Mamba or Transformer for Time Series Forecasting? Mixture of Universals (MoU) Is All You Need
_{Official PyTorch Implementation}

 

Overview of MoU

💥 Our Paper

Our paper has been released on Arxiv. We introduce a new versatile model Mixture of Universals (MoU) to capture both short-term and long-term dynamics for enhancing perfomance in time series forecasting. MoU is composed of two novel designs: Mixture of Feature Extractors (MoF), an adaptive method designed to improve time series patch representations for short-term dependency, and Mixture of Architectures (MoA), which hierarchically integrates Mamba, FeedForward, Convolution, and Self-Attention architectures in a specialized order to model long-term dependency from a hybrid perspective. The proposed approach achieves state-of-the-art performance while maintaining relatively low computational costs.

The overall performance of MoU for long-term forecasting is summarized in the following Table (average performance). More detailed results can be found in our paper.

Model	Ours(MoU)		ModernTCN		PatchTST		DLinear		S-Mamba
Metric	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE	MSE	MAE
ETTh1	0.397	0.423	0.404	0.420	0.413	0.434	0.423	0.437	0.450	0.456
ETTh2	0.317	0.373	0.323	0.378	0.331	0.379	0.431	0.447	0.369	0.405
ETTm1	0.348	0.382	0.354	0.381	0.352	0.382	0.357	0.379	0.366	0.396
ETTm2	0.252	0.315	0.256	0.316	0.256	0.317	0.267	0.332	0.265	0.326
Weather	0.221	0.262	0.224	0.267	0.225	0.264	0.240	0.300	0.236	0.273
illness	1.500	0.784	1.519	0.799	1.513	0.825	2.169	1.041	1.977	0.890
electricity	0.157	0.253	0.157	0.253	0.159	0.253	0.177	0.274	0.166	0.262

⚡️ Preparation

Installation

Download code:

git clone https://github.com/lunaaa95/mou.git
cd mou

A suitable conda environment named mou can be created and activated with:

conda create -n mou python=3.8
conda activate mou
pip install -r requirement.txt

Dataset

Download datasets to folder ./dataset. You can download all datasets from Google Drive provided by Wu, H.

📍 Run

• We provide bash scripts for all datasets. Run bash scripts in folder "./scripts" to start time series long-term forecasting. For example,

bash scripts/MoU/etth1.sh

bash scripts/MoU/etth2.sh

bash scripts/MoU/ettm1.sh

bash scripts/MoU/ettm2.sh

bash scripts/MoU/weather.sh

bash scripts/MoU/electricity.sh

bash scripts/MoU/illness.sh

• We also provide other short-term encoders and long-term encoders to switch the structure of model. Change parameters entype for other short-term encoders and ltencoder for long-term encoders.
• We also give two baseline models of PatchTST and DLinear as well as their runing scripts.

🌟 Citation

@misc{peng2024mambatransformertimeseries,
      title={Mamba or Transformer for Time Series Forecasting? Mixture of Universals (MoU) Is All You Need}, 
      author={Sijia Peng and Yun Xiong and Yangyong Zhu and Zhiqiang Shen},
      year={2024},
      eprint={2408.15997},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2408.15997}, 
}