基于大模型的时序预测新范式，无需微调即可跨数据集预测。
Modèles de Fondation et Ajustement : Vers une Nouvelle Génération de Modèles pour la Prévision des Séries Temporelles
- 用海量时序数据预训练通用模型，学习可迁移的特征表示。
- 微调后在长周期预测上性能显著提升，零样本能力更强。
- 适合需要快速适配新数据集的工业级时序预测场景。
受大型语言模型进展启发，基础模型被用于零样本时序预测，可在预训练时未见的数据集上进行预测。这些大规模模型通过在海量时序数据上训练，学习到适用于点预测和概率预测的通用表征，减少了对特定任务架构和手动调参的需求。本文综述了此类模型的主要架构、预训练策略及优化方法，并研究了预训练后微调对特定数据集性能的影响。实验结果表明，微调通常能提升零样本预测能力，尤其在长周期预测中效果更明显。
原文摘要 · Abstract (English)
Inspired by recent advances in large language models, foundation models have been developed for zero-shot time series forecasting, enabling prediction on datasets unseen during pretraining. These large-scale models, trained on vast collections of time series, learn generalizable representations for both point and probabilistic forecasting, reducing the need for task-specific architectures and manual tuning. In this work, we review the main architectures, pretraining strategies, and optimization methods used in such models, and study the effect of fine-tuning after pretraining to enhance their performance on specific datasets. Our empirical results show that fine-tuning generally improves zero-shot forecasting capabilities, especially for long-term horizons.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。