Research On Adulteration Detection Of Rattan Pepper Oil Based On Near-Infrared Spectroscopy

Compared with other edible vegetable oils,vine pepper oil has good nutritional value,but its production capacity is insufficient and the market is chaotic.Its adulteration detection research can protect the rights and interests of consumers and maintain the normal operation of the market.In this thesis,a qualitative and quantitative research and analysis on the adulteration of vine pepper oil was carried out by using near-infrared spectroscopy(NISR)combined with support vector machine(SVM)method.The main research contents are as follows:The two element system:Spectral data processing adopts multivariate scatter correction(MSC),standard normal variable transformation(SNV),standard normal variable transformation combined with detrending technology(SNV-DT)preprocessing algorithm,competitive adaptive weighted sampling method(CARS),continuous The projection algorithm(SPA)feature optimization algorithm,through the whale algorithm(WOA)optimization,establishes a binary adulteration qualitative and quantitative model of vine pepper oil.The experimental results show that the SNV-DT-CARS-WOA-SVC model has the best prediction performance in the qualitative identification of the authenticity of the binary adulteration of vine pepper oil,and the prediction set accuracy is as high as 94.66%.Among the three binary quantitative prediction models for adulteration content of rattan pepper oil mixed with peanut oil,soybean oil and sunflower oil,model SNV-DT-CARS-WOA-SVR,model SNV-DT-CARS-WOA-SVR and model SNV-CARS-WOA-SVR have the best prediction effect,and their prediction set coefficients of determination(R_p~2)are0.9542,0.9167,0.9634,their root square error(RMSE)are 0.0607,0.0802,0.0533,respectively.The three element system:Spectral data processing adopts MSC,SNV-DT preprocessing algorithm,CARS,SPA feature optimization algorithm,through WOA algorithm,using improved whale algorithm(BAS＿WOA)to optimize parameters,establish a qualitative model of ternary adulteration authenticity of vine pepper oil,experiment The results show that the SNV-DT-CARS-BAS＿WOA-SVC model has the best prediction effect,and the accuracy of the calibration set and prediction set is as high as 100%and 96.42%,respectively.The ternary spectral data processing adopts MSC,SNV,SNV-DT preprocessing algorithm,CARS,SPA feature optimization algorithm,and parameter optimization through WOA and BAS＿WOA algorithm,and establishes a ternary adulteration quantitative prediction model of vine pepper oil,among the SNV-DT-CARS-BAS＿WOA-SVR model has the best quantitative prediction effect,R~2 and the RMSE of the calibration set and prediction set are 0.9996,0.9383 and0.0029,0.0359,respectively.The results show that:in the study of binary and ternary adulteration detection of vine pepper oil,the qualitative identification of the authenticity of the adulteration of vine pepper oil and the quantitative prediction of the adulteration content can be well realized.