Taxonomy-Conditioned Hierarchical Bayesian TSB Models for Heterogeneous Intermittent Demand Forecasting

TL;DR

This paper introduces TSB-HB, a hierarchical Bayesian model for intermittent demand forecasting that improves accuracy and stability by leveraging hierarchical priors and a principled generative framework.

Contribution

The paper presents a novel hierarchical Bayesian extension of the TSB model, providing a coherent generative interpretation and improved forecasting performance for sparse and heterogeneous demand data.

Findings

TSB-HB achieves lowest RMSE and RMSSE on UCI Online Retail dataset.

TSB-HB outperforms classical baselines in MAE and RMSE on a 5,000-series M5 dataset.

Under probabilistic calibration, TSB-HB offers competitive pinball loss and better sharpness-calibration tradeoff.

Abstract

Intermittent demand forecasting poses unique challenges due to sparse observations, cold-start items, and obsolescence. Classical models such as Croston, SBA, and the Teunter--Syntetos--Babai (TSB) method provide simple heuristics but lack a principled generative foundation. We introduce TSB-HB, a hierarchical Bayesian extension of TSB. Demand occurrence is modeled with a Beta--Binomial distribution, while nonzero demand sizes follow a Log-Normal distribution. Crucially, hierarchical priors enable partial pooling across items, stabilizing estimates for sparse or cold-start series while preserving heterogeneity. This framework provides a coherent generative reinterpretation of the classical TSB structure. On the UCI Online Retail dataset, TSB-HB achieves the lowest RMSE and RMSSE among all baselines, while remaining competitive in MAE. On a 5,000-series M5 sample, it improves MAE and RMSE over classical intermittent baselines. Under the calibrated probabilistic configuration, TSB-HB yields competitive pinball loss and a favorable sharpness--calibration tradeoff among the parametric baselines reported in the main text.