Table 4 Pearson correlation coefficients between prediction parameters and song popularity.

From: Quantifying the impact of homophily and influencer networks on song popularity prediction

(i) Previous popularity of artist

0.135

(ii) Genre average

0.069

(iii) Time to reach 200 listenings

\(-0.201\)

(iv) Time between repeated listenings

\(-0.193\)

(v) Number of repeated listeners

0.039

(vi) Area under the trend curve

\(-0.223\)

(vii) Area under the trend curve (normalized)

\(-0.044\)

(viii) Area under the trend curve (subtracted from diagonal)

\(-0.085\)

(ix) Influence 6 h

0.159

(x) Influence 12 h

0.169

(xi) Influence 24 h

0.174

(xii) Homophily

0.081

(xiii) Degree (friendship NW)

0.090

(xiv) Pagerank (friendship NW)

\(-0.054\)

(xv) Nearest neighbor degree (friendship NW)

\(-0.081\)

(xvi) Clustering coefficient (friendship NW)

0.069

(xvii) Degree (influencer NW)

0.132

(xviii) Pagerank (influencer NW)

0.110

(xix) Nearest neighbor degree (influencer NW)

0.090

(xx) Clustering (influencer NW)

0.062

  1. The prediction parameters were computed on the first 200 listenings for each song. Song popularity was defined as the maximum number of times a song was listened to last.fm in its best month. All p-values are below \({10^{-50}}\).
Back to article page