An ICA-based multivariate discretization algorithm

Ye Kang; Shanshan Wang; Xiaoyan Liu; Hokyin Lai; Huaiqing Wang; Baiqi Miao

doi:10.1007/11811220_47

An ICA-based multivariate discretization algorithm

Ye Kang
, Shanshan Wang
, Xiaoyan Liu
, Hokyin Lai
, Huaiqing Wang
, Baiqi Miao

Department of Computer Science

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

7 Citations (Scopus)

Abstract

Discretization is an important preprocessing technique in data mining tasks. Univariate Discretization is the most commonly used method. It discretizes only one single attribute of a dataset at a time, without considering the interaction information with other attributes. Since it is multi-attribute rather than one single attribute determines the targeted class attribute, the result of Univariate Discretization is not optimal. In this paper, a new Multivariate Discretization algorithm is proposed. It uses ICA (Independent Component Analysis) to transform the original attributes into an independent attribute space, and then apply Univariate Discretization to each attribute in the new space. Data mining tasks can be conducted in the new discretized dataset with independent attributes. Themumerical experiment results show that our method improves the discretization performance, especially for the nongaussian datasets, and it is competent compared to PCA-based multivariate method.

Original language	English
Title of host publication	Knowledge Science, Engineering and Management - First International Conference, KSEM 2006, Proceedings
Publisher	Springer Berlin Heidelberg
Pages	556-562
Number of pages	7
ISBN (Print)	3540370331, 9783540370338
DOIs	https://doi.org/10.1007/11811220_47
Publication status	Published - 25 Jul 2006
Event	1st International Conference on Knowledge Science, Engineering and Management - Guilin, China Duration: 5 Aug 2006 → 8 Aug 2006 https://link.springer.com/book/10.1007/11811220 (Conference proceeding)

Publication series

Name	Lecture Notes in Computer Science
Volume	4092
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Name	Lecture Notes in Artificial Intelligence
ISSN (Print)	2945-9133
ISSN (Electronic)	2945-9141

Name	KSEM: International Conference on Knowledge Science, Engineering and Management

Conference

Conference	1st International Conference on Knowledge Science, Engineering and Management
Abbreviated title	KSEM 2006
Country/Territory	China
City	Guilin
Period	5/08/06 → 8/08/06
Internet address	https://link.springer.com/book/10.1007/11811220 (Conference proceeding)

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 9 Industry, Innovation, and Infrastructure

User-Defined Keywords

Data mining
Independent Component Analysis
Multivariate Discretization
Nongaussian

Access to Document

10.1007/11811220_47

Cite this

Kang, Y., Wang, S., Liu, X., Lai, H., Wang, H., & Miao, B. (2006). An ICA-based multivariate discretization algorithm. In Knowledge Science, Engineering and Management - First International Conference, KSEM 2006, Proceedings (pp. 556-562). (Lecture Notes in Computer Science; Vol. 4092), (Lecture Notes in Artificial Intelligence), (KSEM: International Conference on Knowledge Science, Engineering and Management). Springer Berlin Heidelberg. https://doi.org/10.1007/11811220_47

Kang, Ye ; Wang, Shanshan ; Liu, Xiaoyan et al. / An ICA-based multivariate discretization algorithm. Knowledge Science, Engineering and Management - First International Conference, KSEM 2006, Proceedings. Springer Berlin Heidelberg, 2006. pp. 556-562 (Lecture Notes in Computer Science). (Lecture Notes in Artificial Intelligence). (KSEM: International Conference on Knowledge Science, Engineering and Management).

@inproceedings{fb3429e3e5fd4fb097bc69d8e91e75a8,

title = "An ICA-based multivariate discretization algorithm",

abstract = "Discretization is an important preprocessing technique in data mining tasks. Univariate Discretization is the most commonly used method. It discretizes only one single attribute of a dataset at a time, without considering the interaction information with other attributes. Since it is multi-attribute rather than one single attribute determines the targeted class attribute, the result of Univariate Discretization is not optimal. In this paper, a new Multivariate Discretization algorithm is proposed. It uses ICA (Independent Component Analysis) to transform the original attributes into an independent attribute space, and then apply Univariate Discretization to each attribute in the new space. Data mining tasks can be conducted in the new discretized dataset with independent attributes. Themumerical experiment results show that our method improves the discretization performance, especially for the nongaussian datasets, and it is competent compared to PCA-based multivariate method.",

keywords = "Data mining, Independent Component Analysis, Multivariate Discretization, Nongaussian",

author = "Ye Kang and Shanshan Wang and Xiaoyan Liu and Hokyin Lai and Huaiqing Wang and Baiqi Miao",

note = "Supported by a SRG Grant (7001805) from the City University of Hong Kong Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 2006; 1st International Conference on Knowledge Science, Engineering and Management, KSEM 2006 ; Conference date: 05-08-2006 Through 08-08-2006",

