90 工程科学学报，第44卷，第1期 表3（续） Table 3(Continued) Datasets Model G-mean F-measure-micro F-measure-macro Parameter.C KELM 0.841±0.0861 0.9147±0.0433 0.8757±0.066 27 WKELM-WI 0.8765±0.0759 0.9302±0.0347 0.8962±0.0524 26 WKELM-W2 0.8753±0.0662 0.9349±0.0264 0.9±0.0428 20 /-MKELM 0.92±0.0712 0.9651±0.0353 0.9468±0.0507 2 Thyroid -MKELM 0.9051±0.0877 0.9535±0.0329 0.9308±0.051 2, ITDSMM-KELM 0.9139±0.0796 0.9574±0.0373 0.9374±0.0559 24 /p-RWMKE-ELM-W1 0.9231±0.0471 0.9628±0.0127 0.9438±0.0221 22 -RWMKE-ELM-W2 0.9877±0.0217 0.9907±0.0127 0.9875±0.0171 29 0.75 1.00 eG-mean 米F-measure-micro 0.95 -F-measure-macro 0.90 0.85 0.80 0.60 0.75 45 Model Model 0.88 0.98 0.86 0.96 0.84 uonen je 0.94 0.82 0.92 0.78 0.90 2 3 4 567 3 Model Model 0.75 1.00 0.95 0.90 1 米 0.85 0.55 0.80 7 2 3 6 7 Model Model The model names represented by the numbers are as follows:1-KELM;2-WKELM-W1;3-WKELM-W2;4-1-MKELM; 5-/-MKELM;6-TRMKELM;7-/-RWMKEELM-WI;8-/-RWMKEELM-W2 图3各模型在各个UCI数据集上诊断性能比较 Fig.3 Comparison of the diagnostic performance of various models on various UCI data sets表 3 (续) Table 3 (Continued) Datasets Model G-mean F-measure-micro F-measure-macro Parameter，C Thyroid KELM 0.841±0.0861 0.9147±0.0433 0.8757±0.066 2 (7) WKELM-W1 0.8765±0.0759 0.9302±0.0347 0.8962±0.0524 2 (6) WKELM-W2 0.8753±0.0662 0.9349±0.0264 0.9±0.0428 2 (7) l1 -MKELM 0.92±0.0712 0.9651±0.0353 0.9468±0.0507 2 (3) lp -MKELM 0.9051±0.0877 0.9535±0.0329 0.9308±0.051 2 (3) ITDSMM-KELM 0.9139±0.0796 0.9574±0.0373 0.9374±0.0559 2 (6) lp -RWMKE-ELM-W1 0.9231±0.0471 0.9628±0.0127 0.9438±0.0221 2 (2) lp -RWMKE-ELM-W2 0.9877±0.0217 0.9907±0.0127 0.9875±0.0171 2 (4) The model names represented by the numbers are as follows: 1—KELM; 2—WKELM-W1; 3—WKELM-W2; 4−l1 -MKELM; 5—lp -MKELM; 6−TRMKELM; 7—lp -RWMKEELM-W1; 8—lp -RWMKEELM-W2 1 2 3 4 5 6 7 8 Model 1 2 3 4 5 6 7 8 1 Model 1 2 3 4 5 6 7 8 Model 1 2 3 4 5 6 7 8 Model 2 3 4 5 6 7 8 Model 1 2 3 4 5 6 7 8 Model 0.60 0.65 0.70 0.75 Diabetes set evaluation index G-mean F-measure-micro F-measure-macro 0.75 0.80 0.85 0.90 0.95 1.00 Ionosphere set evaluation index 0.78 0.80 0.82 0.84 0.86 0.88 Vowel set evaluation index 0.90 0.92 0.94 0.96 0.98 Cancer set evaluation index 0.55 0.60 0.65 0.70 0.75 Bupa set evaluation index 0.80 0.85 0.90 0.95 1.00 Thyroid set evaluation index 图 3    各模型在各个 UCI 数据集上诊断性能比较 Fig.3    Comparison of the diagnostic performance of various models on various UCI data sets · 90 · 工程科学学报，第 44 卷，第 1 期