Dear R Users,
Once the function svm ( ) in library{e1071} like any function in R is run with
some variables including missing values ( NA ), it is needed to include
na.pass=TRUE arguement in function in order to deal with NAs. The arguement
na.omit is not desired, since it redueces the length of variables.
A problem when running the function svm including na.pass, is that the result
of fitting consists a part of seemingly rigth fiitted values from the first of
the length of fitted values to somewhere in a series and another part of fitted
values which are NAs comes to the end of length of fitted values in a series.
But we expect fitted values without any NA.
How it can be corrected ?
Thank you so much for help.
Amir