TSB-HB: Hierarchical Bayesian TSB Model for Intermittent Demand Forecasting

This is a smarter crystal ball for products that sell only occasionally (like spare parts or niche items). It extends an existing statistical method (TSB) with a hierarchical Bayesian approach so that forecasts for many low-selling items can "learn" from each other, leading to more reliable predictions when historical data is very sparse or intermittent.