Neural Network Based Quantitative Structural Property Relations (QSPRs) for Predicting Boiling Points of Aliphatic Hydrocarbons

Abstract

Quantitative structural property relations (QSPRs) for boiling points of aliphatic hydrocarbons were derived using a back-propagation neural network and a modified Fuzzy ARTMAP architecture. With the backpropagation model, the selected molecular descriptors were capable of distinguishing between diastereomers. The QSPRs were obtained from four valance molecular connectivity indices ( 1 v , 2 v , 3 v , 4 v ), a second-order Kappa shape index ( 2 κ), dipole moment, and molecular weight. The inclusion of dipole moment proved to be particularly useful for distinguishing between cis and trans isomers. A back-propagation 7-4-1 architecture predicted boiling points for the test, validation, and overall data sets of alkanes with average absolute errors of 0.37% (1.65 K), 0.42% (1.73 K), and 0.37% (1.54 K), respectively. The error for the test and overall data sets decreased to 0.19% (0.81 K) and 0.31% (1.30 K), respectively, using the modified Fuzzy ARTMAP network. A back-propagation alkene model, with a 7-10-1 architecture, yielded predictions with average absolute errors for the test, validation, and overall data sets of 1.96% (6.79 K), 1.83% (6.45 K), and 1.25% (4.42 K), respectively. Fuzzy ARTMAP reduced the errors for the test and overall data sets to 0.19% (0.73 K) and 0.25% (0.95 K), respectively. The back-propagation composite model for aliphatic hydrocarbons, with a 7-9-1 architecture, yielded boiling points with average absolute errors for the test, validation, and overall set of 1.74% (6.09 K), 1.25% (4.68 K), and 1.37% (4.85 K), respectively. The error for the test and overall data sets using the Fuzzy ARTMAP composite model decreased to 0.84% (1.15 K) and 0.35% (1.35 K), respectively. Performance of the QSPRs, developed from a simple set of molecular descriptors, displayed accuracy well within the range of expected experimental errors and of better accuracy than other regression analysis and neural network-based boiling points QSPRs previously reported in the literature.

A Fuzzy ARTMAP-Based Quantitative Structure−Property Relationship (QSPR) for the Henry's Law Constant of Organic Compounds

“…The advantage of NNs, over classical regression analysis methods, is their inherent ability to incorporate nonlinear relationships between chemical structural parameters and physicochemical properties. [16][17][18][19][20][21] Neural network/QSPR models for estimating the Henry's Law constant for a data set of 357 organic compounds (-7.08 e logH e 2.32) have been recently reported by English and Carroll (2001). 16 The above authors reported QSPRs based on 12-4-1 and 10-3-1 backpropagation neural network architectures (trained using 303 compounds) that performed with absolute errors of 0.237 and 0.281 logH units for the test set (54 compounds), respectively.…”

“…Recently, neural network-based QSPR models that are based on the cognitive classifier fuzzy ARTMAP have been proposed for the estimation of boiling temperature, 18,19 critical properties, 19 aqueous solubility, 20 and octanol-water partition coefficients 21 of organics. The approach was shown to be superior to the popular back-propagation neural network approach as well as other statistical QSPR correlations reported in the literature.…”

Artificial Neural Network in Drug Delivery and Pharmaceutical Research

“…Several research groups have modeled the normal boiling point of hydrocarbons. Predictive neural network models have been published for alkanes [86,87], alkenes [88] and for diverse hydrocarbons [89]. As expected, the models typically show good fitting and prediction statistics with less than ten simple descriptors.…”

Estimation of infinite dilution activity coefficients of organic compounds in water with neural classifiers

“…However, it is not suitable when pattern recognition or feature extraction capabilities are desired because relationships between variables in such networks are embedded within the weights in a distributed form (Bishop, 1995;Hecht-Nielsen, 1995;Hertz et al, 1991). In difficult problems involving pattern recognition, such as those found in the development of QSPRs for data sets of heterogeneous compound classes, it is advantageous to use neural network classifiers, as shown in a number of recent studies (Espinosa et al, 2000(Espinosa et al, , 2001b(Espinosa et al, , 2002Yaffe et al, 2001Yaffe et al, , 2003 on QSPR development.…”

“…Backpropagation neural networks have recently emerged as an alternative for the development of QSPRs and quantitative structure-activity relationships (QSARs) to predict physicochemical properties and biological activities, respectively (Bünz et al, 1998;Chow et al, 1995;Egolf and Jurs, 1993;Espinosa et al, 2000Espinosa et al, , 2001aGakh et al, 1994;Hall and Story, 1996;Mitchell and Jurs, 1998;Simamoea et al, 1993;Stanton and Jurs, 1990;Stanton et al, 1991;Viswanadhan et al, 2001;Yaffe et al, 2001Yaffe et al, , 2003. This alternative modeling strategy for QSPR development yields significantly higher prediction accuracy compared to that of traditional regressionbased correlations.…”