Repeated Listens in the Music Discovery Process

Brian Manolovitz, Mitsunori Ogihara

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

The long-tail problem for recommender systems is the phenomenon where few items in a system receive most of the interactions, while most items are unpopular and, therefore, are difficult to recommend. In the music domain, this problem is exacerbated by the fact that listeners revisit the songs and artists they enjoy, further widening the gap between popular and unpopular items. This chapter explores many avenues to address the long-tail problem, including a human-subjects study conducted by the authors, which tested the effects of repetition on retention rate. The results of the study indicate that with just one additional listen to a liked unfamiliar song, the listener is 10% more likely to revisit during normal listening.

Original languageEnglish (US)
Title of host publicationStudies in Computational Intelligence
PublisherSpringer Science and Business Media Deutschland GmbH
Pages119-134
Number of pages16
DOIs
StatePublished - 2021

Publication series

NameStudies in Computational Intelligence
Volume946
ISSN (Print)1860-949X
ISSN (Electronic)1860-9503

ASJC Scopus subject areas

  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'Repeated Listens in the Music Discovery Process'. Together they form a unique fingerprint.

Cite this