基于数据挖掘算法的环境毒物QSAR研究
|
摘要
能对人类、畜禽、鱼类或其他生物体显示体内毒性的那些环境污染物叫做环境毒物。近年来，评估化学物质对于人类、野生动物、水生动物和植物的潜在威胁已经成为当前环境科学和毒理学研究的一个热点。在此类研究中，定量结构活性相关（QSAR）方法提供了一种非常有价值的方法来研究有机化合物的毒性，因而开始在环境科学中得到广泛的应用。数据挖掘技术用于在数据中发现潜在有用的知识，在QSAR研究当中，正发挥着越来越重要的作用，而且取得了丰硕的成果。本文应用数据挖掘中的集成学习方法和贝叶斯网络方法来对环境毒物QSAR研究中的若干问题进行讨论。本文主要分为以下几个部分：
1.研究了56个芳香类有机化合物对日本林蛙蝌蚪毒性的QSAR模型。首先利用MODEl和Chemoffice化学软件计算了56个芳香类有机化合物的68个分子描述符，然后用最小冗余最大相关（mRMR）-GA-SVR特征筛选方法得到6个较重要的分子结构参数，再用粒子群优化算法（PSO）得到优化的SVR模型参数，最后将SVR算法用于建立训练集的QSAR模型，并将该模型用于测试集数据的预测。计算结果表明PSO-SVR模型对芳香类有机化合物对日本林蛙蝌蚪的毒性数据有着良好的拟合能力（训练集上的R2和RMSE分别是0.95和0.1）和预测能力（测试集上的q2和RMSE分别为0.9和0.19）。我们还使用贝叶斯网络方法来分析模型中变量之间的关系，找到了模型中可能直接对毒性产生影响的直接变量和可能对毒性产生间接影响的间接变量。
2.使用基于SVM的集成学习算法来研究110个醇类有机小分子化合物对欧洲林蛙毒性的QSAR模型。我们首先使用HyperChem7.5和JChem for Excel计算了66个分子描述符，并且使用基于互信息增益的CFS(Correlation-based FeatureSubset)算法筛选出4个关联化合物毒性最好的分子描述符，然后基于这4个分子描述符用基于Bagging的SVM集成学习算法来建模。最后，又与基于参数变换的SVM集成学习算法的预报性能做了比较。结果表明，基于参数变换的SVR集成学习算法对毒性拟合/预测精度优于PSO-SVR算法和基于Bagging的SVR集成学习算法。
3.使用基于SVM的模型仓库的选择性集成学习算法来研究芳香类化合物对梨形四膜虫毒性的QSAR问题。我们从文献中收集了包含不同取代基团的581种芳香类化合物对水生梨形四膜虫毒性的数值。用Gaussian03，HyperChem和TSAR程序计算了68种分子结构描述符，用CFS算法筛选出6个关联化合物毒性最好的分子描述符。基于这6个描述符，我们用基于SVM的模型仓库的选择性集成学习算法建立了500个芳香类化合物对梨形四膜虫毒性和结构特征参数的关系模型，最后将该模型用于测试集81个芳香类化合物毒性的预测，并将结果与PSO-SVR模型与基于参数变换的SVM集成学习算法模型进行了比较。结果表明：基于SVM模型仓库的选择性集成算法的结果明显优于PSO-SVR模型和基于参数变换的SVM集成学习算法模型。
Environmental toxicants are chemical contaminants that may harm livingorganisms at concentrations found in the environment. In recent years, One of thecurrent interests in environmental sciences and toxicology is the ranking of chemicalsubstances with respect to their potential hazardous effects on humans, wild life,aquatic flora and fauna. Quantitative structure activity relationships (QSAR) hasprovided a valuable approach in research on the toxicity of organic chemicals in suchstudies. Data mining technology is used to extract potential and useful informationfrom the databases, and is playing an increasingly important role in the study ofQSAR. In this paper, ensemble learning method and bayesian networks method wereused to investigate some topics of QSAR of environmental toxicant. The main workof the paper contains following four parts:
Quantitative structure-activity relationship (QSAR) model was developed tocorrelate structures of56organic compounds with their toxicity to tadpoles (Ranajaponica). The68molecular descriptors derived solely from the structures of theorganic compounds were calculated using MODEl and Chemoffice. The descriptorswere screened by the minimum Redundancy Maximum Relevance (mRMR)-geneticalgorithm (GA)-support vector regression (SVR) method. The parameters of SVRmodel was optimized using particle swarm optimization method. The QSAR modelwas developed from a training set consisting of40compounds using SVR methodwith good determination coefficient (R2=0.95). The QSPR model was then testedusing an external test set consisting of16compounds with satisfactory externalpredictive ability (q2=0.90).
110organic compounds with their toxicity to tadpoles (Rana temporaria) fromdifferent references were collected, and66descriptors were calculated with.HyperChem7.5and JChem for Excel. Firstly,4descriptors were selected by usingCFS (Correlation-based Feature Subset) method. Then SVR Ensemble learningapproach based on Bagging was employed to build the model. Finally, the performance of SVR Ensemble learning approach based on Bagging is compared withSVR Ensemble learning approach based on parameter. As a result, SVR Ensemblelearning approach performed better than SVR Ensemble learning approach based onBagging algorithm and molecular descriptors. It can be concluded that SVR Ensemblelearning approach has a potential to improve the performance of SARs analysis.
A quantitative structure-activity relationship (QSAR) study was performed todevelop model for correlating the structures of581aromatic compounds with theiraquatic toxicity to Tetrahymena pyriformis. A set of68molecular descriptors derivedsolely from the structures of the aromatic compounds were calculated based onGaussian03, HyperChem7.5, and TSAR V3.3. A comprehensive feature selectionmethod, CFS (Correlation-based Feature Subset) method, was applied to select thebest descriptor subset in QSAR analysis. The SVR Ensemble Selection from Librariesof Models method was employed to model the toxicity potency from a training set of500compounds. The SVR Ensemble Selection from Libraries of Models was testedusing an external test set of81compounds. A good coefficient of determination(R2=0.79) and external predictive ability (q2=0.69) values were obtained indicatingthe potential of SVR Ensemble Selection from Libraries of Models in facilitating theprediction of toxicity.
