🐍 KARMA: A Multilevel Decomposition Hybrid Mamba Framework for Multivariate Long-Term Time Series Forecasting 🚩

July 15, 2025 · View on GitHub

🌟 Contributions

To the best of our knowledge, KARMA is the first framework to integrate Mamba with multi-level decomposition for multi-scaled hybrid frequency-temporal analysis, fully leveraging Mamba's potential to overcome the limitations of Transformer architectures and existing decomposition methods for efficient time series forecasting.
We propose ATCD for dynamic seasonal-trend decomposition and HFTD for hybrid frequency-temporal analysis, enabling the effective utilization of latent information in time series data and improving predictive accuracy.
Comprehensive experiments on eight real-world datasets across various domains demonstrate that KARMA achieves state-of-the-art performance in most benchmarks.

🧩 Architecture

📑 Full Results

📡 Prerequisites

Ensure you are using Python 3.10 and install the necessary dependencies by running:

pip install -r requirements.txt

📊 Prepare Datastes

Begin by downloading the required datasets. All datasets are conveniently available at Autoformer. Create a separate folder named ./dataset and neatly organize all the csv files as shown below:

dataset
└── electricity.csv
└── Exchange.csv
└── traffic.csv
└── weather.csv
└── ETTh1.csv
└── ETTh2.csv
└── ETTm1.csv
└── ETTm2.csv

💻 Run

sh ./scripts/Karma_all.sh

📚 Citation

If you find this repo useful, please consider citing our paper as follows:

@inproceedings{ye2025karma,
  title={KARMA: A Multilevel Decomposition Hybrid Mamba Framework for Multivariate Long-Term Time Series Forecasting},
  author={Ye, Hang and Duan, Gaoxiang and Zeng, Haoran and Zhu, Yangxin and Meng, Lingxue and Zheng, Xiaoying and Zhu, Yongxin},
  booktitle={International Conference on Wireless Artificial Intelligent Computing Systems and Applications},
  pages={266--276},
  year={2025},
  organization={Springer}
}

🙏 Acknowledgement

Special thanks to the following repositories for their invaluable code and datasets: