Prediction of thermal conductivity detection response factors using an artificial neural network

The main aim of the present work was the development of a quantitative stru cture-activity relationship method using an artificial neural network (ANN) for predicting the thermal conductivity detector response factor. As a fir st step a multiple linear regression (MLR) model was developed and the desc riptors appearing in this model were considered as inputs for the ANN. The descriptors of molecular mass, number of vibrational modes of the molecule, molecular surface area and Balaban index appeared in the MLR model. In agr eement with the molecular diameter approach, molecular mass and molecular s urface area play a major role in estimating the thermal conductivity detect or response factor (TCD-RF). A 4-7-1 neural network was generated for the p rediction of the TCD-RFs of a collection of 110 organic compounds including hydrocarbons, benzene derivatives, esters, alcohols, aldehydes, ketones an d heterocyclics. The mean absolute error between the ANN calculated and the experimental values of the response factors was 0.02 for the prediction se t. (C) 2000 Elsevier Science B.V. All rights reserved.