摘要
运用原子三角法表征 1 -[(2 -hydroxyethoxy) methyl]-6-(phenylthio) thymine(HEPT)类化合物的三维结构信息 .由原子三角的频数计算得到化合物的相似度矩阵 ,并在此基础上衍生出新的变量 .运用多元回归分析和人工神经网络分别构造了预测数学模型 。
The atom triples of some 1 [(2 hydroxyethoxy)methyl] 6 (phenylthio) thymine(HEPT) derivatives containing abundant three dimensional information of molecules were calculated. Molecular similarity matrix was built based on the frequency distributions of the atom triples. These similarities being taken as the new variables were selected by using Leaps and Bounds regression analysis. Multiple regression analysis and neural networks were performed for building the prediction models, and the satisfactory results were obtained by using the neural networks.
出处
《应用化学》
CAS
CSCD
北大核心
2002年第5期420-424,共5页
Chinese Journal of Applied Chemistry
基金
国家自然科学基金资助项目 (2 0 0 770 2 6)
关键词
定量结构-活性相关性
原子三角法
分子相似度
回归分析
人工神经网络
quantitative structure activity relationship,atom triple method,molecular similarity, multiple linear regression,artificial neural network