Improving Clinical Relevance in Ensemble Support Vector Machine Models of Radiation Pneumonitis Risk

Patients undergoing thoracic radiation therapy can develop radiation pneumonitis (RP), a potentially fatal inflammation of the lungs. Support vector machines (SVMs), a statistical machine learning method, have recently been used to build binary-outcome RP prediction models with promising results. In this work, we (1) introduce a feature-ranking selection step to improve the parsimony of our previous ensemble SVM model (2) show that ensembles of SVMs provide a statistically significant performance improvement in the area under the cross-validated receiver operating curve and (3) apply Platt's tuning to the component SVMs to generate probability estimates in order to augment clinical relevance.

• Type: Journal
• Year: 2009
• Where: ICMLA
• Authors: Todd W. Schiller, Yixin Chen, Issam El-Naqa, Joseph O. Deasy
• Keywords: ICMLA 2009, Machine Learning, Receiver Operating Curve, RP Prediction Models, Thoracic Radiation Therapy