powered by
Parameter values x are transformed to 1-10^x. This is helpful for parameters that are likely to be set very close to (but below) a value of 1, such as discount factors in reinforcement learning.
1-10^x
trans_1minus10pow(x)
input
f2 <- function(x){2^x} fn <- c("identity", "exp", "f2") xNat <- diag(3) SPOT::transformX(xNat, fn)
Run the code above in your browser using DataLab