PREDICTION OF GAS-CHROMATOGRAPHIC RETENTION INDEXES OF SOME BENZENE-DERIVATIVES

Gas chromatographic retention indices for some benzene derivatives on Apiezon MH were successfully modelled with the aid of a computer. Numerical descriptors were calculated and multiple linear regression analysis methods were used to generate model equations relating structural features to Kovats retention indices. These descriptors encode topological, geometric, electronic and calculated physical properties of the molecules. A model with R = 0.998 and SE = 10.067 was generated. Calculations of retention indices for a prediction set show that this model has a good predictive ability.