支持向量机对不同物种毒性的qsar分析-qsar analysis of toxicity of support vector machine to different species.docxVIP

下载本文档

2
0
约7.67万字
约 94页
2018-06-09 发布于上海
举报

支持向量机对不同物种毒性的qsar分析-qsar analysis of toxicity of support vector machine to different species.docx

支持向量机对不同物种毒性的qsar分析-qsar analysis of toxicity of support vector machine to different species

QSARStudyofSupportVectorMachinefortheDifferentSpeciesToxicityABSTRACTUsingQSARstudyofcompoundstructureanditstoxicitytoestablishtoxicitypredictionmodel,bywhichwaspredictedandevaluatedthetoxicityofnewcompoundenteringpeople’sbiologicaltoxicantandnotonthemarket,andhasdemonstratedextensiveapplicationinbiologicaltoxicity.Byusingofsupportvectormachineregressionandclassification,thetoxicityoffourvariousspecieswasinvestigated.Studysampleoffeaturein“smallsamplesandmulti-dimensional”datamodelwasadoptedleastsquaressupportvectormachine,andstudysampleoffeaturein“smallsamplesandpoorinformation”datamodelwasadoptedsquaressupportvectorclassification.Thequantumchemicalstructuredescriptorof30nitroaromaticcompoundsacutetoxicitytotetrahymenaperformsandgoldenorfefishwastotallyoptimizedandcalculatedwithB3LYPmethodat6-311G**level.Byvibrationanalysis,thestablegeometrieshasn’timaginaryfrequency.Thecalculatedresultsofseparatemorethan10speciesquantumchemicalstructuredescriptorsare:μ、EHOMO、ELUMO、ENHOMO、ENLUMO、ΔE=ELUMO-EHOMO、Q-NO2、QC-NO2、Vetc.,andcombinedwith30nitroaromaticcompoundsacutetoxicitytotetrahymenapyriformisandgoldenorfefishtoxicitytoconductQSARstudy.Byusingtheleastsquaressupportvectormachinetostudy30nitroaromaticcompoundsacutetoxicitytotetrahymenapyriformis.25nitroaromaticcompoundswastakentoconstructtrainings,5forpredictionones.Atthesametime,liststhestepwiseregression,partialleast-squareregressionandRBFmethodofsimulationresults,foundleastsquaressupportvectormachine(SVM)modelofoptimalperformance.Basedonthemodel,whichwasfoundeachfactorofaffectingcompoundstoxicity.Byusingtheleastsquaressupportvectormachinetostudy30nitroaromaticIIIcompoundsacutetoxicitytogoldenorfefish.23nitroaromaticcompoundswastakentoconstructtrainings,7nitroaromaticcompoundsforpredictionones.Atthesametime,liststhestepwiseregression,partialleast-squareregressionandRBFmethodofsimulationresults,foundleastsquaressupportvectormachine(SVM)modelofoptimalperformance.Basedonthismodel,eachfactorwhichaffectscompoundstoxicitywasfou

您可能关注的文档

文档评论（0）

1亿VIP精品文档

更多 >

支持向量机对不同物种毒性的qsar分析-qsar analysis of toxicity of support vector machine to different species.docxVIP