Using machine learning to prescribe warfarin

Authors:
Brent Martin;Marina Filipovic;Lara Rennie;David Shaw
Affiliations:
Computational Intelligence Research Lab, University of Canterbury, Christchurch, New Zealand;Computational Intelligence Research Lab, University of Canterbury, Christchurch, New Zealand;Computational Intelligence Research Lab, University of Canterbury, Christchurch, New Zealand;Computational Intelligence Research Lab, University of Canterbury, Christchurch, New Zealand
Venue:
AIMSA'10 Proceedings of the 14th international conference on Artificial intelligence: methodology, systems, and applications
Year:
2010

Citing 6
Cited 1

Instance-Based Learning Algorithms

Machine Learning
Original Contribution: Stacked generalization

Neural Networks
C4.5: programs for machine learning

C4.5: programs for machine learning
Bagging predictors

Machine Learning
Data Mining

Data Mining
Explaining the output of ensembles in medical decision support on a case by case basis

Artificial Intelligence in Medicine

Predicting warfarin dosage from clinical data: A supervised learning approach

Artificial Intelligence in Medicine

Quantified Score

Hi-index	0.00

Visualization

Abstract

Predicting the effects of the blood thinner warfarin is very difficult because of its long half-life, interaction with drugs and food, and because every patient has a unique response to a given dose. Previous attempts to use machine learning have shown that no individual learner can accurately predict the drug's effect for all patients. In this paper we present our exploration of this problem using ensemble methods. The resulting system utilizes multiple ML algorithms and input parameters to make multiple predictions, which are then scrutinized by the doctor. Our results indicate that this approach may be a workable solution to the problem of automated warfarin prescription.