《基于氨基酸分子中侧链基极性性质的DNA序列分类》.pdfVIP

752 2014 besolvedthrough modeling：first．to method~”．FISHER method’【，E ，step- Symbolagreement extractthefeaturesofclassA andB： wisediscriminantanalysis[7)；correla- X isfeaturevectorofthe DNA second．to dividethe20 artificialse- tion analysis【6I：comparative analysis sequencefragment； quencestoclassA and B basedon method[1]_fuzzyclusteringmethod㈣ ： x is the percentage ofthe thefeaturesoftheextractedfeatures AHP[51；BPartificiaIneuraInetworkme· aminoacidcontentinthefeaturevec- ofthetwoclasses (theymaycontain thod’【-2】：waveletanalysisandsupport torofthe 『lhDNAsequencefragment． the sequences with the features of vectormachinemethod[11】． equivalenttothek ‘componentofthe neitherclassA norclassB，therefore， Throughoutthesedocuments，few featurevector； itcannotbeclassifiedinanyclass)． havetakenthebiologicalknowledgeof G／istheclassmadeupofthein— Consulting the existing Iiterature feature extraction into consideration． dividualswiththesameproperties． using www．cnki．corn，we found that andtheapplicationofclusteringanaly— Backgroundbiologicalknowledge feature extractionmainly focusingon sisjslimited to fuzzy clustering．while lt is thoughtthat a DNA se- thefollowingaspects： there are even Iess thathave fully quencefragmentreflectsthetwomost Firstis the occurrencefrequen． combined biologicaJknowledge with importantaspects of the sequence ciesofthefourkindsofbases，fOrmina mathematicaImodeling．1nthispaper． feature[13-16】：oneisthecontentofbas- the four-dimensionaIor threedimen． wedividedthe20kindsofaminoacids es，and the otheristhe marshalling sionalfeaturevector[’捌． into4classesbasedonthepolarityof sequence ofthe bases．Many re— Secondiskeywords．includingsin． the side chain radicaIofthe amino searches considerthe 2 separately， g