Internet Electronic Journal of Molecular Design - IEJMD, ISSN 1538-6414, CODEN IEJMAT

ABSTRACT - Internet Electron. J. Mol. Des. May 2003, Volume 2, Number 5, 315-333

Neural Network Modeling of Refractive Indexes of Phosphorus-Containing Organic Compounds Julian Koziol Internet Electron. J. Mol. Des. 2003, 2, 315-333

Free: Download the paper in PDF format

Return to Table of Contents

Get Acrobat Reader to view and print the paper

Abstract:
One of the most intensively explored areas of contemporary computational chemistry is searching for a comprehensive numerical description of chemical structures and for methods that enable to develop efficient and credible QSPR (quantitative structure-property relationships) models. Among these methods artificial neural networks (ANN) turned out to be a very promising methodology in obtaining models converting structural descriptors into different properties of chemicals. Five different models relating structural descriptors to refractive indexes of phosphorus containing organic compounds have been developed using ANN. A newly elaborated set of molecular descriptors is evaluated to determine their usefulness for QSPR studies. Using a data set containing 180 phosphates and diphosphates, ANN trained with the back propagation and conjugated gradient algorithms are able to predict the refractive index with relatively high accuracy. The results obtained show good predictive ability for the ANN models, giving the average prediction error of 0.24% and R²_cv equal to about 0.99. The QSPR studies described in this paper provide strong evidence that the tested structural descriptors are useful and effective for the ANN modeling of the phosphates refractive index.

Free: Download the paper in PDF format

Return to Table of Contents

Get Acrobat Reader to view and print the paper