ai-ml-timeseries_skill

Operational patterns, templates, and decision rules for time series forecasting (modern best practices): tree-based methods (LightGBM), deep learning (Transformers, RNNs), future-guided learning, temporal validation, feature engineering, generative TS (Chronos), and production deployment. Emphasizes explainability, long-term dependency handling, and adaptive forecasting.