getLogLikeRatio: Get the log-likelihood ratio from a binary rule... In VHDClassification: Discrimination/Classification in very high dimension with linear and quadratic rules.

Description

Binary rules can be expressed

Usage

 1 getLogLikeRatio(object)

Arguments

 object an object of type LinearRule or QuadraticRule.

Details

Get everything that defines a log likelihood ratio between two gaussian measures.

Value

A list, see getLogLikeRatio-methods

Robin Girard

References

Fast rate of convergence in high dimensional linear discriminant analysis. R. Girard To appear in Journal of Nonparametric Statistics.\ Very high dimensional discriminant analysis with thresholding estimation. R. Girard. Submitted.

Examples

 1 2 3 4 5 6 7 8 9 10 11 12 p=100; n=20 ; mu=array(0,c(p,4)); mu[1:10,1]=2 ;mu[11:20,2]=2;C=array(c(1,20),p) mu[21:30,3]=2 x=NULL; y=NULL; for (k in 1:4){ x=rbind(x,t(array(C^(1/2),c(p,n))*(matrix(rnorm(p*n),nrow=p,ncol=n))+array(mu[,k],c(p,n)))); y=c(y,array(k,n))} #Learning LearnedLinearPartitionWithLLR=learnPartitionWithLLR(x,y,procedure='FDRThresh') Rule=getBinaryRule(LearnedLinearPartitionWithLLR,1,2) LLR=getLogLikeRatio(Rule) print(LLR)

VHDClassification documentation built on May 2, 2019, 2:38 a.m.