Imputation Methods to Deal with Missing Values when Data Mining Trauma Injury Data

Kay Penny; Thomas Chesney

doi:10.1109/ITI.2006.1708480

Imputation Methods to Deal with Missing Values when Data Mining Trauma Injury Data

Kay Penny^*, Thomas Chesney

^*Corresponding author for this work

Health Services Research Unit (HSRU)

Research output: Chapter in Book/Report/Conference proceeding › Published conference contribution

15 Citations (Scopus)

Abstract

Methods for analysing trauma injury data with missing values, collected at a UK hospital, are reported. One measure of injury severity, the Glasgow coma score, which is known to be associated with patient death, is missing for 12% of patients in the dataset. In order to include these 12% of patients in the analysis, three different data imputation techniques are used to estimate the missing values. The imputed data sets are analysed by an artificial neural network and logistic regression, and their results compared in terms of sensitivity, specificity, positive predictive value and negative predictive value

Original language	English
Title of host publication	28th International Conference on Information Technology Interfaces, 2006.
Publisher	IEEE Explore
Pages	213-218
DOIs	https://doi.org/10.1109/ITI.2006.1708480
Publication status	Published - Jun 2006
Event	Information Technology Interfaces: 2006 - Cavtat, Croatia Duration: 19 Jun 2006 → 22 Jun 2006 Conference number: 28

Conference

Conference	Information Technology Interfaces
Abbreviated title	ITI
Country/Territory	Croatia
City	Cavtat
Period	19/06/06 → 22/06/06

Access to Document

10.1109/ITI.2006.1708480

Cite this

@inproceedings{8f0e3aec5e004c4aaef9ab4c95660995,

title = "Imputation Methods to Deal with Missing Values when Data Mining Trauma Injury Data",

abstract = "Methods for analysing trauma injury data with missing values, collected at a UK hospital, are reported. One measure of injury severity, the Glasgow coma score, which is known to be associated with patient death, is missing for 12% of patients in the dataset. In order to include these 12% of patients in the analysis, three different data imputation techniques are used to estimate the missing values. The imputed data sets are analysed by an artificial neural network and logistic regression, and their results compared in terms of sensitivity, specificity, positive predictive value and negative predictive value",

author = "Kay Penny and Thomas Chesney",

year = "2006",

month = jun,

doi = "10.1109/ITI.2006.1708480",

language = "English",

pages = "213--218",

booktitle = "28th International Conference on Information Technology Interfaces, 2006.",

publisher = "IEEE Explore",

note = "Information Technology Interfaces : 2006, ITI ; Conference date: 19-06-2006 Through 22-06-2006",

}

TY - GEN

T1 - Imputation Methods to Deal with Missing Values when Data Mining Trauma Injury Data

AU - Penny, Kay

AU - Chesney, Thomas

N1 - Conference code: 28

PY - 2006/6

Y1 - 2006/6

N2 - Methods for analysing trauma injury data with missing values, collected at a UK hospital, are reported. One measure of injury severity, the Glasgow coma score, which is known to be associated with patient death, is missing for 12% of patients in the dataset. In order to include these 12% of patients in the analysis, three different data imputation techniques are used to estimate the missing values. The imputed data sets are analysed by an artificial neural network and logistic regression, and their results compared in terms of sensitivity, specificity, positive predictive value and negative predictive value

AB - Methods for analysing trauma injury data with missing values, collected at a UK hospital, are reported. One measure of injury severity, the Glasgow coma score, which is known to be associated with patient death, is missing for 12% of patients in the dataset. In order to include these 12% of patients in the analysis, three different data imputation techniques are used to estimate the missing values. The imputed data sets are analysed by an artificial neural network and logistic regression, and their results compared in terms of sensitivity, specificity, positive predictive value and negative predictive value

U2 - 10.1109/ITI.2006.1708480

DO - 10.1109/ITI.2006.1708480

M3 - Published conference contribution

SP - 213

EP - 218

BT - 28th International Conference on Information Technology Interfaces, 2006.

PB - IEEE Explore

T2 - Information Technology Interfaces

Y2 - 19 June 2006 through 22 June 2006

ER -

Imputation Methods to Deal with Missing Values when Data Mining Trauma Injury Data

Abstract

Conference

Access to Document

Fingerprint

Cite this