SVM predict on dataframe with different factor levels

Question

I have a dataframe I want to make predictions on from an SVM, but the dataframe doesn\'t have all of the levels that the original training dataframe did. Is there an easy way ar

Answer 1

Just be sure to specify the levels in df2

library(e1071)
df = data.frame(y = c(rep(1:3, each = 3)), x = rep(c("A", "B", "C"), each = 3))

m1 = svm(y ~ x, df)
df2 = data.frame(x = factor("B",levels = c("A","B","C")))

predict(m1, df2)