聚類和分類技術(shù)在生物信息學中的應(yīng)用.pdf

1、多前薩力學位論文聚類和分類技術(shù)在生物信息學中的應(yīng)用黃金自日￥m∞自I《撇業(yè)砭方嚏苴蒯蝻劇熹扎塑熊mi握女目自嬲蔓j用21旦口女##日M；腳年‘月t縣剴強●位目日琨曼萎耍橫磐o!篝旦9#￡e$i盎笨皴鋤臥外文摘要Abstract1、housandsuponthousandsbiologydatahavebeenachieVedbybi0109yexperimentsHowtocollect，cleanup，searchandanalyze

2、datae伍caciously，howtopickrulesfmmdata，whicharea11wemustresolVeDataminingisanewtechn0109y，whichisbasedupondatabase，statistjcsandani6cialintelligenceDataminingisausemlandpowerfultooltobiologistInthep印er，wemainlyresearchgen

3、eexpressionandproteinsequencedataWeprovideamethodofpmteinsequencesclassinc“onrWedesignametllodtominecontinuousfrequentpattemsclassifytestdataisbasedonthcse行equentpattemsWedesignamethodofclusteringproteinsequencesInt11ism

4、ethod，weminecontinuous仔equentpatternsatnrst，t11encutsomeoffkquentpattemsandusethembuildf色aturespace，andthenthedatasequencesareprojectedintothefeaturespace，buildsjmilarmatrixofsequences，at1ast，wecondllctclusteringonsimila

5、rmatrixusingk—meanstofirldkclustersWepmposeamethodofgeneexpressionclass湎c“onInthismethod，firstly，wecutgenebyexpectationandVarianceofg芒ne，aJldthentumgeneexpressionintoPtreestructure，atlastweuseP。treeStruc札lretocalculatein

6、formationgainandbuildmulti—decisiontreesWedesignaparallelalgoritmtoclustergeneexpression1nthisalgorithm，wediVidedataimoseveralgroupsa11dpostthemtoSerVerSTheneVerySerVercalculatedata’sdensityandacquirecoregenes，weconductc