Algoritmos de recomendación de nuevas canciones, encuentran más dificultades para encontrar temas para amantes del hip-hop o hard rock

The most used applications for listening to music, such as  Spotify ,  Last.fm or YouTube  itself  , have algorithms capable of predicting and showing you new  music that you may like . In a simple way, it is a recommendation system through  collaborative filtering : the  apps  record the artists and genres that a user listens to and compare these results with similar listeners to find out what others like.

Thus, lovers of Lil Nas X are quite rightly recommended to listen to Post Malone or, if you liked Soleá Morente, they will have recommended Rigoberta Bandini.

But these algorithms are not perfect with something as subjective and human as artistic creation and musical tastes. For this reason, a team of researchers from the Graz University of Technology, the Know-Center GmbH research center, the Johannes Kepler University of Linz, the University of Innsbruck (all from Austria) and the University of Utrecht (the Netherlands) wanted to to test  how accurate the recommendations  generated by these algorithms are, especially for listeners of music that is not very popular or not so well known to the general public.

The main result, published in the latest issue of EPJ Data Science magazine  , is that these algorithms  fail significantly more for hard rock and hip-hop  listeners  than for other music genres.

To verify this, the team took the history of songs listened to by 4,148 users of the Last.fm platform, both listeners who tend to listen to more popular commercial music and those who prefer slightly lesser-known artists (2,074 users in each group).

Based on the artists most listened to by each user, the research used a  computational model  to predict whether they would like a new song or artist using four different recommendation algorithms.

In this way, they confirmed that popular music listeners tend to receive more accurate and precise recommendations than the less commercial group of listeners.

Following this, the authors categorized non-commercial music listeners into  four groups , based on the characteristics of the music they most often listen to.

These groups were: listeners of musical genres that only contain  acoustic instruments , such as folk or singer-songwriters; very energetic music   like punk or hip-hop; very acoustic but voiceless music   like background music; and very energetic but voiceless music   like electronics.

The research was thus able to compare the histories of each group and identify, with the computational model, which users were more likely to listen to music outside their preferences and the diversity of musical genres within each group.

Thanks to this categorization, the study found that listeners of non-voice acoustic music also tended to prefer songs from the other three groups (energetic, non-voice energetic, and acoustic) and received more accurate recommendations from the computational model.

At the time, the group of  energetic music listeners  received the  worst recommendations  from the algorithms, despite the fact that their group featured the widest variety of music genres – hard rock, punk, hardcore, hip-hop and pop rock.

 

Popularity bias in algorithms

Elisabeth Lex , co-author of the paper and Associate Professor of Applied Informatics at the Graz University of Technology, points out that music recommendation algorithms are already «essential» for users who want to search, select and filter the collections of music applications.

Despite this, it indicates that algorithms can fail to recommend for listeners of uncommercial music. “This may be because these systems are  biased towards more popular music , making non- mainstream artists   less listened to,” she notes.

Finally, the authors suggest that their findings could serve as a basis for creating music recommendation systems that offer more precise recommendations. However, they warn that their analysis is based on a sample of Last.fm users, which could be unrepresentative for this or other music platforms.