arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.06603cs.LGcs.AI
收藏

FPN-fusion:增强的线性复杂度时间序列预测模型

FPN-fusion: Enhanced Linear Complexity Time Series Forecasting Model

University of Science and Technology of China(中国科学技术大学);Hefei Institutes of Physical Science, Chinese Academy of Sciences(中国科学院合肥物质科学研究院);Institute of Plasma Physics(等离子体物理研究所)

Chu Li, Pingjia Xiao, Qiping Yuan

AI总结:

本文提出线性复杂度的 FPN-fusion 时间序列预测模型,通过特征金字塔网络和多级融合结构替代传统分解,在多个数据集上以更低计算成本超越 DLiner 和 PatchTST。

摘要:

本研究提出了一种新型时间序列预测模型 FPN-fusion,其设计具有线性计算复杂度,在不增加参数数量或计算需求的情况下,展现出优于 DLiner 的预测性能。我们的模型引入了两项关键创新:第一,采用特征金字塔网络(FPN)有效捕捉时间序列数据特征,绕过了传统的趋势与季节分量分解;第二,开发了一种多级融合结构,以无缝整合深层与浅层特征。实验表明,FPN-fusion 在八个开源数据集上的 32 个测试用例中有 31 个优于 DLiner,均方误差(MSE)平均降低 16.8%,平均绝对误差(MAE)平均降低 11.8%。此外,与基于 Transformer 的 PatchTST 相比,FPN-fusion 在 32 个测试项目中取得了 10 个最佳 MSE 和 15 个最佳 MAE 结果,而总计算负载仅为 PatchTST 的 8%。

英文摘要:

This study presents a novel time series prediction model, FPN-fusion, designed with linear computational complexity, demonstrating superior predictive performance compared to DLiner without increasing parameter count or computational demands. Our model introduces two key innovations: first, a Feature Pyramid Network (FPN) is employed to effectively capture time series data characteristics, bypassing the traditional decomposition into trend and seasonal components. Second, a multi-level fusion structure is developed to integrate deep and shallow features seamlessly. Empirically, FPN-fusion outperforms DLiner in 31 out of 32 test cases on eight open-source datasets, with an average reduction of 16.8% in mean squared error (MSE) and 11.8% in mean absolute error (MAE). Additionally, compared to the transformer-based PatchTST, FPN-fusion achieves 10 best MSE and 15 best MAE results, using only 8% of PatchTST's total computational load in the 32 test projects.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