Study On Identification Of Protein Adulteration And A Variety Of Mixed Adulterations Of Goat Milk By Electronic Nose

In recent years,with the people's attention to nutrition and health,people are more likely to choose goat milk instead of cow milk in dairy products in China.However,some unscrupulous businessman adulterate goat milk with impurity substance for profiteering,especially in order to increase the nitrogen content in dairy products,they adulterate goat milk with urea or non-milk protein such as soy protein,hydrolyzed vegetable protein,or mixed with cheap starch or cream or fake milk,this behavior is bad for the development of dairy industry and is not conducive to the health of consumers.This paper used electronic nose to identify protein adulterations and a variety of mixed adulterations in goat milk.Using principal component analysis and linear discriminant analysis for qualitative analysis,and using linear regression analysis for quantitative analysis,and using Fisher discriminant analysis,K nearest neighbor and multilayer perceptron to establish prediction model.The conclusions are as follows:Through the response of electronic nose to protein adulterations and a variety of mixed adulterations in goat milk could be seen that the response of each kind of adulterations was different,and the response of the corresponding different adulteration concentration also had differences,therefore,it could be concluded that using electronic nose can identify different adulterations in goat milk.Detection of different concentration of protein adulterations in goats' milk,qualitative analysis results showed that the sum of LD1 and LD2 variance contribution rate of different concentration of protein adulteration samples was 94.5%,94.9%,96.7%,97.6% and 93.4%.There was no overlap between the different concentrations,indicating that the electronic nose can be used for the qualitative analysis of different concentration of protein adulterations.The quantitative analysis results showed that regression model coefficients of the different concentration of protein adulterations were 88.2%,95.9%,97.7%,96.9% and 93.9%,the regression fitting degree was relatively high,indicating that electronic nose can be used for quantitative analysis of different concentrations of protein adulterations.Detection of goat milk and protein adulterations,qualitative analysis results showed that there was no overlap between adulteration samples and control samples and each sample group was concentrated,indicating that electronic nose can be used for qualitative analysis of protein adulterations;quantitative analysis results showed that the decision coefficient of regression model which was established by linear regression analysis was 84.5%,showing the higher fitting degree of the regression equation,indicating that electronic nose can be used for quantitative analysis of protein adulterations;correct rates of the classification models established by fisher linear discriminant analysis,K-nearest neighbor and multilayer perceptron were 98.2%,97.1% and 94.4%,showing good prediction results,indicating that electronic nose can be used for predicting all kinds of protein adulterations samples accurately.The goat milk and a variety of mixed adulterations were detected,qualitative analysis results showed that the adulterated samples group and control group have specific regions,and each region had no overlap,indicating that electronic nose can identify goat milk and mixed adulterations;quantitative analysis results showed that the regression determination coefficient of the linear regression analysis was 83.2%,showing that the higher degree of the regression equation,indicating that to a certain extent electronic nose can be used to predict the mixed adulterations;correct rates of the classification models established by fisher linear discriminant analysis,K-nearest neighbor and multilayer perceptron were 100%,88.8% and 95.8%,showing good prediction results,indicating that electronic nose can predict the mixed adulterations samples accurately.