Nondestructive Detection of Authenticity of Thai Jasmine Rice Using Multispectral Imaging
Table 3
Performance of PLS, LS-SVM, and BPNN models for predicting adulteration proportion in Thai jasmine rice (%).
Variety
Chemometrics
RMSEC (%)
Bias
RMSEP (%)
RPD
Jasmine sticky rice
PLS
0.946
3.741
0.873
0. 792
4.679
4.753
BPNN
0.982
2.893
0.923
0. 681
3.827
6.826
LS-SVM
0.990
2.164
0.935
0.382
2.582
9.674
Simiao rice
PLS
0.983
6.259
0.984
0.730
5.780
5.567
BPNN
0.988
4.731
0.984
0.971
5.842
5.477
LS-SVM
0.991
3.961
0.983
ā0.465
3.569
8.791
Northeast Wuchang rice
PLS
0.981
2.924
0.933
2.04
1.10
9.033
BPNN
0.981
1.906
0.980
3.61
1.813
6.528
LS-SVM
0.995
1.398
0.988
2.10
1.603
8.482
, coefficient of determination in calibration; , coefficient of determination in prediction; RMSEC, root mean square error of calibration; RMSEP, root mean square error of prediction; RPD, residual predictive deviation.
