FITS: Modeling Time Series with $10k$ Parameters

Zhijian Xu; Ailing Zeng; Qiang Xu

ICLR 2024spotlight123 citations

FITS: Modeling Time Series with $10k$ Parameters

Abstract

In this paper, we introduce FITS, a lightweight yet powerful model for time series analysis. Unlike existing models that directly process raw time-domain data, FITS operates on the principle that time series can be manipulated through interpolation in the complex frequency domain, achieving performance comparable to state-of-the-art models for time series forecasting and anomaly detection tasks. Notably, FITS accomplishes this with a svelte profile of just about $10k$ parameters, making it ideally suited for edge devices and paving the way for a wide range of applications. The code is available for review at: \url{https://anonymous.4open.science/r/FITS}.

Time series analysisTime series forecastingComplex-valued neural network

BibTeX

@inproceedings{
xu2024fits,
title={{FITS}: Modeling Time Series with \$10k\$ Parameters},
author={Zhijian Xu and Ailing Zeng and Qiang Xu},
booktitle={The Twelfth International Conference on Learning Representations},
year={2024},
url={https://openreview.net/forum?id=bWcnvZ3qMb}
}