基于全局与局部误差融合的集成腐蚀速率预测模型：一种稳健优化策略

郑文培, 刘迎正, 张娅茹, 周涛涛, 李兴涛, 禹胜阳, 蒋璐朦

An Integrated Corrosion Rate Prediction Model Based on Global and Local Error Fusion: A Robust Optimization Strategy

ZHENG Wenpei, LIU Yingzheng, ZHANG Yaru, ZHOU Taotao, LI Xingtao, YU Shengyang, JIANG Lumeng

表2 集成模型预测的比较

Table 2 Comparison of integrated model predictions

Sample MEAE Kriging FLA-BP ELM α = 0.2 α = 0.4 α = 0.5 α = 0.6 α = 0.8 Average 1.0012 0.8812 1.2143 0.7709 0.7703 0.7607 0.7741 0.9021 1 Variance, × 10-3 34.372 4.4142 6.9803 2.9341 2.8674 2.6806 2.9877 3.0785 Average 1.1075 0.7706 2.2398 0.7821 0.7510 0.7442 0.8186 0.8724 2 Variance, × 10-3 15.987 3.1354 10.082 3.2458 3.0584 3.2463 4.2351 4.5526 Average 0.9653 0.7533 1.0436 0.7709 0.7703 0.7607 0.7741 0.9021 3 Variance, × 10-3 1.0359 0.8635 3.6522 4.2536 4.0863 3.9655 3.9763 4.3691 Average 0.0203 0.0031 0.3136 0.0029 0.0028 0.0025 0.0065 0.0098 4 Variance, × 10-3 2.6354 0.6952 1.3655 0.5563 0.5536 0.4695 0.6353 0.7466 Average 0.0012 0.0005 0.0102 0.0009 0.0009 0.0007 0.0008 0.0010 5 Variance, × 10-3 1.3653 0.6654 1.9633 0.7633 0.7464 0.6667 0.9616 1.0153