Subclass-oriented Dimension Reduction with constraint transformation and manifold regularization

Bin Tong, Einoshin Suzuki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

We propose a new method, called Subclass-oriented Dimension Reduction with Pairwise Constraints (SODRPaC), for dimension reduction on high dimensional data. Current linear semi-supervised dimension reduction methods using pairwise constraints, e.g., must-link constraints and cannot-link constraints, can not handle appropriately the data of multiple subclasses where the points of a class are separately distributed in different groups. To illustrate this problem, wparticularly classify the must-link constraint into two categories, which are theinter-subclass must-link constraint and the intra-subclass must-link constraint, respectively. We argue that handling the inter-subclass must-link constraint is challenging for current discriminant criteria. Inspired by the above observation and the cluster assumption that nearby points are possible in the same class, we carefully transform must-link constraints into cannot-link constraints, and then propose a new discriminant criterion by employing the cannot-link constraints and the compactness of shared nearest neighbors. For the reason that the local data structure is one of the most significant features for the data of multiple subclasses, manifold regularization is also incorporated in our dimension reduction framework. Extensive experiments on both synthetic and practical data sets illustrate the effectiveness of our method.

Original languageEnglish
Title of host publicationAdvances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings
Pages1-13
Number of pages13
EditionPART 2
DOIs
Publication statusPublished - Dec 1 2010
Event14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010 - Hyderabad, India
Duration: Jun 21 2010Jun 24 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume6119 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010
CountryIndia
CityHyderabad
Period6/21/106/24/10

Fingerprint

Dimension Reduction
Regularization
Data structures
Discriminant
Pairwise
Local Structure
High-dimensional Data
Experiments
Reduction Method
Compactness
Nearest Neighbor
Data Structures
Classify
Transform

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Tong, B., & Suzuki, E. (2010). Subclass-oriented Dimension Reduction with constraint transformation and manifold regularization. In Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings (PART 2 ed., pp. 1-13). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6119 LNAI, No. PART 2). https://doi.org/10.1007/978-3-642-13672-6_1

Subclass-oriented Dimension Reduction with constraint transformation and manifold regularization. / Tong, Bin; Suzuki, Einoshin.

Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 2. ed. 2010. p. 1-13 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6119 LNAI, No. PART 2).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Tong, B & Suzuki, E 2010, Subclass-oriented Dimension Reduction with constraint transformation and manifold regularization. in Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 6119 LNAI, pp. 1-13, 14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010, Hyderabad, India, 6/21/10. https://doi.org/10.1007/978-3-642-13672-6_1
Tong B, Suzuki E. Subclass-oriented Dimension Reduction with constraint transformation and manifold regularization. In Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 2 ed. 2010. p. 1-13. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). https://doi.org/10.1007/978-3-642-13672-6_1
Tong, Bin ; Suzuki, Einoshin. / Subclass-oriented Dimension Reduction with constraint transformation and manifold regularization. Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 2. ed. 2010. pp. 1-13 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).
@inproceedings{e123d0d293d448f19a0e3e2dd0e300b4,
title = "Subclass-oriented Dimension Reduction with constraint transformation and manifold regularization",
abstract = "We propose a new method, called Subclass-oriented Dimension Reduction with Pairwise Constraints (SODRPaC), for dimension reduction on high dimensional data. Current linear semi-supervised dimension reduction methods using pairwise constraints, e.g., must-link constraints and cannot-link constraints, can not handle appropriately the data of multiple subclasses where the points of a class are separately distributed in different groups. To illustrate this problem, wparticularly classify the must-link constraint into two categories, which are theinter-subclass must-link constraint and the intra-subclass must-link constraint, respectively. We argue that handling the inter-subclass must-link constraint is challenging for current discriminant criteria. Inspired by the above observation and the cluster assumption that nearby points are possible in the same class, we carefully transform must-link constraints into cannot-link constraints, and then propose a new discriminant criterion by employing the cannot-link constraints and the compactness of shared nearest neighbors. For the reason that the local data structure is one of the most significant features for the data of multiple subclasses, manifold regularization is also incorporated in our dimension reduction framework. Extensive experiments on both synthetic and practical data sets illustrate the effectiveness of our method.",
author = "Bin Tong and Einoshin Suzuki",
year = "2010",
month = "12",
day = "1",
doi = "10.1007/978-3-642-13672-6_1",
language = "English",
isbn = "3642136710",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 2",
pages = "1--13",
booktitle = "Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings",
edition = "PART 2",

}

TY - GEN

T1 - Subclass-oriented Dimension Reduction with constraint transformation and manifold regularization

AU - Tong, Bin

AU - Suzuki, Einoshin

PY - 2010/12/1

Y1 - 2010/12/1

N2 - We propose a new method, called Subclass-oriented Dimension Reduction with Pairwise Constraints (SODRPaC), for dimension reduction on high dimensional data. Current linear semi-supervised dimension reduction methods using pairwise constraints, e.g., must-link constraints and cannot-link constraints, can not handle appropriately the data of multiple subclasses where the points of a class are separately distributed in different groups. To illustrate this problem, wparticularly classify the must-link constraint into two categories, which are theinter-subclass must-link constraint and the intra-subclass must-link constraint, respectively. We argue that handling the inter-subclass must-link constraint is challenging for current discriminant criteria. Inspired by the above observation and the cluster assumption that nearby points are possible in the same class, we carefully transform must-link constraints into cannot-link constraints, and then propose a new discriminant criterion by employing the cannot-link constraints and the compactness of shared nearest neighbors. For the reason that the local data structure is one of the most significant features for the data of multiple subclasses, manifold regularization is also incorporated in our dimension reduction framework. Extensive experiments on both synthetic and practical data sets illustrate the effectiveness of our method.

AB - We propose a new method, called Subclass-oriented Dimension Reduction with Pairwise Constraints (SODRPaC), for dimension reduction on high dimensional data. Current linear semi-supervised dimension reduction methods using pairwise constraints, e.g., must-link constraints and cannot-link constraints, can not handle appropriately the data of multiple subclasses where the points of a class are separately distributed in different groups. To illustrate this problem, wparticularly classify the must-link constraint into two categories, which are theinter-subclass must-link constraint and the intra-subclass must-link constraint, respectively. We argue that handling the inter-subclass must-link constraint is challenging for current discriminant criteria. Inspired by the above observation and the cluster assumption that nearby points are possible in the same class, we carefully transform must-link constraints into cannot-link constraints, and then propose a new discriminant criterion by employing the cannot-link constraints and the compactness of shared nearest neighbors. For the reason that the local data structure is one of the most significant features for the data of multiple subclasses, manifold regularization is also incorporated in our dimension reduction framework. Extensive experiments on both synthetic and practical data sets illustrate the effectiveness of our method.

UR - http://www.scopus.com/inward/record.url?scp=79956312031&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79956312031&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-13672-6_1

DO - 10.1007/978-3-642-13672-6_1

M3 - Conference contribution

AN - SCOPUS:79956312031

SN - 3642136710

SN - 9783642136719

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 1

EP - 13

BT - Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings

ER -