QSPR modeling of migration factor in micellar electrokinetic chromatography

Quantitative structure property relationships (QSPR), as one of the most important areas in chemometrics. In this study QSPR as mathematical model was used to investigated the relation between the structural parameters and their properties. Multiple linear regression (MLR) was used for the prediction of migration factor of some compounds in micellar electrokinetic chromatography[1]. Firstly, a collection of compounds were selected as a data set and then they were randomly divided into two groups, training and prediction sets. Important descriptors from calculated descriptors were chosen by stepwise MLR and the modeling was performed using multiple linear regression technique. We constructed four models by different percent of surfactant. The statistical results of whole of them are Q2=0.921, RMSE=0.198 and R=0.960