These tools will no longer be maintained as of December 31, 2024. Archived website can be found here. PubMed4Hh GitHub repository can be found here. Contact NLM Customer Service if you have questions.


Pubmed for Handhelds

PUBMED FOR HANDHELDS

Search MEDLINE/PubMed

  • Title: Quantitative structure-property relationship modelling of the degradability rate constant of alkenes by OH radicals in atmosphere.
    Author: Fatemi MH, Baher E.
    Journal: SAR QSAR Environ Res; 2009; 20(1-2):77-90. PubMed ID: 19343584.
    Abstract:
    In this work, the degradability rate constants of 98 alkenes by OH radicals were predicted from theoretically derived descriptors, which were calculated from the molecular structure alone by applying a quantitative structure-property relationship (QSPR) approach. For the selection of the most relevant descriptors, stepwise multiple linear regression (MLR) and genetic algorithms (GAs) were used. Then some linear and nonlinear techniques were used for the investigation of the relation between selected molecular descriptors and the OH radical degradability rate constant. These methods were MLR, artificial neural networks (ANNs) and support vector machines (SVMs). According to the variable selection method and feature mapping techniques, six QSPR models were constructed which were: stepwise-MLR-MLR, stepwise-MLR-ANN, stepwise-MLR-SVM, GA-MLR, GA-ANN, and GA-SVM. The credibility of these models was evaluated by a leave-24-out cross-validation test. The statistical results are Q(2) = 0.86, SPRESS = 0.16 for GA-ANN, Q(2) = 0.69, SPRESS = 0.20 for GA-SVM, and Q(2) = 0.83, SPRESS = 0.18 for GA-MLR model. Based on these values and other statistical parameters obtained in this work, it was concluded that the GA-ANN model outperformed the other models.
    [Abstract] [Full Text] [Related] [New Search]