表 3. Comparison of recent approaches and our proposed method.
本文方法与新近方法的实验结果对比
| 算法 | 数据集 | Accu | Rec | RMSE |
| user-based 方法 | MATADOR | 89.12% | 85.12% | 1.115 |
| STITCH | 88.54% | 84.62% | 1.295 | |
| item-based 方法 | MATADOR | 88.45% | 84.12% | 1.138 |
| STITCH | 86.41% | 82.89% | 1.345 | |
| Naïve Bayes 方法 | MATADOR | 91.52% | 88.45% | 1.038 |
| STITCH | 89.92% | 86.95% | 1.120 | |
| SVM 方法 | MATADOR | 91.92% | 89.21% | 0.914 |
| STITCH | 90.10% | 88.32% | 0.983 | |
| MTOI 方法 | MATADOR | 91.35% | 90.35% | 0.902 |
| STITCH | 90.65% | 89.65% | 0.972 | |
| 基于二分图的核
回归方法 |
MATADOR | 92.36% | 91.62% | 0.926 |
| STITCH | 91.69% | 90.62% | 0.982 | |
| 基于化学相似度
的方法 |
MATADOR | 90.18% | 89.12% | 0.933 |
| STITCH | 88.64% | 85.64% | 0.950 | |
| 本文方法 | MATADOR | 95.10% | 93.21% | 0.807 |
| STITCH | 94.58% | 91.58% | 0.892 |