Abstract

Predictive and risk stratification models using machine learning algorithms such as Support Vector Machines (SVMs), have been used in cardiology and medicine to improve patient care and prognosis. In this work, we have used SVM based Recursive Feature Elimination (SVM-RFE) methods to select patient attributes/features relevant to the etio-pathogenesis of complications following a drug eluting stent (DES) procedure. With a high dimensional feature space (145 features, in our case), and comparatively few patients, there is a high risk of 'over-fitting'. Also, for the model to be clinically relevant, the number of patient features need to be reduced to a manageable number, to be used in patient care. SVM-RFE selects subsets of patient features that have maximal influence on the risk of a complication. In our results, when compared with our initial model with all the 145 features, we obtained better performance of the classifiers with 75 top ranked patient features, a 50% reduction in the original dimensionality of the data space. There was a universal improvement in performance of all SVMs with different kernels and parameters. This method of feature ranking helps to determine the most informative patient features. Use of these relevant features improves the prediction of complications following a DES procedure.

Original languageEnglish (US)
Title of host publicationComputers in Cardiology 2009, CinC 2009
Pages345-348
Number of pages4
Volume36
StatePublished - 2009
Event36th Annual Conference of Computers in Cardiology, CinC 2009 - Park City, UT, United States
Duration: Sep 13 2009Sep 16 2009

Other

Other36th Annual Conference of Computers in Cardiology, CinC 2009
CountryUnited States
CityPark City, UT
Period9/13/099/16/09

Fingerprint

Drug-Eluting Stents
Stents
Support vector machines
Cardiology
Patient Care
Learning algorithms
Medicine
Learning systems
Classifiers
Support Vector Machine

ASJC Scopus subject areas

  • Computer Science Applications
  • Cardiology and Cardiovascular Medicine

Cite this

Gouripeddi, R. K., Balasubramanian, V. N., Panchanathan, S., Harris, J., Bhaskaran, A., & Siegel, R. M. (2009). Ranking predictors of complications following a drug eluting stent procedure using support vector machines. In Computers in Cardiology 2009, CinC 2009 (Vol. 36, pp. 345-348). [5445398]

Ranking predictors of complications following a drug eluting stent procedure using support vector machines. / Gouripeddi, Ramkiran K.; Balasubramanian, V. N.; Panchanathan, Sethuraman; Harris, J.; Bhaskaran, A.; Siegel, R. M.

Computers in Cardiology 2009, CinC 2009. Vol. 36 2009. p. 345-348 5445398.

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

Gouripeddi, RK, Balasubramanian, VN, Panchanathan, S, Harris, J, Bhaskaran, A & Siegel, RM 2009, Ranking predictors of complications following a drug eluting stent procedure using support vector machines. in Computers in Cardiology 2009, CinC 2009. vol. 36, 5445398, pp. 345-348, 36th Annual Conference of Computers in Cardiology, CinC 2009, Park City, UT, United States, 9/13/09.
Gouripeddi RK, Balasubramanian VN, Panchanathan S, Harris J, Bhaskaran A, Siegel RM. Ranking predictors of complications following a drug eluting stent procedure using support vector machines. In Computers in Cardiology 2009, CinC 2009. Vol. 36. 2009. p. 345-348. 5445398
Gouripeddi, Ramkiran K. ; Balasubramanian, V. N. ; Panchanathan, Sethuraman ; Harris, J. ; Bhaskaran, A. ; Siegel, R. M. / Ranking predictors of complications following a drug eluting stent procedure using support vector machines. Computers in Cardiology 2009, CinC 2009. Vol. 36 2009. pp. 345-348
@inproceedings{f82a931a2df24f6197e069b8222cefcc,
title = "Ranking predictors of complications following a drug eluting stent procedure using support vector machines",
abstract = "Predictive and risk stratification models using machine learning algorithms such as Support Vector Machines (SVMs), have been used in cardiology and medicine to improve patient care and prognosis. In this work, we have used SVM based Recursive Feature Elimination (SVM-RFE) methods to select patient attributes/features relevant to the etio-pathogenesis of complications following a drug eluting stent (DES) procedure. With a high dimensional feature space (145 features, in our case), and comparatively few patients, there is a high risk of 'over-fitting'. Also, for the model to be clinically relevant, the number of patient features need to be reduced to a manageable number, to be used in patient care. SVM-RFE selects subsets of patient features that have maximal influence on the risk of a complication. In our results, when compared with our initial model with all the 145 features, we obtained better performance of the classifiers with 75 top ranked patient features, a 50{\%} reduction in the original dimensionality of the data space. There was a universal improvement in performance of all SVMs with different kernels and parameters. This method of feature ranking helps to determine the most informative patient features. Use of these relevant features improves the prediction of complications following a DES procedure.",
author = "Gouripeddi, {Ramkiran K.} and Balasubramanian, {V. N.} and Sethuraman Panchanathan and J. Harris and A. Bhaskaran and Siegel, {R. M.}",
year = "2009",
language = "English (US)",
isbn = "9781424472819",
volume = "36",
pages = "345--348",
booktitle = "Computers in Cardiology 2009, CinC 2009",

