I created data frame of data (observed and auxiliary data) in R and fit the model on data. Then I predict the dependent variable, but in predict data there are not available (NA). I use code "na.omit (data)" but the problem persists.
Does anyone have suggestions?
Check if the data has infinite values (+Inf, -Inf; possibly coming form "log(0)" or something similar). Remove these infinite values and run the model again.
Example how to remove all infinite values from the data.frame "x":
i = rowSums(!sapply(x, is.finite))==0
x = x[i,]
Sounds like it could be an easy fix (?complete.cases may be helpful), but it's tough to help without knowing what you've done. Can you provide a bit of code that reproduces your problem?
Some good advice for minimal reproducible examples: http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example
Although I strongly advise you follow Troy's advice on reproducible examples, here's something that sometimes (but not always) works:
### remove NAs
row.has.na
Problems such as this often occur when the na datum is encoded as the literal text string "NA." Check to make sure that the data type is NA, with, for example, is.na
- Badges
- Science topic