面向概念漂移数据流的在线集成自适应算法
|
|
崔瑞华, 綦小龙, 刘艳芳, 林玲
|
Online ensemble adaptive algorithm for concept drift of streaming data
|
|
Ruihua Cui, Xiaolong Qi, Yanfang Liu, Ling Lin
|
|
表4 各对比算法在合成数据集上的实验结果对比
|
Table 4 Experimental results of different algorithms on synthetic datasets
|
|
| | ARF⁃HDDM_A | SRP⁃HDDM_W | ARF⁃KSWIN | SRP⁃KSWIN | KSHPR |
|---|
| SEA stream | Accuracy | 88.37% | 86.93% | 88.41% | 87.62% | 89.28% | | Precision | 88.33% | 88.03% | 85.37% | 87.23% | 89.24% | | Recall | 87.77% | 87.47% | 84.85% | 86.77% | 92.49% | | F1⁃Score | 88.02% | 87.71% | 85.05% | 87.02% | 90.83% | | SEA Big | Accuracy | 88.50% | 88.59% | 88.56% | 88.73% | 89.51% | | Precision | 88.58% | 88.32% | 88.43% | 87.44% | 89.43% | | Recall | 88.03% | 87.89% | 87.88% | 89.47% | 92.69% | | F1⁃Score | 88.28% | 88.12% | 88.09% | 87.15% | 91.03% | | Sin incremental | Accuracy | 97.59% | 96.18% | 97.31% | 96.67% | 98.52% | | Precision | 97.43% | 97.01% | 95.49% | 98.11% | 98.79% | | Recall | 96.91% | 96.19% | 93.98% | 97.98% | 99.29% | | F1⁃Score | 97.08% | 96.35% | 94.75% | 97.96% | 99.03% | | Rotating Hyperplane | Accuracy | 78.27% | 78.68% | 79.00% | 76.00% | 83.75% | | Precision | 78.37% | 78.56% | 78.53% | 75.32% | 83.39% | | Recall | 78.97% | 78.97% | 78.66% | 75.21% | 83.66% | | F1⁃Score | 78.93% | 78.48% | 78.58% | 75.27% | 83.52% | | Mixed paper | Accuracy | 98.54% | 96.70% | 97.91% | 96.73% | 99.11% | | Precision | 97.76% | 97.27% | 96.52% | 96.53% | 98.48% | | Recall | 97.91% | 88.76% | 85.84% | 96.32% | 99.32% | | F1⁃Score | 97.19% | 97.19% | 96.45% | 96.48% | 98.89% |
|
|
|