作者: Youcef Saihi , Khairedine Kraim , Fouad Ferkous , Zeineddine Djeghaba , Abdelkader Azzouzi
DOI: 10.5012/BKCS.2013.34.6.1643
关键词: Curcuminoid 、 Xanthone 、 Neural network architecture 、 Artificial neural network 、 Quantitative structure–activity relationship 、 Stereochemistry 、 α glucosidase 、 Chemistry
摘要: Department of Chemistry, Faculty Sciences, University Djelfa, 17000, AlgeriaReceived November 26, 2012, Accepted March 6, 2013A non linear QSAR model was constructed on a series 57 xanthone and curcuminoide derivatives as α-glucosidase inhibitors by back-propagation neural network method. The architecture wasoptimized to obtain three-layer network, composed five descriptors, nine hidden neurons oneoutput neuron. A good predictive determination coefficient obtained (R