Support the Underground: Characteristics of Beyond-Mainstream Music Listeners

TL;DR

The paper tackles the bias in music recommender systems against beyond-mainstream listeners by constructing the LFM-BeyMS dataset, enriching Last.fm histories with acoustic features and genres. It identifies four beyond-mainstream music clusters and maps listeners to four subgroups, then analyzes how openness and diversity relate to recommendation accuracy, finding that openness strongly correlates with improved MAE, with the U_ambi subgroup achieving the best performance. Using four standard algorithms, including NMF, the study shows BeyMS users consistently receive poorer recommendations than mainstream users, yet certain subgroups (notably U_ambi) can surpass mainstream performance under appropriate models. The work provides open data and code, demonstrates the value of cluster-aware user modeling, and outlines future directions toward fairness and more nuanced user models for long-tail music. Overall, it advances understanding of long-tail listener characteristics and informs the design of recommender systems that better serve beyond-mainstream audiences.

Abstract

Music recommender systems have become an integral part of music streaming services such as Spotify and Last.fm to assist users navigating the extensive music collections offered by them. However, while music listeners interested in mainstream music are traditionally served well by music recommender systems, users interested in music beyond the mainstream (i.e., non-popular music) rarely receive relevant recommendations. In this paper, we study the characteristics of beyond-mainstream music and music listeners and analyze to what extent these characteristics impact the quality of music recommendations provided. Therefore, we create a novel dataset consisting of Last.fm listening histories of several thousand beyond-mainstream music listeners, which we enrich with additional metadata describing music tracks and music listeners. Our analysis of this dataset shows four subgroups within the group of beyond-mainstream music listeners that differ not only with respect to their preferred music but also with their demographic characteristics. Furthermore, we evaluate the quality of music recommendations that these subgroups are provided with four different recommendation algorithms where we find significant differences between the groups. Specifically, our results show a positive correlation between a subgroup's openness towards music listened to by members of other subgroups and recommendation accuracy. We believe that our findings provide valuable insights for developing improved user models and recommendation approaches to better serve beyond-mainstream music listeners.

Figures (13)

• Figure 1: Recommendation accuracy measured by the mean absolute error (MAE) of a non-negative matrix factorization-based approach (i.e., NMF luo2014efficient) and a neighborhood-based approach (i.e., UserKNN herlocker2004evaluating) for mainstream and beyond-mainstream user groups in Last.fm. We see that beyond-mainstream users receive a substantially lower recommendation quality (i.e., higher MAE) compared to mainstream music listeners. Thus, for recommender systems, it is harder to provide high-quality recommendations to beyond-mainstream music listeners than to mainstream music listeners.
• Figure 2: Overview of our new LFM-BeyMS dataset and its data sources. We depict the different features, their origin, and relation, and show the feature groups with the number of contained features in brackets. LFM-BeyMS contains BeyMS, i.e., data to study the beyond-mainstream user group, and Recommendation, i.e., data to conduct recommendation experiments of beyond-mainstream and mainstream music listeners.
• Figure 3: Distribution of listening events in our set of Last.fm users. We set the lower and upper bound marked as dashed and dotted lines, respectively based on the gradient, which results in 12,814 users with a similar number of listening events.
• Figure 4: Mainstreaminess distribution of the 12,814 users illustrated in Figure \ref{['fig:le_dist']}. Based on the maximum gradient, we select an upper bound of 0.097732 to identify the 2,074 beyond-mainstream users of the BeyMS user group.
• Figure 5: IDF-score distribution of the top-100 genres in ascending order (i.e., from coarse-grained to fine-grained). The 6 coarse-grained genres below the lower bound of 0.90 are removed from the genre assignments, i.e., "rock", "pop", "electronic", "metal", "alternativerock", "indierock".