1.研究了56个芳香类有机化合物对日本林蛙蝌蚪毒性的QSAR模型。首先利用MODEl和Chemoffice化学软件计算了56个芳香类有机化合物的68个分子描述符，然后用最小冗余最大相关（mRMR）-GA-SVR特征筛选方法得到6个较重要的分子结构参数，再用粒子群优化算法（PSO）得到优化的SVR模型参数，最后将SVR算法用于建立训练集的QSAR模型，并将该模型用于测试集数据的预测。计算结果表明PSO-SVR模型对芳香类有机化合物对日本林蛙蝌蚪的毒性数据有着良好的拟合能力（训练集上的R2和RMSE分别是0.95和0.1）和预测能力（测试集上的q2和RMSE分别为0.9和0.19）。我们还使用贝叶斯网络方法来分析模型中变量之间的关系，找到了模型中可能直接对毒性产生影响的直接变量和可能对毒性产生间接影响的间接变量。
2.使用基于SVM的集成学习算法来研究110个醇类有机小分子化合物对欧洲林蛙毒性的QSAR模型。我们首先使用HyperChem7.5和JChem for Excel计算了66个分子描述符，并且使用基于互信息增益的CFS(Correlation-based FeatureSubset)算法筛选出4个关联化合物毒性最好的分子描述符，然后基于这4个分子描述符用基于Bagging的SVM集成学习算法来建模。最后，又与基于参数变换的SVM集成学习算法的预报性能做了比较。结果表明，基于参数变换的SVR集成学习算法对毒性拟合/预测精度优于PSO-SVR算法和基于Bagging的SVR集成学习算法。
3.使用基于SVM的模型仓库的选择性集成学习算法来研究芳香类化合物对梨形四膜虫毒性的QSAR问题。我们从文献中收集了包含不同取代基团的581种芳香类化合物对水生梨形四膜虫毒性的数值。用Gaussian03，HyperChem和TSAR程序计算了68种分子结构描述符，用CFS算法筛选出6个关联化合物毒性最好的分子描述符。基于这6个描述符，我们用基于SVM的模型仓库的选择性集成学习算法建立了500个芳香类化合物对梨形四膜虫毒性和结构特征参数的关系模型，最后将该模型用于测试集81个芳香类化合物毒性的预测，并将结果与PSO-SVR模型与基于参数变换的SVM集成学习算法模型进行了比较。结果表明：基于SVM模型仓库的选择性集成算法的结果明显优于PSO-SVR模型和基于参数变换的SVM集成学习算法模型。
Environmental toxicants are chemical contaminants that may harm livingorganisms at concentrations found in the environment. In recent years, One of thecurrent interests in environmental sciences and toxicology is the ranking of chemicalsubstances with respect to their potential hazardous effects on humans, wild life,aquatic flora and fauna. Quantitative structure activity relationships (QSAR) hasprovided a valuable approach in research on the toxicity of organic chemicals in suchstudies. Data mining technology is used to extract potential and useful informationfrom the databases, and is playing an increasingly important role in the study ofQSAR. In this paper, ensemble learning method and bayesian networks method wereused to investigate some topics of QSAR of environmental toxicant. The main workof the paper contains following four parts:
Quantitative structure-activity relationship (QSAR) model was developed tocorrelate structures of56organic compounds with their toxicity to tadpoles (Ranajaponica). The68molecular descriptors derived solely from the structures of theorganic compounds were calculated using MODEl and Chemoffice. The descriptorswere screened by the minimum Redundancy Maximum Relevance (mRMR)-geneticalgorithm (GA)-support vector regression (SVR) method. The parameters of SVRmodel was optimized using particle swarm optimization method. The QSAR modelwas developed from a training set consisting of40compounds using SVR methodwith good determination coefficient (R2=0.95). The QSPR model was then testedusing an external test set consisting of16compounds with satisfactory externalpredictive ability (q2=0.90).
110organic compounds with their toxicity to tadpoles (Rana temporaria) fromdifferent references were collected, and66descriptors were calculated with.HyperChem7.5and JChem for Excel. Firstly,4descriptors were selected by usingCFS (Correlation-based Feature Subset) method. Then SVR Ensemble learningapproach based on Bagging was employed to build the model. Finally, the performance of SVR Ensemble learning approach based on Bagging is compared withSVR Ensemble learning approach based on parameter. As a result, SVR Ensemblelearning approach performed better than SVR Ensemble learning approach based onBagging algorithm and molecular descriptors. It can be concluded that SVR Ensemblelearning approach has a potential to improve the performance of SARs analysis.
A quantitative structure-activity relationship (QSAR) study was performed todevelop model for correlating the structures of581aromatic compounds with theiraquatic toxicity to Tetrahymena pyriformis. A set of68molecular descriptors derivedsolely from the structures of the aromatic compounds were calculated based onGaussian03, HyperChem7.5, and TSAR V3.3. A comprehensive feature selectionmethod, CFS (Correlation-based Feature Subset) method, was applied to select thebest descriptor subset in QSAR analysis. The SVR Ensemble Selection from Librariesof Models method was employed to model the toxicity potency from a training set of500compounds. The SVR Ensemble Selection from Libraries of Models was testedusing an external test set of81compounds. A good coefficient of determination(R2=0.79) and external predictive ability (q2=0.69) values were obtained indicatingthe potential of SVR Ensemble Selection from Libraries of Models in facilitating theprediction of toxicity.
引文
[1]魏庆义,金力,周宗灿,李涛,郑玉新,王忠旭, et al.环境化学污染物与人类健康.科学.2007(03):26-28+4.
[2]朱才众and熊鸿燕.环境持久性毒物研究进展概述.疾病控制杂志.2005(04):331-335.
[3]薛峰and薛念涛.环境毒理学与环境监测.环境监测管理与技术.2000(03):17-19.
[4]徐光宪.21世纪的化学是研究泛分子的科学.中国科学基金.2002(02):8-14.
[5]李志良.定量构效关系研究进展.化学通报.1995(09):5-10.
[6]王连生，张爱茜，韩朔睽.定量结构──活性相关研究进展.环境科学进展.1994(04):15-22.
[7]王鹏,苏建成, and单俊杰.用于QSAR的自相关拓扑指数研究.化学通报.1998(10):41-43.
[8]戴家银,靳立军, and王连生.分子拓扑学参数及其在定量结构——活性相关(QSAR)研究中的应用.环境科学进展.1998(04):57-64.
[9]戴树桂,黄国兰,孙红文,张育红,韩朔睽,赵元慧, et al. QSAR计算中的分子连接性指数法.环境化学.1994(03):242-245.
[10] Liu, W.Q., P.G. Yi, and Z.L. Tang. QSPR models for various propertiesof polymethacrylates based on quantum chemical descriptors. Qsar&Combinatorial Science.2006.25(10):936-943.
[11]籍国东,赵元慧, and袁星.量化参数及其在定量结构-活性-性质相关研究中的应用.东北师大学报(自然科学版).1998(04):52-58.
[12] Singh, P.P., H.K. Srivastava, and F.A. Pasha. DFT-based QSAR studyof testosterone and its derivatives. Bioorganic&Medicinal Chemistry.2004.12(1):171-177.
[13] Cho, S.J. and M.A. Hermsmeier. Genetic algorithm guided selection:Variable selection and subset selection. Journal of Chemical Information andComputer Sciences.2002.42(4):927-936.
[14] Whitley, D.C., M.G. Ford, and D.J. Livingstone. Unsupervisedforward selection: A method for eliminating redundant variables. Journal ofChemical Information and Computer Sciences.2000.40(5):1160-1168.
[15] Mao, K.Z. Orthogonal forward selection and backward eliminationalgorithms for feature subset selection. Ieee Transactions on Systems Man andCybernetics Part B-Cybernetics.2004.34(1):629-634.
