Tag: #time-series23 results
AidMap AI fuses open geospatial data to forecast urgency scores for droughts, floods in Central Asia. Uses features like vegetation anomalies and models from XGBoost to transformers. Seeks community feedback on architectures, data bias, and low-resource deployment.
A developer working in the agriculture sector is seeking advice on production-grade time series forecasting for crop volumes and pricing. They are looking for library recommendations and feature engineering strategies for highly seasonal data.
Global Temporal Retriever (GTR) is a plug-and-play module extending MTSF models' context via global pattern retrieval. Uses adaptive embeddings, dynamic alignment, and 2D convolution fusion. SOTA results on six datasets with low overhead; code on GitHub.