checkConvergence.glmnet: Check whether convergence has occurred in the zeroness of...

In EMLasso: Fit (logistic) LASSO when there is missing data in the predictors

Description

Check whether convergence has occurred in the zeroness of coefficients

Usage

  checkConvergence.glmnet(coefs, minIt, maxIt, burnIn = 0,
    verbosity = 0)

  convergenceCheckCreator(minIt=20, maxIt=30, burnIn=0,
    basicCheckFunction=checkConvergence.glmnet)

Arguments

coefs	(sparse) matrix of coefficients (cols=variables, rows=iterations)
minIt	minimum number of iterations before convergence is possible
maxIt	maximum number of iterations before convergence is automatically assumed
burnIn	number of iterations that will be skipped for convergence checking
verbosity	The higher this value, the more levels of progress and debug information is displayed (note: in R for Windows, turn off buffered output)
basicCheckFunction	function (like checkConvergence.glmnet) that will do the actual work

Value

a list holding items:

1. converged: a single logical (TRUE if convergence has happened or maxIt passed)

2. minIt: as passed in

3. maxIt: as passed in

4. burnIn: as passed in

5. iterCount: number of iterations that have already occurred

6. usedFunction: characters, holding "checkConvergence.glmnet"

the result of basicCheckFunction

Note

the first row (initial estimates) and column (intercept) are skipped in the checks

Author(s)

Nick Sabbe [e-mail]

See Also

EMLasso

Examples

curset<-rbinom(11, 1, 0.5)
ret<-matrix(curset, nrow=1)
for(i in 1:10)
{
	chg<-rbinom(11, 1, 0.05)
	curset[chg==1] <- 1-curset[chg==1]
	ret<-rbind(ret, curset)
}
print(ret)
checkConvergence.glmnet(ret, 2, 20, verbosity = 2)

EMLasso documentation built on May 2, 2019, 5:49 p.m.