Research Article

Improving Top-N Recommendation Performance Using Missing Data

Table 1

Performance of different approaches on ML.

ā€‰NDCGRecall1-callCOVCILNDCG+
ā€‰135135

UserCF0.0190.0190.0200.0000.0020.0030.12129.469.80.65
Slope-one0.0260.0330.0330.0010.0030.0050.1819.65.80.68
SVD++0.0300.0360.0430.0010.0030.0060.1920.87.60.71
OrdRec0.0800.0650.0620.0020.0050.0080.2319.04.60.52
Pure0.0370.0330.0320.0010.0020.0040.1134.213.00.66
AllRank0.0630.0580.0550.0020.0040.0060.1921.24.40.70
WSVD++0.1040.0830.0790.0040.0080.0140.2823.68.80.70
RSSVD++0.0950.0740.0730.0060.0070.0130.2724.09.20.69
NSSVD++0.1080.0830.0780.0060.0130.0190.3038.216.20.68