离差最大化结合BP神经网络评价烟叶化学品质
张勇刚, 任志广, 徐志强, 刘建国, 张晓兵, 刘化冰, 夏琛, 程昌合

Chemical Quality Evaluation of Flue-Cured Tobacco Based on Maximization of Deviation and BP Neural Network
Zhang Yonggang, Ren Zhiguang, Xu Zhiqiang, Liu Jianguo, Zhang Xiaobing, Liu Huabing, Xia Chen, Cheng Changhe
表6 预测集数据的预测误差
Table 6 Prediction error of testing samples
样品编号
Sample
number
实测值
Measured
value
预测值
Predicted
value
绝对误差
Absolute
error
相对误差
Relative
error (%)
1 0.5397 0.5378 0.0019 0.3520
2 0.4771 0.4840 0.0069 1.4462
3 0.4721 0.4702 0.0019 0.4025
4 0.6333 0.6239 0.0094 1.4843
5 0.6463 0.6489 0.0026 0.4023
6 0.7056 0.6916 0.0140 1.9841
7 0.3329 0.3373 0.0044 1.3217
8 0.6001 0.6035 0.0034 0.5666
9 0.5013 0.5063 0.0050 0.9974
10 0.4387 0.4396 0.0009 0.2052
11 0.6129 0.6240 0.0111 1.8111
12 0.6345 0.6374 0.0029 0.4571
13 0.4571 0.4582 0.0011 0.2406
14 0.6387 0.6322 0.0065 1.0177
15 0.6864 0.6920 0.0056 0.8159
16 0.6968 0.7008 0.0040 0.5741
17 0.5305 0.5355 0.0050 0.9425
18 0.4662 0.4610 0.0052 1.1154
19 0.6168 0.6148 0.0020 0.3243
20 0.3110 0.3196 0.0086 2.7653
21 0.6696 0.6689 0.0007 0.1045
22 0.3652 0.3641 0.0011 0.3012
23 0.8294 0.8305 0.0011 0.1326
24 0.6416 0.6459 0.0043 0.6702