LoFT-LLM: Low-Frequency Time-Series Forecasting with Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: You, Jiacheng, Yang, Jingcheng, Xie, Yuhang, Wu, Zhongxuan, Li, Xiucheng, Li, Feng, Wang, Pengjie, Xu, Jian, Zheng, Bo, Chen, Xinyang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908761110609920
author You, Jiacheng
Yang, Jingcheng
Xie, Yuhang
Wu, Zhongxuan
Li, Xiucheng
Li, Feng
Wang, Pengjie
Xu, Jian
Zheng, Bo
Chen, Xinyang
author_facet You, Jiacheng
Yang, Jingcheng
Xie, Yuhang
Wu, Zhongxuan
Li, Xiucheng
Li, Feng
Wang, Pengjie
Xu, Jian
Zheng, Bo
Chen, Xinyang
contents Time-series forecasting in real-world applications such as finance and energy often faces challenges due to limited training data and complex, noisy temporal dynamics. Existing deep forecasting models typically supervise predictions using full-length temporal windows, which include substantial high-frequency noise and obscure long-term trends. Moreover, auxiliary variables containing rich domain-specific information are often underutilized, especially in few-shot settings. To address these challenges, we propose LoFT-LLM, a frequency-aware forecasting pipeline that integrates low-frequency learning with semantic calibration via a large language model (LLM). Firstly, a Patch Low-Frequency forecasting Module (PLFM) extracts stable low-frequency trends from localized spectral patches. Secondly, a residual learner then models high-frequency variations. Finally, a fine-tuned LLM refines the predictions by incorporating auxiliary context and domain knowledge through structured natural language prompts. Extensive experiments on financial and energy datasets demonstrate that LoFT-LLM significantly outperforms strong baselines under both full-data and few-shot regimes, delivering superior accuracy, robustness, and interpretability.
format Preprint
id arxiv_https___arxiv_org_abs_2512_20002
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LoFT-LLM: Low-Frequency Time-Series Forecasting with Large Language Models
You, Jiacheng
Yang, Jingcheng
Xie, Yuhang
Wu, Zhongxuan
Li, Xiucheng
Li, Feng
Wang, Pengjie
Xu, Jian
Zheng, Bo
Chen, Xinyang
Machine Learning
Time-series forecasting in real-world applications such as finance and energy often faces challenges due to limited training data and complex, noisy temporal dynamics. Existing deep forecasting models typically supervise predictions using full-length temporal windows, which include substantial high-frequency noise and obscure long-term trends. Moreover, auxiliary variables containing rich domain-specific information are often underutilized, especially in few-shot settings. To address these challenges, we propose LoFT-LLM, a frequency-aware forecasting pipeline that integrates low-frequency learning with semantic calibration via a large language model (LLM). Firstly, a Patch Low-Frequency forecasting Module (PLFM) extracts stable low-frequency trends from localized spectral patches. Secondly, a residual learner then models high-frequency variations. Finally, a fine-tuned LLM refines the predictions by incorporating auxiliary context and domain knowledge through structured natural language prompts. Extensive experiments on financial and energy datasets demonstrate that LoFT-LLM significantly outperforms strong baselines under both full-data and few-shot regimes, delivering superior accuracy, robustness, and interpretability.
title LoFT-LLM: Low-Frequency Time-Series Forecasting with Large Language Models
topic Machine Learning
url https://arxiv.org/abs/2512.20002