kROC: Kernel Estimation for ROC Curves In sROC: Nonparametric Smooth ROC Curves for Continuous Data

Description

To compute the nonparametric kernel estimate of receiver operating characteristic (ROC) Curves for continuous data.

Usage

 ```1 2``` ```kROC(x, y, bw.x="pi_ucv", bw.y="pi_ucv", adjust=1, kernel=c("normal", "epanechnikov"), xgrid, ngrid=256, from, to, cut=3, na.rm = FALSE, ...) ```

Arguments

 `x` numeric vector. `y` numeric vector. `bw.x` the smoothing bandwidth of `x` to be used. `bw` can also be a character string giving a rule to choose the bandwidth. See `bw.CDF` and `bw.CDF.pi`. The default used the Altman and Leger's plug-in approach with an unbiased cross-validation pilot bandwidth. `bw.y` the smoothing bandwidth of `y` to be used. `adjust` the parameter for adjusting the bandwidth. The bandwidth used for the estimate is actually `adjust*bw`. By default, adjust=1. `kernel` a character string giving the smoothing kernel to be used. This must be either “normal” or “epanechnikov”. By default, the normal kernel is used. `xgrid` the user-defined data points at which the CDF is to be evaluated. If missing, the CDF will be evaluated at the equally spaced points defined within the function. `ngrid` the number of equally spaced points at which the density is to be estimated. `from` the left-most points of the grid at which the density is to be estimated. `to` the right-most points of the grid at which the density is to be estimated `cut` by default, the values of from and to are cut bandwidths beyond the extremes of the data. `na.rm` logical; if `TRUE`, missing values are removed from x. If `FALSE` any missing values cause an error. `...` further arguments for methods.

Details

estimate the nonparametric kernel estimate of receiver operating characteristic (ROC) Curves for continuous data

Value

An object of class “ROC”.

 `FPR` the false positive rate. `TPR` the true positive rate. `bw.x, bw.y` the bandwidths used. `nx, ny` the sample sizes after elimination of missing values. `call` the call which produced the result. `x.data.name, y.data.name` the deparsed names of the `x` argument. `x.has.na, y.has.na` logical; if `TRUE`, there are missing values in the original data.

The `print` method reports `summary` values on the `x` and `Fhat` components.

Author(s)

X.F. Wang wangx6@ccf.org

References

Lloyd, C.J. (1998). Using smoothed receiver operating characteristic curves to summarize and compare diagnostic systems. Journal of the American Statistical Association, 93(444): 1356-1364.

Zhou, X.H. and Harezlak, J. (2002). Comparison of bandwidth selection methods for kernel smoothing of ROC curves. Statistics in Medicine, 21, 2045-2055.

Zou, K.H., Hall, W.J., and Shapiro, D.E. (1997). Smooth non-parametric receiver operating characteristic (ROC) curves for continuous diagnostic tests. Statistics in medicine, 16(19): 2143-56.

`bw.CDF`, `bw.CDF.pi.`
 ``` 1 2 3 4 5 6 7 8 9 10``` ```## -------------------- set.seed(100) n <- 200 x <- rgamma(n,2,1) y <- rnorm(n) xy.ROC <- kROC(x,y, bw.x="pi_sj",bw.y="pi_sj") xy.ROC plot(xy.ROC) ```