资源描述：
《machine learning techniques机器学习技术》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。
1、www.nature.com/npjcompumatsARTICLEOPENChemicallyintuited,large-scalescreeningofMOFsbymachinelearningtechniques1,23331,2,4GiorgosBorboudakis,TaxiarchisStergiannakos,MariaFrysali,EmmanuelKlontzas,IoannisTsamardinosand3GeorgeE.FroudakisAnovelcomputationalmethodologyforl
2、arge-scalescreeningofMOFsisappliedtogasstoragewiththeuseofmachinelearningtechnologies.Thisapproachisapromisingtrade-offbetweentheaccuracyofabinitiomethodsandthespeedofclassicalapproaches,strategicallycombinedwithchemicalintuition.Theresultsdemonstratethatthechemicalp
3、ropertiesofMOFsareindeedpredictable(stochastically,notdeterministically)usingmachinelearningmethodsandautomatedanalysisprotocols,withtheaccuracyofpredictionsincreasingwithsamplesize.OurinitialresultsindicatethatthismethodologyispromisingtoapplynotonlytogasstorageinMO
4、Fsbutinmanyothermaterialscienceprojects.npjComputationalMaterials(2017)3:40;doi:10.1038/s41524-017-0045-8INTRODUCTIONsignificantroleinthedevelopmentofthefield,mainlyintwo20Metal–organicframeworks(MOFs)orporouscoordinationpoly-ways:byexplainingtheexperimentalresultsandb
5、yleadingthe21mersarearapidlygrowingfamilyofhybridinorganic–organicexperiments.Intheliterature,thereareseveralmethodologiesnanoporousmaterials,whichbelongtothecategoryofcoordina-investigatingthegasstorageprobleminMOFs.Thereareaccuratetionpolymers.1–3Theserelativelynew
6、materialsconsistofathree-abinitioquantumchemicalapproaches,22computationallightdimensionalperiodicnetwork,constructedfrommolecularbuild-andfastclassicalMonteCarloandmoleculardynamicstechni-2322ingblocks,suchasmetalclustersandorganiclinkers(Fig.1).Thequesand“multi-sca
7、le”methodsthattrytocombineboth.AllpossiblecombinationsofthesenumerousbuildingblocksunderofthemaddressspecificMOFs,eithersynthesizedearlier,ordifferenttopologiesresultisanalmostunlimitednumberofdesignedforaspecificapplicationfollowingchemicalintuition.potentialMOFs!Late
8、ly,acompletelydifferentcomputationalapproachSincetheirdiscovery4MOFshaveattractedsignificantscientificappearedbasedonalarge-scalescre
