Studie zu Machine-Learning-Verfahren im Retourenmanagement veröffentlicht
Ein neuer Artikel zeigt, wie moderne Machine-Learning-Verfahren zur Verbesserung von Retourenprognosen im Online-Handel eingesetzt werden können.
Die Studie analysiert bestehende Forschung zu Retourenprognosen und erweitert diese um eine spezifische Betrachtung des Online-Handels. Die Ergebnisse geben einen Überblick über relevante Algorithmen, Datenanforderungen und Bewertungskriterien.
Abstract des Papers: While e-commerce has recently experienced substantial growth rates, retailers face increasing consumer returns. Machine learning techniques opened up opportunities for improved consumer returns forecasting. In the past, returns forecasting was analysed predominantly from a broader reverse logistics and closed-loop supply chain perspective. This paper extends this view by reviewing the state of research on current algorithms for forecasting returns in e-commerce in particular and integrating it into the body of knowledge regarding forecasting product returns extracted from previous reviews. Methodologically, four reviews were synthesised first. Subsequently, a systematic literature review was conducted, analysing 28 additional publications related to consumer returns and enriching the literature on product returns. Thus, this comprehensive review is the first to analyse current forecasting issues while integrating the e-commerce perspective and emphasising relevant developments regarding advanced algorithms and metrics for their assessment in returns forecasting.
Zitation: Karl, D. (2024): "Evaluating Advanced Product Return Forecasting Algorithms: A (Meta-)Review Integrating Consumer Returns Research", in: International Journal of Business Forecasting and Marketing Intelligence, Vol. 9, No. 2, S. 213-241, DOI: https://doi.org/10.1504/IJBFMI.2024.137648
