Clusterization of pure and formalin fresh noodles with electronic nose based on kernel principal component analysis

Abstract

There are some manufacturers of wet noodles that use hazardous substances to extend shelf life. Detection of formalin content in the noodles is still difficult for ordinary people to do. Standard chemical analysis tends to quantitatively analyze the elements or compounds in the sample. Qualitative analysis using Kernel Principal component analysis (K-PCA) can be done by detecting the aroma of wet noodles using electronic nose. Pure wet noodles and wet noodle mixed with 1%, 5%, 10%, 15% and 20% formalin solutions used as samples. With each sample was weighed with a mass of 20 grams and placed in a closed sample room. Using e-nose system with 5 gas sensors (TGS 2610, TGS 2620, TGS 2611, TGS 2602, and TGS 2600) and K-PCA with linear, polynomial, RBF and sigmoid methods, the samples were clustered. The results proved that pure wet noodles and wet noodles mixed by formalin can be clustered using the K-PCA method with a total variance of 98.83%. So that this e-nose can be used as an alternative instrument to detect the purity of wet noodles from the influence of harmful substances such as formalin.