Time series forecasting with Hahn Kolmogorov-Arnold networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hasan, Md Zahidul, Hamza, A. Ben, Bouguila, Nizar
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917330850676736
author Hasan, Md Zahidul
Hamza, A. Ben
Bouguila, Nizar
author_facet Hasan, Md Zahidul
Hamza, A. Ben
Bouguila, Nizar
contents Recent Transformer- and MLP-based models have demonstrated strong performance in long-term time series forecasting, yet Transformers remain limited by their quadratic complexity and permutation-equivariant attention, while MLPs exhibit spectral bias. We propose HaKAN, a versatile model based on Kolmogorov-Arnold Networks (KANs), leveraging Hahn polynomial-based learnable activation functions and providing a lightweight and interpretable alternative for multivariate time series forecasting. Our model integrates channel independence, patching, a stack of Hahn-KAN blocks with residual connections, and a bottleneck structure comprised of two fully connected layers. The Hahn-KAN block consists of inter- and intra-patch KAN layers to effectively capture both global and local temporal patterns. Extensive experiments on various forecasting benchmarks demonstrate that our model consistently outperforms recent state-of-the-art methods, with ablation studies validating the effectiveness of its core components.
format Preprint
id arxiv_https___arxiv_org_abs_2601_18837
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Time series forecasting with Hahn Kolmogorov-Arnold networks
Hasan, Md Zahidul
Hamza, A. Ben
Bouguila, Nizar
Machine Learning
Recent Transformer- and MLP-based models have demonstrated strong performance in long-term time series forecasting, yet Transformers remain limited by their quadratic complexity and permutation-equivariant attention, while MLPs exhibit spectral bias. We propose HaKAN, a versatile model based on Kolmogorov-Arnold Networks (KANs), leveraging Hahn polynomial-based learnable activation functions and providing a lightweight and interpretable alternative for multivariate time series forecasting. Our model integrates channel independence, patching, a stack of Hahn-KAN blocks with residual connections, and a bottleneck structure comprised of two fully connected layers. The Hahn-KAN block consists of inter- and intra-patch KAN layers to effectively capture both global and local temporal patterns. Extensive experiments on various forecasting benchmarks demonstrate that our model consistently outperforms recent state-of-the-art methods, with ablation studies validating the effectiveness of its core components.
title Time series forecasting with Hahn Kolmogorov-Arnold networks
topic Machine Learning
url https://arxiv.org/abs/2601.18837