Enabling Reproducibility in Group Recommender Systems Joaquin Dario SILVEIRA a , Maria SALAM ´ O b,1 , and Ludovico BORATTO c , a Universitat Polit` ecnica de Catalunya, Barcelona, Spain b Dept. Mathematics and Computer Science, University of Barcelona, Barcelona, Spain Institute of Complex Systems, University of Barcelona, Barcelona, Spain c Dept. Mathematics and Computer Science, University of Cagliari, Cagliari, Italy Abstract. Reproducibility is a challenging aspect that considerably affects the qual- ity of most scientific papers. To deal with this, many open frameworks allow to build, test, and benchmark recommender systems for single users. Group recom- mender systems involve additional tasks w.r.t. those for single users, such as the identification of the groups, or their modeling. While this clearly amplifies the possible reproducibility issues, to date, no framework to benchmark group recom- mender systems exists. In this work, we enable reproducibility in group recom- mender systems by extending the LibRec library, which stands out as one of the richest, with more than 70 different recommender algorithms, good performance and several evaluation metrics. Specifically, we include several approaches for all the stages of group recommender systems: group formation, group modeling strate- gies, and evaluation. To validate our framework, we consider a use-case that com- pares several group building, recommendation, and group modeling approaches. Keywords. Group Recommender Systems, Reproducibility, Algorithms 1. Introduction Enabling reproducibility should be of paramount importance inside the research commu- nity [1]. In fact, it is hard to determine the speed of progress, or even if we are making any, when so much of the newly generated knowledge is not reproducible [2]. The ex- istence of base libraries with known and well studied approaches and algorithms is one of the first steps in any field that seeks to advance on firm knowledge. Moreover, it is in fields in which such frameworks are missing that it is hardest to justify new ideas by benchmarking them against the existing literature. Recommender systems (RSs) support users by suggesting items that might be of interest to them [3]. This is usually done by learning behavioral patterns from historical data, usually in the form of user-item interactions. Nearly any popular programming language has a library or framework for making single recommendations. Despite the amount of papers regarding the problem of generating recommendations for groups of users (group recommender systems, GRSs) [4,5], a firm ground for GRSs does not exist. 1 Corresponding Author: Department of Mathematics and Computer Science, University of Barcelona, Barcelona, Spain; E-mail: maria.salamo@ub.edu Artificial Intelligence Research and Development A. Cortés et al. (Eds.) © 2022 The authors and IOS Press. This article is published online with Open Access by IOS Press and distributed under the terms of the Creative Commons Attribution Non-Commercial License 4.0 (CC BY-NC 4.0). doi:10.3233/FAIA220324 115