we introduce BLAST, a novel pre-training corpus designed to enhance data diversity through a balanced sampling strategy. First, BLAST incorporates 321 billion observations from publicly available datasets and employs a comprehensive suite of statistical metrics to characterize time series patterns. Then, to facilitate pattern-oriented sampling, the data is implicitly clustered using grid-based partitioning. Furthermore, by integrating grid sampling and grid mixup techniques, BLAST ensures a balanced and representative coverage of diverse patterns. Experimental results demonstrate that models pre-trained on BLAST achieve state-of-the-art performance with a fraction of the computational resources and training tokens required by existing methods. Our findings highlight the pivotal role of data diversity in improving both training efficiency and model performance for the universal forecasting task.