Research Article
Improving Top-N Recommendation Performance Using Missing Data
Table 1
Performance of different approaches on ML.
| ā | NDCG | Recall | 1-call | COV | CIL | NDCG+ | ā | 1 | 3 | 5 | 1 | 3 | 5 |
| UserCF | 0.019 | 0.019 | 0.020 | 0.000 | 0.002 | 0.003 | 0.12 | 129.4 | 69.8 | 0.65 | Slope-one | 0.026 | 0.033 | 0.033 | 0.001 | 0.003 | 0.005 | 0.18 | 19.6 | 5.8 | 0.68 | SVD++ | 0.030 | 0.036 | 0.043 | 0.001 | 0.003 | 0.006 | 0.19 | 20.8 | 7.6 | 0.71 | OrdRec | 0.080 | 0.065 | 0.062 | 0.002 | 0.005 | 0.008 | 0.23 | 19.0 | 4.6 | 0.52 | Pure | 0.037 | 0.033 | 0.032 | 0.001 | 0.002 | 0.004 | 0.11 | 34.2 | 13.0 | 0.66 | AllRank | 0.063 | 0.058 | 0.055 | 0.002 | 0.004 | 0.006 | 0.19 | 21.2 | 4.4 | 0.70 | WSVD++ | 0.104 | 0.083 | 0.079 | 0.004 | 0.008 | 0.014 | 0.28 | 23.6 | 8.8 | 0.70 | RSSVD++ | 0.095 | 0.074 | 0.073 | 0.006 | 0.007 | 0.013 | 0.27 | 24.0 | 9.2 | 0.69 | NSSVD++ | 0.108 | 0.083 | 0.078 | 0.006 | 0.013 | 0.019 | 0.30 | 38.2 | 16.2 | 0.68 |
|
|