R How to avoid Warnings: In Ops.factor(y, yHat) : ‘-’ not meaningful for factors

Question

dane <- read.table(file = "http://archive.ics.uci.edu/ml/machine-learning-databases/breast-cancer-wisconsin/wpbc.data",sep=",", dec = ".")

library(randomForest) 
library(cvTools) 
cv <- cvFit(randomForest,  x=dane[,-2], y=dane[,2], R=10, k=100, args=list(ntree=500),foldType = "random",cost = rmspe)

Welcome to Stack Overflow! Please read the info about [how to ask a good question](http://stackoverflow.com/help/how-to-ask) — zx8754, Jun 14 '16 at 13:19
See this example, maybe convert factor to numeric first: `x <- factor(1:3); x - 1` — zx8754, Jun 14 '16 at 13:21

score 1 · Answer 1 · answered Jun 14 '16 at 13:46

1

The cvFit function is designed for numeric outcomes. Although randomForest has no difficulty with factor outcomes (in which case it does classification), the results are being processed by a function that expects a regression model.

answered Jun 14 '16 at 13:46

IRTFM

258,963
21
364
487

R How to avoid Warnings: In Ops.factor(y, yHat) : ‘-’ not meaningful for factors

1 Answers1