Output details
11 - Computer Science and Informatics
Royal Holloway, University of London
Clinical mass spectrometry proteomic diagnosis by conformal predictors
<24> This paper is based on our winning entry for the international classification competition organised by Leiden University Medical Center, Holland and describes the first application of the newly developed conformal predictors to diagnose breast cancer using proteomic mass spectrometry data. The developed conformal predictor allows us not just to classify samples, but to add valid measures of confidence in our predictions for individual patients. This is extremely important in the medical domain where the assessment and control of risk of misclassification is required. The approach has been developed further in MRC grant G0301107 and EU FP7 project 201381.