[16] Lerman, K., S.N. Minton, and C.A. Knoblock. Wrapper maintenance:A machine learning approach. Journal of Artificial Intelligence Research.2003.18:149-181.
[17] Burkholder, T.J. and R.L. Lieber. Stepwise regression is an alternativeto splines for fitting noisy data. Journal of Biomechanics.1996.29(2):235-238.
[18] Eklov, T., P. Martensson, and I. Lundstrom. Selection of variables forinterpreting multivariate gas sensor data. Analytica Chimica Acta.1999.381(2-3):221-232.
[19] Peng, H.C., F.H. Long, and C. Ding. Feature selection based onmutual information: Criteria of max-dependency, max-relevance, andmin-redundancy. Ieee Transactions on Pattern Analysis and Machine Intelligence.2005.27(8):1226-1238.
[20] Hansch C, M.R.M., Fujita T, Maloney P P, Geiger F, Streich M. Thecorrelation of biological activity of plant growth regulators and chloromycetinderivatives with Hammett constants and partition coefficientsJournal of American Chemical Society.1963.85(18):2817-2824.
[21] Hansch C, M.P.P., Fujita T, Muir R M. Correlation of biologicalactivity of phenoxyacetic acids with Hammett substituent constants and partitioncoefficients. Nature.1962.1942(4824):178-180.
[22] Fujita T, I.J., Hansch C. A new substituent constant, π, derived frompartition coefficients. Journal of American Chemical Society.1964.86(23):5175-5180.
[23] Free S M, W.J.M. A mathematical contribution to structure-activitystudies. Journal of Medicinal Chemistry.1964.7(4):395-399.
[24]陈凯先,罗小民, and蒋华良.药物分子设计的发展.中国科学院院刊.2000(04):265-269.
[25] Available from:http://www.oecd.org/chemicalsafety/risk-assessment/37849783.pdf.
[26] Fayyad, U. and P. Stolorz. Data mining and KDD: Promise andchallenges. Future Generation Computer Systems.1997.13(2-3):99-115.
[27] Glymour, C., D. Madigan, D. Pregibon, and P. Smyth. Statisticalthemes and lessons for data mining. Data Mining and Knowledge Discovery.1997.1(1):11-28.
[28]钟晓,马少平,张钹, and俞瑞钊.数据挖掘综述.模式识别与人工智能.2001(01):48-55.
[29]唐晓萍.数据挖掘与知识发现综述.电脑开发与应用.2002(04):31-32+35.
[30]王光宏and蒋平.数据挖掘综述.同济大学学报(自然科学版).2004(02):246-252.
[31] Exner, F. Advances in knowledge discovery and data mining. Journalof the American Society for Information Science.1998.49(4):386-387.
[32] Jain, A.K., R.P.W. Duin, and J.C. Mao. Statistical pattern recognition:A review. Ieee Transactions on Pattern Analysis and Machine Intelligence.2000.22(1):4-37.
[33]吉根林,帅克, and孙志挥.数据挖掘技术及其应用.南京师大学报(自然科学版).2000(02):25-27.
[34]赵丹群.数据挖掘:原理、方法及其应用.现代图书情报技术.2000(06):41-44.
[35] Gaber, M.M., A. Zaslavsky, and S. Krishnaswamy. Mining datastreams: A review. Sigmod Record.2005.34(2):18-26.
[36] Vapnik, V., Statistical Learning Theory.1998, New York: Wiley.
[37]张学工.关于统计学习理论与支持向量机.自动化学报.2000(01):36-46.
[38] Cortes, C. and V. Vapnik. SUPPORT-VECTOR NETWORKS. MachineLearning.1995.20(3):273-297.
[39] A.J. Smola, B.S. NeuroCOLT Technical Report NC-TR-98-030.1998; Royal HollowayCollege,University of London:[Available from: http://www.kernel-machines.org/.
[40]萧嵘,王继成, and张福炎.支持向量机理论综述.计算机科学.2000(03):1-3.
[41]刘江华,程君实, and陈佳品.支持向量机训练算法综述.信息与控制.2002(01):45-50.
[42]祁亨年.支持向量机及其应用研究综述.计算机工程.2004(10):6-9.
[43]许建华,张学工, and李衍达.支持向量机的新发展.控制与决策.2004(05):481-484+495.
[44] Dietterich, T.G. Ensemble methods in machine learning. MultipleClassifier Systems.2000.1857:1-15.
[45] Dietterich, T.G. An experimental comparison of three methods forconstructing ensembles of decision trees: Bagging, boosting, and randomization.Machine Learning.2000.40(2):139-157.
[46] Zhou, Z.H., J.X. Wu, and W. Tang. Ensembling neural networks:Many could be better than all. Artificial Intelligence.2002.137(1-2):239-263.
[47] Schapire, R.E. Theoretical views of boosting. Computational LearningTheory.1999.1572:1-10.
[48] Schapire, R.E. Theoretical views of boosting and applications.Algorithmic Learning Theory, Proceedings.1999.1720:13-25.
[49] Freund, Y. and R.E. Schapire. A decision-theoretic generalization ofon-line learning and an application to boosting. Journal of Computer andSystem Sciences.1997.55(1):119-139.
[50] Breiman, L. Bagging predictors. Machine Learning.1996.24(2):123-140.
[51] Friedman, N., D. Geiger, and M. Goldszmidt. Bayesian networkclassifiers. Machine Learning.1997.29(2-3):131-163.
[52] Heckerman, D. Bayesian networks for data mining. Data Mining andKnowledge Discovery.1997.1(1):79-119.
[53] Jordan, M.I., Z. Ghahramani, T.S. Jaakkola, and L.K. Saul. Anintroduction to variational methods for graphical models. Machine Learning.1999.37(2):183-233.
[54] Friedman, N., M. Linial, I. Nachman, and D. Pe'er. Using Bayesiannetworks to analyze expression data. Journal of Computational Biology.2000.7(3-4):601-620.
[55] Cheng, J., R. Greiner, J. Kelly, D. Bell, and W. Liu. Learning Bayesiannetworks from data: An information-theory based approach. ArtificialIntelligence.2002.137(1-2):43-90.
[56] Shahlaei, M., A. Fassihi, and L. Saghaie. Application of PC-ANN andPC-LS-SVM in QSAR of CCR1antagonist compounds: A comparative study.European Journal of Medicinal Chemistry.2010.45(4):1572-1582.
[57] Doucet, J.P., F. Barbault, H.R. Xia, A. Panaye, and B. Fan. NonlinearSVM approaches to QSPR/QSAR studies and drug design. CurrentComputer-Aided Drug Design.2007.3(4):263-289.
[58] Emhrechts, M.J. and C.M. Breneman. K-PLS for QSAR: A nonlinearSVM-like approach to PLS. Abstracts of Papers of the American ChemicalSociety.2003.225:U777-U777.
