Reconstruction of Cross-Sectional Missing Data Using Neural Networks

Iffat Gheyas, Leslie S. Smith

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

Abstract

The treatment of incomplete data is an important step in the pre-processing of data. We propose a non-parametric multiple imputation algorithm (GMI) for the reconstruction of missing data, based on Generalized Regression Neural Networks (GRNN). We compare GMI with popular missing data imputation algorithms: EM (Expectation Maximization) MI (Multiple Imputation), MCMC (Markov Chain Monte Carlo) MI, and hot deck MI. A separate GRNN classifier is trained and tested on the dataset imputed with each imputation algorithm. The imputation algorithms are evaluated based on the accuracy of the GRNN classifier after the imputation process. We show the effectiveness of our proposed algorithm on twenty-six real datasets.

LanguageEnglish
Title of host publicationEngineering Applications of Neural Networks - 11th International Conference, EANN 2009, Proceedings
Pages28-34
Number of pages7
DOIs
Publication statusPublished - 1 Dec 2009
Externally publishedYes

Publication series

NameCommunications in Computer and Information Science
Volume43 CCIS
ISSN (Print)1865-0929

Fingerprint

Neural networks
Classifiers
Markov processes
Processing

Cite this

Gheyas, I., & Smith, L. S. (2009). Reconstruction of Cross-Sectional Missing Data Using Neural Networks. In Engineering Applications of Neural Networks - 11th International Conference, EANN 2009, Proceedings (pp. 28-34). (Communications in Computer and Information Science; Vol. 43 CCIS). https://doi.org/10.1007/978-3-642-03969-0_3
Gheyas, Iffat ; Smith, Leslie S. / Reconstruction of Cross-Sectional Missing Data Using Neural Networks. Engineering Applications of Neural Networks - 11th International Conference, EANN 2009, Proceedings. 2009. pp. 28-34 (Communications in Computer and Information Science).
@inproceedings{05dccbfc620a434c8554a4dcfde17663,
title = "Reconstruction of Cross-Sectional Missing Data Using Neural Networks",
abstract = "The treatment of incomplete data is an important step in the pre-processing of data. We propose a non-parametric multiple imputation algorithm (GMI) for the reconstruction of missing data, based on Generalized Regression Neural Networks (GRNN). We compare GMI with popular missing data imputation algorithms: EM (Expectation Maximization) MI (Multiple Imputation), MCMC (Markov Chain Monte Carlo) MI, and hot deck MI. A separate GRNN classifier is trained and tested on the dataset imputed with each imputation algorithm. The imputation algorithms are evaluated based on the accuracy of the GRNN classifier after the imputation process. We show the effectiveness of our proposed algorithm on twenty-six real datasets.",
keywords = "Generalized Regression Neural Networks, Imputation, Missing Values, Multiple Imputation",
author = "Iffat Gheyas and Smith, {Leslie S.}",
year = "2009",
month = "12",
day = "1",
doi = "10.1007/978-3-642-03969-0_3",
language = "English",
isbn = "3642039685",
series = "Communications in Computer and Information Science",
pages = "28--34",
booktitle = "Engineering Applications of Neural Networks - 11th International Conference, EANN 2009, Proceedings",

}

Gheyas, I & Smith, LS 2009, Reconstruction of Cross-Sectional Missing Data Using Neural Networks. in Engineering Applications of Neural Networks - 11th International Conference, EANN 2009, Proceedings. Communications in Computer and Information Science, vol. 43 CCIS, pp. 28-34. https://doi.org/10.1007/978-3-642-03969-0_3

Reconstruction of Cross-Sectional Missing Data Using Neural Networks. / Gheyas, Iffat; Smith, Leslie S.

Engineering Applications of Neural Networks - 11th International Conference, EANN 2009, Proceedings. 2009. p. 28-34 (Communications in Computer and Information Science; Vol. 43 CCIS).

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

TY - GEN

T1 - Reconstruction of Cross-Sectional Missing Data Using Neural Networks

AU - Gheyas, Iffat

AU - Smith, Leslie S.

PY - 2009/12/1

Y1 - 2009/12/1

N2 - The treatment of incomplete data is an important step in the pre-processing of data. We propose a non-parametric multiple imputation algorithm (GMI) for the reconstruction of missing data, based on Generalized Regression Neural Networks (GRNN). We compare GMI with popular missing data imputation algorithms: EM (Expectation Maximization) MI (Multiple Imputation), MCMC (Markov Chain Monte Carlo) MI, and hot deck MI. A separate GRNN classifier is trained and tested on the dataset imputed with each imputation algorithm. The imputation algorithms are evaluated based on the accuracy of the GRNN classifier after the imputation process. We show the effectiveness of our proposed algorithm on twenty-six real datasets.

AB - The treatment of incomplete data is an important step in the pre-processing of data. We propose a non-parametric multiple imputation algorithm (GMI) for the reconstruction of missing data, based on Generalized Regression Neural Networks (GRNN). We compare GMI with popular missing data imputation algorithms: EM (Expectation Maximization) MI (Multiple Imputation), MCMC (Markov Chain Monte Carlo) MI, and hot deck MI. A separate GRNN classifier is trained and tested on the dataset imputed with each imputation algorithm. The imputation algorithms are evaluated based on the accuracy of the GRNN classifier after the imputation process. We show the effectiveness of our proposed algorithm on twenty-six real datasets.

KW - Generalized Regression Neural Networks

KW - Imputation

KW - Missing Values

KW - Multiple Imputation

UR - http://www.scopus.com/inward/record.url?scp=78049397186&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-03969-0_3

DO - 10.1007/978-3-642-03969-0_3

M3 - Conference contribution

SN - 3642039685

SN - 9783642039683

T3 - Communications in Computer and Information Science

SP - 28

EP - 34

BT - Engineering Applications of Neural Networks - 11th International Conference, EANN 2009, Proceedings

ER -

Gheyas I, Smith LS. Reconstruction of Cross-Sectional Missing Data Using Neural Networks. In Engineering Applications of Neural Networks - 11th International Conference, EANN 2009, Proceedings. 2009. p. 28-34. (Communications in Computer and Information Science). https://doi.org/10.1007/978-3-642-03969-0_3