[R] How to select a reasonable shrinkage coefficient in stepplr?

Dear R-users, I am using stepplr for L2 regularized logistic regression. Since number of attribute is too large i discarded interaction terms. Everything is fine but only problem i have faced that i cannot choose a good shrinkage coefficient (lambda). If CV is the best way to estimate, can you pl

[R] Algorithm for penalized regression in design package

Dear R-users, I'm using lrm() in from the design package for l2-regularized logistic regression. Does anyone know which algorithm lrm() uses for this? Thanks, Tirtha -- View this message in context: http://www.nabble.com/Algorithm-for-penalized-regression-in-design-package-tp14585913p145859

[R] Algorithm for penalized regression in design package

Dear R-users, I'm using lrm() in from the design package for l2-regularized logistic regression. Does anyone know which algorithm lrm() uses for this? An article by Cessie and Houwelingen (Ridge estimators in logistic regression; Applied Statistics, 1992) is cited in the reference manual. Is th

[R] 'Best penalty' in design package

Dear Users, In case of ridge logistic regression, i want to calculate the optimum penalty using aic and bic criteria. Here is the sample code: fit <- lrm(RES ~CAT01+NUM01+NUM02+CAT02+CAT03+CAT04+NUM03+CAT05+CAT06+NUM04+ CAT07+CAT08+NUM05+NUM06, data = train.data, x = TRUE, y = T

[R] 'Best penalty' in design package

Dear Users, In case of ridge logistic regression, i want to calculate the optimum penalty using aic and bic criteria. Here is the sample code: fit <- lrm(RES ~CAT01+NUM01+NUM02+CAT02+CAT03+CAT04+NUM03+CAT05+CAT06+NUM04+ CAT07+CAT08+NUM05+NUM06, data = train.data, x = TRUE, y = TR

[R] Stepwise logistic model selection using Cp and BIC criteria

Hi, Is there any package for logistic model selection using BIC and Mallow's Cp statistic? If not, then kindly suggest me some ways to deal with these problems. Thanks. -- View this message in context: http://www.nabble.com/Stepwise-logistic-model-selection-using-Cp-and-BIC-criteria-tf4464430.

[R] glmpath error

Hi, I am using glampath package for L1 regularized logistic regression. I got the following error messege. > model.fit <- glmpath(train.data[,1:20], train.data$RES, family=binomial) Error in one %*% x : requires numeric matrix/vector arguments where train.data is a 700X21 matrix and 21st colum

Re: [R] glmpath error

Then what is the solution? Duncan Murdoch-2 wrote: > > Tirthadeep wrote: >> Hi, >> >> I am using glampath package for L1 regularized logistic regression. I got >> the following error messege. >> >> >>> model.fit <- glmp

[R] glmpath: how to choose best lambda

Hi all, I am using glampath package for L1 regularized logistic regression. I have read the article " L1 regularization path algorithm for GLM" by park and Hastie (2006). One thing I can't understand that how to find best lambda for my prediction. I want to use that lambda for the prediction not