[59] Wang, X., Y. Dong, S. Xu, L. Wang, and S. Han. Quantitativestructure-activity relationships for the toxicity to the tadpole Rana japonica ofselected phenols. Bulletin of Environmental Contamination and Toxicology.2000.64(6):859-865.
[60] Wang, X., Y. Dong, L. Wang, and S. Han. Acute toxicity of substitutedphenols to Rana japonica tadpoles and mechanism-based quantitativestructure-activity relationship (QSAR) study. Chemosphere.2001.44(3):447-455.
[61] Huang, H., X. Wang, W. Ou, J. Zhao, Y. Shao, and L. Wang. Acutetoxicity of benzene derivatives to the tadpoles (Rana japonica) and QSARanalyses. Chemosphere.2003.53(8):963-970.
[62] Huang, H., X. Wang, Y. Shao, D. Chen, X. Dai, and L. Wang. QSARfor prediction of joint toxicity of substituted phenols to tadpoles (Rana japonica).Bulletin of Environmental Contamination and Toxicology.2003.71(6):1124-1130.
[63] Gramatica, P. Principles of QSAR models validation: internal andexternal. Qsar&Combinatorial Science.2007.26(5):694-701.
[64] Ding, C. and H.C. Peng. Minimum redundancy feature selection frommicroarray gene expression data. Proceedings of the2003Ieee BioinformaticsConference.2003:523-528.
[65] Shen, Q., W.M. Shi, X.P. Yang, and B.X. Ye. Particle swarm algorithmtrained neural network for QSAR studies of inhibitors of platelet-derived growthfactor receptor phosphorylation. European Journal Of Pharmaceutical Sciences.2006.28(5):369-376.
[66] Brits, R., A.P. Engelbrecht, and F. van den Bergh. Locating multipleoptima using particle swarm optimization. Applied Mathematics andComputation.2007.189(2):1859-1883.
[67] Liu, B., L. Wang, and Y.-H. Jin. An effective hybrid PSO-basedalgorithm for flow shop scheduling with limited buffers. Computers&Operations Research.2008.35(9):2791-2806.
[68] Pedrycz, W., B.J. Park, and N.J. Pizzi. Identifying core sets ofdiscriminatory features using particle swarm optimization. Expert Systems withApplications.2009.36(3, Part1):4610-4616.
[69] Larranaga, P., H. Karshenas, C. Bielza, and R. Santana. A review onevolutionary algorithms in Bayesian network learning and inference tasks.Information Sciences.2013.233:109-125.
[70] Ibezim, E., P.R. Duchowicz, E.V. Ortiz, and E.A. Castro. QSAR onaryl-piperazine derivatives with activity on malaria. Chemometrics andIntelligent Laboratory Systems.2012.110(1):81-88.
[71] Toropova, A.P., A.A. Toropov, A. Lombardo, A. Roncaglioni, E.Benfenati, and G. Gini. CORAL: QSAR Models for Acute Toxicity in FatheadMinnow (Pimephales promelas). Journal of Computational Chemistry.2012.33(12):1218-1223.
[72] Toropova, A.P., A.A. Toropov, S.E. Martyanov, E. Benfenati, G. Gini,D. Leszczynska, et al. CORAL: QSAR modeling of toxicity of organic chemicalstowards Daphnia magna. Chemometrics and Intelligent Laboratory Systems.2012.110(1):177-181.
[73] Liu, J.N., D.L. Fan, L.L. Shi, L. Wang, L.J. Zhou, Y.W. Chen, et al.2Dand3D-QSAR study on the fish embryo Toxicity test of EC50. Research Journalof Chemistry and Environment.2013.17(5):52-59.
[74] Qiu, J., Y. Dai, X.S. Zhang, and G.S. Chen. QSAR Modeling of Toxicityof Acyclic Quaternary Ammonium Compounds on Scenedesmus QuadricaudaUsing2D and3D Descriptors. Bulletin of Environmental Contamination andToxicology.2013.91(1):83-88.
[75] Joshi, V.Y., M.M. Kadam, and M.R. Sawant. Comparison of QSARand QSPR based aquatic toxicity for mixed surfactants. Journal of Surfactantsand Detergents.2007.10(1):25-34.
[76] Colombo, A., E. Benfenati, M. Karelson, and U. Maran. The proposalof architecture for chemical splitting to optimize QSAR models for aquatictoxicity. Chemosphere.2008.72(5):772-780.
[77] Zvinavashe, E., H. van den Berg, A.E.M.F. Soffers, J. Vervoort, A.Freidig, A.J. Murk, et al. QSAR models for predicting in vivo aquatic toxicity ofchlorinated alkanes to fish. Chemical Research in Toxicology.2008.21(3):739-745.
[78]Polishchuk, P.G., E.N. Muratov, A.G. Artemenko, O.G. Kolumbin, N.N.Muratov, and V.E. Kuz'min. Application of Random Forest Approach to QSARPrediction of Aquatic Toxicity. Journal of Chemical Information and Modeling.2009.49(11):2481-2488.
[79] Lanevskij, K., L. Juska, R. Didziapetris, and P. Japertas. Novel QSARmodels for predicting toxicity of chemicals to aquatic organisms and identifyingthe mode of action. Abstracts of Papers of the American Chemical Society.2011.241.
[80] Hernandez, N., R. Kiralj, M.M.C. Ferreira, and I. Talavera. Criticalcomparative analysis, validation and interpretation of SVM and PLS regressionmodels in a QSAR study on HIV-1protease inhibitors. Chemometrics andIntelligent Laboratory Systems.2009.98(1):65-77.
[81] Norinder, U. and M.E. Ek. QSAR investigation of NaV1.7activecompounds using the SVM/Signature approach and the Bioclipse Modelingplatform. Bioorganic&Medicinal Chemistry Letters.2013.23(1):261-263.
[82] Wang, Z.M., N. Han, Z.M. Yuan, and Z.H. Wu. Feature Selection forHigh-Dimensional Data Based on Ridge Regression and SVM and ItsApplication in Peptide QSAR Modeling. Acta Physico-Chimica Sinica.2013.29(3):498-507.
[83]Abraham, M.H. and C. Rafols. Factors that influence tadpole narcosis.An LFER analysis. Journal of the Chemical Society, Perkin Transactions2.1995(10):1843-1851.
[84] Agrawal, V.K., S. Chaturvedi, M.H. Abraham, and P.V. Khadikar.QSAR Study on tadpole narcosis. Bioorganic&Medicinal Chemistry.2003.11(20):4523-4533.
[85] Bowen, K.R., K.B. Flanagan, W.E. Acree Jr, M.H. Abraham, and C.Rafols. Correlation of the toxicity of organic compounds to tadpoles using theAbraham model. Science of the Total Environment.2006.371(1–3):99-109.
[86] Hall, M.A. and L.A. Smith. Feature selection for machine learning:comparing a correlation-based filter approach to the wrapper. Proceedings ofthe Twelfth International Florida AI Research Society Conference|Proceedingsof the Twelfth International Florida AI Research Society Conference.1999:235-9|xx+540.
