Concept learning and feature selection based on square-error clustering

被引：43

作者：

Mirkin, B

机构：

[1] Rutgers State Univ, Ctr Discrete Math & Theoret Comp Sci, DIMACS, Piscataway, NJ 08854 USA

[2] Cent Econ Math Inst, Moscow, Russia

来源：

MACHINE LEARNING | 1999年 / 35卷 / 01期

关键词：

clustering; variable weights; conjunctive concepts; feature selection; feature space transformation;

D O I：

10.1023/A:1007567018844

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 [模式识别与智能系统]; 0812 [计算机科学与技术]; 0835 [软件工程]; 1405 [智能科学与技术];

摘要：

Based on a reinterpretation of the square-error criterion for classical clustering, a "separate-and-conquer" version of K-Means clustering is presented and a contribution weight is determined for each variable of every cluster. The weight is used to produce conjunctive concepts that describe clusters and to reduce or transform the variable (feature) space.

引用

页码：25 / 39

页数：15

共 20 条

[1]

AHA DW, 1995, LEARNING DATA AI STA

[2]

[Anonymous], TAXONOMY BEHAV SCI

[3]

CRC press, 1987, MULTIVARIATE ANAL VA

[4]

DEVANEY M, 1997, P 14 INT C MACH LEAR

[5]

FAYYAD UM, 1996, ADV KNOWLEDGE DISCOV

[6]

Iterative optimization and simplification of hierarchical clusterings [J].

Fisher, D .

JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH, 1996, 4 :147-179

[7]

APPLYING AL CLUSTERING TO ENGINEERING TASKS [J].

FISHER, D ;

XU, L ;

CARNES, JR ;

REICH, Y ;

FENVES, SJ ;

CHEN, J ;

SHIAVI, R ;

BISWAS, G ;

WEINBERG, J .

IEEE EXPERT-INTELLIGENT SYSTEMS & THEIR APPLICATIONS, 1993, 8 (06) :51-60

[8]

GENNARI JH, 1989, P 6 INT WORKSH MACH, P379

[9]

Hanson R., 1991, P 12 INT JOINT C ART, P692

[10]

Jain K, 1988, Algorithms for clustering data

← 1 2 →