Quantitative Structure–Property Relationship Studies for Predicting Flash Points of Organic Compounds using Support Vector Machines

Abstract

A Quantitative Structure -Property Relationship (QSPR) model was developed to predict the flash points of organic compounds. The widely used group contribution method was employed, and a new collection of 57 functional groups were selected as the molecular descriptors. The new chemometrics method of Support Vector Machine (SVM) was employed for fitting the possible quantitative relationship that existed between these functional groups and flash points. A total of 1282 organic compounds of various chemical families were used and randomly divided into a training set (1026) and an external prediction set (256).The optimum parameters of the SVM were obtained by employing the leave-one-out cross-validation method. Simulated with the final optimum SVM, the results show that most of the predicted flash point values are in good agreement with the experimental data, with the average absolute error being 6.894 K, and the root mean square error being 11.367 for the whole dataset, which are lower than those obtained by previous works. Moreover, by employing the convenient group contribution method as well as the large modeling dataset, the presented model is also expected to be simple to apply and with a wide applicability range.

Quantitative structure–property relationship studies for predicting flash points of alkanes using group bond contribution method with back-propagation neural network

“…The optimal number of neurons in the hidden layer was determined by varying the number of hidden neurons and observing the root mean square error (RMS) [8], which was used as a measure of the prediction error of the trained model and was calculated with the following equation:…”

“…In recent years, the modeling technique of artificial neural network (ANN) has been widely used in the field of QSPR [4,8,13,14]. ANN is a powerful tool for correlating and estimating chemical properties and one of a group of intelligence technologies for data analysis that differ from other classical analysis techniques.…”

“…Quantitative structure-property relationships (QSPR) method which relates descriptors of the molecular structure to the properties of chemical compounds, has been reported quite extensively in the literature for the prediction of flash point [3][4][5][6][7][8][9][10]. For example, the first method for estimating the flash point of organic compounds from their molecular structure was developed by Suzuki et al The 25 atomic and group contributions were employed for predicting the flash points of 33 aliphatic and 26 aromatic hydrocarbons with an average absolute deviation of 12.2 and 6.1 • C, respectively.…”

Prediction of the flash points of alkanes by group bond contribution method using artificial neural networks

“…The optimal number of neurons in the hidden layer was determined by varying the number of hidden neurons and observing the root mean square error (RMS) [14]. The number of neurons that gave the lowest RMS was chosen.…”

A novel QSPR model for prediction of lower flammability limits of organic compounds based on support vector machine

“…Originally, SVM was developed for classification problems, and has demonstrated a good performance in solving these problems by numerous successful applications [17][18][19][20][21][22]. In recent years, with the introduction of ε-insensitive loss function, SVM has also been extended to solve regression problems, and has shown great performance in QSPR studies due to its remarkable ability to interpret the nonlinear relationships between molecular structure and properties [23][24][25][26][27][28]. In the most of these cases, the performance of SVM modeling either matches or is significantly better than that of traditional machine learning approaches [25].…”