Breaking the Context Bottleneck on Long Time Series Forecasting

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Ma, Chao, Hou, Yikai, Li, Xiang, Sun, Yinggang, Yu, Haining, Fang, Zhou, Qu, Jiaxing
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914020950278144
author Ma, Chao
Hou, Yikai
Li, Xiang
Sun, Yinggang
Yu, Haining
Fang, Zhou
Qu, Jiaxing
author_facet Ma, Chao
Hou, Yikai
Li, Xiang
Sun, Yinggang
Yu, Haining
Fang, Zhou
Qu, Jiaxing
contents Long-term time-series forecasting is essential for planning and decision-making in economics, energy, and transportation, where long foresight is required. To obtain such long foresight, models must be both efficient and effective in processing long sequence. Recent advancements have enhanced the efficiency of these models; however, the challenge of effectively leveraging longer sequences persists. This is primarily due to the tendency of these models to overfit when presented with extended inputs, necessitating the use of shorter input lengths to maintain tolerable error margins. In this work, we investigate the multiscale modeling method and propose the Logsparse Decomposable Multiscaling (LDM) framework for the efficient and effective processing of long sequences. We demonstrate that by decoupling patterns at different scales in time series, we can enhance predictability by reducing non-stationarity, improve efficiency through a compact long input representation, and simplify the architecture by providing clear task assignments. Experimental results demonstrate that LDM not only outperforms all baselines in long-term forecasting benchmarks, but also reducing both training time and memory costs.
format Preprint
id arxiv_https___arxiv_org_abs_2412_16572
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Breaking the Context Bottleneck on Long Time Series Forecasting
Ma, Chao
Hou, Yikai
Li, Xiang
Sun, Yinggang
Yu, Haining
Fang, Zhou
Qu, Jiaxing
Machine Learning
Artificial Intelligence
Long-term time-series forecasting is essential for planning and decision-making in economics, energy, and transportation, where long foresight is required. To obtain such long foresight, models must be both efficient and effective in processing long sequence. Recent advancements have enhanced the efficiency of these models; however, the challenge of effectively leveraging longer sequences persists. This is primarily due to the tendency of these models to overfit when presented with extended inputs, necessitating the use of shorter input lengths to maintain tolerable error margins. In this work, we investigate the multiscale modeling method and propose the Logsparse Decomposable Multiscaling (LDM) framework for the efficient and effective processing of long sequences. We demonstrate that by decoupling patterns at different scales in time series, we can enhance predictability by reducing non-stationarity, improve efficiency through a compact long input representation, and simplify the architecture by providing clear task assignments. Experimental results demonstrate that LDM not only outperforms all baselines in long-term forecasting benchmarks, but also reducing both training time and memory costs.
title Breaking the Context Bottleneck on Long Time Series Forecasting
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2412.16572