Prediction of GC Retention Indexes for Insect‐Produced Methyl‐Substituted Alkanes Using a Wavelet Neural Network
Morteza Atabati, Kobra Zarei · Journal of the Chinese Chemical Society · 2008
Abstract A quantitative structure‐property relationship (QSPR) study based on the wavelet neural network (WNN) technique was performed for the prediction of gas chromatography retention indexes of methyl‐substituted alkanes produced by insects. In addition to the simple structural descriptors, semi‐empirical quantum chemical calculations at the AM1 (Austin Model 1) level were used to find the optimum 3D geometry of the studied molecules and a numbers of descriptors were calculated with HyperChem and Dragon software. A stepwise MLR (Multiple Linear Regression) method was used to select the best descriptors, and the selected descriptors were used as input neurons in a wavelet neural network model. The average relative error was 2.2%.