[87] QSAR+descriptors]. Available from:http://www.ifm.liu.se/compchem/msi/doc/life/cerius46/qsar/theory_descriptors.html.
[88] Medven, Z., H. Gusten, and A. Sabljic. Comparative QSAR study onhydroxyl radical reactivity with unsaturated hydrocarbons: PLS versus MLR.Journal of Chemometrics.1996.10(2):135-147.
[89] Hemmateenejad, B., R. Miri, M. Akhond, and M. Shamsipur. QSARstudy of the calcium channel antagonist activity of some recently synthesizeddihydropyridine derivatives. An application of genetic algorithm for variableselection in MLR and PLS methods. Chemometrics and Intelligent LaboratorySystems.2002.64(1):91-99.
[90] Shen, Q., J.H. Jiang, G.L. Shen, and R.Q. Yu. Variable selection by anevolution algorithm using modified Cp based on MLR and PLS modeling: QSARstudies of carcinogenicity of aromatic amines. Analytical and BioanalyticalChemistry.2003.375(2):248-254.
[91] He, M., T.H. Li, and P.S. Cong. Nonlinear PLS improved by numericgenetic algorithm for QSAR modelling. Chemical Journal of ChineseUniversities-Chinese.1997.18(6):854-859.
[92] Devillers, J., A. Chezeau, and E. Thybaud. PLS-QSAR of the adult anddevelopmental toxicity of chemicals to Hydra attenuata. Sar and Qsar inEnvironmental Research.2002.13(7-8):705-712.
[93] Leonard, J.T. and K. Roy. QSAR by LFER model of HIV proteaseinhibitor mannitol derivatives using FA-MLR, PCRA, and PLS techniques.Bioorganic&Medicinal Chemistry.2006.14(4):1039-1046.
[94] Avery, M.A. and N. Sapre.2D and3d QSAr modeling of anti-HIV-IITIBO derivatives: A comparative study using ANN and multiple linearregression methods. Abstracts of Papers of the American Chemical Society.2003.225:U753-U754.
[95] Su, Q. and L. Zhou. QSAR modeling of AT1receptor antagonists usingANN. Journal of Molecular Modeling.2006.12(6):869-875.
[96] Deeb, O. and B. Hemmateenejad. ANN-QSAR model of drug-bindingto human serum albumin. Chemical Biology&Drug Design.2007.70(1):19-29.
[97] Gonzalez-Diaz, H., I. Bonet, C. Teran, E. De Clercq, R. Bello, M.M.Garcia, et al. ANN-QSAR model for selection of anticancer leads fromstructurally heterogeneous series of compounds. European Journal of MedicinalChemistry.2007.42(5):580-585.
[98] Goodarzi, M., M.P. Freitas, and R. Jensen. Ant colony optimization asa feature selection method in the QSAR modeling of anti-HIV-1activities of3-(3,5-dimethylbenzyl)uracil derivatives using MLR, PLS and SVM regressions.Chemometrics and Intelligent Laboratory Systems.2009.98(2):123-129.
[99] Katritzky, A.R., Y. Ren, S.H. Slavov, and M. Karelson. A ComparativeQsar Study of Svm and Ppr in the Correlation of Lithium Cation Basicities.Collection of Czechoslovak Chemical Communications.2009.74(1):217-241.
[100] Caruana, R., A. Niculescu-Mizil, G. Crew, and A. Ksikes,Ensemble selection from libraries of models, in Proceedings of the twenty-firstinternational conference on Machine learning.2004, ACM: Banff, Alberta,Canada. p.18.
[101] Schultz, T.W. Structure-toxicity relationships for benzenesevaluated with Tetrahymena pyriformis. Chemical Research in Toxicology.1999.12(12):1262-1267.
[102] Cronin, M.T.D., N. Manga, J.R. Seward, G.D. Sinks, and T.W.Schultz. Parametrization of electrophilicity for the prediction of the toxicity ofaromatic compounds. Chemical Research in Toxicology.2001.14(11):1498-1505.
[103] Cronin, M.T.D. and T.W. Schultz. Development of quantitativestructure-activity relationships for the toxicity of aromatic compounds toTetrahymena pyriformis: Comparative assessment of the methodologies.Chemical Research in Toxicology.2001.14(9):1284-1295.
[104] Serra, J.R., P.C. Jurs, and K.L.E. Kaiser. Linear regression andcomputational neural network prediction of Tetrahymena acute toxicity foraromatic compounds from molecular structure. Chemical Research inToxicology.2001.14(11):1535-1545.
[105] Cronin, M.T.D., A.O. Aptula, J.C. Duffy, T.I. Netzeva, P.H. Rowe,I.V. Valkova, et al. Comparative assessment of methods to develop QSARs forthe prediction of the toxicity of phenols to Tetrahymena pyriformis.Chemosphere.2002.49(10):1201-1221.
[106] Wang, X.D., C. Sun, Y. Wang, and L.S. Wang. Quantitativestructure-activity relationships for the inhibition toxicity to root elongation ofCucumis sativus of selected phenols and interspecies correlation withTetrahymena pyriformis. Chemosphere.2002.46(2):153-161.
[107] Cronin, M.T.D., T.I. Netzeva, J.C. Dearden, R. Edwards, andA.D.P. Worgan. Assessment and modeling of the toxicity of organic chemicals toChlorella vulgaris: Development of a novel database. Chemical Research inToxicology.2004.17(4):545-554.
[108] Aptula, A.O., D.W. Roberts, M.T.D. Cronin, and T.W. Schultz.Chemistry-toxicity relationships for the effects of Di-and trihydroxybenzenes toTetrahymena pyriformis. Chemical Research in Toxicology.2005.18(5):844-854.
[109] Netzeva, T.I. and T.W. Schultz. QSARs for the aquatic toxicity ofaromatic aldehydes from Tetrahymena data. Chemosphere.2005.61(11):1632-1643.
[2]朱才众and熊鸿燕.环境持久性毒物研究进展概述.疾病控制杂志.2005(04):331-335.
[3]薛峰and薛念涛.环境毒理学与环境监测.环境监测管理与技术.2000(03):17-19.
[4]徐光宪.21世纪的化学是研究泛分子的科学.中国科学基金.2002(02):8-14.
[5]李志良.定量构效关系研究进展.化学通报.1995(09):5-10.
[6]王连生，张爱茜，韩朔睽.定量结构──活性相关研究进展.环境科学进展.1994(04):15-22.
[7]王鹏,苏建成, and单俊杰.用于QSAR的自相关拓扑指数研究.化学通报.1998(10):41-43.
[8]戴家银,靳立军, and王连生.分子拓扑学参数及其在定量结构——活性相关(QSAR)研究中的应用.环境科学进展.1998(04):57-64.
[9]戴树桂,黄国兰,孙红文,张育红,韩朔睽,赵元慧, et al. QSAR计算中的分子连接性指数法.环境化学.1994(03):242-245.
