Skip to main content

Table 2 Performance comparison on eight datasets in terms of RMSE

From: Evaluation of recent advances in recommender systems on Arabic content

Datasets ALFM A3NCF PARL CARL CARP q%
(English vs. Arabic)
ALFM (%) A3NCF (%) PARL (%) CARL (%) CARP (%)
MI 0.881 0.975 0.891 0.882 0.881 0.21 0.21 0 0 0
Arabic_MI 0.883 0.977 0.891 0.882 0.881
PLG 0.973 1.041 0.980 0.985 0.967 0.21 1.06 0 0 0
Arabic_PLG 0.975 1.052 0.980 0.985 0.967
Auto 0.894 0.982 0.914 0.901 0.894 0.11 1.12 0 0 0
Arabic_Auto 0.895 0.993 0.914 0.901 0.894
IV 0.976 1.023 1.020 0.990 0.986 0.20 0.78 0 0 0
Arabic_IV 0.978 1.031 1.020 0.990 0.986
  1. q%: accuracy decrease (pair of English-Arabic datasets)