考虑时间动态性和序列模式的个性化推荐算法*
|
|
李杰, 杨芳, 徐晨曦
|
A Personalized Recommendation Algorithm with Temporal Dynamics and Sequential Patterns
|
|
Li Jie,Yang Fang,Xu Chenxi
|
|
| 表1 相似度计算函数比较 |
| |
|
| TopN | 相似度函数 | 准确率 | 召回率 | F值 | | 5 | 皮尔逊 | 7.25% | 6.33% | 6.76% | | 余弦相似 | 7.98% | 8.06% | 8.02% | | 修正余弦 | 8.26% | 8.11% | 8.18% | | 本文改进函数 | 8.24% | 9.34% | 8.76% | | 10 | 皮尔逊 | 10.38% | 9.96% | 10.17% | | 余弦相似 | 11.24% | 10.01% | 10.59% | | 修正余弦 | 12.22% | 11.88% | 12.05% | | 本文改进函数 | 13.58% | 11.79% | 12.62% | | 20 | 皮尔逊 | 10.38% | 10.48% | 10.43% | | 余弦相似 | 11.24% | 10.84% | 11.04% | | 修正余弦 | 11.98% | 12.65% | 12.43% | | 本文改进函数 | 11.79% | 13.05% | 12.39% |
|
|
|