[10] Liu, W.Q., P.G. Yi, and Z.L. Tang. QSPR models for various propertiesof polymethacrylates based on quantum chemical descriptors. Qsar&Combinatorial Science.2006.25(10):936-943.
[11]籍国东,赵元慧, and袁星.量化参数及其在定量结构-活性-性质相关研究中的应用.东北师大学报(自然科学版).1998(04):52-58.
[12] Singh, P.P., H.K. Srivastava, and F.A. Pasha. DFT-based QSAR studyof testosterone and its derivatives. Bioorganic&Medicinal Chemistry.2004.12(1):171-177.
[13] Cho, S.J. and M.A. Hermsmeier. Genetic algorithm guided selection:Variable selection and subset selection. Journal of Chemical Information andComputer Sciences.2002.42(4):927-936.
[14] Whitley, D.C., M.G. Ford, and D.J. Livingstone. Unsupervisedforward selection: A method for eliminating redundant variables. Journal ofChemical Information and Computer Sciences.2000.40(5):1160-1168.
[15] Mao, K.Z. Orthogonal forward selection and backward eliminationalgorithms for feature subset selection. Ieee Transactions on Systems Man andCybernetics Part B-Cybernetics.2004.34(1):629-634.
[16] Lerman, K., S.N. Minton, and C.A. Knoblock. Wrapper maintenance:A machine learning approach. Journal of Artificial Intelligence Research.2003.18:149-181.
[17] Burkholder, T.J. and R.L. Lieber. Stepwise regression is an alternativeto splines for fitting noisy data. Journal of Biomechanics.1996.29(2):235-238.
[18] Eklov, T., P. Martensson, and I. Lundstrom. Selection of variables forinterpreting multivariate gas sensor data. Analytica Chimica Acta.1999.381(2-3):221-232.
[19] Peng, H.C., F.H. Long, and C. Ding. Feature selection based onmutual information: Criteria of max-dependency, max-relevance, andmin-redundancy. Ieee Transactions on Pattern Analysis and Machine Intelligence.2005.27(8):1226-1238.
[20] Hansch C, M.R.M., Fujita T, Maloney P P, Geiger F, Streich M. Thecorrelation of biological activity of plant growth regulators and chloromycetinderivatives with Hammett constants and partition coefficientsJournal of American Chemical Society.1963.85(18):2817-2824.
[21] Hansch C, M.P.P., Fujita T, Muir R M. Correlation of biologicalactivity of phenoxyacetic acids with Hammett substituent constants and partitioncoefficients. Nature.1962.1942(4824):178-180.
[22] Fujita T, I.J., Hansch C. A new substituent constant, π, derived frompartition coefficients. Journal of American Chemical Society.1964.86(23):5175-5180.
[23] Free S M, W.J.M. A mathematical contribution to structure-activitystudies. Journal of Medicinal Chemistry.1964.7(4):395-399.
[24]陈凯先,罗小民, and蒋华良.药物分子设计的发展.中国科学院院刊.2000(04):265-269.
[25] Available from:http://www.oecd.org/chemicalsafety/risk-assessment/37849783.pdf.
[26] Fayyad, U. and P. Stolorz. Data mining and KDD: Promise andchallenges. Future Generation Computer Systems.1997.13(2-3):99-115.
[27] Glymour, C., D. Madigan, D. Pregibon, and P. Smyth. Statisticalthemes and lessons for data mining. Data Mining and Knowledge Discovery.1997.1(1):11-28.
[28]钟晓,马少平,张钹, and俞瑞钊.数据挖掘综述.模式识别与人工智能.2001(01):48-55.
[29]唐晓萍.数据挖掘与知识发现综述.电脑开发与应用.2002(04):31-32+35.
[30]王光宏and蒋平.数据挖掘综述.同济大学学报(自然科学版).2004(02):246-252.
[31] Exner, F. Advances in knowledge discovery and data mining. Journalof the American Society for Information Science.1998.49(4):386-387.
[32] Jain, A.K., R.P.W. Duin, and J.C. Mao. Statistical pattern recognition:A review. Ieee Transactions on Pattern Analysis and Machine Intelligence.2000.22(1):4-37.
[33]吉根林,帅克, and孙志挥.数据挖掘技术及其应用.南京师大学报(自然科学版).2000(02):25-27.
[34]赵丹群.数据挖掘:原理、方法及其应用.现代图书情报技术.2000(06):41-44.
[35] Gaber, M.M., A. Zaslavsky, and S. Krishnaswamy. Mining datastreams: A review. Sigmod Record.2005.34(2):18-26.
[36] Vapnik, V., Statistical Learning Theory.1998, New York: Wiley.
[37]张学工.关于统计学习理论与支持向量机.自动化学报.2000(01):36-46.
[38] Cortes, C. and V. Vapnik. SUPPORT-VECTOR NETWORKS. MachineLearning.1995.20(3):273-297.
[39] A.J. Smola, B.S. NeuroCOLT Technical Report NC-TR-98-030.1998; Royal HollowayCollege,University of London:[Available from: http://www.kernel-machines.org/.
[40]萧嵘,王继成, and张福炎.支持向量机理论综述.计算机科学.2000(03):1-3.
[41]刘江华,程君实, and陈佳品.支持向量机训练算法综述.信息与控制.2002(01):45-50.
[42]祁亨年.支持向量机及其应用研究综述.计算机工程.2004(10):6-9.
[43]许建华,张学工, and李衍达.支持向量机的新发展.控制与决策.2004(05):481-484+495.
[44] Dietterich, T.G. Ensemble methods in machine learning. MultipleClassifier Systems.2000.1857:1-15.
[45] Dietterich, T.G. An experimental comparison of three methods forconstructing ensembles of decision trees: Bagging, boosting, and randomization.Machine Learning.2000.40(2):139-157.
[46] Zhou, Z.H., J.X. Wu, and W. Tang. Ensembling neural networks:Many could be better than all. Artificial Intelligence.2002.137(1-2):239-263.
[47] Schapire, R.E. Theoretical views of boosting. Computational LearningTheory.1999.1572:1-10.
[48] Schapire, R.E. Theoretical views of boosting and applications.Algorithmic Learning Theory, Proceedings.1999.1720:13-25.
[49] Freund, Y. and R.E. Schapire. A decision-theoretic generalization ofon-line learning and an application to boosting. Journal of Computer andSystem Sciences.1997.55(1):119-139.
[50] Breiman, L. Bagging predictors. Machine Learning.1996.24(2):123-140.
[51] Friedman, N., D. Geiger, and M. Goldszmidt. Bayesian networkclassifiers. Machine Learning.1997.29(2-3):131-163.
[52] Heckerman, D. Bayesian networks for data mining. Data Mining andKnowledge Discovery.1997.1(1):79-119.
[53] Jordan, M.I., Z. Ghahramani, T.S. Jaakkola, and L.K. Saul. Anintroduction to variational methods for graphical models. Machine Learning.1999.37(2):183-233.
