Identification of early moldy rice samples by PCA and PNN

Abstract

A method of identifying early moldy rice by principal component analysis (PCA) and probabilistic neural network (PNN) was presented in this paper. In the experiment, eight gas sensors were chosen to compose the electronic nose's array, which was used to gather early different level mildew data of rice samples. These gathered data were reduced dimensions by PCA and then were passed through PNN to identify their categories. The rate of identification was 91.67%. Compared with the method of PNN only used, the identification method of PCA and PNN has higher recognition accuracy and less classification time. Thus the experimental results of this paper showed that the method of PCA and PNN in classifying early different degrees moldy rice was effective. © 2012 Springer-Verlag Berlin Heidelberg.