Prediction of programmed-temperature retention values of naphthas by artificial neural networks

It is proposed for the first time a method of prediction of the programmed- temperature retention times of components of naphthas in capillary gas chro matography using artificial neural networks. People are used to predict the programmed-temperature retention rime using many formulas such as the inte gral formula, which requires that four parameters must be determined by cal culation or experiments. However the results obtained by the formula are no t so good to meet the demand of industry. In order to predict retention tim e accurately and conveniently, artificial neural networks using five-fold c ross-validation and leave-20%-out methods have been applied. Only two param eters: density and isothermal retention index were used as input vectors. T he average RMS error for predicted values of five different networks was 0. 18, whereas the RMS error of predictions by the integral formula was 0.69. Obviously, the predictions by neural networks: were much better than predic tions by the formula, and neural networks need fewer parameters than the fo rmula. So neural networks can successfully and conveniently solve the probl em of predictions of programmed-temperature retention times, and provide us eful data for analysis of naphthas in petrochemical industry.