[54] Friedman, N., M. Linial, I. Nachman, and D. Pe'er. Using Bayesiannetworks to analyze expression data. Journal of Computational Biology.2000.7(3-4):601-620.
[55] Cheng, J., R. Greiner, J. Kelly, D. Bell, and W. Liu. Learning Bayesiannetworks from data: An information-theory based approach. ArtificialIntelligence.2002.137(1-2):43-90.
[56] Shahlaei, M., A. Fassihi, and L. Saghaie. Application of PC-ANN andPC-LS-SVM in QSAR of CCR1antagonist compounds: A comparative study.European Journal of Medicinal Chemistry.2010.45(4):1572-1582.
[57] Doucet, J.P., F. Barbault, H.R. Xia, A. Panaye, and B. Fan. NonlinearSVM approaches to QSPR/QSAR studies and drug design. CurrentComputer-Aided Drug Design.2007.3(4):263-289.
[58] Emhrechts, M.J. and C.M. Breneman. K-PLS for QSAR: A nonlinearSVM-like approach to PLS. Abstracts of Papers of the American ChemicalSociety.2003.225:U777-U777.
[59] Wang, X., Y. Dong, S. Xu, L. Wang, and S. Han. Quantitativestructure-activity relationships for the toxicity to the tadpole Rana japonica ofselected phenols. Bulletin of Environmental Contamination and Toxicology.2000.64(6):859-865.
[60] Wang, X., Y. Dong, L. Wang, and S. Han. Acute toxicity of substitutedphenols to Rana japonica tadpoles and mechanism-based quantitativestructure-activity relationship (QSAR) study. Chemosphere.2001.44(3):447-455.
[61] Huang, H., X. Wang, W. Ou, J. Zhao, Y. Shao, and L. Wang. Acutetoxicity of benzene derivatives to the tadpoles (Rana japonica) and QSARanalyses. Chemosphere.2003.53(8):963-970.
[62] Huang, H., X. Wang, Y. Shao, D. Chen, X. Dai, and L. Wang. QSARfor prediction of joint toxicity of substituted phenols to tadpoles (Rana japonica).Bulletin of Environmental Contamination and Toxicology.2003.71(6):1124-1130.
[63] Gramatica, P. Principles of QSAR models validation: internal andexternal. Qsar&Combinatorial Science.2007.26(5):694-701.
[64] Ding, C. and H.C. Peng. Minimum redundancy feature selection frommicroarray gene expression data. Proceedings of the2003Ieee BioinformaticsConference.2003:523-528.
[65] Shen, Q., W.M. Shi, X.P. Yang, and B.X. Ye. Particle swarm algorithmtrained neural network for QSAR studies of inhibitors of platelet-derived growthfactor receptor phosphorylation. European Journal Of Pharmaceutical Sciences.2006.28(5):369-376.
[66] Brits, R., A.P. Engelbrecht, and F. van den Bergh. Locating multipleoptima using particle swarm optimization. Applied Mathematics andComputation.2007.189(2):1859-1883.
[67] Liu, B., L. Wang, and Y.-H. Jin. An effective hybrid PSO-basedalgorithm for flow shop scheduling with limited buffers. Computers&Operations Research.2008.35(9):2791-2806.
[68] Pedrycz, W., B.J. Park, and N.J. Pizzi. Identifying core sets ofdiscriminatory features using particle swarm optimization. Expert Systems withApplications.2009.36(3, Part1):4610-4616.
[69] Larranaga, P., H. Karshenas, C. Bielza, and R. Santana. A review onevolutionary algorithms in Bayesian network learning and inference tasks.Information Sciences.2013.233:109-125.
[70] Ibezim, E., P.R. Duchowicz, E.V. Ortiz, and E.A. Castro. QSAR onaryl-piperazine derivatives with activity on malaria. Chemometrics andIntelligent Laboratory Systems.2012.110(1):81-88.
[71] Toropova, A.P., A.A. Toropov, A. Lombardo, A. Roncaglioni, E.Benfenati, and G. Gini. CORAL: QSAR Models for Acute Toxicity in FatheadMinnow (Pimephales promelas). Journal of Computational Chemistry.2012.33(12):1218-1223.
[72] Toropova, A.P., A.A. Toropov, S.E. Martyanov, E. Benfenati, G. Gini,D. Leszczynska, et al. CORAL: QSAR modeling of toxicity of organic chemicalstowards Daphnia magna. Chemometrics and Intelligent Laboratory Systems.2012.110(1):177-181.
[73] Liu, J.N., D.L. Fan, L.L. Shi, L. Wang, L.J. Zhou, Y.W. Chen, et al.2Dand3D-QSAR study on the fish embryo Toxicity test of EC50. Research Journalof Chemistry and Environment.2013.17(5):52-59.
[74] Qiu, J., Y. Dai, X.S. Zhang, and G.S. Chen. QSAR Modeling of Toxicityof Acyclic Quaternary Ammonium Compounds on Scenedesmus QuadricaudaUsing2D and3D Descriptors. Bulletin of Environmental Contamination andToxicology.2013.91(1):83-88.
[75] Joshi, V.Y., M.M. Kadam, and M.R. Sawant. Comparison of QSARand QSPR based aquatic toxicity for mixed surfactants. Journal of Surfactantsand Detergents.2007.10(1):25-34.
[76] Colombo, A., E. Benfenati, M. Karelson, and U. Maran. The proposalof architecture for chemical splitting to optimize QSAR models for aquatictoxicity. Chemosphere.2008.72(5):772-780.
[77] Zvinavashe, E., H. van den Berg, A.E.M.F. Soffers, J. Vervoort, A.Freidig, A.J. Murk, et al. QSAR models for predicting in vivo aquatic toxicity ofchlorinated alkanes to fish. Chemical Research in Toxicology.2008.21(3):739-745.
[78]Polishchuk, P.G., E.N. Muratov, A.G. Artemenko, O.G. Kolumbin, N.N.Muratov, and V.E. Kuz'min. Application of Random Forest Approach to QSARPrediction of Aquatic Toxicity. Journal of Chemical Information and Modeling.2009.49(11):2481-2488.
[79] Lanevskij, K., L. Juska, R. Didziapetris, and P. Japertas. Novel QSARmodels for predicting toxicity of chemicals to aquatic organisms and identifyingthe mode of action. Abstracts of Papers of the American Chemical Society.2011.241.
[80] Hernandez, N., R. Kiralj, M.M.C. Ferreira, and I. Talavera. Criticalcomparative analysis, validation and interpretation of SVM and PLS regressionmodels in a QSAR study on HIV-1protease inhibitors. Chemometrics andIntelligent Laboratory Systems.2009.98(1):65-77.
[81] Norinder, U. and M.E. Ek. QSAR investigation of NaV1.7activecompounds using the SVM/Signature approach and the Bioclipse Modelingplatform. Bioorganic&Medicinal Chemistry Letters.2013.23(1):261-263.
[82] Wang, Z.M., N. Han, Z.M. Yuan, and Z.H. Wu. Feature Selection forHigh-Dimensional Data Based on Ridge Regression and SVM and ItsApplication in Peptide QSAR Modeling. Acta Physico-Chimica Sinica.2013.29(3):498-507.
