Font Size: a A A

The Research Of 3D-QSPR On Ketones

Posted on:2012-07-19Degree:MasterType:Thesis
Country:ChinaCandidate:X F ChengFull Text:PDF
GTID:2211330368493381Subject:Inorganic Chemistry
Abstract/Summary:PDF Full Text Request
QSPR(Quantitative Structure Property Relationship) mainly applied practices combining of theoretical calculation method and various of statistical analysis tools to study quantitative function relation between compound structure and biological, chemical and physical properties of them. Which is a very active field of Chemical Informatics Research. QSPR research has the important theoretical and practical value in chemistry, life sciences and environmental science in all.This paper customizes the new 3D topology index(Yg, Yh and Yt), according to the molecules of 245 ketones three-dimensional configuration atomic space location, using Chem Windows or Chem Office 2008 software collecte atomic and molecular coordinate data obtained data matrix, Again through MATLAB 7.0 programming calculation, through the linear multivariate regression analysis method, the artificial neural network based on BP algorithm such calculations to study and establish 245 kind of ketone molecular 3D topology index and the chemical and physical properties (melting point, boiling point, the critical temperature and critical. strong, etc.) between the linear and nonlinear relevance. This study sets up two kinds of correlation optimal models:one is the obtained MLR, which has the good correlation through 3D topology index (Yj, Yh and Yg), and the number of carbon atoms (Cn) with 245 kind of methadone in nature (BP, MP, CT, CP), and these correlation coefficient R get above 0.8810, achieved good level, this elucidation ketone molecular 3D topology index (Yj, Yh and Yg) reflects the space characteristics and physicochemical properties of parameters have certain correlation. It is by BP ANN method calculated the nature of 245 ketones predicted results and experimental value basic anastomosis, which means that there is the good correlation between the molecular 3D topology index (Yj, Yh and Yg) and compound physicochemical properties. Through the linear and nonlinear method to deal with all get good correlation, this study shows that it will provides new ideas and methods and the custom 3D topology index (Yj, Yh and Yg) can reflect the space structure feature, molecular for 3D space QSPR study. It can undertake organic ketone structures and properties of prediction, provides a new idea for 3D-QSPR/QSAR research. And provides relevant information for organic molecules identification, organic polymer simulation, organic drug synthesis,which has certain theoretical and practical significance.
Keywords/Search Tags:ketone, multivariate linear regression analysis, BP artificial neural network, quantitative structure property relationship
PDF Full Text Request
Related items