A quantitative structure–retention relationship model is built up by CORAl software based on representation of the molecular structure by simplified molecular input line entry system and graphical representation. This model was used for the prediction of linear retention indices of 96 volatiles in cooking rice. The hybrid version of the molecular structure representation by combination of SMIlES and the molecular graph provided the best correlation of the prediction for the considered retention indices. The reliability of established model was evaluated using the leave-many-out cross-validation method (Q = 0.932) as well as by y scrambling that reveals the suitability of the developed model. The results of the present study show that CORAl software can be used to accurately predict of linear retention indices of chemicals in gas chromatography. 2
