Interval-valued data prediction via regularized artificial neural network

Zebin Yang, Dennis K.J. Lin, Aijun Zhang

Research output: Contribution to journalArticle

Abstract

The prediction of interval-valued data is a challenging task as the predicted lower bounds of intervals should not cross over the corresponding upper bounds. In this paper, a regularized artificial neural network (RANN) is proposed to address this difficult problem. It provides a flexible trade-off between prediction accuracy and interval crossing. Compared to existing hard-constrained methods, the RANN has the advantage that it does not necessarily reduce the prediction accuracy while preventing interval crossing. Extensive experiments are conducted based on both simulation and real-life datasets, with comparison to multiple traditional models, including the linear constrained center and range method, the least absolute shrinkage and selection operator-based interval-valued regression, the nonlinear interval kernel regression, the interval multi-layer perceptron and the multi-output support vector regression. Experimental results show that the proposed RANN model is an effective tool for interval-valued data prediction tasks with high prediction accuracy.

Original languageEnglish (US)
Pages (from-to)336-345
Number of pages10
JournalNeurocomputing
Volume331
DOIs
StatePublished - Feb 28 2019

Fingerprint

Neural Networks (Computer)
Neural networks
Linear Models
Multilayer neural networks
Experiments
Datasets

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Cognitive Neuroscience
  • Artificial Intelligence

Cite this

@article{466d570a961e40018a46004b76f97c41,
title = "Interval-valued data prediction via regularized artificial neural network",
abstract = "The prediction of interval-valued data is a challenging task as the predicted lower bounds of intervals should not cross over the corresponding upper bounds. In this paper, a regularized artificial neural network (RANN) is proposed to address this difficult problem. It provides a flexible trade-off between prediction accuracy and interval crossing. Compared to existing hard-constrained methods, the RANN has the advantage that it does not necessarily reduce the prediction accuracy while preventing interval crossing. Extensive experiments are conducted based on both simulation and real-life datasets, with comparison to multiple traditional models, including the linear constrained center and range method, the least absolute shrinkage and selection operator-based interval-valued regression, the nonlinear interval kernel regression, the interval multi-layer perceptron and the multi-output support vector regression. Experimental results show that the proposed RANN model is an effective tool for interval-valued data prediction tasks with high prediction accuracy.",
author = "Zebin Yang and Lin, {Dennis K.J.} and Aijun Zhang",
year = "2019",
month = "2",
day = "28",
doi = "10.1016/j.neucom.2018.11.063",
language = "English (US)",
volume = "331",
pages = "336--345",
journal = "Neurocomputing",
issn = "0925-2312",
publisher = "Elsevier",

}

Interval-valued data prediction via regularized artificial neural network. / Yang, Zebin; Lin, Dennis K.J.; Zhang, Aijun.

In: Neurocomputing, Vol. 331, 28.02.2019, p. 336-345.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Interval-valued data prediction via regularized artificial neural network

AU - Yang, Zebin

AU - Lin, Dennis K.J.

AU - Zhang, Aijun

PY - 2019/2/28

Y1 - 2019/2/28

N2 - The prediction of interval-valued data is a challenging task as the predicted lower bounds of intervals should not cross over the corresponding upper bounds. In this paper, a regularized artificial neural network (RANN) is proposed to address this difficult problem. It provides a flexible trade-off between prediction accuracy and interval crossing. Compared to existing hard-constrained methods, the RANN has the advantage that it does not necessarily reduce the prediction accuracy while preventing interval crossing. Extensive experiments are conducted based on both simulation and real-life datasets, with comparison to multiple traditional models, including the linear constrained center and range method, the least absolute shrinkage and selection operator-based interval-valued regression, the nonlinear interval kernel regression, the interval multi-layer perceptron and the multi-output support vector regression. Experimental results show that the proposed RANN model is an effective tool for interval-valued data prediction tasks with high prediction accuracy.

AB - The prediction of interval-valued data is a challenging task as the predicted lower bounds of intervals should not cross over the corresponding upper bounds. In this paper, a regularized artificial neural network (RANN) is proposed to address this difficult problem. It provides a flexible trade-off between prediction accuracy and interval crossing. Compared to existing hard-constrained methods, the RANN has the advantage that it does not necessarily reduce the prediction accuracy while preventing interval crossing. Extensive experiments are conducted based on both simulation and real-life datasets, with comparison to multiple traditional models, including the linear constrained center and range method, the least absolute shrinkage and selection operator-based interval-valued regression, the nonlinear interval kernel regression, the interval multi-layer perceptron and the multi-output support vector regression. Experimental results show that the proposed RANN model is an effective tool for interval-valued data prediction tasks with high prediction accuracy.

UR - http://www.scopus.com/inward/record.url?scp=85057866531&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057866531&partnerID=8YFLogxK

U2 - 10.1016/j.neucom.2018.11.063

DO - 10.1016/j.neucom.2018.11.063

M3 - Article

VL - 331

SP - 336

EP - 345

JO - Neurocomputing

JF - Neurocomputing

SN - 0925-2312

ER -