Maziar Soleymani Ardejani - Principal component analysis-ranking as method for the simulation of 13 C nuclear a variable selection ma netic resonance spectra of xanthones using artificial neural networks

Document created by Maziar Soleymani Ardejani on Aug 22, 2014
Version 1Show Document
  • View in full screen mode

  Publication Details (including relevant citation   information):

  Source: Qsar & Combinatorial Science Volume: 26 Issue: 6   Pages: 764-772 Published: JUN 2007

  DOI: 10.1002/qsar.200630111


  A Quantitative Structure-Property Relationship (QSPR) relating   atom-based calculated descriptors to 13C NMR chemical   shifts was developed to accurately simulate 13C NMR   spectra of polyhydroxy and methoxy substituted dibenzo pyrons   (xanthones). A dataset consisting of 35 xanthones was employed   for the present analysis. A set of 132 topological, geometrical,   and electronic descriptors representing various structural   characteristics was calculated for each of 497 unique carbon   atoms in the dataset. Principal Component Analysis (PCA)-ranking   and Artificial Neural Networks (ANNs) were used as descriptor   selection and model building methods, respectively. Analyses of   the results revealed a correlation coefficient and Root Mean   Square Error (RMSE) of 0.998 and 1.42 ppm, respectively, for the   prediction set.

  Address (URL):