How to calculate AIC for a SVM model built using "ksvm" in "kernlab" package?

Question

I built a SVM using "ksvm" from "kernlab" package. Here is my code:

library(kernlab)
load("landslides2.Rd")
fit<-ksvm(landslides.2006~slope+aspect+swi+distriver+ndvi+spi,data=d,kernel="polydot",kpar=list(scale=0.2,degree=3,offset=0.05),C=0.01,prob.model=TRUE)

Is there anyway to calculate the AIC (Akaike information criterion) for this model? Thank you!