DiagTest3Grp (version 1.6)

Sp.Sm.Se: Obtain empirical estimation of the three correct classification probabilities

Description

once identified the optimal cut-points from VUS or Youden3Grp analyses, use the function to obtain the three empirical correct classification probabilities associated with each diagnosis group.

Usage

Sp.Sm.Se(x, y, z, t.minus, t.plus)

Arguments

x

A numeric vector, a diagnostic test's measurements in the D- (usually healthy subjects).

y

A numeric vector, a diagnostic test's measurements in the D0 (usually mildly diseased subjects).

z

A numeric vector, a diagnostic test's measurements in the D+ (usually severely diseased subjects).

t.minus

A numeric value, the lower optimal cut-point identified from VUS or Youden3Grp analyses.

t.plus

A numeric value, the upper optimal cut-point identified from VUS or Youden3Grp analyses.

Value

Return a numeric vector with three components Sp, Sm and Se, the three correct classification probabilities.

Warning

The bootstrapping to obtain the variance on the nonparametric VUS estimate may take a while.

Details

Specificity: $Sp=Pr(x\le t_-)$ for $D^-$ group; Sensitivity: $Se=Pr(z\ge t_+)$ for $D^+$ group and the probability of the diagnostic test for the $D^0$ group fall between the two cut points: $Sm=Pr(t_- \le y \le t_+)$. These three probabilities will be estimated empirically.

References

Xiong, C. and van Belle, G. and Miller, J.P. and Morris, J.C. (2006) Measuring and Estimating Diagnostic Accuracy When There Are Three Ordinal Diagnostic Groups. Statistics In Medicine 25 7 1251--1273.

Ferri, C. and Hernandez-Orallo, J. and Salido, M.A. (2003) Volume under the ROC Surface for Multi-class Problems LECTURE NOTES IN COMPUTER SCIENCE 108--120.

See Also

VUS Normal.VUS NonParametric.VUS.var

Examples

 data(AL)
 group <- AL$group

 ##take the negated kfront marker measurements
 kfront <- -AL$kfront

 x <- kfront[group=="D-"]
 y <- kfront[group=="D0"]
 z <- kfront[group=="D+"]
 class.prob <- Sp.Sm.Se(x=x,y=y,z=z,t.minus=-1.68,t.plus=0.91)