Conclusion
The aim of this study is to develop and implement an automated
method for detection and quantification of adulteration
of sesame oil in routine food analysis laboratories. In this paper,
the PCA and SOM are used for screening of the data. It is shown
that different types of mixture of oils are clustered in different
regions based on their intrinsic correlation. Then, the detection
model is designed as a hierarchical approach using the SVM algorithm
and the PLS algorithm. For the prediction of level of sesame
oil in mixture, this hierarchical approach firstly attempts to identify
the type of adulterant using SVM classification. The results
show that the authentic sesame oil can be correctly discriminated
from the adulterated ones, and the adulterated ones also can be
classified with correct rate 100% according to the type of adulterant.
Then, several PLS-based models are built for the mixtures
with specific adulterant. The RMSEs for prediction of sesame oil
range from 1.19% to 4.29%, which meet the requirements of
routine food control.