Research Article

Improving Top-N Recommendation Performance Using Missing Data

Table 2

Performance of different approaches on EM.

ā€‰NDCGRecall1-callCOVCILNDCG+
ā€‰135135

UserCF0.0100.0140.0190.0020.0080.0170.12174.0116.20.65
Slope-one0.0350.0340.0340.0080.0190.0270.1511.07.00.71
SVD++0.1120.0910.0880.0310.0580.0800.2830.612.60.83
OrdRec0.0630.0570.0490.0150.0400.0440.1910.66.80.66
Pure0.0740.0710.0710.0180.0390.0520.2738.815.00.69
AllRank0.1110.1040.1030.0310.0730.1030.3713.44.40.83
WSVD++0.0810.0980.1000.0200.0650.0900.3656.026.00.80
RSSVD++0.0950.0960.1000.0270.0610.0860.3452.823.60.79
NSSVD++0.1530.1480.1500.0420.1020.1430.5058.223.80.81