摘要
综述了Q SRR中的多元线性回归(M ult ple L inear Reg resio n,M L R)、偏最小二乘回归(Part ial L east Squares Regr ession,P LSR)、人工神经网络(Ar tificial N eural N et wor ks,AN N)几种常用建模方法、分子描述符的选取的要点及原则及相关模型的建立和评价等方法在化学与环境化合物分析中的应用现状,并对QSRR在该领域的研究方向和发展前景进行了展望,以期为今后更深层次的研究作综合性的参考。
The paper reviewed the application status of QSRR in the analysis of chemistryand environmental compound, including the following main elements, several kinds of common modeling method such as multple linear regression (MLR), partial least squares regression (PLSR), artificial neural networks (ANN), points and principles about molecular descriptors selection and methods of correlation model establishment and evaluation. In addition, the development prospect and research direction of QSRR in the fields were also discussed with aview to provide the all-round references to subsequent deep study.
出处
《光谱实验室》
CAS
2013年第5期2287-2293,共7页
Chinese Journal of Spectroscopy Laboratory