Sean Porter
2014-Mar-20 07:26 UTC
randomForest warning: The response has five or fewer unique values. Are you sure you want to do regression?
Hello everyone, Im relatively new to R and new to the randomForest package and have scoured the archives for help with no luck. I am trying to perform a regression on a set of predictors and response variables to determine the most important predictors. I have 100 response variables collected from 14 sites and 8 predictor variables from the same 14 sites. I run the code to perform the randomForest regression given by Pitcher et al 2011 ( http://gradientforest.r-forge.r-project.org/biodiversity-survey.pdf ). However, after running the code I get the warning: " In randomForest.default(m, y, ...) : The response has five or fewer unique values. Are you sure you want to do regression?" And it produces a set of 500 regression trees for each of 3 species only when the number of species in the response file is 100. I noticed that in the example by Pitcher they get 500 trees from only 90 species even though they input 110 species in the response data. Why am I getting the warning/how do I solve it, and why is randomForest producing trees for only 3 species when I am looking at 100 species (response variables)? Many thanks Sean [[alternative HTML version deleted]]