Skip to main content

Table 4 Mean absolute error (MAE) measurements for the four subgroups and four personalized recommendation algorithms. NMF (in bold) outperforms all other algorithms for all subgroups. Among the subgroups, the best accuracy results (i.e., lowest MAE scores) are reached by \(U_{\text{ambi}}\), while the worst accuracy results (i.e., highest MAE scores) are reached by \(U_{\text{hard}}\). To facilitate comparison, we also show the MAE measurements for the BeyMS and MS user groups

From: Support the underground: characteristics of beyond-mainstream music listeners

Subgroup UserItemAvg UserKNN UserKNNAvg NMF
\(U_{\text{folk}}\) 63.2143 70.3049 67.4406 57.2278
\(U_{\text{hard}}\) 65.1464 73.1949 69.2855 59.6887
\(U_{\text{ambi}}\) 60.5558 69.8315 65.5708 54.2073
\(U_{\text{elec}}\) 62.2894 71.0387 66.1499 56.6209
BeyMS 63.4608 71.6694 67.5856 57.7703
MS 61.2562 68.4894 63.3985 54.8182