Ӧо
APPLICATION RESEARCH OF COMPUTERS
2000 Vol.17 No.3 P.91-92




ذھ
衡
ժҪоϢϵרھذķģͣʵַ԰ʮҪʵ塣
ؼʣ֪ʶ ھ ݿ Ϣ
1
żķչݿ⼼Ĺ㷺Ӧãиҵ˴ݣЩԴı̲мֵ֪ʶݿзֲȡЩ֪ʶʮҪʵ塣ݿ֪ʶ(Knowledge Discovery in Database,KDD)ھо˵ǰݿоȵ֮һ
ھ(Data Mining)ǴӴݿȡġδ֪ġԾǱڼֵ֪ʶ͹Ĺ̡
ھõ֪ʶʾΪ(Conception)(Rules)(Regularities)ģʽ(Patterns)ȡ
ھڰϢеӦãϢָйְϢݿ⣬ĻۣĿǰڹҵȫݿ⣬Դɶáݵı̲ʵʹõ֪ʶھЩ֪ʶҪھ
ǹҵкҪĹһҪֹͨʽ԰ϽвѯٸݹȷһרҵǿӡĹĵĹͨھԶءܵھذЩذṩרңȷͿԴؽ͹߹Чʡ
2ذھ
ذֵָȷģͬʵİذ롣ڹҵнذ(ͬһﰸΪ)һ𣬶Է󰸡ڡͳƷߵȹкܴ
2.1ذĶ
ϿɳΪһϢϵͳԪʾ
S=<U, A, V, F >
УUһ()޼ϣΪnUɱʾΪU={u1, u2, ..., un}A޸Ե޼ϣmԣAɱʾΪA={a1, a2, ..., am}VԵֵ򼯣V={v1, v2, ..., vm}viaiֵFϢ(Information function)FUAVF(xi, aj)vj2.1 ԣUAUԼϣǸ֪ʶȷģбݶͬʵԼAӼеÿa(aA)ԡ
磺ֶΡԡ
2.2 ضȣaֵΪVa={Va1, Va2, ..., Van}һֵVaiһس̶ȵĶ {c1, c2, ..., cn} cjVajVaiس̶ȣVajVaiضȡ
磺ֶ={β棬ţð䣬Ǳ}ӦһضΪ{100.80.200}
2.3 ذ𰸼Եض֮CڸֵCp𰸼Ϊذ
磺ΪѡֶΣضΪ0.8ضֵΪ2.5𰸼
𡡡ѡֶ
1ҵ԰լ
2ҵ԰լ
𰸼ذ
2.2ذķ
壬ذҪ֪ԵضȣضȵֵרҸ֪ʶΪضȣضȾMa

CijaĵjԵiضȡ
MaһԳƾֻȡǼɡ
磺ֶεضȾΪ
β桡šð䡡Ǳ
100000
β桡010000
š0.80.51000
00.90100
ð䡡00.80.3010
Ǳ000001
ʱҪ֪ԵضȾԼضȾͿضȵĺͣݸֵCpȫذ
2.3ذ㷨
 ϢϵͳUn¼ԵضȾ
Ma1--Mam
 Lذļ
1) FOR i=1 TO n-1 (nΪ¼)
2) FOR j=i+1 TO n
3) A[i,j]=0
4) FOR k=1 TO m
5) A[i,j]=A[i,j]+Mak[i,j] (Mak[i,j]ؾȡض)
6) NEXT
7) IF A[i,j]>=Cp
8) 
9) ENDIF
10) NEXT
11) 
12) NEXT
13) L
1)ѭȡ2)Ƚϣ4)6)ض֮ͣ7)жǷأ11)ذ13)
3ʵ
ĵойְϢϵͳнʵ飬㷨ȷԼִЧʡ
Ϣݿйа100,671谸ص㡢ֶΪԣֵΪ2.5ھذ324飬ֶβ棬ҵԡɱˣص尮гذ6𣻵ͨѶģߵذ13...
ͨ㷨ҳذЩɹרҲοרҷۣͿɰֹݾ԰ϢݿѯҪ࣬ЧҲࡣ
4
ھϵͳ࣬ͨϵͳרϵͳͨϵͳӦǿ㷨ִЧʲߣרϵͳխ֪ʶ㷨ִЧʡھһԵǴģݿ⣬㷨ִЧʺҪĴھ֪ʶģֽ˷оִЧʽϸߵִ㷨
ǹҵһˮƽҪߡӵĹذھʵ֣˴ЧʡĿǰڰϢϵͳھӦհףⷽ̽ʵ塣
(ҵѧϢѧ빤ѧԺ  110023)
(ҵѧϢѧ빤ѧԺ  110023)
ο
1  , ռ, . ƹķ㷨. С΢ͼϵͳ, 1999.9
2  , , ռ. ھ. ҵѧѧ, 1999.9
3  . ݿ֪ʶֵо. йѧԺо, ʿ, 1997
4  . ھоӦ. ҵѧ, ˶ʿ, 1999
5 Han J, Conference Tutorial Notes: Data Mining Techniques, In Proceedings of ACM SIGMOD International Conference'96 on Management of Data (SIGMOD'96), Montreal, Canada, June, 1996
6 Fayyed U, Hausller D, Stolorz P. Mining Scientific Data, Comm. ACM 39(11), 1996
ոڣ1999922
