Conformity-Based Source Subset Selection for Instance Transfer

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Instance transfer aims at improving prediction models for a target domain by transferring data from related source domains. The effectiveness of instance transfer depends on the relevance of source data to the target domain. When the relevance of source data is limited, the only option is to select a subset of source data of which the relevance is acceptable. In this paper, we introduce three algorithms that perform source-subset selection prior to model training. The algorithms employ a conformity-based test that estimates the source-subset relevance based on individual instances or on subsets as a whole. Experiments conducted on four real-world data sets demonstrated the effectiveness of the proposed algorithms. Especially, it was shown that pre-training subset-selection based on set relevance is capable of outperforming the existing instance-transfer techniques. (C) 2017 Elsevier B.V. All rights reserved.

Original languageEnglish
Pages (from-to)41-51
Number of pages11
JournalNeurocomputing
Volume258
DOIs
Publication statusPublished - 4 Oct 2017

Keywords

  • Conformal test
  • Instance-transfer learning
  • Source-subset selection

Cite this

@article{2fd5074d76e04a569f3af2c6f2da519d,
title = "Conformity-Based Source Subset Selection for Instance Transfer",
abstract = "Instance transfer aims at improving prediction models for a target domain by transferring data from related source domains. The effectiveness of instance transfer depends on the relevance of source data to the target domain. When the relevance of source data is limited, the only option is to select a subset of source data of which the relevance is acceptable. In this paper, we introduce three algorithms that perform source-subset selection prior to model training. The algorithms employ a conformity-based test that estimates the source-subset relevance based on individual instances or on subsets as a whole. Experiments conducted on four real-world data sets demonstrated the effectiveness of the proposed algorithms. Especially, it was shown that pre-training subset-selection based on set relevance is capable of outperforming the existing instance-transfer techniques. (C) 2017 Elsevier B.V. All rights reserved.",
keywords = "Conformal test, Instance-transfer learning, Source-subset selection",
author = "Shuang Zhou and Evgueni Smirnov and Gijsbertus Schoenmakers and Ralf Peeters",
year = "2017",
month = "10",
day = "4",
doi = "10.1016/j.neucom.2016.11.071",
language = "English",
volume = "258",
pages = "41--51",
journal = "Neurocomputing",
issn = "0925-2312",
publisher = "Elsevier Science",

}

Conformity-Based Source Subset Selection for Instance Transfer. / Zhou, Shuang; Smirnov, Evgueni; Schoenmakers, Gijsbertus; Peeters, Ralf.

In: Neurocomputing, Vol. 258, 04.10.2017, p. 41-51.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Conformity-Based Source Subset Selection for Instance Transfer

AU - Zhou, Shuang

AU - Smirnov, Evgueni

AU - Schoenmakers, Gijsbertus

AU - Peeters, Ralf

PY - 2017/10/4

Y1 - 2017/10/4

N2 - Instance transfer aims at improving prediction models for a target domain by transferring data from related source domains. The effectiveness of instance transfer depends on the relevance of source data to the target domain. When the relevance of source data is limited, the only option is to select a subset of source data of which the relevance is acceptable. In this paper, we introduce three algorithms that perform source-subset selection prior to model training. The algorithms employ a conformity-based test that estimates the source-subset relevance based on individual instances or on subsets as a whole. Experiments conducted on four real-world data sets demonstrated the effectiveness of the proposed algorithms. Especially, it was shown that pre-training subset-selection based on set relevance is capable of outperforming the existing instance-transfer techniques. (C) 2017 Elsevier B.V. All rights reserved.

AB - Instance transfer aims at improving prediction models for a target domain by transferring data from related source domains. The effectiveness of instance transfer depends on the relevance of source data to the target domain. When the relevance of source data is limited, the only option is to select a subset of source data of which the relevance is acceptable. In this paper, we introduce three algorithms that perform source-subset selection prior to model training. The algorithms employ a conformity-based test that estimates the source-subset relevance based on individual instances or on subsets as a whole. Experiments conducted on four real-world data sets demonstrated the effectiveness of the proposed algorithms. Especially, it was shown that pre-training subset-selection based on set relevance is capable of outperforming the existing instance-transfer techniques. (C) 2017 Elsevier B.V. All rights reserved.

KW - Conformal test

KW - Instance-transfer learning

KW - Source-subset selection

U2 - 10.1016/j.neucom.2016.11.071

DO - 10.1016/j.neucom.2016.11.071

M3 - Article

VL - 258

SP - 41

EP - 51

JO - Neurocomputing

JF - Neurocomputing

SN - 0925-2312

ER -