Hierarchical forecast reconciliation with machine learning

Over the last 15 years, studies on hierarchical forecasting have moved away from single-level approaches towards proposing linear combination approaches across multiple levels of the hierarchy. Such combinations offer coherent reconciled forecasts, improved forecasting performance and aligned decision-making. This paper proposes a novel hierarchical forecasting approach based on machine learning. The proposed method allows for non-linear combinations of the base forecasts, thus being more general than linear approaches. We structurally combine the objectives of improved post-sample empirical forecasting accuracy and coherence. Due to its non-linear nature, our approach selectively combines the base forecasts in a direct and automated way without requiring that the complete information must be used for producing reconciled forecasts for each series and level. The proposed method is evaluated both in terms of accuracy and bias using two different data sets coming from the tourism and retail industries. Our results suggest that the proposed method gives superior point forecasts than existing approaches, especially when the series comprising the hierarchy are not characterized by the same patterns. •We use machine learning approaches for hierarchical reconciliation.•We offer a non-linear approach to the problem of hierarchical coherence.•Our approach explicitly focuses on post-sample in estimating the combination weights.•Our approach offers improved forecasting performance on two real-life data sets.