Hello R users!
I have several data frames where some of the variables have many missing
observations. For example, Q1 in one of my data frames has over 66% of its
observations missing. I have tried imputation with mice but it does not work
for all the data frames and I get the following message or a similar message to
this:
iter imp variable
1 1 Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Q15 Q19
Q36 Q47 Q52 Q79 Q80 Q94 Q97 Q104 Q108 Q122 Q131 Q134 P1 P2 P3 P4
P5 P6Error in solve.default(xtx + diag(pen)) :
system is computationally singular: reciprocal condition number = 1.83044e-16
In addition: Warning messages:
1: In sqrt((sum(residuals^2))/(sum(ry) - ncol(x) - 1)) : NaNs produced
...
7: In sqrt((sum(residuals^2))/(sum(ry) - ncol(x) - 1)) : NaNs produced
Note: warnings 2 to 6 suppressed by me.
I would like to try a different approach where I delete the variables that have
more than 50% missing observations from the data frame (well, the actual
percentage might change). I have already deleted from the data frame the
variables that were all missing and for this I used the following code, which
was kindly suggested by one of you:
## Data frame after removing any blank columns:dfQ <- dfQtemp[ ,
sapply(dfQtemp, function(x) !all(is.na(x)))]
Any ideas or suggestons for deleting variables with partially missing data?
Thanks and have a great weekend!
Rita ===================================== "If you think education is
expensive, try ignorance."--Derek Bok
[[alternative HTML version deleted]]
______________________________________________
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.