year = "2006",

month = jul,

day = "25",

doi = "10.1007/11811220\_47",

language = "English",

isbn = "3540370331",

series = "Lecture Notes in Computer Science",

publisher = "Springer Berlin Heidelberg",

pages = "556--562",

booktitle = "Knowledge Science, Engineering and Management - First International Conference, KSEM 2006, Proceedings",

url = "https://link.springer.com/book/10.1007/11811220",

}

Kang, Y, Wang, S, Liu, X, Lai, H, Wang, H & Miao, B 2006, An ICA-based multivariate discretization algorithm. in Knowledge Science, Engineering and Management - First International Conference, KSEM 2006, Proceedings. Lecture Notes in Computer Science, vol. 4092, Lecture Notes in Artificial Intelligence, KSEM: International Conference on Knowledge Science, Engineering and Management, Springer Berlin Heidelberg, pp. 556-562, 1st International Conference on Knowledge Science, Engineering and Management, Guilin, China, 5/08/06. https://doi.org/10.1007/11811220_47

An ICA-based multivariate discretization algorithm. / Kang, Ye; Wang, Shanshan; Liu, Xiaoyan et al.
Knowledge Science, Engineering and Management - First International Conference, KSEM 2006, Proceedings. Springer Berlin Heidelberg, 2006. p. 556-562 (Lecture Notes in Computer Science; Vol. 4092), (Lecture Notes in Artificial Intelligence), (KSEM: International Conference on Knowledge Science, Engineering and Management).

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

TY - GEN

T1 - An ICA-based multivariate discretization algorithm

AU - Kang, Ye

AU - Wang, Shanshan

AU - Liu, Xiaoyan

AU - Lai, Hokyin

AU - Wang, Huaiqing

AU - Miao, Baiqi

PY - 2006/7/25

Y1 - 2006/7/25

N2 - Discretization is an important preprocessing technique in data mining tasks. Univariate Discretization is the most commonly used method. It discretizes only one single attribute of a dataset at a time, without considering the interaction information with other attributes. Since it is multi-attribute rather than one single attribute determines the targeted class attribute, the result of Univariate Discretization is not optimal. In this paper, a new Multivariate Discretization algorithm is proposed. It uses ICA (Independent Component Analysis) to transform the original attributes into an independent attribute space, and then apply Univariate Discretization to each attribute in the new space. Data mining tasks can be conducted in the new discretized dataset with independent attributes. Themumerical experiment results show that our method improves the discretization performance, especially for the nongaussian datasets, and it is competent compared to PCA-based multivariate method.

AB - Discretization is an important preprocessing technique in data mining tasks. Univariate Discretization is the most commonly used method. It discretizes only one single attribute of a dataset at a time, without considering the interaction information with other attributes. Since it is multi-attribute rather than one single attribute determines the targeted class attribute, the result of Univariate Discretization is not optimal. In this paper, a new Multivariate Discretization algorithm is proposed. It uses ICA (Independent Component Analysis) to transform the original attributes into an independent attribute space, and then apply Univariate Discretization to each attribute in the new space. Data mining tasks can be conducted in the new discretized dataset with independent attributes. Themumerical experiment results show that our method improves the discretization performance, especially for the nongaussian datasets, and it is competent compared to PCA-based multivariate method.

KW - Data mining

KW - Independent Component Analysis

KW - Multivariate Discretization

KW - Nongaussian

UR - https://www.scopus.com/pages/publications/33749402115

UR - https://link.springer.com/chapter/10.1007/11811220_47#Abs1

U2 - 10.1007/11811220_47

DO - 10.1007/11811220_47

M3 - Conference proceeding

AN - SCOPUS:33749402115

SN - 3540370331

SN - 9783540370338

T3 - Lecture Notes in Computer Science

SP - 556

EP - 562

BT - Knowledge Science, Engineering and Management - First International Conference, KSEM 2006, Proceedings

PB - Springer Berlin Heidelberg

T2 - 1st International Conference on Knowledge Science, Engineering and Management

Y2 - 5 August 2006 through 8 August 2006

ER -

Kang Y, Wang S, Liu X, Lai H, Wang H, Miao B. An ICA-based multivariate discretization algorithm. In Knowledge Science, Engineering and Management - First International Conference, KSEM 2006, Proceedings. Springer Berlin Heidelberg. 2006. p. 556-562. (Lecture Notes in Computer Science). (Lecture Notes in Artificial Intelligence). (KSEM: International Conference on Knowledge Science, Engineering and Management). doi: 10.1007/11811220_47

An ICA-based multivariate discretization algorithm

Abstract

Publication series

Conference

UN SDGs

User-Defined Keywords

Access to Document

Other files and links

Fingerprint

Cite this