power.target.Nclasses: Compute the alternative hypothesis' PDs by increasing the PDs...

In validateRS: One-Sided Multivariate Testing Procedures for Rating Systems

Description

Compute the alternative hypothesis' PDs by increasing the PDs under the null hypothesis towards 1.

One can either increase all the PDs (i.e. altenative H1A in the ECB Working series' paper) or one at a time (scenario H1B).

The alternative hypothesis H1C can be computed using sample.knowledge.H1.

Usage

power.target.Nclasses(p.0, size, N , target = 0.5, alpha = 0.05, precision = 1e-07)

Arguments

p.0	A Vector with the PD values under H0
size	A Vector with the size of the rating classes
N	The number of PDs in which H1 should be different from the PDs under H0. For the alternative hypothesis H1A of the ECB Working paper, N should be set to length(p.0), for the alternative hypothesis H1B N should be 1.
target	The target power for the benchmark test
alpha	The significance level chosen
precision	The precision for the PDs under H1

Value

A vector with PD values under H1, such that the power of the benchmark equals the target power.

See Also

sample.knowledge.H1

Examples

data(ratingData)

p.0<-ratingData$p.0
sizes<-ratingData$sizes

h1A<-power.target.Nclasses(p.0=p.0, size=sizes[1,], N=length(p.0), 
                                 target=0.50)
h1A                                 

h1B<-power.target.Nclasses(p.0=p.0, size=sizes[,], N=1, 
                                 target=0.30)
h1B                                 

validateRS documentation built on May 2, 2019, 3:47 p.m.