}

TY - GEN

T1 - Ranking predictors of complications following a drug eluting stent procedure using support vector machines

AU - Gouripeddi, Ramkiran K.

AU - Balasubramanian, V. N.

AU - Panchanathan, Sethuraman

AU - Harris, J.

AU - Bhaskaran, A.

AU - Siegel, R. M.

PY - 2009

Y1 - 2009

N2 - Predictive and risk stratification models using machine learning algorithms such as Support Vector Machines (SVMs), have been used in cardiology and medicine to improve patient care and prognosis. In this work, we have used SVM based Recursive Feature Elimination (SVM-RFE) methods to select patient attributes/features relevant to the etio-pathogenesis of complications following a drug eluting stent (DES) procedure. With a high dimensional feature space (145 features, in our case), and comparatively few patients, there is a high risk of 'over-fitting'. Also, for the model to be clinically relevant, the number of patient features need to be reduced to a manageable number, to be used in patient care. SVM-RFE selects subsets of patient features that have maximal influence on the risk of a complication. In our results, when compared with our initial model with all the 145 features, we obtained better performance of the classifiers with 75 top ranked patient features, a 50% reduction in the original dimensionality of the data space. There was a universal improvement in performance of all SVMs with different kernels and parameters. This method of feature ranking helps to determine the most informative patient features. Use of these relevant features improves the prediction of complications following a DES procedure.

AB - Predictive and risk stratification models using machine learning algorithms such as Support Vector Machines (SVMs), have been used in cardiology and medicine to improve patient care and prognosis. In this work, we have used SVM based Recursive Feature Elimination (SVM-RFE) methods to select patient attributes/features relevant to the etio-pathogenesis of complications following a drug eluting stent (DES) procedure. With a high dimensional feature space (145 features, in our case), and comparatively few patients, there is a high risk of 'over-fitting'. Also, for the model to be clinically relevant, the number of patient features need to be reduced to a manageable number, to be used in patient care. SVM-RFE selects subsets of patient features that have maximal influence on the risk of a complication. In our results, when compared with our initial model with all the 145 features, we obtained better performance of the classifiers with 75 top ranked patient features, a 50% reduction in the original dimensionality of the data space. There was a universal improvement in performance of all SVMs with different kernels and parameters. This method of feature ranking helps to determine the most informative patient features. Use of these relevant features improves the prediction of complications following a DES procedure.

UR - http://www.scopus.com/inward/record.url?scp=77952717199&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77952717199&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:77952717199

SN - 9781424472819

VL - 36

SP - 345

EP - 348

BT - Computers in Cardiology 2009, CinC 2009

ER -