Exploration of Linear and Nonlinear Modeling Techniques to Predict of Retention Index of Essential Oils

Abstract

AbstractThe essential oils are widely used in pharmaceutical, cosmetic and perfume industry, and for flavouring and preservation of several food products. GC and GC‐MS is one of the most powerful tools in analytical volatile compound (such as essential oils). Genetic algorithm and multiple linear regression (GA‐MLR), partial least square (GA‐PLS), kernel PLS (GA‐KPLS) and Levenberg‐Marquardt artificial neural network (L‐M ANN) techniques were used to investigate the correlation between retention index (RI) and descriptors for 113 diverse compounds in essential oils of four Teucrium species which obtained by GC and GC‐MS. Five simple one‐ and two‐dimensional descriptors were selected by GA‐KPLS and considered as input for developing L‐M ANN. The applied internal (leave‐group‐out cross validation (LGO‐CV)) and external (test set) validation methods were used for the predictive power of four models. The correlation coefficient LGO‐CV (Q2) between experimental and predicted RI for training and test sets by GA‐MLR, GA‐PLS, GA‐KPLS and L‐M ANN was 0.91, 0.92, 0.96 and 0.99 (for 88 compounds), 0.88, 0.91, 0.94 and 0.97 (for 25 compounds), respectively. This indicates that L‐M ANN can be used as an alternative modeling tool for quantitative structure‐property/retention relationship (QSPR/QSRR) studies. This is the first research on the QSRR of the essential oil compounds against the RI using the GA‐KPLS and L‐M ANN.