On 11/11/16 17:53, James Duffy wrote:
Thanks for the suggestion. Removing that column worked. Could I suggest adding in a check either pre reading into R or pre-analysis in R to drop any character columns from the modelling but keep them in the attribute table?
Well, as I said, I'm not sure that character columns (aka factors in R) are really a problem as such. At least some of the classifiers should be able to deal with them. This would be important, for example, for hierarchical classification, where part of the information concerning a segment might be the class of the encompassing segment in a coarser segmentation. I'll have to look into this.
The issue here was the incoherence between training data and non-classified data in terms of column names.
Moritz _______________________________________________ grass-user mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/grass-user
