A Flexible Method for Estimating the ROC Curve |
| |
Authors: | Haobo Ren Xiao-Hua Zhou Hua Liang |
| |
Institution: |
a Department of Statistics, Texas A&M University, USA.
b Department of Biostatistics, University of Washington, USA.
c St Jude Children's Research Hospital, Memphis, USA. |
| |
Abstract: | In this paper we propose a flexible method for estimating a receiver operating characteristic (ROC) curve that is based on a continuous-scale test. The approach is easily understood and efficiently computed, and robust to the smooth parameter selection, which needs intensive computation when using local polynomial and smoothing spline techniques. The results from our simulation experiment indicate that the moderate-sample numerical performance of our estimator is better than the empirical ROC curve estimator and comparable to the local linear estimator. The availability of easy implementation is also illustrated by our simulation. We apply the proposed method to two real data sets. |
| |
Keywords: | Penalized Spline Kernel Smoothing Local Polynomial Bandwidth Selection |
本文献已被 InformaWorld 等数据库收录! |