ai-ml-timeseries

What problem does it solve?

Time series forecasting is challenging due to temporal dependencies, seasonality, and the need for robust validation. This Skill provides modern, production-ready patterns to build accurate and explainable forecasting models.

Core Features & Use Cases

• Advanced Forecasting Models: Implement tree-based methods (LightGBM), deep learning (Transformers, RNNs), and Future-Guided Learning for event prediction.
• Robust Validation: Design temporal validation, backtesting, and multi-step forecasting strategies to prevent data leakage and ensure reliability.
• Production Best Practices: Covers feature engineering (lags, rolling windows), model selection, explainability (SHAP), and deployment with drift monitoring.

Quick Start

Use the ai-ml-timeseries skill to generate a LightGBM model for daily sales forecasting, including lag and rolling features.