KN.test: Estimation of the number of spikes K in a spiked covariance...

In trambakbanerjee/casp: Coordinate-wise Adaptive Shrinkage Prediction

Estimation of the number of spikes K in a spiked covariance model

Description

Estimate the number of spikes in a spiked covariance model using the procedure described in Kritchman and Nadler, (2009)

Usage

KN.test(S, numObs, alpha = 0.05)

Arguments

S	the sample covariance matrix
numObs	number of data observations
alpha	significance level for the test

Details

This method calculates the number of spikes in a spiked covariance model by iteratively comparing a standardized eigenvalue to the quantile of Tracy-Wisdom distribution for given significance level α. The procedure is described in Kritchman and Nadler, (2009).

Value

1. numOfSpikes - The number of spikes K

2. sigma.sq - Estimated noise variance

References

Kritchman, S. and Nadler, B., 2009. Non-parametric detection of the number of signals: Hypothesis testing and random matrix theory. IEEE Transactions on Signal Processing, 57(10), pp.3930-3941.

See Also

sigmasq.est

Examples

library(casp)
S<-diag(c(10,8,6,4,rep(1,6)))
KN.test.out<-KN.test(S,50)

trambakbanerjee/casp documentation built on Nov. 22, 2022, 7:24 p.m.