《基于算法适用知识的挖掘算法选择交互系统_朱恒民》由会员分享，可在线阅读，更多相关《基于算法适用知识的挖掘算法选择交互系统_朱恒民（4页珍藏版）》请在金锄头文库上搜索。
1、2862v(1S) Vol.28No.62008M12JournalofNanjingUniversityofPostsandTelecommunications(NaturalScience) Dec.2008cI|:1673-5439(2008)06-0065-04EaM!E4d朱恒民,黄卫东2v65,2210003K1:为具体挖掘任务选择合适的挖掘算法需要用户对挖掘任务、各种挖掘算法和数据特征都非常熟悉,一般用户是很难达到这个要求的。针对以往研究的系统存在实现困难、不能适应动态添加算法等不足,文中形式化表示数据挖掘算法的适用知识,并基于此设计算法选择交互问题和选择逻辑,实现了一个易于实现
2、的数据挖掘算法选择交互系统。实例验证了系统的有效性。1oM:算法选择;数据挖掘;算法适用知识;交互系统ms|:TP39DSM:AASystemofMiningAlgorithmSelectionBasedonAlgorithmSuitingKnowledgeZHUHeng-min, HUANGWei-dongColegeofEconomicsandManagement, NanjingUniversityofPostsandTelecommunications, Nanjing210003, ChinaAbstract:Appropriatealgorithmcouldbeselectedby
3、datamininguseronlywhenheisveryfamiliarwithmin-ingtask, alkindsofalgorithmsanddatafeatures.It sdificultforordinaryusertodoit.Aimingatshortcomingsofpastworks, algorithmsuitingknowledgeisrepresentedformaly.Basedonthealgorithm, interactivequestionsandse-lectionlogicaredesigned.Aneasy-to-implementedsyste
4、mofdataminingalgorithmselectionisrealized.Anexampleshowstheeficiencyofthealgorithm.Keywords:Algorithmselection;Datamining;Algorithmsuitingknowledge;Interactivesysteml:2007-11-14;:2008-05-27:8n1S$9(KJD520151)YT:(025)85866161E-mail:0!/5L;E%。P9?1)!EaM,V|Vr/T5M:ECissuitforA1 opa1;B1 opb1;,A1、B1aHq, a1、b
5、1sYHq, opT。aMB#3T:B“ =”,cl:HqA1Hqa1, A1 =a1VUECaA1 =a1,7aA1 =ai(i1)f;=“%=”,cl:A1 %=a1VUECaA1 =a1,TESBt),9?PaA1=ai(i1)f;“ S =”, A1 =a1VUECA1 =a1fa,A1 =ai(i1)f。, ID3EaMVVr:ID3 issuitforLVU=;cpS =;c8S =;E=%。3E54 !9E54 !9E4d,!9YdTrq。 CONSULTANTd 4-5E,sE54 !9,yNiLC4、?aFE。|EaM!954 。!E4d5#!、1p+3Z。,1!5e,Vvd
6、!9。11p+5!EaM7。55&VaMHq7。,aHq“LVU=”,V35“LVU?”。5!sMaHqHq7,MaHqHq!s。,5aMcaHqA,sYA=a1A=a2,5A5!sBa1,!662v(1S)2008Ms=a2。TaHqAHqa1, a2, , anbW,51aHqA9FB,a1 a2 an。1FEaMMT,MTiBHcEaMH,daHq,1XHq;MTXHqbW,V7%1Fa1a2an,F,5|a1a2 anVr“”、“”“”,V7ZLs5。dE4 N5:B,aMMl!E;=,SE。4 X:d!5ss!E,iN35;5,1sEiHq,BOTop“ =”,5i,E=-i,i5;
7、BOTop“%=”,5EHqS“M”。,518。dvS:“M”Vlvwi!E。8E4 E1U。E1Algorithm select():dEaM,!E1,5s;:dwia!E;(1)s1!5,dVs!S;(2)SE;!E17(3)i35T,id9a(Ti);a(Ti)35TiE,aMcTiMaHqE(4)foreach5Tido(5)sTi,sai;(6)foreachEjdo(7)ifjaMcTiMHqAkthen(8)ifai! =val(j, Ak)thenTsaijaMHq(9)if(op(j, Ak)=“ =”)thenjaMHqAkT“=”(10):=-j;VE1Ej(11)T:=T- a(Th)=jTh; a(Th)=jThVUjaMHqM5(12)foreach5Tkdo(13)if(j a(Tk)thena(Tk):=a(Tk)&j;M15TkE(14)elseif(op(j, Ak)=“%=”)thenmark(j, Ak):=“M”;1ES:(15)j:=j+1;(16)i:=i+1;(17)if(=)thenreturn;(18)elsed9ES:“M”Q,vQVlvwi!E。E1(1)(3)3E#M5;(7)(10)aMMl!E;(11)13)5,hA14;(14)S:TE
