2 Citations (Scopus)

Abstract

Multiple types of heterogeneity including label heterogeneity and feature heterogeneity often co-exist in many real-world data mining applications, such as diabetes treatment classification, gene functionality prediction, and brain image analysis. To effectively leverage such heterogeneity, in this article, we propose a novel graph-based model for Learning with both Label and Feature heterogeneity, namely L2F. It models the label correlation by requiring that any two label-specific classifiers behave similarly on the same views if the associated labels are similar, and imposes the view consistency by requiring that view-based classifiers generate similar predictions on the same examples. The objective function for L2F is jointly convex. To solve the optimization problem, we propose an iterative algorithm, which is guaranteed to converge to the global optimum. One appealing feature of L2F is that it is capable of handling data with missing views and labels. Furthermore, we analyze its generalization performance based on Rademacher complexity, which sheds light on the benefits of jointly modeling the label and feature heterogeneity. Experimental results on various biomedical datasets show the effectiveness of the proposed approach.

Original languageEnglish (US)
Article number39
JournalACM Transactions on Knowledge Discovery from Data
Volume10
Issue number4
DOIs
StatePublished - May 1 2016

Fingerprint

Labels
Classifiers
Data handling
Medical problems
Image analysis
Data mining
Brain
Genes

Keywords

  • Heterogeneous learning
  • Medical informatics
  • Multi-label learning
  • Multi-view learning

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Jointly modeling label and feature heterogeneity in medical informatics. / Yang, Pei; Yang, Hongxia; Fu, Haoda; Zhou, Dawei; Ye, Jieping; Lappas, Theodoros; He, Jingrui.

In: ACM Transactions on Knowledge Discovery from Data, Vol. 10, No. 4, 39, 01.05.2016.

Research output: Contribution to journalArticle

Yang, Pei ; Yang, Hongxia ; Fu, Haoda ; Zhou, Dawei ; Ye, Jieping ; Lappas, Theodoros ; He, Jingrui. / Jointly modeling label and feature heterogeneity in medical informatics. In: ACM Transactions on Knowledge Discovery from Data. 2016 ; Vol. 10, No. 4.
@article{0da817de47e34205852e6039539c0b94,
title = "Jointly modeling label and feature heterogeneity in medical informatics",
abstract = "Multiple types of heterogeneity including label heterogeneity and feature heterogeneity often co-exist in many real-world data mining applications, such as diabetes treatment classification, gene functionality prediction, and brain image analysis. To effectively leverage such heterogeneity, in this article, we propose a novel graph-based model for Learning with both Label and Feature heterogeneity, namely L2F. It models the label correlation by requiring that any two label-specific classifiers behave similarly on the same views if the associated labels are similar, and imposes the view consistency by requiring that view-based classifiers generate similar predictions on the same examples. The objective function for L2F is jointly convex. To solve the optimization problem, we propose an iterative algorithm, which is guaranteed to converge to the global optimum. One appealing feature of L2F is that it is capable of handling data with missing views and labels. Furthermore, we analyze its generalization performance based on Rademacher complexity, which sheds light on the benefits of jointly modeling the label and feature heterogeneity. Experimental results on various biomedical datasets show the effectiveness of the proposed approach.",
keywords = "Heterogeneous learning, Medical informatics, Multi-label learning, Multi-view learning",
author = "Pei Yang and Hongxia Yang and Haoda Fu and Dawei Zhou and Jieping Ye and Theodoros Lappas and Jingrui He",
year = "2016",
month = "5",
day = "1",
doi = "10.1145/2768831",
language = "English (US)",
volume = "10",
journal = "ACM Transactions on Knowledge Discovery from Data",
issn = "1556-4681",
publisher = "Association for Computing Machinery (ACM)",
number = "4",

}

TY - JOUR

T1 - Jointly modeling label and feature heterogeneity in medical informatics

AU - Yang, Pei

AU - Yang, Hongxia

AU - Fu, Haoda

AU - Zhou, Dawei

AU - Ye, Jieping

AU - Lappas, Theodoros

AU - He, Jingrui

PY - 2016/5/1

Y1 - 2016/5/1

N2 - Multiple types of heterogeneity including label heterogeneity and feature heterogeneity often co-exist in many real-world data mining applications, such as diabetes treatment classification, gene functionality prediction, and brain image analysis. To effectively leverage such heterogeneity, in this article, we propose a novel graph-based model for Learning with both Label and Feature heterogeneity, namely L2F. It models the label correlation by requiring that any two label-specific classifiers behave similarly on the same views if the associated labels are similar, and imposes the view consistency by requiring that view-based classifiers generate similar predictions on the same examples. The objective function for L2F is jointly convex. To solve the optimization problem, we propose an iterative algorithm, which is guaranteed to converge to the global optimum. One appealing feature of L2F is that it is capable of handling data with missing views and labels. Furthermore, we analyze its generalization performance based on Rademacher complexity, which sheds light on the benefits of jointly modeling the label and feature heterogeneity. Experimental results on various biomedical datasets show the effectiveness of the proposed approach.

AB - Multiple types of heterogeneity including label heterogeneity and feature heterogeneity often co-exist in many real-world data mining applications, such as diabetes treatment classification, gene functionality prediction, and brain image analysis. To effectively leverage such heterogeneity, in this article, we propose a novel graph-based model for Learning with both Label and Feature heterogeneity, namely L2F. It models the label correlation by requiring that any two label-specific classifiers behave similarly on the same views if the associated labels are similar, and imposes the view consistency by requiring that view-based classifiers generate similar predictions on the same examples. The objective function for L2F is jointly convex. To solve the optimization problem, we propose an iterative algorithm, which is guaranteed to converge to the global optimum. One appealing feature of L2F is that it is capable of handling data with missing views and labels. Furthermore, we analyze its generalization performance based on Rademacher complexity, which sheds light on the benefits of jointly modeling the label and feature heterogeneity. Experimental results on various biomedical datasets show the effectiveness of the proposed approach.

KW - Heterogeneous learning

KW - Medical informatics

KW - Multi-label learning

KW - Multi-view learning

UR - http://www.scopus.com/inward/record.url?scp=84973444975&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84973444975&partnerID=8YFLogxK

U2 - 10.1145/2768831

DO - 10.1145/2768831

M3 - Article

VL - 10

JO - ACM Transactions on Knowledge Discovery from Data

JF - ACM Transactions on Knowledge Discovery from Data

SN - 1556-4681

IS - 4

M1 - 39

ER -