基于自适应环境因子熵权决策的多目标特征选择
|
|
李涛, 李佳霖, 阮宁, 徐久成
|
Adaptive environmental factor entropy weight decision⁃making⁃based multi⁃objective feature selection
|
|
Tao Li, Jialin Li, Ning Ruan, Jiucheng Xu
|
|
表3 AMFS与五种多目标进化算法的性能比较
|
Table 3 Performance of AMFS and other five multi?objective evolutionary algorithms
|
|
| 数据集 | 性能指标 | MOFSBDE | FSMOPSO | NSGAFS | DCDREA | NSGA⁃Ⅱ | AMFS |
|---|
| Vehicle | Best | 97.81% | 95.31% | 98.21% | 97.82% | 81.81% | 98.99% | | Avg | 93.48% | 91.22% | 97.11% | 95.99% | 82.82% | 97.66% | | Wine | Best | 89.32% | 90.89% | 97.92% | 96.21% | 84.42% | 97.78% | | Avg | 86.56% | 87.01% | 94.24% | 93.37% | 83.32% | 96.63% | | Sonar | Best | 83.29% | 84.10% | 91.59% | 89.23% | 77.32% | 93.31% | | Avg | 79.99% | 80.11% | 87.31% | 88.42% | 77.39% | 91.99% | | Ionosphere | Best | 94.59% | 100% | 89.95% | 100% | 65.12% | 99.82% | | Avg | 91.70% | 97.89% | 86.67% | 98.78% | 64.31% | 98.99% | | Vowel | Best | 97.57% | 85.68% | 75.79% | 97.77% | 66.83% | 99.13% | | Avg | 96.62% | 84.70% | 73.99% | 96.10% | 57.83% | 97.75% | | Segmentation | Best | 97.73% | 97.22% | 98.91% | 100% | 89.80% | 100% | | Avg | 96.86% | 96.61% | 97.45% | 97.65% | 87.79% | 98.89% | | AD_Best | 93.82% | 92.68% | 93.97% | 97.91% | 77.55% | 98.22% | | AD_Avg | 91.35% | 90.55% | 91.41% | 95.58% | 75.57% | 97.11% | | T⁃test_Best | 0.02639 （+） | 0.03083 （+） | 0.16461 （-） | 0.08581 （-） | 0.00444 （+） | | | T⁃test_Avg | 0.01919 （+） | 0.01489 （+） | 0.09646 （-） | 0.01350 （+） | 0.00851 （+） | |
|
|
|