119th General Meeting of the KCS

Type Oral Presentation
Area Oral Presentation of Young Analytical Chemists Ⅱ
Room No. 303호
Time FRI 10:24-:
Code ANAL2.O-39
Subject MATLAB-based Statistical Analysis Software for Edible Oils Classification
Authors 손민희, 오한빈*
서강대학교 화학과, Korea
Abstract Edible oils contain a variety of lipids, particularly diacylglycerols (DAGs) and triacylglycerols (TAGs). Different edible oils have their own signature lipid distribution. In this study, we acquired MALDI-TOF (Tinkerbell, Asta, Korea) mass spectra of 9 species of edible oils, such as sesame, perilla, olive, canola, grape-seed, sunflower-seed, corn, soybean, coconut oils. For the classification of edible oils using statistical analysis, we used principle component analysis (PCA), partial least square-determinant analysis (PLS-DA) and support vector machine (SVM). Using these statistical methods, statistical models were established. Based on those statistical analysis results, an edible oil analysis software was constructed using MATLAB. This software can not only visualize the acquired MALDI mass spectrum, but also readily classify unknown edile oils based on the PCA/PLS-DA/SVM statistical models by projecting the acquired spectrum components onto the established statistical models for the classification of edile oils.
E-mail mh122192@gmail.com