Hi all, I have two questions about variables in glmnet: 1. We are doing a logistic regression with binary outcome variable using a set of predictors that include continuous and binary predictors(coded 0 and 1). If the latter are centered and standardized, they will be transformed into negative and positive numbers; when multiplied by a single beta, I believe they will have undesirable effects. Is there a way to standardize only specified variables? Alternatively, should glmnet be run with manually centered and standardized continuous variables, binary variables coded 0 and 1, and with standardize = FALSE. 2. We have predictors with missing values. We have been handling these by creating a dummy variable for the predictor with a value of 0 if a value is present and 1 if a value is absent. If the model is forced to include both the predictor and the dummy variable, the model-assigned coefficient will effectively "interpolate" for the missing value. How can I force the dummy variable to be included in glmnet whenever the predictor variable is included?
-- View this message in context: http://r.789695.n4.nabble.com/Glmnet-Variable-Questions-tp3620379p3620379.html Sent from the R help mailing list archive at Nabble.com. ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
