MP-MoE: Matrix Profile-Guided Mixture of Experts for Precipitation Forecasting

1. 1Identify complex topography and convective instability in tropical regions as challenges for precipitation forecasting
2. 2Develop a data-driven post-processing framework using matrix profiles to mitigate biases in NWP models
3. 3Implement a mixture of experts (MoE) model guided by matrix profiles to improve forecasting accuracy
4. 4Evaluate the performance of the MP-MoE model under minor temporal misalignments to assess its robustness

Data scientists and researchers working on weather forecasting models can benefit from this approach, as it addresses the limitations of traditional NWP models and provides a more accurate forecasting method

💡 Using matrix profiles to guide a mixture of experts can improve the accuracy of precipitation forecasting in tropical regions

💡 Improving precipitation forecasting in tropical regions with MP-MoE, a matrix profile-guided mixture of experts model