Optimal Combination Forecasts on Retail Multi-Dimensional Sales Data

TL;DR

This paper explores hierarchical time series forecasting in retail, demonstrating that optimal combination methods, especially Weighted Least Squares, improve accuracy by leveraging nested sales data structures.

Contribution

It introduces an optimal combination approach for hierarchical retail sales forecasting, showing improved accuracy over existing reconciliation methods.

Findings

WLS outperforms other reconciliation methods.

Hierarchical models improve forecast accuracy for aggregated sales.

Nested information enhances predictive performance.

Abstract

Time series data in the retail world are particularly rich in terms of dimensionality, and these dimensions can be aggregated in groups or hierarchies. Valuable information is nested in these complex structures, which helps to predict the aggregated time series data. From a portfolio of brands under HUUB's monitoring, we selected two to explore their sales behaviour, leveraging the grouping properties of their product structure. Using statistical models, namely SARIMA, to forecast each level of the hierarchy, an optimal combination approach was used to generate more consistent forecasts in the higher levels. Our results show that the proposed methods can indeed capture nested information in the more granular series, helping to improve the forecast accuracy of the aggregated series. The Weighted Least Squares (WLS) method surpasses all other methods proposed in the study, including the Minimum Trace (MinT) reconciliation.