Revisiting neighbourhood-based recommenders for temporal scenarios

A. Bellogín, P. Sánchez

1st Workshop on Temporal Reasoning in Recommender System - RecTemp 2017, Como (Italia). 27-31 agosto 2017

Resumen:

Modelling the temporal context efficiently and effectively is essential to provide useful recommendations to users. Methods such as matrix factorisation and Markov chains have been combined recently to model the temporal preferences of users in a sequential basis. In this work, we focus on Neighbourhood-based Collaborative Filtering and propose a simple technique that incorporates interaction sequences when producing a personalised ranking. We show the efficiency of this method when compared against other
sequence- and time-aware recommendation methods under two classical temporal evaluation methodologies.

Publicado en CEUR Workshop Proceedings, pp: 40-44

Fecha de publicación: 2017-08-27.

Cita:
A. Bellogín, P. Sánchez, Revisiting neighbourhood-based recommenders for temporal scenarios, 1st Workshop on Temporal Reasoning in Recommender System - RecTemp 2017, Como (Italia). 27-31 agosto 2017. En: CEUR Workshop Proceedings, vol. 1922, e-ISSN: 1613-0073