Abstract—In this paper we use two industry datasets and apply two embedding methods and their combination for developing and testing session-based recommenders. The first dataset corresponds to an e-commerce scenario demanding effective next-item recommendation. The second dataset represents a last-basket prediction setting. Experimental results show that in the next-item task, the content-based approach utilizing textual descriptors of the items extracted by Doc2Vec and collaborative item embedding approach Item2Vec, which is trained upon item sequences, have comparable performance. When combined, they produce the best next-item predictor. In the last-basket recommendation scenario, Item2Vec significantly outperforms the Doc2Vec embedding method. Finally, we report on experiments with reranking methods that demonstrate the effectiveness of simple and practical methods, using item categories, to improve the recommendations.
Index Terms—Session-based recommenders, embeddings methods, next-item recommendations, last-basket prediction.
All authors are with the International Hellenic University, Thessaloniki, Greece (e-mail of the corresponding author: firstname.lastname@example.org).
Cite: Embeddings Methods for Next-Item and Last-Basket Session-Based Recommendations, "Embeddings Methods for Next-Item and Last-Basket Session-Based Recommendations," International Journal of Machine Learning and Computing vol. 12, no. 4, pp. 136-142, 2022.Copyright © 2022 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).