Document Type : Research Paper
Quantitative Structure-Property Relationship (QSPR) models for modeling and predicting thermodynamic properties such as the enthalpy of vaporization at standard condition (ΔH˚vap kJ mol-1) and normal temperature of boiling points (T˚bp K) of 57 mono and Polycyclic Aromatic Hydrocarbons (PAHs) have been investigated. The PAHs were randomly separated into 2 groups: training and test sets. A set of molecular descriptors was calculated for selected compounds using the Dragon software. The Genetic Algorithm (GA) method and backward stepwise regression were used to select the suitable descriptors. Multiple Linear Regression (MLR) technique was used to obtain a linear relationship between descriptors and chemical properties. The predictive ability of the GA-MLR models was implemented using squared cross-validation and external validation methods. The aforementioned results and discussion lead us to conclude that the training set models established by GA-MLR method have good correlation of thermodynamic properties, which means QSPR models could be efficiently used for estimating and predicting of the above mentioned properties of the mono and PAHs.