Modeling Popularity and Temporal Drift of Music Genre Preferences

Dataset statistics for the LowMS, MedMS, and HighMS Last.fm user groups. Here, |U| is the number of distinct users, |A| is the number of distinct artists, |G| is the number of distinct genres, |LE| is the number of listening events, |GA| is the number of genre assignments, |GA|/|LE| is the number of genre assignments per listening event, $\overline{G_u}$ is the average number of genres a user u has listened to, $\overline{\mathit{\text{MS}}}$ is the average mainstreaminess value, and $\overline{\mathit{\text{Age}}}$ is the average age of users in the group.

Comparison of our five baselines as well as our approach based on the BLL equation for modeling and predicting music genre preferences. In this table, a “✔” indicates that a specific approach covers a specific feature. While TOP, CF_u and CF_i also consider collaboration among users (i.e., investigate listening events of all users), our BLL_u approach is the only one that is personalized and accounts for the features of popularity as well as temporal drifts.

Genre prediction accuracy results of our study comparing our BLL_u approach with a group-based baseline (TOP), a user-based collaborative filtering baseline (CF_u), an item-based collaborative filtering baseline (CF_i), a frequency-based baseline (POP_u) and a recency-based baseline (TIME_u). For all three user groups (i.e., LowMS, MedMS, and HighMS), the combination of popularity and temporal drift of music genre preferences in the form of BLL_u provides the best results for all metrics. According to a t-test with α = .001, “***” indicates statistically significant differences between BLL_u and all other approaches for all user groups.