A white-boxed ISSM approach to estimate uncertainty distributions of Walmart sales

TL;DR

This paper introduces a white-boxed ISSM approach combining state-space models and Monte Carlo simulations to accurately estimate uncertainty in Walmart sales forecasting, excelling at granular SKU-level predictions.

Contribution

It presents a novel hybrid model integrating state-space methods with Monte Carlo simulations for detailed sales uncertainty estimation, achieving top-tier performance in a major forecasting competition.

Findings

Ranked 6th overall among 909 submissions.

Achieved 1st place at SKU-level predictions.

Effectively modeled over-dispersed sales with negative binomial distributions.

Abstract

We present our solution for the M5 Forecasting - Uncertainty competition. Our solution ranked 6\ts{th} out of 909 submissions across all hierarchical levels and ranked first for prediction at the finest level of granularity (product-store sales, i.e.\ SKUs). The model combines a multi-stage state-space model and Monte Carlo simulations to generate the forecasting scenarios (trajectories). Observed sales are modelled with negative binomial distributions to represent discrete over-dispersed sales. Seasonal factors are hand-crafted and modelled with linear coefficients that are calculated at the store-department level.