ROC curve estimator based on kernel estimates of the constituent distributions

Rufibach, K. (2011).

Code to compute the kernel density estimate of the ROC curve and the corresponding AUC proposed in Hall & Hyndman (2003). This code was used to compare the log-concave ROC curve estimate to this kernel based estimate in Rufibach (2012).
This code contains a small example using data available in logcondens. The methodology is also available in the function smooth() in the R package pROC .


References

Hall, P.G., Hyndman, R.J. (2003). Improved methods for bandwidth selection when estimating ROC curves, Statist. Probab. Lett. 64, 181-189. doi

Rufibach, K. (2012). A smooth ROC curve estimator based on log-concave density estimates. Int. J. Biostat., accepted. arxiv