Predicting the category of customers' next product to buy in web shops
Laura Rekasiute, Alvaro Jose Jimenez Palenzuela, Nijole Salnaite, Ramon Carrera Cuenca, Flavius Frăsincar · Proceedings of the 37th ACM/SIGAPP Symposium on Applied Computing · 2022
Recommender systems are widely used by online retailers to entice customers into making new purchases. Understanding and predicting customer behavior is thus of utmost importance to retailers. In this paper our main goal is to predict the next product category that a certain customer will buy given his/her purchase history. We propose a Sequential Event Prediction model that captures both general and customer-specific consumption behavior through confidence rules. We use anonymized purchasing data from a Web shop in the Netherlands to show empirically that our approach outperforms several models proposed in the literature.