[83]Abraham, M.H. and C. Rafols. Factors that influence tadpole narcosis.An LFER analysis. Journal of the Chemical Society, Perkin Transactions2.1995(10):1843-1851.
[84] Agrawal, V.K., S. Chaturvedi, M.H. Abraham, and P.V. Khadikar.QSAR Study on tadpole narcosis. Bioorganic&Medicinal Chemistry.2003.11(20):4523-4533.
[85] Bowen, K.R., K.B. Flanagan, W.E. Acree Jr, M.H. Abraham, and C.Rafols. Correlation of the toxicity of organic compounds to tadpoles using theAbraham model. Science of the Total Environment.2006.371(1–3):99-109.
[86] Hall, M.A. and L.A. Smith. Feature selection for machine learning:comparing a correlation-based filter approach to the wrapper. Proceedings ofthe Twelfth International Florida AI Research Society Conference|Proceedingsof the Twelfth International Florida AI Research Society Conference.1999:235-9|xx+540.
[87] QSAR+descriptors]. Available from:http://www.ifm.liu.se/compchem/msi/doc/life/cerius46/qsar/theory_descriptors.html.
[88] Medven, Z., H. Gusten, and A. Sabljic. Comparative QSAR study onhydroxyl radical reactivity with unsaturated hydrocarbons: PLS versus MLR.Journal of Chemometrics.1996.10(2):135-147.
[89] Hemmateenejad, B., R. Miri, M. Akhond, and M. Shamsipur. QSARstudy of the calcium channel antagonist activity of some recently synthesizeddihydropyridine derivatives. An application of genetic algorithm for variableselection in MLR and PLS methods. Chemometrics and Intelligent LaboratorySystems.2002.64(1):91-99.
[90] Shen, Q., J.H. Jiang, G.L. Shen, and R.Q. Yu. Variable selection by anevolution algorithm using modified Cp based on MLR and PLS modeling: QSARstudies of carcinogenicity of aromatic amines. Analytical and BioanalyticalChemistry.2003.375(2):248-254.
[91] He, M., T.H. Li, and P.S. Cong. Nonlinear PLS improved by numericgenetic algorithm for QSAR modelling. Chemical Journal of ChineseUniversities-Chinese.1997.18(6):854-859.
[92] Devillers, J., A. Chezeau, and E. Thybaud. PLS-QSAR of the adult anddevelopmental toxicity of chemicals to Hydra attenuata. Sar and Qsar inEnvironmental Research.2002.13(7-8):705-712.
[93] Leonard, J.T. and K. Roy. QSAR by LFER model of HIV proteaseinhibitor mannitol derivatives using FA-MLR, PCRA, and PLS techniques.Bioorganic&Medicinal Chemistry.2006.14(4):1039-1046.
[94] Avery, M.A. and N. Sapre.2D and3d QSAr modeling of anti-HIV-IITIBO derivatives: A comparative study using ANN and multiple linearregression methods. Abstracts of Papers of the American Chemical Society.2003.225:U753-U754.
[95] Su, Q. and L. Zhou. QSAR modeling of AT1receptor antagonists usingANN. Journal of Molecular Modeling.2006.12(6):869-875.
[96] Deeb, O. and B. Hemmateenejad. ANN-QSAR model of drug-bindingto human serum albumin. Chemical Biology&Drug Design.2007.70(1):19-29.
[97] Gonzalez-Diaz, H., I. Bonet, C. Teran, E. De Clercq, R. Bello, M.M.Garcia, et al. ANN-QSAR model for selection of anticancer leads fromstructurally heterogeneous series of compounds. European Journal of MedicinalChemistry.2007.42(5):580-585.
[98] Goodarzi, M., M.P. Freitas, and R. Jensen. Ant colony optimization asa feature selection method in the QSAR modeling of anti-HIV-1activities of3-(3,5-dimethylbenzyl)uracil derivatives using MLR, PLS and SVM regressions.Chemometrics and Intelligent Laboratory Systems.2009.98(2):123-129.
[99] Katritzky, A.R., Y. Ren, S.H. Slavov, and M. Karelson. A ComparativeQsar Study of Svm and Ppr in the Correlation of Lithium Cation Basicities.Collection of Czechoslovak Chemical Communications.2009.74(1):217-241.
[100] Caruana, R., A. Niculescu-Mizil, G. Crew, and A. Ksikes,Ensemble selection from libraries of models, in Proceedings of the twenty-firstinternational conference on Machine learning.2004, ACM: Banff, Alberta,Canada. p.18.
[101] Schultz, T.W. Structure-toxicity relationships for benzenesevaluated with Tetrahymena pyriformis. Chemical Research in Toxicology.1999.12(12):1262-1267.
[102] Cronin, M.T.D., N. Manga, J.R. Seward, G.D. Sinks, and T.W.Schultz. Parametrization of electrophilicity for the prediction of the toxicity ofaromatic compounds. Chemical Research in Toxicology.2001.14(11):1498-1505.
[103] Cronin, M.T.D. and T.W. Schultz. Development of quantitativestructure-activity relationships for the toxicity of aromatic compounds toTetrahymena pyriformis: Comparative assessment of the methodologies.Chemical Research in Toxicology.2001.14(9):1284-1295.
[104] Serra, J.R., P.C. Jurs, and K.L.E. Kaiser. Linear regression andcomputational neural network prediction of Tetrahymena acute toxicity foraromatic compounds from molecular structure. Chemical Research inToxicology.2001.14(11):1535-1545.
[105] Cronin, M.T.D., A.O. Aptula, J.C. Duffy, T.I. Netzeva, P.H. Rowe,I.V. Valkova, et al. Comparative assessment of methods to develop QSARs forthe prediction of the toxicity of phenols to Tetrahymena pyriformis.Chemosphere.2002.49(10):1201-1221.
[106] Wang, X.D., C. Sun, Y. Wang, and L.S. Wang. Quantitativestructure-activity relationships for the inhibition toxicity to root elongation ofCucumis sativus of selected phenols and interspecies correlation withTetrahymena pyriformis. Chemosphere.2002.46(2):153-161.
[107] Cronin, M.T.D., T.I. Netzeva, J.C. Dearden, R. Edwards, andA.D.P. Worgan. Assessment and modeling of the toxicity of organic chemicals toChlorella vulgaris: Development of a novel database. Chemical Research inToxicology.2004.17(4):545-554.
[108] Aptula, A.O., D.W. Roberts, M.T.D. Cronin, and T.W. Schultz.Chemistry-toxicity relationships for the effects of Di-and trihydroxybenzenes toTetrahymena pyriformis. Chemical Research in Toxicology.2005.18(5):844-854.
[109] Netzeva, T.I. and T.W. Schultz. QSARs for the aquatic toxicity ofaromatic aldehydes from Tetrahymena data. Chemosphere.2005.61(11):1632-1643.