一种基于嵌入式的弱标记分类算法
|
李亚重,杨有龙,仇海全
|
Label embedding for weak label classification
|
Yachong Li,Youlong Yang,Haiquan Qiu
|
|
表6 测试数据在平均精度上的预测结果
|
Table 6 Prediction results for test data on Average Precision
|
|
Data set | | LEWL | MLML | LRML | CPLST | BR |
---|
Emotions | 0.3 | 0.7320.026 | 0.7160.01999 | 0.7290.0222 | 0.6620.038 | 0.6450.020 | 0.7 | 0.7080.012 | 0.6870.039 | 0.6820.016 | 0.5580.023 | 0.5680.021 | Yeast | 0.3 | 0.7080.006 | 0.6540.005 | 0.4410.024 | 0.5870.004 | 0.5730.005 | 0.7 | 0.6980.005 | 0.6480.008 | 0.440.026 | 0.4480.008 | 0.4500.003 | CAL500 | 0.3 | 0.3650.017 | 0.310.013 | 0.2020.011 | 0.2480.012 | 0.2480.004 | 0.7 | 0.3690.013 | 0.3070.013 | 0.1970.009 | 0.1700.008 | 0.1750.009 | Medical | 0.3 | 0.8050.025 | 0.7340.009 | 0.7750.008 | 0.6750.049 | 0.2890.029 | 0.7 | 0.7180.039 | 0.6850.018 | 0.7210.038 | 0.3320.030 | 0.1120.011 | Langlog | 0.3 | 0.5430.019 | 0.4190.016 | 0.4100.017 | 0.4610.048 | 0.4690.019 | 0.7 | 0.5300.016 | 0.5310.019 | 0.3990.017 | 0.3080.029 | 0.3090.011 | Enron | 0.3 | 0.5360.041 | 0.3950.051 | 0.3480.007 | 0.3310.045 | 0.2960.043 | 0.7 | 0.4930.054 | 0.3460.055 | 0.2520.011 | 0.2130.020 | 0.1970.025 | Corel5k | 0.3 | 0.0380.003 | 0.0270.008 | 0.0270.005 | 0.0300.007 | 0.0270.007 | 0.7 | 0.0220.002 | 0.0250.007 | 0.0260.004 | 0.0270.007 | 0.0230.007 |
|
|
|