[R] error message when using mice: missing observations in cov/cor
n_haag at web.de
Sun Mar 14 12:38:42 CET 2010
I am trying to do multiple imputations with the mice package. My dataset consists of 132 variables and 92 data rows. I specified the predictor matrix using quickpred, which resulted in about 30 predictors per variable. I am using norm as imputation method for continuous variables and logreg for categorical variables.
At various points during the imputation I get the error message: error in FUN(newX[, i], ...) : missing observations in cov/cor . I tried to exclude some variables but cannot trace down the problem, as it reappears with other variables. It seems as if mice imputes NaN for some variables, but I don't know why. It also doesn't seem to depend on the iteration number. Can anyone help me?
More information about the R-help