<ul><li>A new framework called KARMA has been developed for multivariate long-term time series forecasting.</li><li>KARMA uses an Adaptive Time Channel Decomposition module (ATCD) and a Hybrid Frequency-Time Decomposition module (HFTD) to extract trend and seasonal components.</li><li>The framework integrates a multi-scale Mamba-based KarmaBlock to process global and local information efficiently.</li><li>Experiments on real-world datasets show that KARMA outperforms mainstream baseline methods in predictive accuracy and computational efficiency.</li></ul>

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

Discover more