Selection for variables in quantitative structure-activity/property relationship studies - Comparison between methods of orthogonal descriptors and leaps-and-bounds regression analysis

In this paper, the comparison of orthogonal descriptors and Leaps-and-Bounds regression analysis is performed. The results obtained by using orthogonal descriptors are better than that obtained by using Leaps-and-Bounds regression for the data set of nitrobenzenes used in this study. Leaps-and-Bounds regression can be used effectively for selection of variables in quantitative structure-activity/property relationship(QSAR/QSPR) studies. Consequently, orthogonalisation of descriptors is also a good method for variable selection for studies on QSAR/QSPR.

Studies on the method of leaps and bounds regression

In this paper, we introduce the method of leaps and bounds regression which can be used to select variables quickly and obtain the best regression models. These models contain one variable, two variables, three variables and so on. The results obtained by using leaps and bounds regression were compared with those achieved by using stepwise regression to lead to the conclusion that leaps and bounds regression is an effective method.

小波变换和神经网络在化学中的应用

本论文分为两部分：小波变换和神经网络在化学中的应用。小波变换是新近出现的数学方法。近年来，在化学中得到广泛的应用，本论文介绍小波变换和我分辨分析的原理和方法，并将其应用到信号的压缩和滤噪。在研究中提出了常用小波变换数据压缩的三种方法，将紧支集小波和正交三次B－样条小波压缩4－苯乙基邻苯二甲酸酐的红外光谱数据进行了对比，计算表明正交三次B－样条小波变换方法效果较好，而在全部保留模糊信号及只保留锐化信号中数值较大的系数时，压缩比大而重建光谱数据与原始光谱数据间的均方差较小。应用小波变换对表面等离子体子共振（Surface Plasmon Resonance，SPR）仪的信号进行滤噪处理，利用SPR仪器信号和噪音的频率特性而将其分离，取得良好效果。本文对神经网络在化学中应用进行了较深入的研究，并对影响神经网络的诸多因素进行了探讨。在神经网络和多元回归等在化学应用中过多的变量会导致数学模型的预测结果变差，因而选择合适变量是很重要的。本文对比了传统的统计方法（前进选择法，后退剔除法，逐步回归法），Leaps-and-Bounds回归法，正交变换法，主成分分析以及最新的优化技术遗传算法，得到了一些有意义的结果。同时提出了组合算法和前进选择法的得合算法，结果表明这种算法在一定程度上避免了局部最优且减少了计算量。本论文还利用上述方法进行了一些定量结构活性相关性研究，主要内容：1）环境中有毒有害有机物（苯酚，苯胺，硝基苯）的定量结构活性相性研究。2）抗艾滋病类药物（HEPT）定量结构活性相关性研究。3）抗肿瘤类药物（氮芥子气类化合物和2－甲酸吡啶缩氨基硫脲类化合物定量结构活性相关性研究。4）苯酚和苯胺类衍生物色谱比移值预测。5）将神经网络用于茶叶的分类。

Applications of the method of three-dimensional projection of a molecule in quantitative structure-activity relationship of phenol derivatives

Five variables for phenol derivatives were calculated by molecular projection in three-dimensional space which were combined with eight quantum-chemical parameters and three Am indices. These variables were selected by using leaps-and-bounds regression analysis. Multiple linear regression analysis and artificial neural networks' were performed, and the results obtained by using. artificial neural networks are superior than that obtained by using multiple linear regression.

Correlation analysis of the structures and pK(a) values of carboxylic acids

The quantum-chemical descriptors were used for QSPR study of the structures of carboxylic acids and their pK(a) values. The algorithm of "Leaps and Bounds" regression was performed for selection of the variables. The CoMFA method was carried out for 3D-QSPR. As the introduction of the charge of oxygen atom(Q(2)), the results obtained by CoMFA were improved greatly.

Applications of generalized torsion angles in studies on quantitative structure-activity relationship

In this article, generalized torsion angles of derivatives of 1-[(2-hydroxyethoxy)methy1]-6(phenylthio)thymine(HEPT) were calculated, which include abundant three dimensional information of molecules. Molecular similarity matrix was built based on the calculated generalized torsion angles. These similarities were taken as the new variables, and the new variables were selected by using Leaps-and-Bounds regression analysis. Multiple regression analysis and neural networks were performed, and the satisfactory results were achieved by using the neural networks.

Correlation analysis of the structures and gas-liquid chromatographic retention indices of amines

Quantitative structure-retention relationship(QSRR) was studied for amines to gas-liquid chromatography on three stationary phases of different polarities with the topological indices A(m) (A(m1), A(m2), A(m3)) and gravitational index GI. The algorithm of "Leaps and Bounds" was performed for selection of the variables. And the multi-regression and the quasi-Newton neural networks were employed for the calculation with better results.

Comparison of different methods for variable selection

In chemistry for chemical analysis of a multi-component sample or quantitative structure-activity/property relationship (QSAR/QSPR) studies, variable selection is a key step. In this study, comparisons between different methods were performed. These methods include three classical methods such as forward selection, backward elimination and stepwise regression; orthogonal descriptors; leaps-and-bounds regression and genetic algorithm. Thirty-five nitrobenzenes were taken as the data set. From these structures quantum chemical parameters, topological indices and indicator variable were extracted as the descriptors for the comparisons of variable selections. The interesting results have been obtained. (C) 2001 Elsevier Science B.V. All rights reserved.

Prediction of C-13-nuclear magnetic resonance chemical shifts for aliphatic amines

Prediction of C-13-nuclear magnetic resonance chemical shifts for aliphatic amines is performed. The topological, geological and electronic descriptors are generated. To reduce the variables, the best subsets of the descriptors are obtained by using leaps-and-bounds regression analysis. The model is achieved using multiple regression with satisfactory results.

HEPT类化合物的QSAR研究

为定量结构 /活性相关性研究提取了量子化学参数 ,拓扑指数 Am,分子连接性指数 mxt及疏水性常数 ,同时应用正交变换和最佳变量子集算法 (Leaps-and-Bonds)进行了变量压缩和选择 ,进而实施了多元回归分析 ,并由此结果进行了 HEPT 类化合物 (1 -[(2 -hydroxyethoxy) methyl]-6-(phenylthio) -thyminederivatives)的结构 /活性关系的理论解释 .进行了人工神经网络法对于该类化合物的活性预测 ,其结果明显好于多元回归法

Classification of valence changes of trivalent rare earth ions in alkaline earth borates using artificial neural networks

The investigations of classification on the valence changes from RE3+ to RE2+ (RE = Eu, Sm, Yb, Tm) in host compounds of alkaline earth berate were performed using artificial neural networks (ANNs). For comparison, the common methods of pattern recognition, such as SIMCA, KNN, Fisher discriminant analysis and stepwise discriminant analysis were adopted. A learning set consisting of 24 host compounds and a test set consisting of 12 host compounds were characterized by eight crystal structure parameters. These parameters were reduced from 8 to 4 by leaps and bounds algorithm. The recognition rates from 87.5 to 95.8% and prediction capabilities from 75.0 to 91.7% were obtained. The results provided by ANN method were better than that achieved by the other four methods. (C) 1999 Elsevier Science B.V. All rights reserved.