Time series forecasting improves with smarter data balancing and pattern grouping
QUALS: Corpus Equilibrium for Universal Forecasting via Pattern Quantization and Learnability Synchronization
Machine LearningArtificial Intelligence
Summary
Forecasting future events from time series data is important for many fields but training models well is hard because data varies a lot. The authors created QUALS, a method that groups data patterns and carefully balances their use during training. This approach helps models learn better from less data and perform well on new, unseen data. Their tests show QUALS improves forecasting performance even when training time and data are limited.
What this means in practice
- •For transportation system operators: Improve traffic flow prediction by training forecasting models more efficiently using diverse traffic pattern data.
- •For power grid managers: Enhance power demand forecasting accuracy with less training data by balancing complex and simple consumption patterns.
Authors
Yujie Li, Zezhi Shao, Chengqing Yu, Yisong Fu, Weijie Zhu, Yifan Du, Jilin Hu, Bin Yang, Yongjun Xu, Fei Wang
Abstract
Ubiquitous time series data across diverse domains enables critical applications in areas such as transportation systems and power grids. Recently, training foundation models on massive datasets to achieve accurate zero-shot forecasting has emerged as a major research focus. However, current studies predominantly prioritize architectural innovations while insufficiently addressing data diversity, often relying on simple data sampling strategies that fail to manage complex data distributions effectively, leading to inefficient use of training data and suboptimal performance. To address this, we propose QUALS, a large-scale time series corpus equilibrium framework. QUALS significantly enhances data efficiency, i.e., enabling existing models to achieve superior performance using only a small fraction of the original training data. Specifically, QUALS operates through two core mechanisms. First, a pattern quantization framework systematically decodes heterogeneous patterns from mixed corpora via vector quantization and uniform binning. Second, a learnability synchronization framework calibrates sampling weights for heterogeneous patterns, bridging the optimization gap between simple and complex motifs to maximize overall training efficiency. Extensive benchmarks demonstrate that pre-training on QUALS consistently achieves superior zero-shot performance, even under substantially reduced training budgets.