Abstract:Platform businesses operate on a digital core and their decision making requires high-dimensional accurate forecast streams at different levels of cross-sectional (e.g., geographical regions) and temporal aggregation (e.g., minutes to days). It also necessitates coherent forecasts across all levels of the hierarchy to ensure aligned decision making across different planning units such as pricing, product, controlling and strategy. Given that platform data streams feature complex characteristics and interdependencies, we introduce a non-linear hierarchical forecast reconciliation method that produces cross-temporal reconciled forecasts in a direct and automated way through the use of popular machine learning methods. The method is sufficiently fast to allow forecast-based high-frequency decision making that platforms require. We empirically test our framework on a unique, large-scale streaming dataset from a leading on-demand delivery platform in Europe.
Abstract:Data stream forecasts are essential inputs for decision making at digital platforms. Machine learning algorithms are appealing candidates to produce such forecasts. Yet, digital platforms require a large-scale forecast framework that can flexibly respond to sudden performance drops. Re-training ML algorithms at the same speed as new data batches enter is usually computationally too costly. On the other hand, infrequent re-training requires specifying the re-training frequency and typically comes with a severe cost of forecast deterioration. To ensure accurate and stable forecasts, we propose a simple data-driven monitoring procedure to answer the question when the ML algorithm should be re-trained. Instead of investigating instability of the data streams, we test if the incoming streaming forecast loss batch differs from a well-defined reference batch. Using a novel dataset constituting 15-min frequency data streams from an on-demand logistics platform operating in London, we apply the monitoring procedure to popular ML algorithms including random forest, XGBoost and lasso. We show that monitor-based re-training produces accurate forecasts compared to viable benchmarks while preserving computational feasibility. Moreover, the choice of monitoring procedure is more important than the choice of ML algorithm, thereby permitting practitioners to combine the proposed monitoring procedure with one's favorite forecasting algorithm.