1st Edition

ROC Analysis for Classification and Prediction in Practice

    234 Pages 18 Color & 20 B/W Illustrations
    by Chapman & Hall

    This book presents a unified and up-to-date introduction to ROC methodologies, covering both diagnosis (classification) and prediction. The emphasis is on the conceptual underpinning of ROC analysis and the practical implementation in diverse scientific fields. A plethora of examples accompany the methodologic discussion using standard statistical software such as R and STATA. The book arrives after two decades of intensive growth in both the methods and the applications of ROC analysis and presents a new synthesis. The authors provide a contemporary, integrated exposition of ROC methodology for both classification and prediction and include material on multiple-class ROC. This book avoids lengthy technical exposition and provides code and datasets in each chapter. ROC Analysis for Classification and Prediction in Practice is intended for researchers and graduate students, but will also be useful for those that use ROC analysis in diverse disciplines such as diagnostic medicine, bioinformatics, medical physics, and perception psychology.

    1. Introduction  2. Measures of Diagnostic and Predictive Performance  3. Statistical inference for the ROC curve  4. Comparing ROC curves  5. The ROC surface and k-class classification for k > 2  6. ROC regression  7. Missing data and errors-in-variables in ROC analysis


    Christos T Nakas is Full Professor in Biometry at the University of Thessaly, Volos, Greece, and Primary Investigator/Consultant for Biostatistics and Data Science at the Department of Clinical Chemistry (UKC), Inselspital, University Hospital of the University of Bern, Bern, Switzerland. His research revolves around ROC analysis, Statistical testing/modeling, methods of Agreement, and their applications in Medicine, and Life Sciences disciplines in general.

    Leonidas E Bantis is Assistant Professor in Biostatistics at the Department of Biostatistics and Data Science, University of Kansas Medical Center, and a member of the University of Kansas Cancer Center, Kansas City, KS, USA. His research focus lies on the development of methods related to marker discovery, evaluation, modeling, and comparisons. He is primarily interested in the mathematical aspects and different metrics that are involved in the receiver operating characteristic (ROC) space.

    Constantine A Gatsonis is Henry Ledyard Goddard University Professor of Biostatistics, at Brown University School of Public Health, Providence, RI, U.S.A. He is the founding Chair of the Department of Biostatistics and founding Director of the Center for Statistical Sciences at Brown. Dr. Gatsonis is a leading authority on the evaluation of diagnostic and screening tests, and has made major contributions to the development of methods for medical technology assessment and health services and outcomes research. He is a world leader in methods for applying and synthesizing evidence on diagnostic tests in medicine and is currently developing methods for Comparative Effectiveness Research in diagnosis and prediction, and radiomics.