[1]乐耀佳,许建华.一种高维数据集的子空间聚类算法[J].南京师范大学学报(工程技术版),2009,09(03):055-63.
 Yue Yaojia,Xu Jianhua.A Subspace Clustering Algorithm for High-dimensional Data[J].Journal of Nanjing Normal University(Engineering and Technology),2009,09(03):055-63.
点击复制

一种高维数据集的子空间聚类算法
分享到:

南京师范大学学报(工程技术版)[ISSN:1006-6977/CN:61-1281/TN]

卷:
09卷
期数:
2009年03期
页码:
055-63
栏目:
出版日期:
2009-09-30

文章信息/Info

Title:
A Subspace Clustering Algorithm for High-dimensional Data
作者:
乐耀佳;许建华;
南京师范大学计算机科学与技术学院, 江苏南京210097
Author(s):
Yue YaojiaXu Jianhua
School of Computer Science and Technology,Nanjing Normal University,Nanjing 210097,China
关键词:
聚类算法 子空间聚类 基因芯片
Keywords:
c lustering a lgo rithm subspace c luster ing gene CMOS ch ip
分类号:
TP18
摘要:
提出了一个基于密度和网格的子空间聚类算法.该算法运用启发式的密度连通思想来确定一维空间初始簇的生成,使用自底向上的搜索策略来发现存在子空间中的簇.实验结果表明,在处理高维数据时,在不牺牲算法的其他性能的同时提高了聚类的有效性,降低了对输入数据顺序及噪音数据的敏感性.
Abstract:
A new subspace cluster ing algorithm based on gr id and density is proposed in th is paper. The a lgo rithm m akes use o f heuristic density-connected idea to genera te the initial c lusters in the first dim ension, and applies bottomup strategy to search the subspace c luste rs. W ith the exper im ents on rea-l wo rld g ene expression da ta, the resu lts show tha t our a lgo rithm is effective w ithout sacr ific ing othe r perform ances and reduces the sensitiv ity to the da ta order and to the noise data in dealing w ith h igh-d im ensiona l data.

参考文献/References:

[ 1] Agraw a l R, Gehrke J, Gunopulos D. Autom a tic subspace cluster ing of h igh d im ensiona l data fo r data m ining app lications
[ C] / / Proceed ings o f the 1998 ACM-SIGMOD Interna tiona l Conference onM anagement o f Data. Sea ttle, W ash ing ton: ACM
Press, 1998, 6: 94-105.
[ 2] G runwa ld P D. M ode l selection based on m in imum description length[ J]. Journal ofM athem atica l Psycho logy, 2000, 44:
133-152.
[ 3] Cheng C H, Fu A W-C, ZhangY. Entropy-based subspace c luste ring for m in ing num er ica l data[ C] / / Proceed ings of the 5 th
ACM SIGKDD Interna tiona l Conference on Know ledge D iscov ery and DataM in ing. San D iego, USA: ACM Press, 1999: 84-
93.
[ 4] Nagesh H, Go il S, Choudhary A. Adaptive g rids for c lusteringm assive da ta sets[ C] / / Proceed ing s o f SIAM Internationa lConference
on DataM in ing. SIAM, 2001: 477- 493.
[ 5] Ka iling K, K riege lH-P, Kr-ge r P. Density-connected subspace cluster ing for high-d imensional data[ C ] / / Proceed ings of the
4th SIAM In ternational Conference on DataM ining. Lake Buena V ia ta, FL, 2004: 246-257.
[ 6] Procopiuc CM, JonesM, Aga rw al P K, et a.l A monte car lo a lgo rithm fo r fast pro jective c luster ing [ C ] / / Proceed ings o f the
2002 ACM SIGMOD In ternational Conference onM anag em ent of Data. M ad ison: ACM Press, 2002: 418- 427.
[ 7] G lom baM, Urszu laM-K. IBUSCA: a gr id-based bo ttom-up subspace cluster ing algor ithm [ C] / / Proceed ing s of the 6 th International
Conference on Inte lligent Sy stem s Design and Applications. USA: IEEE Compu ter Soc iety, 2006: 671-676.
[ 8] Agga rwa l C C, Procop iucC, Wo lf J L, et a.l Fast a lgor ithm s for projected c lustering[ C] / / Proceed ings o f the 1999 ACM SIG-MOD International Confe rence onM anagem en t of Data. New York: ACM Press, 1999: 61-72.
[ 9] Aggarw a lC C, P S Yu. F ind ing genera lized pro jected clusters in h igh d im ensiona l spaces[ C] / / Proceed ing s of the 2000 ACM
SIGMOD inte rnational con ference onM anagem ent of data. Oallas, Texas: ACM Press, 2000: 70-81.
[ 10] L iu J, Strohma ierK, W angW. Revea ling true subspace clusters in h igh dim ensions[ C] / / Pro ceedings o f the 4th IEEE Internationa
l Conference on Da taM ining. USA: IEEE Com puter So ciety, 2004: 463- 466.
[ 11] UCIM ach ine Learn ing Ropository[ EB /OL] . http: / /a rchive. ics. uc.i edu /m l/m ach ine- lea rning-databases/breast- cancer-w iscons
in /.
[ 12] UCIM ach ine Learn ing Ropository[ EB /OL] . http: / /a rchive. ics. uc.i edu /m l/m ach ine- lea rning-databases/ ir is /.
[ 13] UCIM ach ine Learn ing Ropository[ EB /OL] . http: / /a rchive. ics. uc.i edu /m l/m ach ine- lea rning-databases/yeast/.
[ 14] H e J, LanM, Tan C L, e t a.l In itia lization of c luster re finem ent a lgor ithm s: a rev iew and com pa rative study[ C] / / Proceedings
o f IEEE Internationa l Jo int Conference on Neural Netwo rks. USA: IEEE Com puter So ciety, 2004: 297-302.
[ 15] Bohm C, Ka iling K, Kr iege lH P, e t a.l Density connected cluster ing w ith local subspace prefe rences[ C ] / / Proceed ings o f
the 4th IEEE Interna tiona l Conference on Da taM ining. USA: IEEE Com puter Soc iety, 2004: 27-34.

备注/Memo

备注/Memo:
通讯联系人: 许建华, 教授, 研究方向: 模式识别、神经网络、机器学习、信号处理等. E-ma il: xu jianhua@ n jnu. edu. cn
更新日期/Last Update: 2013-04-23