R-CNR-Tree, R-NNet Neural Network, junk characters
When I run an analysis using the R-CNR-Tree or R-NNet Neural Network algorithm, I get junk characters in the newly generated column of the result.
R algorithms does not support unicode characters. Therefore, if the input data contains unicode characters, the result of R algorithms will contain junk characters.
This is a limitation with the R 2.11.1 version that Predictive Analysis supports. Therefore, if you don't want to see junk characters in the result, ensure that the input data does not contain unicode characters.