Computational estimation of soybean oil adulteration in Nepalese mustard seed oil based on fatty acid composition

The experiment was carried out for the computational estimation of soybean oil adulteration in the mustard seed oil using chemometric technique based on fatty acid composition. Principal component analysis and K-mean clustering of fatty acid composition data showed 4 major mustard/rapeseed clusters, two of high erucic and two of low erucic mustard type. Soybean and other possible adulterants made a distinct cluster from them. The methodology for estimation of soybean oil adulteration was developed based on weighted least square error principle using Microsoft Excel Solver program. This principle was successfully validated on the real soybean mustard oil blends. Moreover, the method was further evaluated on the 4000 training set and 3999 validation set simulated blends of mustard and soybean oil based on 212 fatty acid composition data. The blending was simulated with data from different literature from different part of the world along with data collected in our laboratory. The simulated blend consisted of random mixture of up to 16 samples. These are more extreme conditions than that can usually be expected in real scenario.

Jaar van publicatie:2011

Toegankelijkheid:Open