Hacking the TSB Model for Intermittent Time Series to Accommodate for Availability Constraints

In many demand forecasting problems, we face the challenge of predicting the demand for a product or service with very sparse data. This is especially true in retail, where many products are only sold occasionally. This sparsity can come from many sources, as greatly described in the (recommended!) blog post “Why zeroes happen” by Ivan Svetunkov. On one hand, the sparsity can simply come from the lack of demand for the product. On the other hand, the sparsity can also come from the lack of availability of the product. For example, a product might be only available for purchase during certain times of the year or only in certain geographical locations. There could be other reasons as well. In this notebook, we experiment with an extension on a classical TSB time-series model for intermittent time series to accommodate availability constraints.