TimeFound: A Foundation Model for Time Series Forecasting

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Xiao, Congxi, Zhou, Jingbo, Xiao, Yixiong, Lu, Xinjiang, Zhang, Le, Xiong, Hui
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915184076914688
author Xiao, Congxi
Zhou, Jingbo
Xiao, Yixiong
Lu, Xinjiang
Zhang, Le
Xiong, Hui
author_facet Xiao, Congxi
Zhou, Jingbo
Xiao, Yixiong
Lu, Xinjiang
Zhang, Le
Xiong, Hui
contents We present TimeFound, an encoder-decoder transformer-based time series foundation model for out-of-the-box zero-shot forecasting. To handle time series data from various domains, TimeFound employs a multi-resolution patching strategy to capture complex temporal patterns at multiple scales. We pre-train our model with two sizes (200M and 710M parameters) on a large time-series corpus comprising both real-world and synthetic datasets. Over a collection of unseen datasets across diverse domains and forecasting horizons, our empirical evaluations suggest that TimeFound can achieve superior or competitive zero-shot forecasting performance, compared to state-of-the-art time series foundation models.
format Preprint
id arxiv_https___arxiv_org_abs_2503_04118
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle TimeFound: A Foundation Model for Time Series Forecasting
Xiao, Congxi
Zhou, Jingbo
Xiao, Yixiong
Lu, Xinjiang
Zhang, Le
Xiong, Hui
Machine Learning
We present TimeFound, an encoder-decoder transformer-based time series foundation model for out-of-the-box zero-shot forecasting. To handle time series data from various domains, TimeFound employs a multi-resolution patching strategy to capture complex temporal patterns at multiple scales. We pre-train our model with two sizes (200M and 710M parameters) on a large time-series corpus comprising both real-world and synthetic datasets. Over a collection of unseen datasets across diverse domains and forecasting horizons, our empirical evaluations suggest that TimeFound can achieve superior or competitive zero-shot forecasting performance, compared to state-of-the-art time series foundation models.
title TimeFound: A Foundation Model for Time Series Forecasting
topic Machine Learning
url https://arxiv.org/abs/2503.04118