第6期 程龙，等：弱标记不完备决策系统的增量式属性约简算法 ·1085· 0.9 1.0 0.8d 0.9 0.7 0.8 0.6 0.7 0.5 0.6 0.4 -◆-WIDAR -◆-VIDAR 0.3 -Labeled data 0.5 Labeled data o-Semi D -o-Semi D 0.2 0.4 40 50 60 70 80 90 40 50 60 70 8090 标记缺失比例% 标记缺失比例% (c)Dermatology (d)Cylinder bands 1.000◆ 1.0r 0.998 0.9 0.996 0.994 0.8 0.992 $0.990 点0.7 0.988 ◆WIDAR 0.6 ◆-WIDAR 0.986 ▲-Labeled data ▲Labeled data o-Semi D o-Semi D 0.984 0. 0 6070 80 90 30 40 50 60708090 标记缺失比例% 标记缺失比例% (e)Mushroom (①Letter recognition 图33N分类器的分类精度 Fig.3 Classification accuracy with the 3NN classifier 0.9 1.0 0.9 0.8 0.7 0.4 0.6 ◆-WIDAR 0.3 ◆WIDAR 0.5 Labeled data ▲-Labeled data 02 0- Semi D 0- Semi D 0.4 0 50 60 70 80 90 0 50 60 70 80 90 标记缺失比例/% 标记缺失比例% (a)Automobile (b)Soybean (Large) 1.0 0.9 1.0◆ 0.8 0.9 0.7 0.8 0 05 0.7 0.4 0.6 ●-NIDAR WIDAR 0.3 Labeled data 0.5 Labeled data 0- Semi D -o-Semi D 02 90 0.4 0 50 6070 50 60 70 80 90 标记缺失比例% 标记缺失比例% (c)Dermatology (d)Cylinder bands40 50 60 70 80 90 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 分类精度 标记缺失比例/% (c) Dermatology 40 50 60 70 80 90 0.4 0.5 0.6 0.7 0.8 0.9 1.0 分类精度 标记缺失比例/% (d) Cylinder bands 40 50 60 70 80 90 0.984 0.986 0.988 0.990 0.992 0.994 0.996 0.998 1.000 分类精度 标记缺失比例/% (e) Mushroom 30 40 50 60 70 80 90 0.5 0.6 0.7 0.8 0.9 1.0 分类精度 标记缺失比例/% (f) Letter recognition WIDAR Labeled data Semi_D WIDAR Labeled data Semi_D WIDAR Labeled data Semi_D WIDAR Labeled data Semi_D 图 3 3NN 分类器的分类精度 Fig. 3 Classification accuracy with the 3NN classifier 40 50 60 70 80 90 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 分类精度 标记缺失比例/% WIDAR Labeled data Semi_D WIDAR Labeled data Semi_D WIDAR Labeled data Semi_D WIDAR Labeled data Semi_D (a) Automobile 40 50 60 70 80 90 0.4 0.5 0.6 0.7 0.8 0.9 1.0 分类精度 标记缺失比例/% (b) Soybean (Large) 40 50 60 70 80 90 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 分类精度 标记缺失比例/% (c) Dermatology 40 50 60 70 80 90 0.4 0.5 0.6 0.7 0.8 0.9 1.0 分类精度 标记缺失比例/% (d) Cylinder bands 第 6 期 程龙，等：弱标记不完备决策系统的增量式属性约简算法 ·1085·